diff options
author | Stephen Hemminger <shemminger@linux-foundation.org> | 2007-10-03 23:41:36 (GMT) |
---|---|---|
committer | David S. Miller <davem@sunset.davemloft.net> | 2007-10-10 23:47:45 (GMT) |
commit | bea3348eef27e6044b6161fd04c3152215f96411 (patch) | |
tree | f0990b263e5ce42505d290a4c346fe990bcd4c33 | |
parent | dde4e47e8fe333a5649a3fa0e7db1fa7c08d6158 (diff) | |
download | linux-bea3348eef27e6044b6161fd04c3152215f96411.tar.xz |
[NET]: Make NAPI polling independent of struct net_device objects.
Several devices have multiple independant RX queues per net
device, and some have a single interrupt doorbell for several
queues.
In either case, it's easier to support layouts like that if the
structure representing the poll is independant from the net
device itself.
The signature of the ->poll() call back goes from:
int foo_poll(struct net_device *dev, int *budget)
to
int foo_poll(struct napi_struct *napi, int budget)
The caller is returned the number of RX packets processed (or
the number of "NAPI credits" consumed if you want to get
abstract). The callee no longer messes around bumping
dev->quota, *budget, etc. because that is all handled in the
caller upon return.
The napi_struct is to be embedded in the device driver private data
structures.
Furthermore, it is the driver's responsibility to disable all NAPI
instances in it's ->stop() device close handler. Since the
napi_struct is privatized into the driver's private data structures,
only the driver knows how to get at all of the napi_struct instances
it may have per-device.
With lots of help and suggestions from Rusty Russell, Roland Dreier,
Michael Chan, Jeff Garzik, and Jamal Hadi Salim.
Bug fixes from Thomas Graf, Roland Dreier, Peter Zijlstra,
Joseph Fannin, Scott Wood, Hans J. Koch, and Michael Chan.
[ Ported to current tree and all drivers converted. Integrated
Stephen's follow-on kerneldoc additions, and restored poll_list
handling to the old style to fix mutual exclusion issues. -DaveM ]
Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
91 files changed, 1758 insertions, 2375 deletions
diff --git a/Documentation/DocBook/kernel-api.tmpl b/Documentation/DocBook/kernel-api.tmpl index b886f52..e5da4f2 100644 --- a/Documentation/DocBook/kernel-api.tmpl +++ b/Documentation/DocBook/kernel-api.tmpl @@ -240,17 +240,23 @@ X!Ilib/string.c <sect1><title>Driver Support</title> !Enet/core/dev.c !Enet/ethernet/eth.c +!Enet/sched/sch_generic.c !Iinclude/linux/etherdevice.h +!Iinclude/linux/netdevice.h + </sect1> + <sect1><title>PHY Support</title> !Edrivers/net/phy/phy.c !Idrivers/net/phy/phy.c !Edrivers/net/phy/phy_device.c !Idrivers/net/phy/phy_device.c !Edrivers/net/phy/mdio_bus.c !Idrivers/net/phy/mdio_bus.c + </sect1> <!-- FIXME: Removed for now since no structured comments in source + <sect1><title>Wireless</title> X!Enet/core/wireless.c ---> </sect1> +--> <sect1><title>Synchronous PPP</title> !Edrivers/net/wan/syncppp.c </sect1> diff --git a/Documentation/networking/NAPI_HOWTO.txt b/Documentation/networking/NAPI_HOWTO.txt deleted file mode 100644 index 7907435..0000000 --- a/Documentation/networking/NAPI_HOWTO.txt +++ /dev/null @@ -1,766 +0,0 @@ -HISTORY: -February 16/2002 -- revision 0.2.1: -COR typo corrected -February 10/2002 -- revision 0.2: -some spell checking ;-> -January 12/2002 -- revision 0.1 -This is still work in progress so may change. -To keep up to date please watch this space. - -Introduction to NAPI -==================== - -NAPI is a proven (www.cyberus.ca/~hadi/usenix-paper.tgz) technique -to improve network performance on Linux. For more details please -read that paper. -NAPI provides a "inherent mitigation" which is bound by system capacity -as can be seen from the following data collected by Robert on Gigabit -ethernet (e1000): - - Psize Ipps Tput Rxint Txint Done Ndone - --------------------------------------------------------------- - 60 890000 409362 17 27622 7 6823 - 128 758150 464364 21 9301 10 7738 - 256 445632 774646 42 15507 21 12906 - 512 232666 994445 241292 19147 241192 1062 - 1024 119061 1000003 872519 19258 872511 0 - 1440 85193 1000003 946576 19505 946569 0 - - -Legend: -"Ipps" stands for input packets per second. -"Tput" == packets out of total 1M that made it out. -"txint" == transmit completion interrupts seen -"Done" == The number of times that the poll() managed to pull all -packets out of the rx ring. Note from this that the lower the -load the more we could clean up the rxring -"Ndone" == is the converse of "Done". Note again, that the higher -the load the more times we couldn't clean up the rxring. - -Observe that: -when the NIC receives 890Kpackets/sec only 17 rx interrupts are generated. -The system cant handle the processing at 1 interrupt/packet at that load level. -At lower rates on the other hand, rx interrupts go up and therefore the -interrupt/packet ratio goes up (as observable from that table). So there is -possibility that under low enough input, you get one poll call for each -input packet caused by a single interrupt each time. And if the system -cant handle interrupt per packet ratio of 1, then it will just have to -chug along .... - - -0) Prerequisites: -================== -A driver MAY continue using the old 2.4 technique for interfacing -to the network stack and not benefit from the NAPI changes. -NAPI additions to the kernel do not break backward compatibility. -NAPI, however, requires the following features to be available: - -A) DMA ring or enough RAM to store packets in software devices. - -B) Ability to turn off interrupts or maybe events that send packets up -the stack. - -NAPI processes packet events in what is known as dev->poll() method. -Typically, only packet receive events are processed in dev->poll(). -The rest of the events MAY be processed by the regular interrupt handler -to reduce processing latency (justified also because there are not that -many of them). -Note, however, NAPI does not enforce that dev->poll() only processes -receive events. -Tests with the tulip driver indicated slightly increased latency if -all of the interrupt handler is moved to dev->poll(). Also MII handling -gets a little trickier. -The example used in this document is to move the receive processing only -to dev->poll(); this is shown with the patch for the tulip driver. -For an example of code that moves all the interrupt driver to -dev->poll() look at the ported e1000 code. - -There are caveats that might force you to go with moving everything to -dev->poll(). Different NICs work differently depending on their status/event -acknowledgement setup. -There are two types of event register ACK mechanisms. - I) what is known as Clear-on-read (COR). - when you read the status/event register, it clears everything! - The natsemi and sunbmac NICs are known to do this. - In this case your only choice is to move all to dev->poll() - - II) Clear-on-write (COW) - i) you clear the status by writing a 1 in the bit-location you want. - These are the majority of the NICs and work the best with NAPI. - Put only receive events in dev->poll(); leave the rest in - the old interrupt handler. - ii) whatever you write in the status register clears every thing ;-> - Cant seem to find any supported by Linux which do this. If - someone knows such a chip email us please. - Move all to dev->poll() - -C) Ability to detect new work correctly. -NAPI works by shutting down event interrupts when there's work and -turning them on when there's none. -New packets might show up in the small window while interrupts were being -re-enabled (refer to appendix 2). A packet might sneak in during the period -we are enabling interrupts. We only get to know about such a packet when the -next new packet arrives and generates an interrupt. -Essentially, there is a small window of opportunity for a race condition -which for clarity we'll refer to as the "rotting packet". - -This is a very important topic and appendix 2 is dedicated for more -discussion. - -Locking rules and environmental guarantees -========================================== - --Guarantee: Only one CPU at any time can call dev->poll(); this is because -only one CPU can pick the initial interrupt and hence the initial -netif_rx_schedule(dev); -- The core layer invokes devices to send packets in a round robin format. -This implies receive is totally lockless because of the guarantee that only -one CPU is executing it. -- contention can only be the result of some other CPU accessing the rx -ring. This happens only in close() and suspend() (when these methods -try to clean the rx ring); -****guarantee: driver authors need not worry about this; synchronization -is taken care for them by the top net layer. --local interrupts are enabled (if you dont move all to dev->poll()). For -example link/MII and txcomplete continue functioning just same old way. -This improves the latency of processing these events. It is also assumed that -the receive interrupt is the largest cause of noise. Note this might not -always be true. -[according to Manfred Spraul, the winbond insists on sending one -txmitcomplete interrupt for each packet (although this can be mitigated)]. -For these broken drivers, move all to dev->poll(). - -For the rest of this text, we'll assume that dev->poll() only -processes receive events. - -new methods introduce by NAPI -============================= - -a) netif_rx_schedule(dev) -Called by an IRQ handler to schedule a poll for device - -b) netif_rx_schedule_prep(dev) -puts the device in a state which allows for it to be added to the -CPU polling list if it is up and running. You can look at this as -the first half of netif_rx_schedule(dev) above; the second half -being c) below. - -c) __netif_rx_schedule(dev) -Add device to the poll list for this CPU; assuming that _prep above -has already been called and returned 1. - -d) netif_rx_reschedule(dev, undo) -Called to reschedule polling for device specifically for some -deficient hardware. Read Appendix 2 for more details. - -e) netif_rx_complete(dev) - -Remove interface from the CPU poll list: it must be in the poll list -on current cpu. This primitive is called by dev->poll(), when -it completes its work. The device cannot be out of poll list at this -call, if it is then clearly it is a BUG(). You'll know ;-> - -All of the above methods are used below, so keep reading for clarity. - -Device driver changes to be made when porting NAPI -================================================== - -Below we describe what kind of changes are required for NAPI to work. - -1) introduction of dev->poll() method -===================================== - -This is the method that is invoked by the network core when it requests -for new packets from the driver. A driver is allowed to send upto -dev->quota packets by the current CPU before yielding to the network -subsystem (so other devices can also get opportunity to send to the stack). - -dev->poll() prototype looks as follows: -int my_poll(struct net_device *dev, int *budget) - -budget is the remaining number of packets the network subsystem on the -current CPU can send up the stack before yielding to other system tasks. -*Each driver is responsible for decrementing budget by the total number of -packets sent. - Total number of packets cannot exceed dev->quota. - -dev->poll() method is invoked by the top layer, the driver just sends if it -can to the stack the packet quantity requested. - -more on dev->poll() below after the interrupt changes are explained. - -2) registering dev->poll() method -=================================== - -dev->poll should be set in the dev->probe() method. -e.g: -dev->open = my_open; -. -. -/* two new additions */ -/* first register my poll method */ -dev->poll = my_poll; -/* next register my weight/quanta; can be overridden in /proc */ -dev->weight = 16; -. -. -dev->stop = my_close; - - - -3) scheduling dev->poll() -============================= -This involves modifying the interrupt handler and the code -path which takes the packet off the NIC and sends them to the -stack. - -it's important at this point to introduce the classical D Becker -interrupt processor: - ------------------- -static irqreturn_t -netdevice_interrupt(int irq, void *dev_id, struct pt_regs *regs) -{ - - struct net_device *dev = (struct net_device *)dev_instance; - struct my_private *tp = (struct my_private *)dev->priv; - - int work_count = my_work_count; - status = read_interrupt_status_reg(); - if (status == 0) - return IRQ_NONE; /* Shared IRQ: not us */ - if (status == 0xffff) - return IRQ_HANDLED; /* Hot unplug */ - if (status & error) - do_some_error_handling() - - do { - acknowledge_ints_ASAP(); - - if (status & link_interrupt) { - spin_lock(&tp->link_lock); - do_some_link_stat_stuff(); - spin_lock(&tp->link_lock); - } - - if (status & rx_interrupt) { - receive_packets(dev); - } - - if (status & rx_nobufs) { - make_rx_buffs_avail(); - } - - if (status & tx_related) { - spin_lock(&tp->lock); - tx_ring_free(dev); - if (tx_died) - restart_tx(); - spin_unlock(&tp->lock); - } - - status = read_interrupt_status_reg(); - - } while (!(status & error) || more_work_to_be_done); - return IRQ_HANDLED; -} - ----------------------------------------------------------------------- - -We now change this to what is shown below to NAPI-enable it: - ----------------------------------------------------------------------- -static irqreturn_t -netdevice_interrupt(int irq, void *dev_id, struct pt_regs *regs) -{ - struct net_device *dev = (struct net_device *)dev_instance; - struct my_private *tp = (struct my_private *)dev->priv; - - status = read_interrupt_status_reg(); - if (status == 0) - return IRQ_NONE; /* Shared IRQ: not us */ - if (status == 0xffff) - return IRQ_HANDLED; /* Hot unplug */ - if (status & error) - do_some_error_handling(); - - do { -/************************ start note *********************************/ - acknowledge_ints_ASAP(); // dont ack rx and rxnobuff here -/************************ end note *********************************/ - - if (status & link_interrupt) { - spin_lock(&tp->link_lock); - do_some_link_stat_stuff(); - spin_unlock(&tp->link_lock); - } -/************************ start note *********************************/ - if (status & rx_interrupt || (status & rx_nobuffs)) { - if (netif_rx_schedule_prep(dev)) { - - /* disable interrupts caused - * by arriving packets */ - disable_rx_and_rxnobuff_ints(); - /* tell system we have work to be done. */ - __netif_rx_schedule(dev); - } else { - printk("driver bug! interrupt while in poll\n"); - /* FIX by disabling interrupts */ - disable_rx_and_rxnobuff_ints(); - } - } -/************************ end note note *********************************/ - - if (status & tx_related) { - spin_lock(&tp->lock); - tx_ring_free(dev); - - if (tx_died) - restart_tx(); - spin_unlock(&tp->lock); - } - - status = read_interrupt_status_reg(); - -/************************ start note *********************************/ - } while (!(status & error) || more_work_to_be_done(status)); -/************************ end note note *********************************/ - return IRQ_HANDLED; -} - ---------------------------------------------------------------------- - - -We note several things from above: - -I) Any interrupt source which is caused by arriving packets is now -turned off when it occurs. Depending on the hardware, there could be -several reasons that arriving packets would cause interrupts; these are the -interrupt sources we wish to avoid. The two common ones are a) a packet -arriving (rxint) b) a packet arriving and finding no DMA buffers available -(rxnobuff) . -This means also acknowledge_ints_ASAP() will not clear the status -register for those two items above; clearing is done in the place where -proper work is done within NAPI; at the poll() and refill_rx_ring() -discussed further below. -netif_rx_schedule_prep() returns 1 if device is in running state and -gets successfully added to the core poll list. If we get a zero value -we can _almost_ assume are already added to the list (instead of not running. -Logic based on the fact that you shouldn't get interrupt if not running) -We rectify this by disabling rx and rxnobuf interrupts. - -II) that receive_packets(dev) and make_rx_buffs_avail() may have disappeared. -These functionalities are still around actually...... - -infact, receive_packets(dev) is very close to my_poll() and -make_rx_buffs_avail() is invoked from my_poll() - -4) converting receive_packets() to dev->poll() -=============================================== - -We need to convert the classical D Becker receive_packets(dev) to my_poll() - -First the typical receive_packets() below: -------------------------------------------------------------------- - -/* this is called by interrupt handler */ -static void receive_packets (struct net_device *dev) -{ - - struct my_private *tp = (struct my_private *)dev->priv; - rx_ring = tp->rx_ring; - cur_rx = tp->cur_rx; - int entry = cur_rx % RX_RING_SIZE; - int received = 0; - int rx_work_limit = tp->dirty_rx + RX_RING_SIZE - tp->cur_rx; - - while (rx_ring_not_empty) { - u32 rx_status; - unsigned int rx_size; - unsigned int pkt_size; - struct sk_buff *skb; - /* read size+status of next frame from DMA ring buffer */ - /* the number 16 and 4 are just examples */ - rx_status = le32_to_cpu (*(u32 *) (rx_ring + ring_offset)); - rx_size = rx_status >> 16; - pkt_size = rx_size - 4; - - /* process errors */ - if ((rx_size > (MAX_ETH_FRAME_SIZE+4)) || - (!(rx_status & RxStatusOK))) { - netdrv_rx_err (rx_status, dev, tp, ioaddr); - return; - } - - if (--rx_work_limit < 0) - break; - - /* grab a skb */ - skb = dev_alloc_skb (pkt_size + 2); - if (skb) { - . - . - netif_rx (skb); - . - . - } else { /* OOM */ - /*seems very driver specific ... some just pass - whatever is on the ring already. */ - } - - /* move to the next skb on the ring */ - entry = (++tp->cur_rx) % RX_RING_SIZE; - received++ ; - - } - - /* store current ring pointer state */ - tp->cur_rx = cur_rx; - - /* Refill the Rx ring buffers if they are needed */ - refill_rx_ring(); - . - . - -} -------------------------------------------------------------------- -We change it to a new one below; note the additional parameter in -the call. - -------------------------------------------------------------------- - -/* this is called by the network core */ -static int my_poll (struct net_device *dev, int *budget) -{ - - struct my_private *tp = (struct my_private *)dev->priv; - rx_ring = tp->rx_ring; - cur_rx = tp->cur_rx; - int entry = cur_rx % RX_BUF_LEN; - /* maximum packets to send to the stack */ -/************************ note note *********************************/ - int rx_work_limit = dev->quota; - -/************************ end note note *********************************/ - do { // outer beginning loop starts here - - clear_rx_status_register_bit(); - - while (rx_ring_not_empty) { - u32 rx_status; - unsigned int rx_size; - unsigned int pkt_size; - struct sk_buff *skb; - /* read size+status of next frame from DMA ring buffer */ - /* the number 16 and 4 are just examples */ - rx_status = le32_to_cpu (*(u32 *) (rx_ring + ring_offset)); - rx_size = rx_status >> 16; - pkt_size = rx_size - 4; - - /* process errors */ - if ((rx_size > (MAX_ETH_FRAME_SIZE+4)) || - (!(rx_status & RxStatusOK))) { - netdrv_rx_err (rx_status, dev, tp, ioaddr); - return 1; - } - -/************************ note note *********************************/ - if (--rx_work_limit < 0) { /* we got packets, but no quota */ - /* store current ring pointer state */ - tp->cur_rx = cur_rx; - - /* Refill the Rx ring buffers if they are needed */ - refill_rx_ring(dev); - goto not_done; - } -/********************** end note **********************************/ - - /* grab a skb */ - skb = dev_alloc_skb (pkt_size + 2); - if (skb) { - . - . -/************************ note note *********************************/ - netif_receive_skb (skb); -/********************** end note **********************************/ - . - . - } else { /* OOM */ - /*seems very driver specific ... common is just pass - whatever is on the ring already. */ - } - - /* move to the next skb on the ring */ - entry = (++tp->cur_rx) % RX_RING_SIZE; - received++ ; - - } - - /* store current ring pointer state */ - tp->cur_rx = cur_rx; - - /* Refill the Rx ring buffers if they are needed */ - refill_rx_ring(dev); - - /* no packets on ring; but new ones can arrive since we last - checked */ - status = read_interrupt_status_reg(); - if (rx status is not set) { - /* If something arrives in this narrow window, - an interrupt will be generated */ - goto done; - } - /* done! at least that's what it looks like ;-> - if new packets came in after our last check on status bits - they'll be caught by the while check and we go back and clear them - since we havent exceeded our quota */ - } while (rx_status_is_set); - -done: - -/************************ note note *********************************/ - dev->quota -= received; - *budget -= received; - - /* If RX ring is not full we are out of memory. */ - if (tp->rx_buffers[tp->dirty_rx % RX_RING_SIZE].skb == NULL) - goto oom; - - /* we are happy/done, no more packets on ring; put us back - to where we can start processing interrupts again */ - netif_rx_complete(dev); - enable_rx_and_rxnobuf_ints(); - - /* The last op happens after poll completion. Which means the following: - * 1. it can race with disabling irqs in irq handler (which are done to - * schedule polls) - * 2. it can race with dis/enabling irqs in other poll threads - * 3. if an irq raised after the beginning of the outer beginning - * loop (marked in the code above), it will be immediately - * triggered here. - * - * Summarizing: the logic may result in some redundant irqs both - * due to races in masking and due to too late acking of already - * processed irqs. The good news: no events are ever lost. - */ - - return 0; /* done */ - -not_done: - if (tp->cur_rx - tp->dirty_rx > RX_RING_SIZE/2 || - tp->rx_buffers[tp->dirty_rx % RX_RING_SIZE].skb == NULL) - refill_rx_ring(dev); - - if (!received) { - printk("received==0\n"); - received = 1; - } - dev->quota -= received; - *budget -= received; - return 1; /* not_done */ - -oom: - /* Start timer, stop polling, but do not enable rx interrupts. */ - start_poll_timer(dev); - return 0; /* we'll take it from here so tell core "done"*/ - -/************************ End note note *********************************/ -} -------------------------------------------------------------------- - -From above we note that: -0) rx_work_limit = dev->quota -1) refill_rx_ring() is in charge of clearing the bit for rxnobuff when -it does the work. -2) We have a done and not_done state. -3) instead of netif_rx() we call netif_receive_skb() to pass the skb. -4) we have a new way of handling oom condition -5) A new outer for (;;) loop has been added. This serves the purpose of -ensuring that if a new packet has come in, after we are all set and done, -and we have not exceeded our quota that we continue sending packets up. - - ------------------------------------------------------------ -Poll timer code will need to do the following: - -a) - - if (tp->cur_rx - tp->dirty_rx > RX_RING_SIZE/2 || - tp->rx_buffers[tp->dirty_rx % RX_RING_SIZE].skb == NULL) - refill_rx_ring(dev); - - /* If RX ring is not full we are still out of memory. - Restart the timer again. Else we re-add ourselves - to the master poll list. - */ - - if (tp->rx_buffers[tp->dirty_rx % RX_RING_SIZE].skb == NULL) - restart_timer(); - - else netif_rx_schedule(dev); /* we are back on the poll list */ - -5) dev->close() and dev->suspend() issues -========================================== -The driver writer needn't worry about this; the top net layer takes -care of it. - -6) Adding new Stats to /proc -============================= -In order to debug some of the new features, we introduce new stats -that need to be collected. -TODO: Fill this later. - -APPENDIX 1: discussion on using ethernet HW FC -============================================== -Most chips with FC only send a pause packet when they run out of Rx buffers. -Since packets are pulled off the DMA ring by a softirq in NAPI, -if the system is slow in grabbing them and we have a high input -rate (faster than the system's capacity to remove packets), then theoretically -there will only be one rx interrupt for all packets during a given packetstorm. -Under low load, we might have a single interrupt per packet. -FC should be programmed to apply in the case when the system cant pull out -packets fast enough i.e send a pause only when you run out of rx buffers. -Note FC in itself is a good solution but we have found it to not be -much of a commodity feature (both in NICs and switches) and hence falls -under the same category as using NIC based mitigation. Also, experiments -indicate that it's much harder to resolve the resource allocation -issue (aka lazy receiving that NAPI offers) and hence quantify its usefulness -proved harder. In any case, FC works even better with NAPI but is not -necessary. - - -APPENDIX 2: the "rotting packet" race-window avoidance scheme -============================================================= - -There are two types of associations seen here - -1) status/int which honors level triggered IRQ - -If a status bit for receive or rxnobuff is set and the corresponding -interrupt-enable bit is not on, then no interrupts will be generated. However, -as soon as the "interrupt-enable" bit is unmasked, an immediate interrupt is -generated. [assuming the status bit was not turned off]. -Generally the concept of level triggered IRQs in association with a status and -interrupt-enable CSR register set is used to avoid the race. - -If we take the example of the tulip: -"pending work" is indicated by the status bit(CSR5 in tulip). -the corresponding interrupt bit (CSR7 in tulip) might be turned off (but -the CSR5 will continue to be turned on with new packet arrivals even if -we clear it the first time) -Very important is the fact that if we turn on the interrupt bit on when -status is set that an immediate irq is triggered. - -If we cleared the rx ring and proclaimed there was "no more work -to be done" and then went on to do a few other things; then when we enable -interrupts, there is a possibility that a new packet might sneak in during -this phase. It helps to look at the pseudo code for the tulip poll -routine: - --------------------------- - do { - ACK; - while (ring_is_not_empty()) { - work-work-work - if quota is exceeded: exit, no touching irq status/mask - } - /* No packets, but new can arrive while we are doing this*/ - CSR5 := read - if (CSR5 is not set) { - /* If something arrives in this narrow window here, - * where the comments are ;-> irq will be generated */ - unmask irqs; - exit poll; - } - } while (rx_status_is_set); ------------------------- - -CSR5 bit of interest is only the rx status. -If you look at the last if statement: -you just finished grabbing all the packets from the rx ring .. you check if -status bit says there are more packets just in ... it says none; you then -enable rx interrupts again; if a new packet just came in during this check, -we are counting that CSR5 will be set in that small window of opportunity -and that by re-enabling interrupts, we would actually trigger an interrupt -to register the new packet for processing. - -[The above description nay be very verbose, if you have better wording -that will make this more understandable, please suggest it.] - -2) non-capable hardware - -These do not generally respect level triggered IRQs. Normally, -irqs may be lost while being masked and the only way to leave poll is to do -a double check for new input after netif_rx_complete() is invoked -and re-enable polling (after seeing this new input). - -Sample code: - ---------- - . - . -restart_poll: - while (ring_is_not_empty()) { - work-work-work - if quota is exceeded: exit, not touching irq status/mask - } - . - . - . - enable_rx_interrupts() - netif_rx_complete(dev); - if (ring_has_new_packet() && netif_rx_reschedule(dev, received)) { - disable_rx_and_rxnobufs() - goto restart_poll - } while (rx_status_is_set); ---------- - -Basically netif_rx_complete() removes us from the poll list, but because a -new packet which will never be caught due to the possibility of a race -might come in, we attempt to re-add ourselves to the poll list. - - - - -APPENDIX 3: Scheduling issues. -============================== -As seen NAPI moves processing to softirq level. Linux uses the ksoftirqd as the -general solution to schedule softirq's to run before next interrupt and by putting -them under scheduler control. Also this prevents consecutive softirq's from -monopolize the CPU. This also have the effect that the priority of ksoftirq needs -to be considered when running very CPU-intensive applications and networking to -get the proper balance of softirq/user balance. Increasing ksoftirq priority to 0 -(eventually more) is reported cure problems with low network performance at high -CPU load. - -Most used processes in a GIGE router: -USER PID %CPU %MEM SIZE RSS TTY STAT START TIME COMMAND -root 3 0.2 0.0 0 0 ? RWN Aug 15 602:00 (ksoftirqd_CPU0) -root 232 0.0 7.9 41400 40884 ? S Aug 15 74:12 gated - --------------------------------------------------------------------- - -relevant sites: -================== -ftp://robur.slu.se/pub/Linux/net-development/NAPI/ - - --------------------------------------------------------------------- -TODO: Write net-skeleton.c driver. -------------------------------------------------------------- - -Authors: -======== -Alexey Kuznetsov <kuznet@ms2.inr.ac.ru> -Jamal Hadi Salim <hadi@cyberus.ca> -Robert Olsson <Robert.Olsson@data.slu.se> - -Acknowledgements: -================ -People who made this document better: - -Lennert Buytenhek <buytenh@gnu.org> -Andrew Morton <akpm@zip.com.au> -Manfred Spraul <manfred@colorfullife.com> -Donald Becker <becker@scyld.com> -Jeff Garzik <jgarzik@pobox.com> diff --git a/Documentation/networking/netdevices.txt b/Documentation/networking/netdevices.txt index 3786929..9f7be9b 100644 --- a/Documentation/networking/netdevices.txt +++ b/Documentation/networking/netdevices.txt @@ -95,9 +95,13 @@ dev->set_multicast_list: Synchronization: netif_tx_lock spinlock. Context: BHs disabled -dev->poll: - Synchronization: __LINK_STATE_RX_SCHED bit in dev->state. See - dev_close code and comments in net/core/dev.c for more info. +struct napi_struct synchronization rules +======================================== +napi->poll: + Synchronization: NAPI_STATE_SCHED bit in napi->state. Device + driver's dev->close method will invoke napi_disable() on + all NAPI instances which will do a sleeping poll on the + NAPI_STATE_SCHED napi->state bit, waiting for all pending + NAPI activity to cease. Context: softirq will be called with interrupts disabled by netconsole. - diff --git a/drivers/infiniband/ulp/ipoib/ipoib.h b/drivers/infiniband/ulp/ipoib/ipoib.h index 285c143..35f3ca4 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib.h +++ b/drivers/infiniband/ulp/ipoib/ipoib.h @@ -228,6 +228,8 @@ struct ipoib_dev_priv { struct net_device *dev; + struct napi_struct napi; + unsigned long flags; struct mutex mcast_mutex; @@ -351,7 +353,7 @@ extern struct workqueue_struct *ipoib_workqueue; /* functions */ -int ipoib_poll(struct net_device *dev, int *budget); +int ipoib_poll(struct napi_struct *napi, int budget); void ipoib_ib_completion(struct ib_cq *cq, void *dev_ptr); struct ipoib_ah *ipoib_create_ah(struct net_device *dev, diff --git a/drivers/infiniband/ulp/ipoib/ipoib_ib.c b/drivers/infiniband/ulp/ipoib/ipoib_ib.c index 1094488..481e4b6 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_ib.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_ib.c @@ -281,63 +281,58 @@ static void ipoib_ib_handle_tx_wc(struct net_device *dev, struct ib_wc *wc) wc->status, wr_id, wc->vendor_err); } -int ipoib_poll(struct net_device *dev, int *budget) +int ipoib_poll(struct napi_struct *napi, int budget) { - struct ipoib_dev_priv *priv = netdev_priv(dev); - int max = min(*budget, dev->quota); + struct ipoib_dev_priv *priv = container_of(napi, struct ipoib_dev_priv, napi); + struct net_device *dev = priv->dev; int done; int t; - int empty; int n, i; done = 0; - empty = 0; - while (max) { +poll_more: + while (done < budget) { + int max = (budget - done); + t = min(IPOIB_NUM_WC, max); n = ib_poll_cq(priv->cq, t, priv->ibwc); - for (i = 0; i < n; ++i) { + for (i = 0; i < n; i++) { struct ib_wc *wc = priv->ibwc + i; if (wc->wr_id & IPOIB_CM_OP_SRQ) { ++done; - --max; ipoib_cm_handle_rx_wc(dev, wc); } else if (wc->wr_id & IPOIB_OP_RECV) { ++done; - --max; ipoib_ib_handle_rx_wc(dev, wc); } else ipoib_ib_handle_tx_wc(dev, wc); } - if (n != t) { - empty = 1; + if (n != t) break; - } } - dev->quota -= done; - *budget -= done; - - if (empty) { - netif_rx_complete(dev); + if (done < budget) { + netif_rx_complete(dev, napi); if (unlikely(ib_req_notify_cq(priv->cq, IB_CQ_NEXT_COMP | IB_CQ_REPORT_MISSED_EVENTS)) && - netif_rx_reschedule(dev, 0)) - return 1; - - return 0; + netif_rx_reschedule(dev, napi)) + goto poll_more; } - return 1; + return done; } void ipoib_ib_completion(struct ib_cq *cq, void *dev_ptr) { - netif_rx_schedule(dev_ptr); + struct net_device *dev = dev_ptr; + struct ipoib_dev_priv *priv = netdev_priv(dev); + + netif_rx_schedule(dev, &priv->napi); } static inline int post_send(struct ipoib_dev_priv *priv, @@ -577,7 +572,6 @@ int ipoib_ib_dev_stop(struct net_device *dev, int flush) int i; clear_bit(IPOIB_FLAG_INITIALIZED, &priv->flags); - netif_poll_disable(dev); ipoib_cm_dev_stop(dev); @@ -660,7 +654,6 @@ timeout: msleep(1); } - netif_poll_enable(dev); ib_req_notify_cq(priv->cq, IB_CQ_NEXT_COMP); return 0; diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c b/drivers/infiniband/ulp/ipoib/ipoib_main.c index 894b1dcd..a59ff07 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_main.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_main.c @@ -98,16 +98,20 @@ int ipoib_open(struct net_device *dev) ipoib_dbg(priv, "bringing up interface\n"); + napi_enable(&priv->napi); set_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags); if (ipoib_pkey_dev_delay_open(dev)) return 0; - if (ipoib_ib_dev_open(dev)) + if (ipoib_ib_dev_open(dev)) { + napi_disable(&priv->napi); return -EINVAL; + } if (ipoib_ib_dev_up(dev)) { ipoib_ib_dev_stop(dev, 1); + napi_disable(&priv->napi); return -EINVAL; } @@ -140,6 +144,7 @@ static int ipoib_stop(struct net_device *dev) ipoib_dbg(priv, "stopping interface\n"); clear_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags); + napi_disable(&priv->napi); netif_stop_queue(dev); @@ -948,8 +953,8 @@ static void ipoib_setup(struct net_device *dev) dev->hard_header = ipoib_hard_header; dev->set_multicast_list = ipoib_set_mcast_list; dev->neigh_setup = ipoib_neigh_setup_dev; - dev->poll = ipoib_poll; - dev->weight = 100; + + netif_napi_add(dev, &priv->napi, ipoib_poll, 100); dev->watchdog_timeo = HZ; diff --git a/drivers/net/8139cp.c b/drivers/net/8139cp.c index a79f28c..7f18ca23 100644 --- a/drivers/net/8139cp.c +++ b/drivers/net/8139cp.c @@ -334,6 +334,8 @@ struct cp_private { spinlock_t lock; u32 msg_enable; + struct napi_struct napi; + struct pci_dev *pdev; u32 rx_config; u16 cpcmd; @@ -501,12 +503,12 @@ static inline unsigned int cp_rx_csum_ok (u32 status) return 0; } -static int cp_rx_poll (struct net_device *dev, int *budget) +static int cp_rx_poll(struct napi_struct *napi, int budget) { - struct cp_private *cp = netdev_priv(dev); - unsigned rx_tail = cp->rx_tail; - unsigned rx_work = dev->quota; - unsigned rx; + struct cp_private *cp = container_of(napi, struct cp_private, napi); + struct net_device *dev = cp->dev; + unsigned int rx_tail = cp->rx_tail; + int rx; rx_status_loop: rx = 0; @@ -588,33 +590,28 @@ rx_next: desc->opts1 = cpu_to_le32(DescOwn | cp->rx_buf_sz); rx_tail = NEXT_RX(rx_tail); - if (!rx_work--) + if (rx >= budget) break; } cp->rx_tail = rx_tail; - dev->quota -= rx; - *budget -= rx; - /* if we did not reach work limit, then we're done with * this round of polling */ - if (rx_work) { + if (rx < budget) { unsigned long flags; if (cpr16(IntrStatus) & cp_rx_intr_mask) goto rx_status_loop; - local_irq_save(flags); + spin_lock_irqsave(&cp->lock, flags); cpw16_f(IntrMask, cp_intr_mask); - __netif_rx_complete(dev); - local_irq_restore(flags); - - return 0; /* done */ + __netif_rx_complete(dev, napi); + spin_unlock_irqrestore(&cp->lock, flags); } - return 1; /* not done */ + return rx; } static irqreturn_t cp_interrupt (int irq, void *dev_instance) @@ -647,9 +644,9 @@ static irqreturn_t cp_interrupt (int irq, void *dev_instance) } if (status & (RxOK | RxErr | RxEmpty | RxFIFOOvr)) - if (netif_rx_schedule_prep(dev)) { + if (netif_rx_schedule_prep(dev, &cp->napi)) { cpw16_f(IntrMask, cp_norx_intr_mask); - __netif_rx_schedule(dev); + __netif_rx_schedule(dev, &cp->napi); } if (status & (TxOK | TxErr | TxEmpty | SWInt)) @@ -1175,6 +1172,8 @@ static int cp_open (struct net_device *dev) if (rc) return rc; + napi_enable(&cp->napi); + cp_init_hw(cp); rc = request_irq(dev->irq, cp_interrupt, IRQF_SHARED, dev->name, dev); @@ -1188,6 +1187,7 @@ static int cp_open (struct net_device *dev) return 0; err_out_hw: + napi_disable(&cp->napi); cp_stop_hw(cp); cp_free_rings(cp); return rc; @@ -1198,6 +1198,8 @@ static int cp_close (struct net_device *dev) struct cp_private *cp = netdev_priv(dev); unsigned long flags; + napi_disable(&cp->napi); + if (netif_msg_ifdown(cp)) printk(KERN_DEBUG "%s: disabling interface\n", dev->name); @@ -1933,11 +1935,10 @@ static int cp_init_one (struct pci_dev *pdev, const struct pci_device_id *ent) dev->hard_start_xmit = cp_start_xmit; dev->get_stats = cp_get_stats; dev->do_ioctl = cp_ioctl; - dev->poll = cp_rx_poll; #ifdef CONFIG_NET_POLL_CONTROLLER dev->poll_controller = cp_poll_controller; #endif - dev->weight = 16; /* arbitrary? from NAPI_HOWTO.txt. */ + netif_napi_add(dev, &cp->napi, cp_rx_poll, 16); #ifdef BROKEN dev->change_mtu = cp_change_mtu; #endif diff --git a/drivers/net/8139too.c b/drivers/net/8139too.c index f4e4298..20af6ba 100644 --- a/drivers/net/8139too.c +++ b/drivers/net/8139too.c @@ -573,6 +573,8 @@ struct rtl8139_private { int drv_flags; struct pci_dev *pci_dev; u32 msg_enable; + struct napi_struct napi; + struct net_device *dev; struct net_device_stats stats; unsigned char *rx_ring; unsigned int cur_rx; /* Index into the Rx buffer of next Rx pkt. */ @@ -625,10 +627,10 @@ static void rtl8139_tx_timeout (struct net_device *dev); static void rtl8139_init_ring (struct net_device *dev); static int rtl8139_start_xmit (struct sk_buff *skb, struct net_device *dev); -static int rtl8139_poll(struct net_device *dev, int *budget); #ifdef CONFIG_NET_POLL_CONTROLLER static void rtl8139_poll_controller(struct net_device *dev); #endif +static int rtl8139_poll(struct napi_struct *napi, int budget); static irqreturn_t rtl8139_interrupt (int irq, void *dev_instance); static int rtl8139_close (struct net_device *dev); static int netdev_ioctl (struct net_device *dev, struct ifreq *rq, int cmd); @@ -963,6 +965,7 @@ static int __devinit rtl8139_init_one (struct pci_dev *pdev, assert (dev != NULL); tp = netdev_priv(dev); + tp->dev = dev; ioaddr = tp->mmio_addr; assert (ioaddr != NULL); @@ -976,8 +979,7 @@ static int __devinit rtl8139_init_one (struct pci_dev *pdev, /* The Rtl8139-specific entries in the device structure. */ dev->open = rtl8139_open; dev->hard_start_xmit = rtl8139_start_xmit; - dev->poll = rtl8139_poll; - dev->weight = 64; + netif_napi_add(dev, &tp->napi, rtl8139_poll, 64); dev->stop = rtl8139_close; dev->get_stats = rtl8139_get_stats; dev->set_multicast_list = rtl8139_set_rx_mode; @@ -1332,6 +1334,8 @@ static int rtl8139_open (struct net_device *dev) } + napi_enable(&tp->napi); + tp->mii.full_duplex = tp->mii.force_media; tp->tx_flag = (TX_FIFO_THRESH << 11) & 0x003f0000; @@ -2103,39 +2107,32 @@ static void rtl8139_weird_interrupt (struct net_device *dev, } } -static int rtl8139_poll(struct net_device *dev, int *budget) +static int rtl8139_poll(struct napi_struct *napi, int budget) { - struct rtl8139_private *tp = netdev_priv(dev); + struct rtl8139_private *tp = container_of(napi, struct rtl8139_private, napi); + struct net_device *dev = tp->dev; void __iomem *ioaddr = tp->mmio_addr; - int orig_budget = min(*budget, dev->quota); - int done = 1; + int work_done; spin_lock(&tp->rx_lock); - if (likely(RTL_R16(IntrStatus) & RxAckBits)) { - int work_done; - - work_done = rtl8139_rx(dev, tp, orig_budget); - if (likely(work_done > 0)) { - *budget -= work_done; - dev->quota -= work_done; - done = (work_done < orig_budget); - } - } + work_done = 0; + if (likely(RTL_R16(IntrStatus) & RxAckBits)) + work_done += rtl8139_rx(dev, tp, budget); - if (done) { + if (work_done < budget) { unsigned long flags; /* * Order is important since data can get interrupted * again when we think we are done. */ - local_irq_save(flags); + spin_lock_irqsave(&tp->lock, flags); RTL_W16_F(IntrMask, rtl8139_intr_mask); - __netif_rx_complete(dev); - local_irq_restore(flags); + __netif_rx_complete(dev, napi); + spin_unlock_irqrestore(&tp->lock, flags); } spin_unlock(&tp->rx_lock); - return !done; + return work_done; } /* The interrupt handler does all of the Rx thread work and cleans up @@ -2180,9 +2177,9 @@ static irqreturn_t rtl8139_interrupt (int irq, void *dev_instance) /* Receive packets are processed by poll routine. If not running start it now. */ if (status & RxAckBits){ - if (netif_rx_schedule_prep(dev)) { + if (netif_rx_schedule_prep(dev, &tp->napi)) { RTL_W16_F (IntrMask, rtl8139_norx_intr_mask); - __netif_rx_schedule (dev); + __netif_rx_schedule(dev, &tp->napi); } } @@ -2223,7 +2220,8 @@ static int rtl8139_close (struct net_device *dev) void __iomem *ioaddr = tp->mmio_addr; unsigned long flags; - netif_stop_queue (dev); + netif_stop_queue(dev); + napi_disable(&tp->napi); if (netif_msg_ifdown(tp)) printk(KERN_DEBUG "%s: Shutting down ethercard, status was 0x%4.4x.\n", diff --git a/drivers/net/amd8111e.c b/drivers/net/amd8111e.c index a61b2f8..cf06fc0 100644 --- a/drivers/net/amd8111e.c +++ b/drivers/net/amd8111e.c @@ -723,9 +723,10 @@ static int amd8111e_tx(struct net_device *dev) #ifdef CONFIG_AMD8111E_NAPI /* This function handles the driver receive operation in polling mode */ -static int amd8111e_rx_poll(struct net_device *dev, int * budget) +static int amd8111e_rx_poll(struct napi_struct *napi, int budget) { - struct amd8111e_priv *lp = netdev_priv(dev); + struct amd8111e_priv *lp = container_of(napi, struct amd8111e_priv, napi); + struct net_device *dev = lp->amd8111e_net_dev; int rx_index = lp->rx_idx & RX_RING_DR_MOD_MASK; void __iomem *mmio = lp->mmio; struct sk_buff *skb,*new_skb; @@ -737,7 +738,7 @@ static int amd8111e_rx_poll(struct net_device *dev, int * budget) #if AMD8111E_VLAN_TAG_USED short vtag; #endif - int rx_pkt_limit = dev->quota; + int rx_pkt_limit = budget; unsigned long flags; do{ @@ -838,21 +839,14 @@ static int amd8111e_rx_poll(struct net_device *dev, int * budget) } while(intr0 & RINT0); /* Receive descriptor is empty now */ - dev->quota -= num_rx_pkt; - *budget -= num_rx_pkt; - spin_lock_irqsave(&lp->lock, flags); - netif_rx_complete(dev); + __netif_rx_complete(dev, napi); writel(VAL0|RINTEN0, mmio + INTEN0); writel(VAL2 | RDMD0, mmio + CMD0); spin_unlock_irqrestore(&lp->lock, flags); - return 0; rx_not_empty: - /* Do not call a netif_rx_complete */ - dev->quota -= num_rx_pkt; - *budget -= num_rx_pkt; - return 1; + return num_rx_pkt; } #else @@ -1287,11 +1281,11 @@ static irqreturn_t amd8111e_interrupt(int irq, void *dev_id) /* Check if Receive Interrupt has occurred. */ #ifdef CONFIG_AMD8111E_NAPI if(intr0 & RINT0){ - if(netif_rx_schedule_prep(dev)){ + if(netif_rx_schedule_prep(dev, &lp->napi)){ /* Disable receive interupts */ writel(RINTEN0, mmio + INTEN0); /* Schedule a polling routine */ - __netif_rx_schedule(dev); + __netif_rx_schedule(dev, &lp->napi); } else if (intren0 & RINTEN0) { printk("************Driver bug! \ @@ -1345,6 +1339,8 @@ static int amd8111e_close(struct net_device * dev) struct amd8111e_priv *lp = netdev_priv(dev); netif_stop_queue(dev); + napi_disable(&lp->napi); + spin_lock_irq(&lp->lock); amd8111e_disable_interrupt(lp); @@ -1375,12 +1371,15 @@ static int amd8111e_open(struct net_device * dev ) dev->name, dev)) return -EAGAIN; + napi_enable(&lp->napi); + spin_lock_irq(&lp->lock); amd8111e_init_hw_default(lp); if(amd8111e_restart(dev)){ spin_unlock_irq(&lp->lock); + napi_disable(&lp->napi); if (dev->irq) free_irq(dev->irq, dev); return -ENOMEM; @@ -2031,8 +2030,7 @@ static int __devinit amd8111e_probe_one(struct pci_dev *pdev, dev->tx_timeout = amd8111e_tx_timeout; dev->watchdog_timeo = AMD8111E_TX_TIMEOUT; #ifdef CONFIG_AMD8111E_NAPI - dev->poll = amd8111e_rx_poll; - dev->weight = 32; + netif_napi_add(dev, &lp->napi, amd8111e_rx_poll, 32); #endif #ifdef CONFIG_NET_POLL_CONTROLLER dev->poll_controller = amd8111e_poll; diff --git a/drivers/net/amd8111e.h b/drivers/net/amd8111e.h index e65080a..612e653 100644 --- a/drivers/net/amd8111e.h +++ b/drivers/net/amd8111e.h @@ -763,6 +763,8 @@ struct amd8111e_priv{ /* Reg memory mapped address */ void __iomem *mmio; + struct napi_struct napi; + spinlock_t lock; /* Guard lock */ unsigned long rx_idx, tx_idx; /* The next free ring entry */ unsigned long tx_complete_idx; diff --git a/drivers/net/arm/ep93xx_eth.c b/drivers/net/arm/ep93xx_eth.c index f6ece1d..7f016f3 100644 --- a/drivers/net/arm/ep93xx_eth.c +++ b/drivers/net/arm/ep93xx_eth.c @@ -169,6 +169,9 @@ struct ep93xx_priv spinlock_t tx_pending_lock; unsigned int tx_pending; + struct net_device *dev; + struct napi_struct napi; + struct net_device_stats stats; struct mii_if_info mii; @@ -190,15 +193,11 @@ static struct net_device_stats *ep93xx_get_stats(struct net_device *dev) return &(ep->stats); } -static int ep93xx_rx(struct net_device *dev, int *budget) +static int ep93xx_rx(struct net_device *dev, int processed, int budget) { struct ep93xx_priv *ep = netdev_priv(dev); - int rx_done; - int processed; - rx_done = 0; - processed = 0; - while (*budget > 0) { + while (processed < budget) { int entry; struct ep93xx_rstat *rstat; u32 rstat0; @@ -211,10 +210,8 @@ static int ep93xx_rx(struct net_device *dev, int *budget) rstat0 = rstat->rstat0; rstat1 = rstat->rstat1; - if (!(rstat0 & RSTAT0_RFP) || !(rstat1 & RSTAT1_RFP)) { - rx_done = 1; + if (!(rstat0 & RSTAT0_RFP) || !(rstat1 & RSTAT1_RFP)) break; - } rstat->rstat0 = 0; rstat->rstat1 = 0; @@ -275,8 +272,6 @@ static int ep93xx_rx(struct net_device *dev, int *budget) err: ep->rx_pointer = (entry + 1) & (RX_QUEUE_ENTRIES - 1); processed++; - dev->quota--; - (*budget)--; } if (processed) { @@ -284,7 +279,7 @@ err: wrw(ep, REG_RXSTSENQ, processed); } - return !rx_done; + return processed; } static int ep93xx_have_more_rx(struct ep93xx_priv *ep) @@ -293,36 +288,32 @@ static int ep93xx_have_more_rx(struct ep93xx_priv *ep) return !!((rstat->rstat0 & RSTAT0_RFP) && (rstat->rstat1 & RSTAT1_RFP)); } -static int ep93xx_poll(struct net_device *dev, int *budget) +static int ep93xx_poll(struct napi_struct *napi, int budget) { - struct ep93xx_priv *ep = netdev_priv(dev); - - /* - * @@@ Have to stop polling if device is downed while we - * are polling. - */ + struct ep93xx_priv *ep = container_of(napi, struct ep93xx_priv, napi); + struct net_device *dev = ep->dev; + int rx = 0; poll_some_more: - if (ep93xx_rx(dev, budget)) - return 1; - - netif_rx_complete(dev); - - spin_lock_irq(&ep->rx_lock); - wrl(ep, REG_INTEN, REG_INTEN_TX | REG_INTEN_RX); - if (ep93xx_have_more_rx(ep)) { - wrl(ep, REG_INTEN, REG_INTEN_TX); - wrl(ep, REG_INTSTSP, REG_INTSTS_RX); + rx = ep93xx_rx(dev, rx, budget); + if (rx < budget) { + int more = 0; + + spin_lock_irq(&ep->rx_lock); + __netif_rx_complete(dev, napi); + wrl(ep, REG_INTEN, REG_INTEN_TX | REG_INTEN_RX); + if (ep93xx_have_more_rx(ep)) { + wrl(ep, REG_INTEN, REG_INTEN_TX); + wrl(ep, REG_INTSTSP, REG_INTSTS_RX); + more = 1; + } spin_unlock_irq(&ep->rx_lock); - if (netif_rx_reschedule(dev, 0)) + if (more && netif_rx_reschedule(dev, napi)) goto poll_some_more; - - return 0; } - spin_unlock_irq(&ep->rx_lock); - return 0; + return rx; } static int ep93xx_xmit(struct sk_buff *skb, struct net_device *dev) @@ -426,9 +417,9 @@ static irqreturn_t ep93xx_irq(int irq, void *dev_id) if (status & REG_INTSTS_RX) { spin_lock(&ep->rx_lock); - if (likely(__netif_rx_schedule_prep(dev))) { + if (likely(__netif_rx_schedule_prep(dev, &ep->napi))) { wrl(ep, REG_INTEN, REG_INTEN_TX); - __netif_rx_schedule(dev); + __netif_rx_schedule(dev, &ep->napi); } spin_unlock(&ep->rx_lock); } @@ -648,7 +639,10 @@ static int ep93xx_open(struct net_device *dev) dev->dev_addr[4], dev->dev_addr[5]); } + napi_enable(&ep->napi); + if (ep93xx_start_hw(dev)) { + napi_disable(&ep->napi); ep93xx_free_buffers(ep); return -EIO; } @@ -662,6 +656,7 @@ static int ep93xx_open(struct net_device *dev) err = request_irq(ep->irq, ep93xx_irq, IRQF_SHARED, dev->name, dev); if (err) { + napi_disable(&ep->napi); ep93xx_stop_hw(dev); ep93xx_free_buffers(ep); return err; @@ -678,6 +673,7 @@ static int ep93xx_close(struct net_device *dev) { struct ep93xx_priv *ep = netdev_priv(dev); + napi_disable(&ep->napi); netif_stop_queue(dev); wrl(ep, REG_GIINTMSK, 0); @@ -788,14 +784,12 @@ struct net_device *ep93xx_dev_alloc(struct ep93xx_eth_data *data) dev->get_stats = ep93xx_get_stats; dev->ethtool_ops = &ep93xx_ethtool_ops; - dev->poll = ep93xx_poll; dev->hard_start_xmit = ep93xx_xmit; dev->open = ep93xx_open; dev->stop = ep93xx_close; dev->do_ioctl = ep93xx_ioctl; dev->features |= NETIF_F_SG | NETIF_F_HW_CSUM; - dev->weight = 64; return dev; } @@ -847,6 +841,8 @@ static int ep93xx_eth_probe(struct platform_device *pdev) goto err_out; } ep = netdev_priv(dev); + ep->dev = dev; + netif_napi_add(dev, &ep->napi, ep93xx_poll, 64); platform_set_drvdata(pdev, dev); diff --git a/drivers/net/b44.c b/drivers/net/b44.c index 0795df2..b92b3e2 100644 --- a/drivers/net/b44.c +++ b/drivers/net/b44.c @@ -848,10 +848,11 @@ static int b44_rx(struct b44 *bp, int budget) return received; } -static int b44_poll(struct net_device *netdev, int *budget) +static int b44_poll(struct napi_struct *napi, int budget) { - struct b44 *bp = netdev_priv(netdev); - int done; + struct b44 *bp = container_of(napi, struct b44, napi); + struct net_device *netdev = bp->dev; + int work_done; spin_lock_irq(&bp->lock); @@ -862,22 +863,9 @@ static int b44_poll(struct net_device *netdev, int *budget) } spin_unlock_irq(&bp->lock); - done = 1; - if (bp->istat & ISTAT_RX) { - int orig_budget = *budget; - int work_done; - - if (orig_budget > netdev->quota) - orig_budget = netdev->quota; - - work_done = b44_rx(bp, orig_budget); - - *budget -= work_done; - netdev->quota -= work_done; - - if (work_done >= orig_budget) - done = 0; - } + work_done = 0; + if (bp->istat & ISTAT_RX) + work_done += b44_rx(bp, budget); if (bp->istat & ISTAT_ERRORS) { unsigned long flags; @@ -888,15 +876,15 @@ static int b44_poll(struct net_device *netdev, int *budget) b44_init_hw(bp, B44_FULL_RESET_SKIP_PHY); netif_wake_queue(bp->dev); spin_unlock_irqrestore(&bp->lock, flags); - done = 1; + work_done = 0; } - if (done) { - netif_rx_complete(netdev); + if (work_done < budget) { + netif_rx_complete(netdev, napi); b44_enable_ints(bp); } - return (done ? 0 : 1); + return work_done; } static irqreturn_t b44_interrupt(int irq, void *dev_id) @@ -924,13 +912,13 @@ static irqreturn_t b44_interrupt(int irq, void *dev_id) goto irq_ack; } - if (netif_rx_schedule_prep(dev)) { + if (netif_rx_schedule_prep(dev, &bp->napi)) { /* NOTE: These writes are posted by the readback of * the ISTAT register below. */ bp->istat = istat; __b44_disable_ints(bp); - __netif_rx_schedule(dev); + __netif_rx_schedule(dev, &bp->napi); } else { printk(KERN_ERR PFX "%s: Error, poll already scheduled\n", dev->name); @@ -1420,6 +1408,8 @@ static int b44_open(struct net_device *dev) if (err) goto out; + napi_enable(&bp->napi); + b44_init_rings(bp); b44_init_hw(bp, B44_FULL_RESET); @@ -1427,6 +1417,7 @@ static int b44_open(struct net_device *dev) err = request_irq(dev->irq, b44_interrupt, IRQF_SHARED, dev->name, dev); if (unlikely(err < 0)) { + napi_disable(&bp->napi); b44_chip_reset(bp); b44_free_rings(bp); b44_free_consistent(bp); @@ -1609,7 +1600,7 @@ static int b44_close(struct net_device *dev) netif_stop_queue(dev); - netif_poll_disable(dev); + napi_disable(&bp->napi); del_timer_sync(&bp->timer); @@ -1626,8 +1617,6 @@ static int b44_close(struct net_device *dev) free_irq(dev->irq, dev); - netif_poll_enable(dev); - if (bp->flags & B44_FLAG_WOL_ENABLE) { b44_init_hw(bp, B44_PARTIAL_RESET); b44_setup_wol(bp); @@ -2194,8 +2183,7 @@ static int __devinit b44_init_one(struct pci_dev *pdev, dev->set_mac_address = b44_set_mac_addr; dev->do_ioctl = b44_ioctl; dev->tx_timeout = b44_tx_timeout; - dev->poll = b44_poll; - dev->weight = 64; + netif_napi_add(dev, &bp->napi, b44_poll, 64); dev->watchdog_timeo = B44_TX_TIMEOUT; #ifdef CONFIG_NET_POLL_CONTROLLER dev->poll_controller = b44_poll_controller; diff --git a/drivers/net/b44.h b/drivers/net/b44.h index e537e63..63c55a4 100644 --- a/drivers/net/b44.h +++ b/drivers/net/b44.h @@ -423,6 +423,8 @@ struct b44 { struct ring_info *rx_buffers; struct ring_info *tx_buffers; + struct napi_struct napi; + u32 dma_offset; u32 flags; #define B44_FLAG_B0_ANDLATER 0x00000001 diff --git a/drivers/net/bnx2.c b/drivers/net/bnx2.c index 66eed22..ab028ad 100644 --- a/drivers/net/bnx2.c +++ b/drivers/net/bnx2.c @@ -428,7 +428,7 @@ bnx2_netif_stop(struct bnx2 *bp) { bnx2_disable_int_sync(bp); if (netif_running(bp->dev)) { - netif_poll_disable(bp->dev); + napi_disable(&bp->napi); netif_tx_disable(bp->dev); bp->dev->trans_start = jiffies; /* prevent tx timeout */ } @@ -440,7 +440,7 @@ bnx2_netif_start(struct bnx2 *bp) if (atomic_dec_and_test(&bp->intr_sem)) { if (netif_running(bp->dev)) { netif_wake_queue(bp->dev); - netif_poll_enable(bp->dev); + napi_enable(&bp->napi); bnx2_enable_int(bp); } } @@ -2551,7 +2551,7 @@ bnx2_msi(int irq, void *dev_instance) if (unlikely(atomic_read(&bp->intr_sem) != 0)) return IRQ_HANDLED; - netif_rx_schedule(dev); + netif_rx_schedule(dev, &bp->napi); return IRQ_HANDLED; } @@ -2568,7 +2568,7 @@ bnx2_msi_1shot(int irq, void *dev_instance) if (unlikely(atomic_read(&bp->intr_sem) != 0)) return IRQ_HANDLED; - netif_rx_schedule(dev); + netif_rx_schedule(dev, &bp->napi); return IRQ_HANDLED; } @@ -2604,9 +2604,9 @@ bnx2_interrupt(int irq, void *dev_instance) if (unlikely(atomic_read(&bp->intr_sem) != 0)) return IRQ_HANDLED; - if (netif_rx_schedule_prep(dev)) { + if (netif_rx_schedule_prep(dev, &bp->napi)) { bp->last_status_idx = sblk->status_idx; - __netif_rx_schedule(dev); + __netif_rx_schedule(dev, &bp->napi); } return IRQ_HANDLED; @@ -2632,12 +2632,14 @@ bnx2_has_work(struct bnx2 *bp) } static int -bnx2_poll(struct net_device *dev, int *budget) +bnx2_poll(struct napi_struct *napi, int budget) { - struct bnx2 *bp = netdev_priv(dev); + struct bnx2 *bp = container_of(napi, struct bnx2, napi); + struct net_device *dev = bp->dev; struct status_block *sblk = bp->status_blk; u32 status_attn_bits = sblk->status_attn_bits; u32 status_attn_bits_ack = sblk->status_attn_bits_ack; + int work_done = 0; if ((status_attn_bits & STATUS_ATTN_EVENTS) != (status_attn_bits_ack & STATUS_ATTN_EVENTS)) { @@ -2655,23 +2657,14 @@ bnx2_poll(struct net_device *dev, int *budget) if (bp->status_blk->status_tx_quick_consumer_index0 != bp->hw_tx_cons) bnx2_tx_int(bp); - if (bp->status_blk->status_rx_quick_consumer_index0 != bp->hw_rx_cons) { - int orig_budget = *budget; - int work_done; - - if (orig_budget > dev->quota) - orig_budget = dev->quota; - - work_done = bnx2_rx_int(bp, orig_budget); - *budget -= work_done; - dev->quota -= work_done; - } + if (bp->status_blk->status_rx_quick_consumer_index0 != bp->hw_rx_cons) + work_done = bnx2_rx_int(bp, budget); bp->last_status_idx = bp->status_blk->status_idx; rmb(); if (!bnx2_has_work(bp)) { - netif_rx_complete(dev); + netif_rx_complete(dev, napi); if (likely(bp->flags & USING_MSI_FLAG)) { REG_WR(bp, BNX2_PCICFG_INT_ACK_CMD, BNX2_PCICFG_INT_ACK_CMD_INDEX_VALID | @@ -2686,10 +2679,9 @@ bnx2_poll(struct net_device *dev, int *budget) REG_WR(bp, BNX2_PCICFG_INT_ACK_CMD, BNX2_PCICFG_INT_ACK_CMD_INDEX_VALID | bp->last_status_idx); - return 0; } - return 1; + return work_done; } /* Called with rtnl_lock from vlan functions and also netif_tx_lock @@ -5039,6 +5031,8 @@ bnx2_open(struct net_device *dev) if (rc) return rc; + napi_enable(&bp->napi); + if ((bp->flags & MSI_CAP_FLAG) && !disable_msi) { if (pci_enable_msi(bp->pdev) == 0) { bp->flags |= USING_MSI_FLAG; @@ -5049,6 +5043,7 @@ bnx2_open(struct net_device *dev) rc = bnx2_request_irq(bp); if (rc) { + napi_disable(&bp->napi); bnx2_free_mem(bp); return rc; } @@ -5056,6 +5051,7 @@ bnx2_open(struct net_device *dev) rc = bnx2_init_nic(bp); if (rc) { + napi_disable(&bp->napi); bnx2_free_irq(bp); bnx2_free_skbs(bp); bnx2_free_mem(bp); @@ -5088,6 +5084,7 @@ bnx2_open(struct net_device *dev) rc = bnx2_request_irq(bp); if (rc) { + napi_disable(&bp->napi); bnx2_free_skbs(bp); bnx2_free_mem(bp); del_timer_sync(&bp->timer); @@ -5301,7 +5298,8 @@ bnx2_close(struct net_device *dev) while (bp->in_reset_task) msleep(1); - bnx2_netif_stop(bp); + bnx2_disable_int_sync(bp); + napi_disable(&bp->napi); del_timer_sync(&bp->timer); if (bp->flags & NO_WOL_FLAG) reset_code = BNX2_DRV_MSG_CODE_UNLOAD_LNK_DN; @@ -6858,11 +6856,10 @@ bnx2_init_one(struct pci_dev *pdev, const struct pci_device_id *ent) #ifdef BCM_VLAN dev->vlan_rx_register = bnx2_vlan_rx_register; #endif - dev->poll = bnx2_poll; dev->ethtool_ops = &bnx2_ethtool_ops; - dev->weight = 64; bp = netdev_priv(dev); + netif_napi_add(dev, &bp->napi, bnx2_poll, 64); #if defined(HAVE_POLL_CONTROLLER) || defined(CONFIG_NET_POLL_CONTROLLER) dev->poll_controller = poll_bnx2; diff --git a/drivers/net/bnx2.h b/drivers/net/bnx2.h index 102adfe..fbae439 100644 --- a/drivers/net/bnx2.h +++ b/drivers/net/bnx2.h @@ -6473,6 +6473,8 @@ struct bnx2 { struct net_device *dev; struct pci_dev *pdev; + struct napi_struct napi; + atomic_t intr_sem; struct status_block *status_blk; diff --git a/drivers/net/cassini.c b/drivers/net/cassini.c index f6e4030..13f14df 100644 --- a/drivers/net/cassini.c +++ b/drivers/net/cassini.c @@ -2485,7 +2485,7 @@ static irqreturn_t cas_interruptN(int irq, void *dev_id) if (status & INTR_RX_DONE_ALT) { /* handle rx separately */ #ifdef USE_NAPI cas_mask_intr(cp); - netif_rx_schedule(dev); + netif_rx_schedule(dev, &cp->napi); #else cas_rx_ringN(cp, ring, 0); #endif @@ -2536,7 +2536,7 @@ static irqreturn_t cas_interrupt1(int irq, void *dev_id) if (status & INTR_RX_DONE_ALT) { /* handle rx separately */ #ifdef USE_NAPI cas_mask_intr(cp); - netif_rx_schedule(dev); + netif_rx_schedule(dev, &cp->napi); #else cas_rx_ringN(cp, 1, 0); #endif @@ -2592,7 +2592,7 @@ static irqreturn_t cas_interrupt(int irq, void *dev_id) if (status & INTR_RX_DONE) { #ifdef USE_NAPI cas_mask_intr(cp); - netif_rx_schedule(dev); + netif_rx_schedule(dev, &cp->napi); #else cas_rx_ringN(cp, 0, 0); #endif @@ -2607,9 +2607,10 @@ static irqreturn_t cas_interrupt(int irq, void *dev_id) #ifdef USE_NAPI -static int cas_poll(struct net_device *dev, int *budget) +static int cas_poll(struct napi_struct *napi, int budget) { - struct cas *cp = netdev_priv(dev); + struct cas *cp = container_of(napi, struct cas, napi); + struct net_device *dev = cp->dev; int i, enable_intr, todo, credits; u32 status = readl(cp->regs + REG_INTR_STATUS); unsigned long flags; @@ -2620,20 +2621,18 @@ static int cas_poll(struct net_device *dev, int *budget) /* NAPI rx packets. we spread the credits across all of the * rxc rings - */ - todo = min(*budget, dev->quota); - - /* to make sure we're fair with the work we loop through each + * + * to make sure we're fair with the work we loop through each * ring N_RX_COMP_RING times with a request of - * todo / N_RX_COMP_RINGS + * budget / N_RX_COMP_RINGS */ enable_intr = 1; credits = 0; for (i = 0; i < N_RX_COMP_RINGS; i++) { int j; for (j = 0; j < N_RX_COMP_RINGS; j++) { - credits += cas_rx_ringN(cp, j, todo / N_RX_COMP_RINGS); - if (credits >= todo) { + credits += cas_rx_ringN(cp, j, budget / N_RX_COMP_RINGS); + if (credits >= budget) { enable_intr = 0; goto rx_comp; } @@ -2641,9 +2640,6 @@ static int cas_poll(struct net_device *dev, int *budget) } rx_comp: - *budget -= credits; - dev->quota -= credits; - /* final rx completion */ spin_lock_irqsave(&cp->lock, flags); if (status) @@ -2674,11 +2670,10 @@ rx_comp: #endif spin_unlock_irqrestore(&cp->lock, flags); if (enable_intr) { - netif_rx_complete(dev); + netif_rx_complete(dev, napi); cas_unmask_intr(cp); - return 0; } - return 1; + return credits; } #endif @@ -4351,6 +4346,9 @@ static int cas_open(struct net_device *dev) goto err_spare; } +#ifdef USE_NAPI + napi_enable(&cp->napi); +#endif /* init hw */ cas_lock_all_save(cp, flags); cas_clean_rings(cp); @@ -4376,6 +4374,9 @@ static int cas_close(struct net_device *dev) unsigned long flags; struct cas *cp = netdev_priv(dev); +#ifdef USE_NAPI + napi_enable(&cp->napi); +#endif /* Make sure we don't get distracted by suspend/resume */ mutex_lock(&cp->pm_mutex); @@ -5062,8 +5063,7 @@ static int __devinit cas_init_one(struct pci_dev *pdev, dev->watchdog_timeo = CAS_TX_TIMEOUT; dev->change_mtu = cas_change_mtu; #ifdef USE_NAPI - dev->poll = cas_poll; - dev->weight = 64; + netif_napi_add(dev, &cp->napi, cas_poll, 64); #endif #ifdef CONFIG_NET_POLL_CONTROLLER dev->poll_controller = cas_netpoll; diff --git a/drivers/net/cassini.h b/drivers/net/cassini.h index a970804..2f93f83 100644 --- a/drivers/net/cassini.h +++ b/drivers/net/cassini.h @@ -4280,6 +4280,8 @@ struct cas { int rx_cur[N_RX_COMP_RINGS], rx_new[N_RX_COMP_RINGS]; int rx_last[N_RX_DESC_RINGS]; + struct napi_struct napi; + /* Set when chip is actually in operational state * (ie. not power managed) */ int hw_running; diff --git a/drivers/net/chelsio/common.h b/drivers/net/chelsio/common.h index 8ba702c..b5de445 100644 --- a/drivers/net/chelsio/common.h +++ b/drivers/net/chelsio/common.h @@ -278,6 +278,7 @@ struct adapter { struct peespi *espi; struct petp *tp; + struct napi_struct napi; struct port_info port[MAX_NPORTS]; struct delayed_work stats_update_task; struct timer_list stats_update_timer; diff --git a/drivers/net/chelsio/cxgb2.c b/drivers/net/chelsio/cxgb2.c index 231ce43..593736c 100644 --- a/drivers/net/chelsio/cxgb2.c +++ b/drivers/net/chelsio/cxgb2.c @@ -255,8 +255,11 @@ static int cxgb_open(struct net_device *dev) struct adapter *adapter = dev->priv; int other_ports = adapter->open_device_map & PORT_MASK; - if (!adapter->open_device_map && (err = cxgb_up(adapter)) < 0) + napi_enable(&adapter->napi); + if (!adapter->open_device_map && (err = cxgb_up(adapter)) < 0) { + napi_disable(&adapter->napi); return err; + } __set_bit(dev->if_port, &adapter->open_device_map); link_start(&adapter->port[dev->if_port]); @@ -274,6 +277,7 @@ static int cxgb_close(struct net_device *dev) struct cmac *mac = p->mac; netif_stop_queue(dev); + napi_disable(&adapter->napi); mac->ops->disable(mac, MAC_DIRECTION_TX | MAC_DIRECTION_RX); netif_carrier_off(dev); @@ -1113,8 +1117,7 @@ static int __devinit init_one(struct pci_dev *pdev, netdev->poll_controller = t1_netpoll; #endif #ifdef CONFIG_CHELSIO_T1_NAPI - netdev->weight = 64; - netdev->poll = t1_poll; + netif_napi_add(netdev, &adapter->napi, t1_poll, 64); #endif SET_ETHTOOL_OPS(netdev, &t1_ethtool_ops); diff --git a/drivers/net/chelsio/sge.c b/drivers/net/chelsio/sge.c index e4f874a..ffa7e64 100644 --- a/drivers/net/chelsio/sge.c +++ b/drivers/net/chelsio/sge.c @@ -1620,23 +1620,20 @@ static int process_pure_responses(struct adapter *adapter) * or protection from interrupts as data interrupts are off at this point and * other adapter interrupts do not interfere. */ -int t1_poll(struct net_device *dev, int *budget) +int t1_poll(struct napi_struct *napi, int budget) { - struct adapter *adapter = dev->priv; + struct adapter *adapter = container_of(napi, struct adapter, napi); + struct net_device *dev = adapter->port[0].dev; int work_done; - work_done = process_responses(adapter, min(*budget, dev->quota)); - *budget -= work_done; - dev->quota -= work_done; - - if (unlikely(responses_pending(adapter))) - return 1; - - netif_rx_complete(dev); - writel(adapter->sge->respQ.cidx, adapter->regs + A_SG_SLEEPING); - - return 0; + work_done = process_responses(adapter, budget); + if (likely(!responses_pending(adapter))) { + netif_rx_complete(dev, napi); + writel(adapter->sge->respQ.cidx, + adapter->regs + A_SG_SLEEPING); + } + return work_done; } /* @@ -1653,13 +1650,13 @@ irqreturn_t t1_interrupt(int irq, void *data) writel(F_PL_INTR_SGE_DATA, adapter->regs + A_PL_CAUSE); - if (__netif_rx_schedule_prep(dev)) { + if (napi_schedule_prep(&adapter->napi)) { if (process_pure_responses(adapter)) - __netif_rx_schedule(dev); + __netif_rx_schedule(dev, &adapter->napi); else { /* no data, no NAPI needed */ writel(sge->respQ.cidx, adapter->regs + A_SG_SLEEPING); - netif_poll_enable(dev); /* undo schedule_prep */ + napi_enable(&adapter->napi); /* undo schedule_prep */ } } return IRQ_HANDLED; diff --git a/drivers/net/chelsio/sge.h b/drivers/net/chelsio/sge.h index d132a0ef..713d9c5 100644 --- a/drivers/net/chelsio/sge.h +++ b/drivers/net/chelsio/sge.h @@ -77,7 +77,7 @@ int t1_sge_configure(struct sge *, struct sge_params *); int t1_sge_set_coalesce_params(struct sge *, struct sge_params *); void t1_sge_destroy(struct sge *); irqreturn_t t1_interrupt(int irq, void *cookie); -int t1_poll(struct net_device *, int *); +int t1_poll(struct napi_struct *, int); int t1_start_xmit(struct sk_buff *skb, struct net_device *dev); void t1_set_vlan_accel(struct adapter *adapter, int on_off); diff --git a/drivers/net/cxgb3/adapter.h b/drivers/net/cxgb3/adapter.h index 20e887d..0442617 100644 --- a/drivers/net/cxgb3/adapter.h +++ b/drivers/net/cxgb3/adapter.h @@ -49,11 +49,13 @@ typedef irqreturn_t(*intr_handler_t) (int, void *); struct vlan_group; - struct adapter; +struct sge_qset; + struct port_info { struct adapter *adapter; struct vlan_group *vlan_grp; + struct sge_qset *qs; const struct port_type_info *port_type; u8 port_id; u8 rx_csum_offload; @@ -173,10 +175,12 @@ enum { /* per port SGE statistics */ }; struct sge_qset { /* an SGE queue set */ + struct adapter *adap; + struct napi_struct napi; struct sge_rspq rspq; struct sge_fl fl[SGE_RXQ_PER_SET]; struct sge_txq txq[SGE_TXQ_PER_SET]; - struct net_device *netdev; /* associated net device */ + struct net_device *netdev; unsigned long txq_stopped; /* which Tx queues are stopped */ struct timer_list tx_reclaim_timer; /* reclaims TX buffers */ unsigned long port_stats[SGE_PSTAT_MAX]; @@ -221,12 +225,6 @@ struct adapter { struct delayed_work adap_check_task; struct work_struct ext_intr_handler_task; - /* - * Dummy netdevices are needed when using multiple receive queues with - * NAPI as each netdevice can service only one queue. - */ - struct net_device *dummy_netdev[SGE_QSETS - 1]; - struct dentry *debugfs_root; struct mutex mdio_lock; @@ -253,12 +251,6 @@ static inline struct port_info *adap2pinfo(struct adapter *adap, int idx) return netdev_priv(adap->port[idx]); } -/* - * We use the spare atalk_ptr to map a net device to its SGE queue set. - * This is a macro so it can be used as l-value. - */ -#define dev2qset(netdev) ((netdev)->atalk_ptr) - #define OFFLOAD_DEVMAP_BIT 15 #define tdev2adap(d) container_of(d, struct adapter, tdev) @@ -284,7 +276,7 @@ int t3_mgmt_tx(struct adapter *adap, struct sk_buff *skb); void t3_update_qset_coalesce(struct sge_qset *qs, const struct qset_params *p); int t3_sge_alloc_qset(struct adapter *adapter, unsigned int id, int nports, int irq_vec_idx, const struct qset_params *p, - int ntxq, struct net_device *netdev); + int ntxq, struct net_device *dev); int t3_get_desc(const struct sge_qset *qs, unsigned int qnum, unsigned int idx, unsigned char *data); irqreturn_t t3_sge_intr_msix(int irq, void *cookie); diff --git a/drivers/net/cxgb3/cxgb3_main.c b/drivers/net/cxgb3/cxgb3_main.c index 5ab319c..5db7d4e 100644 --- a/drivers/net/cxgb3/cxgb3_main.c +++ b/drivers/net/cxgb3/cxgb3_main.c @@ -339,49 +339,17 @@ static void setup_rss(struct adapter *adap) V_RRCPLCPUSIZE(6), cpus, rspq_map); } -/* - * If we have multiple receive queues per port serviced by NAPI we need one - * netdevice per queue as NAPI operates on netdevices. We already have one - * netdevice, namely the one associated with the interface, so we use dummy - * ones for any additional queues. Note that these netdevices exist purely - * so that NAPI has something to work with, they do not represent network - * ports and are not registered. - */ -static int init_dummy_netdevs(struct adapter *adap) +static void init_napi(struct adapter *adap) { - int i, j, dummy_idx = 0; - struct net_device *nd; - - for_each_port(adap, i) { - struct net_device *dev = adap->port[i]; - const struct port_info *pi = netdev_priv(dev); - - for (j = 0; j < pi->nqsets - 1; j++) { - if (!adap->dummy_netdev[dummy_idx]) { - struct port_info *p; - - nd = alloc_netdev(sizeof(*p), "", ether_setup); - if (!nd) - goto free_all; + int i; - p = netdev_priv(nd); - p->adapter = adap; - nd->weight = 64; - set_bit(__LINK_STATE_START, &nd->state); - adap->dummy_netdev[dummy_idx] = nd; - } - strcpy(adap->dummy_netdev[dummy_idx]->name, dev->name); - dummy_idx++; - } - } - return 0; + for (i = 0; i < SGE_QSETS; i++) { + struct sge_qset *qs = &adap->sge.qs[i]; -free_all: - while (--dummy_idx >= 0) { - free_netdev(adap->dummy_netdev[dummy_idx]); - adap->dummy_netdev[dummy_idx] = NULL; + if (qs->adap) + netif_napi_add(qs->netdev, &qs->napi, qs->napi.poll, + 64); } - return -ENOMEM; } /* @@ -392,20 +360,18 @@ free_all: static void quiesce_rx(struct adapter *adap) { int i; - struct net_device *dev; - for_each_port(adap, i) { - dev = adap->port[i]; - while (test_bit(__LINK_STATE_RX_SCHED, &dev->state)) - msleep(1); - } + for (i = 0; i < SGE_QSETS; i++) + if (adap->sge.qs[i].adap) + napi_disable(&adap->sge.qs[i].napi); +} - for (i = 0; i < ARRAY_SIZE(adap->dummy_netdev); i++) { - dev = adap->dummy_netdev[i]; - if (dev) - while (test_bit(__LINK_STATE_RX_SCHED, &dev->state)) - msleep(1); - } +static void enable_all_napi(struct adapter *adap) +{ + int i; + for (i = 0; i < SGE_QSETS; i++) + if (adap->sge.qs[i].adap) + napi_enable(&adap->sge.qs[i].napi); } /** @@ -418,7 +384,7 @@ static void quiesce_rx(struct adapter *adap) */ static int setup_sge_qsets(struct adapter *adap) { - int i, j, err, irq_idx = 0, qset_idx = 0, dummy_dev_idx = 0; + int i, j, err, irq_idx = 0, qset_idx = 0; unsigned int ntxq = SGE_TXQ_PER_SET; if (adap->params.rev > 0 && !(adap->flags & USING_MSI)) @@ -426,15 +392,14 @@ static int setup_sge_qsets(struct adapter *adap) for_each_port(adap, i) { struct net_device *dev = adap->port[i]; - const struct port_info *pi = netdev_priv(dev); + struct port_info *pi = netdev_priv(dev); + pi->qs = &adap->sge.qs[pi->first_qset]; for (j = 0; j < pi->nqsets; ++j, ++qset_idx) { err = t3_sge_alloc_qset(adap, qset_idx, 1, (adap->flags & USING_MSIX) ? qset_idx + 1 : irq_idx, - &adap->params.sge.qset[qset_idx], ntxq, - j == 0 ? dev : - adap-> dummy_netdev[dummy_dev_idx++]); + &adap->params.sge.qset[qset_idx], ntxq, dev); if (err) { t3_free_sge_resources(adap); return err; @@ -845,21 +810,18 @@ static int cxgb_up(struct adapter *adap) goto out; } - err = init_dummy_netdevs(adap); - if (err) - goto out; - err = t3_init_hw(adap, 0); if (err) goto out; t3_write_reg(adap, A_ULPRX_TDDP_PSZ, V_HPZ0(PAGE_SHIFT - 12)); - + err = setup_sge_qsets(adap); if (err) goto out; setup_rss(adap); + init_napi(adap); adap->flags |= FULL_INIT_DONE; } @@ -886,6 +848,7 @@ static int cxgb_up(struct adapter *adap) adap->name, adap))) goto irq_err; + enable_all_napi(adap); t3_sge_start(adap); t3_intr_enable(adap); @@ -1012,8 +975,10 @@ static int cxgb_open(struct net_device *dev) int other_ports = adapter->open_device_map & PORT_MASK; int err; - if (!adapter->open_device_map && (err = cxgb_up(adapter)) < 0) + if (!adapter->open_device_map && (err = cxgb_up(adapter)) < 0) { + quiesce_rx(adapter); return err; + } set_bit(pi->port_id, &adapter->open_device_map); if (is_offload(adapter) && !ofld_disable) { @@ -2524,7 +2489,6 @@ static int __devinit init_one(struct pci_dev *pdev, #ifdef CONFIG_NET_POLL_CONTROLLER netdev->poll_controller = cxgb_netpoll; #endif - netdev->weight = 64; SET_ETHTOOL_OPS(netdev, &cxgb_ethtool_ops); } @@ -2625,12 +2589,6 @@ static void __devexit remove_one(struct pci_dev *pdev) t3_free_sge_resources(adapter); cxgb_disable_msi(adapter); - for (i = 0; i < ARRAY_SIZE(adapter->dummy_netdev); i++) - if (adapter->dummy_netdev[i]) { - free_netdev(adapter->dummy_netdev[i]); - adapter->dummy_netdev[i] = NULL; - } - for_each_port(adapter, i) if (adapter->port[i]) free_netdev(adapter->port[i]); diff --git a/drivers/net/cxgb3/sge.c b/drivers/net/cxgb3/sge.c index 58a5f60..069c1ac 100644 --- a/drivers/net/cxgb3/sge.c +++ b/drivers/net/cxgb3/sge.c @@ -591,9 +591,6 @@ void t3_free_qset(struct adapter *adapter, struct sge_qset *q) q->rspq.desc, q->rspq.phys_addr); } - if (q->netdev) - q->netdev->atalk_ptr = NULL; - memset(q, 0, sizeof(*q)); } @@ -1074,7 +1071,7 @@ int t3_eth_xmit(struct sk_buff *skb, struct net_device *dev) unsigned int ndesc, pidx, credits, gen, compl; const struct port_info *pi = netdev_priv(dev); struct adapter *adap = pi->adapter; - struct sge_qset *qs = dev2qset(dev); + struct sge_qset *qs = pi->qs; struct sge_txq *q = &qs->txq[TXQ_ETH]; /* @@ -1326,13 +1323,12 @@ static void restart_ctrlq(unsigned long data) struct sk_buff *skb; struct sge_qset *qs = (struct sge_qset *)data; struct sge_txq *q = &qs->txq[TXQ_CTRL]; - const struct port_info *pi = netdev_priv(qs->netdev); - struct adapter *adap = pi->adapter; spin_lock(&q->lock); again:reclaim_completed_tx_imm(q); - while (q->in_use < q->size && (skb = __skb_dequeue(&q->sendq)) != NULL) { + while (q->in_use < q->size && + (skb = __skb_dequeue(&q->sendq)) != NULL) { write_imm(&q->desc[q->pidx], skb, skb->len, q->gen); @@ -1354,7 +1350,7 @@ static void restart_ctrlq(unsigned long data) } spin_unlock(&q->lock); - t3_write_reg(adap, A_SG_KDOORBELL, + t3_write_reg(qs->adap, A_SG_KDOORBELL, F_SELEGRCNTX | V_EGRCNTX(q->cntxt_id)); } @@ -1638,8 +1634,7 @@ static inline void offload_enqueue(struct sge_rspq *q, struct sk_buff *skb) else { struct sge_qset *qs = rspq_to_qset(q); - if (__netif_rx_schedule_prep(qs->netdev)) - __netif_rx_schedule(qs->netdev); + napi_schedule(&qs->napi); q->rx_head = skb; } q->rx_tail = skb; @@ -1675,34 +1670,30 @@ static inline void deliver_partial_bundle(struct t3cdev *tdev, * receive handler. Batches need to be of modest size as we do prefetches * on the packets in each. */ -static int ofld_poll(struct net_device *dev, int *budget) +static int ofld_poll(struct napi_struct *napi, int budget) { - const struct port_info *pi = netdev_priv(dev); - struct adapter *adapter = pi->adapter; - struct sge_qset *qs = dev2qset(dev); + struct sge_qset *qs = container_of(napi, struct sge_qset, napi); struct sge_rspq *q = &qs->rspq; - int work_done, limit = min(*budget, dev->quota), avail = limit; + struct adapter *adapter = qs->adap; + int work_done = 0; - while (avail) { + while (work_done < budget) { struct sk_buff *head, *tail, *skbs[RX_BUNDLE_SIZE]; int ngathered; spin_lock_irq(&q->lock); head = q->rx_head; if (!head) { - work_done = limit - avail; - *budget -= work_done; - dev->quota -= work_done; - __netif_rx_complete(dev); + napi_complete(napi); spin_unlock_irq(&q->lock); - return 0; + return work_done; } tail = q->rx_tail; q->rx_head = q->rx_tail = NULL; spin_unlock_irq(&q->lock); - for (ngathered = 0; avail && head; avail--) { + for (ngathered = 0; work_done < budget && head; work_done++) { prefetch(head->data); skbs[ngathered] = head; head = head->next; @@ -1724,10 +1715,8 @@ static int ofld_poll(struct net_device *dev, int *budget) } deliver_partial_bundle(&adapter->tdev, q, skbs, ngathered); } - work_done = limit - avail; - *budget -= work_done; - dev->quota -= work_done; - return 1; + + return work_done; } /** @@ -2071,50 +2060,47 @@ static inline int is_pure_response(const struct rsp_desc *r) /** * napi_rx_handler - the NAPI handler for Rx processing - * @dev: the net device + * @napi: the napi instance * @budget: how many packets we can process in this round * * Handler for new data events when using NAPI. */ -static int napi_rx_handler(struct net_device *dev, int *budget) +static int napi_rx_handler(struct napi_struct *napi, int budget) { - const struct port_info *pi = netdev_priv(dev); - struct adapter *adap = pi->adapter; - struct sge_qset *qs = dev2qset(dev); - int effective_budget = min(*budget, dev->quota); - - int work_done = process_responses(adap, qs, effective_budget); - *budget -= work_done; - dev->quota -= work_done; + struct sge_qset *qs = container_of(napi, struct sge_qset, napi); + struct adapter *adap = qs->adap; + int work_done = process_responses(adap, qs, budget); - if (work_done >= effective_budget) - return 1; - - netif_rx_complete(dev); + if (likely(work_done < budget)) { + napi_complete(napi); - /* - * Because we don't atomically flush the following write it is - * possible that in very rare cases it can reach the device in a way - * that races with a new response being written plus an error interrupt - * causing the NAPI interrupt handler below to return unhandled status - * to the OS. To protect against this would require flushing the write - * and doing both the write and the flush with interrupts off. Way too - * expensive and unjustifiable given the rarity of the race. - * - * The race cannot happen at all with MSI-X. - */ - t3_write_reg(adap, A_SG_GTS, V_RSPQ(qs->rspq.cntxt_id) | - V_NEWTIMER(qs->rspq.next_holdoff) | - V_NEWINDEX(qs->rspq.cidx)); - return 0; + /* + * Because we don't atomically flush the following + * write it is possible that in very rare cases it can + * reach the device in a way that races with a new + * response being written plus an error interrupt + * causing the NAPI interrupt handler below to return + * unhandled status to the OS. To protect against + * this would require flushing the write and doing + * both the write and the flush with interrupts off. + * Way too expensive and unjustifiable given the + * rarity of the race. + * + * The race cannot happen at all with MSI-X. + */ + t3_write_reg(adap, A_SG_GTS, V_RSPQ(qs->rspq.cntxt_id) | + V_NEWTIMER(qs->rspq.next_holdoff) | + V_NEWINDEX(qs->rspq.cidx)); + } + return work_done; } /* * Returns true if the device is already scheduled for polling. */ -static inline int napi_is_scheduled(struct net_device *dev) +static inline int napi_is_scheduled(struct napi_struct *napi) { - return test_bit(__LINK_STATE_RX_SCHED, &dev->state); + return test_bit(NAPI_STATE_SCHED, &napi->state); } /** @@ -2197,8 +2183,7 @@ static inline int handle_responses(struct adapter *adap, struct sge_rspq *q) V_NEWTIMER(q->holdoff_tmr) | V_NEWINDEX(q->cidx)); return 0; } - if (likely(__netif_rx_schedule_prep(qs->netdev))) - __netif_rx_schedule(qs->netdev); + napi_schedule(&qs->napi); return 1; } @@ -2209,8 +2194,7 @@ static inline int handle_responses(struct adapter *adap, struct sge_rspq *q) irqreturn_t t3_sge_intr_msix(int irq, void *cookie) { struct sge_qset *qs = cookie; - const struct port_info *pi = netdev_priv(qs->netdev); - struct adapter *adap = pi->adapter; + struct adapter *adap = qs->adap; struct sge_rspq *q = &qs->rspq; spin_lock(&q->lock); @@ -2229,13 +2213,11 @@ irqreturn_t t3_sge_intr_msix(int irq, void *cookie) irqreturn_t t3_sge_intr_msix_napi(int irq, void *cookie) { struct sge_qset *qs = cookie; - const struct port_info *pi = netdev_priv(qs->netdev); - struct adapter *adap = pi->adapter; struct sge_rspq *q = &qs->rspq; spin_lock(&q->lock); - if (handle_responses(adap, q) < 0) + if (handle_responses(qs->adap, q) < 0) q->unhandled_irqs++; spin_unlock(&q->lock); return IRQ_HANDLED; @@ -2278,11 +2260,13 @@ static irqreturn_t t3_intr_msi(int irq, void *cookie) return IRQ_HANDLED; } -static int rspq_check_napi(struct net_device *dev, struct sge_rspq *q) +static int rspq_check_napi(struct sge_qset *qs) { - if (!napi_is_scheduled(dev) && is_new_response(&q->desc[q->cidx], q)) { - if (likely(__netif_rx_schedule_prep(dev))) - __netif_rx_schedule(dev); + struct sge_rspq *q = &qs->rspq; + + if (!napi_is_scheduled(&qs->napi) && + is_new_response(&q->desc[q->cidx], q)) { + napi_schedule(&qs->napi); return 1; } return 0; @@ -2303,10 +2287,9 @@ irqreturn_t t3_intr_msi_napi(int irq, void *cookie) spin_lock(&q->lock); - new_packets = rspq_check_napi(adap->sge.qs[0].netdev, q); + new_packets = rspq_check_napi(&adap->sge.qs[0]); if (adap->params.nports == 2) - new_packets += rspq_check_napi(adap->sge.qs[1].netdev, - &adap->sge.qs[1].rspq); + new_packets += rspq_check_napi(&adap->sge.qs[1]); if (!new_packets && t3_slow_intr_handler(adap) == 0) q->unhandled_irqs++; @@ -2409,9 +2392,9 @@ static irqreturn_t t3b_intr(int irq, void *cookie) static irqreturn_t t3b_intr_napi(int irq, void *cookie) { u32 map; - struct net_device *dev; struct adapter *adap = cookie; - struct sge_rspq *q0 = &adap->sge.qs[0].rspq; + struct sge_qset *qs0 = &adap->sge.qs[0]; + struct sge_rspq *q0 = &qs0->rspq; t3_write_reg(adap, A_PL_CLI, 0); map = t3_read_reg(adap, A_SG_DATA_INTR); @@ -2424,18 +2407,11 @@ static irqreturn_t t3b_intr_napi(int irq, void *cookie) if (unlikely(map & F_ERRINTR)) t3_slow_intr_handler(adap); - if (likely(map & 1)) { - dev = adap->sge.qs[0].netdev; - - if (likely(__netif_rx_schedule_prep(dev))) - __netif_rx_schedule(dev); - } - if (map & 2) { - dev = adap->sge.qs[1].netdev; + if (likely(map & 1)) + napi_schedule(&qs0->napi); - if (likely(__netif_rx_schedule_prep(dev))) - __netif_rx_schedule(dev); - } + if (map & 2) + napi_schedule(&adap->sge.qs[1].napi); spin_unlock(&q0->lock); return IRQ_HANDLED; @@ -2514,8 +2490,7 @@ static void sge_timer_cb(unsigned long data) { spinlock_t *lock; struct sge_qset *qs = (struct sge_qset *)data; - const struct port_info *pi = netdev_priv(qs->netdev); - struct adapter *adap = pi->adapter; + struct adapter *adap = qs->adap; if (spin_trylock(&qs->txq[TXQ_ETH].lock)) { reclaim_completed_tx(adap, &qs->txq[TXQ_ETH]); @@ -2526,9 +2501,9 @@ static void sge_timer_cb(unsigned long data) spin_unlock(&qs->txq[TXQ_OFLD].lock); } lock = (adap->flags & USING_MSIX) ? &qs->rspq.lock : - &adap->sge.qs[0].rspq.lock; + &adap->sge.qs[0].rspq.lock; if (spin_trylock_irq(lock)) { - if (!napi_is_scheduled(qs->netdev)) { + if (!napi_is_scheduled(&qs->napi)) { u32 status = t3_read_reg(adap, A_SG_RSPQ_FL_STATUS); if (qs->fl[0].credits < qs->fl[0].size) @@ -2562,12 +2537,9 @@ static void sge_timer_cb(unsigned long data) */ void t3_update_qset_coalesce(struct sge_qset *qs, const struct qset_params *p) { - if (!qs->netdev) - return; - qs->rspq.holdoff_tmr = max(p->coalesce_usecs * 10, 1U);/* can't be 0 */ qs->rspq.polling = p->polling; - qs->netdev->poll = p->polling ? napi_rx_handler : ofld_poll; + qs->napi.poll = p->polling ? napi_rx_handler : ofld_poll; } /** @@ -2587,7 +2559,7 @@ void t3_update_qset_coalesce(struct sge_qset *qs, const struct qset_params *p) */ int t3_sge_alloc_qset(struct adapter *adapter, unsigned int id, int nports, int irq_vec_idx, const struct qset_params *p, - int ntxq, struct net_device *netdev) + int ntxq, struct net_device *dev) { int i, ret = -ENOMEM; struct sge_qset *q = &adapter->sge.qs[id]; @@ -2708,16 +2680,10 @@ int t3_sge_alloc_qset(struct adapter *adapter, unsigned int id, int nports, } spin_unlock(&adapter->sge.reg_lock); - q->netdev = netdev; - t3_update_qset_coalesce(q, p); - /* - * We use atalk_ptr as a backpointer to a qset. In case a device is - * associated with multiple queue sets only the first one sets - * atalk_ptr. - */ - if (netdev->atalk_ptr == NULL) - netdev->atalk_ptr = q; + q->adap = adapter; + q->netdev = dev; + t3_update_qset_coalesce(q, p); refill_fl(adapter, &q->fl[0], q->fl[0].size, GFP_KERNEL); refill_fl(adapter, &q->fl[1], q->fl[1].size, GFP_KERNEL); diff --git a/drivers/net/e100.c b/drivers/net/e100.c index 280313b..e25f5ec 100644 --- a/drivers/net/e100.c +++ b/drivers/net/e100.c @@ -539,6 +539,7 @@ struct nic { struct csr __iomem *csr; enum scb_cmd_lo cuc_cmd; unsigned int cbs_avail; + struct napi_struct napi; struct cb *cbs; struct cb *cb_to_use; struct cb *cb_to_send; @@ -1974,35 +1975,31 @@ static irqreturn_t e100_intr(int irq, void *dev_id) if(stat_ack & stat_ack_rnr) nic->ru_running = RU_SUSPENDED; - if(likely(netif_rx_schedule_prep(netdev))) { + if(likely(netif_rx_schedule_prep(netdev, &nic->napi))) { e100_disable_irq(nic); - __netif_rx_schedule(netdev); + __netif_rx_schedule(netdev, &nic->napi); } return IRQ_HANDLED; } -static int e100_poll(struct net_device *netdev, int *budget) +static int e100_poll(struct napi_struct *napi, int budget) { - struct nic *nic = netdev_priv(netdev); - unsigned int work_to_do = min(netdev->quota, *budget); - unsigned int work_done = 0; + struct nic *nic = container_of(napi, struct nic, napi); + struct net_device *netdev = nic->netdev; + int work_done = 0; int tx_cleaned; - e100_rx_clean(nic, &work_done, work_to_do); + e100_rx_clean(nic, &work_done, budget); tx_cleaned = e100_tx_clean(nic); /* If no Rx and Tx cleanup work was done, exit polling mode. */ if((!tx_cleaned && (work_done == 0)) || !netif_running(netdev)) { - netif_rx_complete(netdev); + netif_rx_complete(netdev, napi); e100_enable_irq(nic); - return 0; } - *budget -= work_done; - netdev->quota -= work_done; - - return 1; + return work_done; } #ifdef CONFIG_NET_POLL_CONTROLLER @@ -2071,7 +2068,7 @@ static int e100_up(struct nic *nic) nic->netdev->name, nic->netdev))) goto err_no_irq; netif_wake_queue(nic->netdev); - netif_poll_enable(nic->netdev); + napi_enable(&nic->napi); /* enable ints _after_ enabling poll, preventing a race between * disable ints+schedule */ e100_enable_irq(nic); @@ -2089,7 +2086,7 @@ err_rx_clean_list: static void e100_down(struct nic *nic) { /* wait here for poll to complete */ - netif_poll_disable(nic->netdev); + napi_disable(&nic->napi); netif_stop_queue(nic->netdev); e100_hw_reset(nic); free_irq(nic->pdev->irq, nic->netdev); @@ -2572,14 +2569,13 @@ static int __devinit e100_probe(struct pci_dev *pdev, SET_ETHTOOL_OPS(netdev, &e100_ethtool_ops); netdev->tx_timeout = e100_tx_timeout; netdev->watchdog_timeo = E100_WATCHDOG_PERIOD; - netdev->poll = e100_poll; - netdev->weight = E100_NAPI_WEIGHT; #ifdef CONFIG_NET_POLL_CONTROLLER netdev->poll_controller = e100_netpoll; #endif strncpy(netdev->name, pci_name(pdev), sizeof(netdev->name) - 1); nic = netdev_priv(netdev); + netif_napi_add(netdev, &nic->napi, e100_poll, E100_NAPI_WEIGHT); nic->netdev = netdev; nic->pdev = pdev; nic->msg_enable = (1 << debug) - 1; @@ -2733,7 +2729,7 @@ static int e100_suspend(struct pci_dev *pdev, pm_message_t state) struct nic *nic = netdev_priv(netdev); if (netif_running(netdev)) - netif_poll_disable(nic->netdev); + napi_disable(&nic->napi); del_timer_sync(&nic->watchdog); netif_carrier_off(nic->netdev); netif_device_detach(netdev); @@ -2779,7 +2775,7 @@ static void e100_shutdown(struct pci_dev *pdev) struct nic *nic = netdev_priv(netdev); if (netif_running(netdev)) - netif_poll_disable(nic->netdev); + napi_disable(&nic->napi); del_timer_sync(&nic->watchdog); netif_carrier_off(nic->netdev); @@ -2804,12 +2800,13 @@ static void e100_shutdown(struct pci_dev *pdev) static pci_ers_result_t e100_io_error_detected(struct pci_dev *pdev, pci_channel_state_t state) { struct net_device *netdev = pci_get_drvdata(pdev); + struct nic *nic = netdev_priv(netdev); /* Similar to calling e100_down(), but avoids adpater I/O. */ netdev->stop(netdev); /* Detach; put netif into state similar to hotplug unplug. */ - netif_poll_enable(netdev); + napi_enable(&nic->napi); netif_device_detach(netdev); pci_disable_device(pdev); diff --git a/drivers/net/e1000/e1000.h b/drivers/net/e1000/e1000.h index 16a6edf..781ed99 100644 --- a/drivers/net/e1000/e1000.h +++ b/drivers/net/e1000/e1000.h @@ -300,6 +300,7 @@ struct e1000_adapter { int cleaned_count); struct e1000_rx_ring *rx_ring; /* One per active queue */ #ifdef CONFIG_E1000_NAPI + struct napi_struct napi; struct net_device *polling_netdev; /* One per active queue */ #endif int num_tx_queues; diff --git a/drivers/net/e1000/e1000_main.c b/drivers/net/e1000/e1000_main.c index e7c8951..723568d 100644 --- a/drivers/net/e1000/e1000_main.c +++ b/drivers/net/e1000/e1000_main.c @@ -166,7 +166,7 @@ static irqreturn_t e1000_intr_msi(int irq, void *data); static boolean_t e1000_clean_tx_irq(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring); #ifdef CONFIG_E1000_NAPI -static int e1000_clean(struct net_device *poll_dev, int *budget); +static int e1000_clean(struct napi_struct *napi, int budget); static boolean_t e1000_clean_rx_irq(struct e1000_adapter *adapter, struct e1000_rx_ring *rx_ring, int *work_done, int work_to_do); @@ -545,7 +545,7 @@ int e1000_up(struct e1000_adapter *adapter) clear_bit(__E1000_DOWN, &adapter->flags); #ifdef CONFIG_E1000_NAPI - netif_poll_enable(adapter->netdev); + napi_enable(&adapter->napi); #endif e1000_irq_enable(adapter); @@ -634,7 +634,7 @@ e1000_down(struct e1000_adapter *adapter) set_bit(__E1000_DOWN, &adapter->flags); #ifdef CONFIG_E1000_NAPI - netif_poll_disable(netdev); + napi_disable(&adapter->napi); #endif e1000_irq_disable(adapter); @@ -936,8 +936,7 @@ e1000_probe(struct pci_dev *pdev, netdev->tx_timeout = &e1000_tx_timeout; netdev->watchdog_timeo = 5 * HZ; #ifdef CONFIG_E1000_NAPI - netdev->poll = &e1000_clean; - netdev->weight = 64; + netif_napi_add(netdev, &adapter->napi, e1000_clean, 64); #endif netdev->vlan_rx_register = e1000_vlan_rx_register; netdev->vlan_rx_add_vid = e1000_vlan_rx_add_vid; @@ -1151,9 +1150,6 @@ e1000_probe(struct pci_dev *pdev, /* tell the stack to leave us alone until e1000_open() is called */ netif_carrier_off(netdev); netif_stop_queue(netdev); -#ifdef CONFIG_E1000_NAPI - netif_poll_disable(netdev); -#endif strcpy(netdev->name, "eth%d"); if ((err = register_netdev(netdev))) @@ -1222,12 +1218,13 @@ e1000_remove(struct pci_dev *pdev) * would have already happened in close and is redundant. */ e1000_release_hw_control(adapter); - unregister_netdev(netdev); #ifdef CONFIG_E1000_NAPI for (i = 0; i < adapter->num_rx_queues; i++) dev_put(&adapter->polling_netdev[i]); #endif + unregister_netdev(netdev); + if (!e1000_check_phy_reset_block(&adapter->hw)) e1000_phy_hw_reset(&adapter->hw); @@ -1325,8 +1322,6 @@ e1000_sw_init(struct e1000_adapter *adapter) #ifdef CONFIG_E1000_NAPI for (i = 0; i < adapter->num_rx_queues; i++) { adapter->polling_netdev[i].priv = adapter; - adapter->polling_netdev[i].poll = &e1000_clean; - adapter->polling_netdev[i].weight = 64; dev_hold(&adapter->polling_netdev[i]); set_bit(__LINK_STATE_START, &adapter->polling_netdev[i].state); } @@ -1443,7 +1438,7 @@ e1000_open(struct net_device *netdev) clear_bit(__E1000_DOWN, &adapter->flags); #ifdef CONFIG_E1000_NAPI - netif_poll_enable(netdev); + napi_enable(&adapter->napi); #endif e1000_irq_enable(adapter); @@ -3786,12 +3781,12 @@ e1000_intr_msi(int irq, void *data) } #ifdef CONFIG_E1000_NAPI - if (likely(netif_rx_schedule_prep(netdev))) { + if (likely(netif_rx_schedule_prep(netdev, &adapter->napi))) { adapter->total_tx_bytes = 0; adapter->total_tx_packets = 0; adapter->total_rx_bytes = 0; adapter->total_rx_packets = 0; - __netif_rx_schedule(netdev); + __netif_rx_schedule(netdev, &adapter->napi); } else e1000_irq_enable(adapter); #else @@ -3871,12 +3866,12 @@ e1000_intr(int irq, void *data) E1000_WRITE_REG(hw, IMC, ~0); E1000_WRITE_FLUSH(hw); } - if (likely(netif_rx_schedule_prep(netdev))) { + if (likely(netif_rx_schedule_prep(netdev, &adapter->napi))) { adapter->total_tx_bytes = 0; adapter->total_tx_packets = 0; adapter->total_rx_bytes = 0; adapter->total_rx_packets = 0; - __netif_rx_schedule(netdev); + __netif_rx_schedule(netdev, &adapter->napi); } else /* this really should not happen! if it does it is basically a * bug, but not a hard error, so enable ints and continue */ @@ -3924,10 +3919,10 @@ e1000_intr(int irq, void *data) **/ static int -e1000_clean(struct net_device *poll_dev, int *budget) +e1000_clean(struct napi_struct *napi, int budget) { - struct e1000_adapter *adapter; - int work_to_do = min(*budget, poll_dev->quota); + struct e1000_adapter *adapter = container_of(napi, struct e1000_adapter, napi); + struct net_device *poll_dev = adapter->netdev; int tx_cleaned = 0, work_done = 0; /* Must NOT use netdev_priv macro here. */ @@ -3948,23 +3943,19 @@ e1000_clean(struct net_device *poll_dev, int *budget) } adapter->clean_rx(adapter, &adapter->rx_ring[0], - &work_done, work_to_do); - - *budget -= work_done; - poll_dev->quota -= work_done; + &work_done, budget); /* If no Tx and not enough Rx work done, exit the polling mode */ - if ((!tx_cleaned && (work_done == 0)) || + if ((!tx_cleaned && (work_done < budget)) || !netif_running(poll_dev)) { quit_polling: if (likely(adapter->itr_setting & 3)) e1000_set_itr(adapter); - netif_rx_complete(poll_dev); + netif_rx_complete(poll_dev, napi); e1000_irq_enable(adapter); - return 0; } - return 1; + return work_done; } #endif diff --git a/drivers/net/ehea/ehea.h b/drivers/net/ehea/ehea.h index 8d58be5..a154681 100644 --- a/drivers/net/ehea/ehea.h +++ b/drivers/net/ehea/ehea.h @@ -351,6 +351,7 @@ struct ehea_q_skb_arr { * Port resources */ struct ehea_port_res { + struct napi_struct napi; struct port_stats p_stats; struct ehea_mr send_mr; /* send memory region */ struct ehea_mr recv_mr; /* receive memory region */ @@ -362,7 +363,6 @@ struct ehea_port_res { struct ehea_cq *send_cq; struct ehea_cq *recv_cq; struct ehea_eq *eq; - struct net_device *d_netdev; struct ehea_q_skb_arr rq1_skba; struct ehea_q_skb_arr rq2_skba; struct ehea_q_skb_arr rq3_skba; diff --git a/drivers/net/ehea/ehea_main.c b/drivers/net/ehea/ehea_main.c index 717b129..5ebd545 100644 --- a/drivers/net/ehea/ehea_main.c +++ b/drivers/net/ehea/ehea_main.c @@ -393,9 +393,9 @@ static int ehea_treat_poll_error(struct ehea_port_res *pr, int rq, return 0; } -static struct ehea_cqe *ehea_proc_rwqes(struct net_device *dev, - struct ehea_port_res *pr, - int *budget) +static int ehea_proc_rwqes(struct net_device *dev, + struct ehea_port_res *pr, + int budget) { struct ehea_port *port = pr->port; struct ehea_qp *qp = pr->qp; @@ -408,18 +408,16 @@ static struct ehea_cqe *ehea_proc_rwqes(struct net_device *dev, int skb_arr_rq2_len = pr->rq2_skba.len; int skb_arr_rq3_len = pr->rq3_skba.len; int processed, processed_rq1, processed_rq2, processed_rq3; - int wqe_index, last_wqe_index, rq, my_quota, port_reset; + int wqe_index, last_wqe_index, rq, port_reset; processed = processed_rq1 = processed_rq2 = processed_rq3 = 0; last_wqe_index = 0; - my_quota = min(*budget, dev->quota); cqe = ehea_poll_rq1(qp, &wqe_index); - while ((my_quota > 0) && cqe) { + while ((processed < budget) && cqe) { ehea_inc_rq1(qp); processed_rq1++; processed++; - my_quota--; if (netif_msg_rx_status(port)) ehea_dump(cqe, sizeof(*cqe), "CQE"); @@ -434,14 +432,14 @@ static struct ehea_cqe *ehea_proc_rwqes(struct net_device *dev, if (netif_msg_rx_err(port)) ehea_error("LL rq1: skb=NULL"); - skb = netdev_alloc_skb(port->netdev, + skb = netdev_alloc_skb(dev, EHEA_L_PKT_SIZE); if (!skb) break; } skb_copy_to_linear_data(skb, ((char*)cqe) + 64, cqe->num_bytes_transfered - 4); - ehea_fill_skb(port->netdev, skb, cqe); + ehea_fill_skb(dev, skb, cqe); } else if (rq == 2) { /* RQ2 */ skb = get_skb_by_index(skb_arr_rq2, skb_arr_rq2_len, cqe); @@ -450,7 +448,7 @@ static struct ehea_cqe *ehea_proc_rwqes(struct net_device *dev, ehea_error("rq2: skb=NULL"); break; } - ehea_fill_skb(port->netdev, skb, cqe); + ehea_fill_skb(dev, skb, cqe); processed_rq2++; } else { /* RQ3 */ skb = get_skb_by_index(skb_arr_rq3, @@ -460,7 +458,7 @@ static struct ehea_cqe *ehea_proc_rwqes(struct net_device *dev, ehea_error("rq3: skb=NULL"); break; } - ehea_fill_skb(port->netdev, skb, cqe); + ehea_fill_skb(dev, skb, cqe); processed_rq3++; } @@ -471,7 +469,7 @@ static struct ehea_cqe *ehea_proc_rwqes(struct net_device *dev, else netif_receive_skb(skb); - port->netdev->last_rx = jiffies; + dev->last_rx = jiffies; } else { pr->p_stats.poll_receive_errors++; port_reset = ehea_treat_poll_error(pr, rq, cqe, @@ -484,14 +482,12 @@ static struct ehea_cqe *ehea_proc_rwqes(struct net_device *dev, } pr->rx_packets += processed; - *budget -= processed; ehea_refill_rq1(pr, last_wqe_index, processed_rq1); ehea_refill_rq2(pr, processed_rq2); ehea_refill_rq3(pr, processed_rq3); - cqe = ehea_poll_rq1(qp, &wqe_index); - return cqe; + return processed; } static struct ehea_cqe *ehea_proc_cqes(struct ehea_port_res *pr, int my_quota) @@ -554,22 +550,27 @@ static struct ehea_cqe *ehea_proc_cqes(struct ehea_port_res *pr, int my_quota) } #define EHEA_NAPI_POLL_NUM_BEFORE_IRQ 16 +#define EHEA_POLL_MAX_CQES 65535 -static int ehea_poll(struct net_device *dev, int *budget) +static int ehea_poll(struct napi_struct *napi, int budget) { - struct ehea_port_res *pr = dev->priv; + struct ehea_port_res *pr = container_of(napi, struct ehea_port_res, napi); + struct net_device *dev = pr->port->netdev; struct ehea_cqe *cqe; struct ehea_cqe *cqe_skb = NULL; int force_irq, wqe_index; - - cqe = ehea_poll_rq1(pr->qp, &wqe_index); - cqe_skb = ehea_poll_cq(pr->send_cq); + int rx = 0; force_irq = (pr->poll_counter > EHEA_NAPI_POLL_NUM_BEFORE_IRQ); + cqe_skb = ehea_proc_cqes(pr, EHEA_POLL_MAX_CQES); + + if (!force_irq) + rx += ehea_proc_rwqes(dev, pr, budget - rx); - if ((!cqe && !cqe_skb) || force_irq) { + while ((rx != budget) || force_irq) { pr->poll_counter = 0; - netif_rx_complete(dev); + force_irq = 0; + netif_rx_complete(dev, napi); ehea_reset_cq_ep(pr->recv_cq); ehea_reset_cq_ep(pr->send_cq); ehea_reset_cq_n1(pr->recv_cq); @@ -578,43 +579,35 @@ static int ehea_poll(struct net_device *dev, int *budget) cqe_skb = ehea_poll_cq(pr->send_cq); if (!cqe && !cqe_skb) - return 0; + return rx; - if (!netif_rx_reschedule(dev, dev->quota)) - return 0; - } - - cqe = ehea_proc_rwqes(dev, pr, budget); - cqe_skb = ehea_proc_cqes(pr, 300); + if (!netif_rx_reschedule(dev, napi)) + return rx; - if (cqe || cqe_skb) - pr->poll_counter++; + cqe_skb = ehea_proc_cqes(pr, EHEA_POLL_MAX_CQES); + rx += ehea_proc_rwqes(dev, pr, budget - rx); + } - return 1; + pr->poll_counter++; + return rx; } #ifdef CONFIG_NET_POLL_CONTROLLER static void ehea_netpoll(struct net_device *dev) { struct ehea_port *port = netdev_priv(dev); + int i; - netif_rx_schedule(port->port_res[0].d_netdev); + for (i = 0; i < port->num_def_qps; i++) + netif_rx_schedule(dev, &port->port_res[i].napi); } #endif -static int ehea_poll_firstqueue(struct net_device *dev, int *budget) -{ - struct ehea_port *port = netdev_priv(dev); - struct net_device *d_dev = port->port_res[0].d_netdev; - - return ehea_poll(d_dev, budget); -} - static irqreturn_t ehea_recv_irq_handler(int irq, void *param) { struct ehea_port_res *pr = param; - netif_rx_schedule(pr->d_netdev); + netif_rx_schedule(pr->port->netdev, &pr->napi); return IRQ_HANDLED; } @@ -1236,14 +1229,7 @@ static int ehea_init_port_res(struct ehea_port *port, struct ehea_port_res *pr, kfree(init_attr); - pr->d_netdev = alloc_netdev(0, "", ether_setup); - if (!pr->d_netdev) - goto out_free; - pr->d_netdev->priv = pr; - pr->d_netdev->weight = 64; - pr->d_netdev->poll = ehea_poll; - set_bit(__LINK_STATE_START, &pr->d_netdev->state); - strcpy(pr->d_netdev->name, port->netdev->name); + netif_napi_add(pr->port->netdev, &pr->napi, ehea_poll, 64); ret = 0; goto out; @@ -1266,8 +1252,6 @@ static int ehea_clean_portres(struct ehea_port *port, struct ehea_port_res *pr) { int ret, i; - free_netdev(pr->d_netdev); - ret = ehea_destroy_qp(pr->qp); if (!ret) { @@ -2248,6 +2232,22 @@ out: return ret; } +static void port_napi_disable(struct ehea_port *port) +{ + int i; + + for (i = 0; i < port->num_def_qps; i++) + napi_disable(&port->port_res[i].napi); +} + +static void port_napi_enable(struct ehea_port *port) +{ + int i; + + for (i = 0; i < port->num_def_qps; i++) + napi_enable(&port->port_res[i].napi); +} + static int ehea_open(struct net_device *dev) { int ret; @@ -2259,8 +2259,10 @@ static int ehea_open(struct net_device *dev) ehea_info("enabling port %s", dev->name); ret = ehea_up(dev); - if (!ret) + if (!ret) { + port_napi_enable(port); netif_start_queue(dev); + } up(&port->port_lock); @@ -2269,7 +2271,7 @@ static int ehea_open(struct net_device *dev) static int ehea_down(struct net_device *dev) { - int ret, i; + int ret; struct ehea_port *port = netdev_priv(dev); if (port->state == EHEA_PORT_DOWN) @@ -2278,10 +2280,7 @@ static int ehea_down(struct net_device *dev) ehea_drop_multicast_list(dev); ehea_free_interrupts(dev); - for (i = 0; i < port->num_def_qps; i++) - while (test_bit(__LINK_STATE_RX_SCHED, - &port->port_res[i].d_netdev->state)) - msleep(1); + port_napi_disable(port); port->state = EHEA_PORT_DOWN; @@ -2319,7 +2318,8 @@ static void ehea_reset_port(struct work_struct *work) port->resets++; down(&port->port_lock); netif_stop_queue(dev); - netif_poll_disable(dev); + + port_napi_disable(port); ehea_down(dev); @@ -2330,7 +2330,8 @@ static void ehea_reset_port(struct work_struct *work) if (netif_msg_timer(port)) ehea_info("Device %s resetted successfully", dev->name); - netif_poll_enable(dev); + port_napi_enable(port); + netif_wake_queue(dev); out: up(&port->port_lock); @@ -2358,7 +2359,9 @@ static void ehea_rereg_mrs(struct work_struct *work) dev->name); down(&port->port_lock); netif_stop_queue(dev); - netif_poll_disable(dev); + + port_napi_disable(port); + ehea_down(dev); up(&port->port_lock); } @@ -2406,7 +2409,7 @@ static void ehea_rereg_mrs(struct work_struct *work) ret = ehea_up(dev); if (!ret) { - netif_poll_enable(dev); + port_napi_enable(port); netif_wake_queue(dev); } @@ -2644,11 +2647,9 @@ struct ehea_port *ehea_setup_single_port(struct ehea_adapter *adapter, memcpy(dev->dev_addr, &port->mac_addr, ETH_ALEN); dev->open = ehea_open; - dev->poll = ehea_poll_firstqueue; #ifdef CONFIG_NET_POLL_CONTROLLER dev->poll_controller = ehea_netpoll; #endif - dev->weight = 64; dev->stop = ehea_stop; dev->hard_start_xmit = ehea_start_xmit; dev->get_stats = ehea_get_stats; diff --git a/drivers/net/epic100.c b/drivers/net/epic100.c index 1197784..f8446e3 100644 --- a/drivers/net/epic100.c +++ b/drivers/net/epic100.c @@ -262,6 +262,7 @@ struct epic_private { /* Ring pointers. */ spinlock_t lock; /* Group with Tx control cache line. */ spinlock_t napi_lock; + struct napi_struct napi; unsigned int reschedule_in_poll; unsigned int cur_tx, dirty_tx; @@ -294,7 +295,7 @@ static void epic_tx_timeout(struct net_device *dev); static void epic_init_ring(struct net_device *dev); static int epic_start_xmit(struct sk_buff *skb, struct net_device *dev); static int epic_rx(struct net_device *dev, int budget); -static int epic_poll(struct net_device *dev, int *budget); +static int epic_poll(struct napi_struct *napi, int budget); static irqreturn_t epic_interrupt(int irq, void *dev_instance); static int netdev_ioctl(struct net_device *dev, struct ifreq *rq, int cmd); static const struct ethtool_ops netdev_ethtool_ops; @@ -487,8 +488,7 @@ static int __devinit epic_init_one (struct pci_dev *pdev, dev->ethtool_ops = &netdev_ethtool_ops; dev->watchdog_timeo = TX_TIMEOUT; dev->tx_timeout = &epic_tx_timeout; - dev->poll = epic_poll; - dev->weight = 64; + netif_napi_add(dev, &ep->napi, epic_poll, 64); ret = register_netdev(dev); if (ret < 0) @@ -660,8 +660,11 @@ static int epic_open(struct net_device *dev) /* Soft reset the chip. */ outl(0x4001, ioaddr + GENCTL); - if ((retval = request_irq(dev->irq, &epic_interrupt, IRQF_SHARED, dev->name, dev))) + napi_enable(&ep->napi); + if ((retval = request_irq(dev->irq, &epic_interrupt, IRQF_SHARED, dev->name, dev))) { + napi_disable(&ep->napi); return retval; + } epic_init_ring(dev); @@ -1103,9 +1106,9 @@ static irqreturn_t epic_interrupt(int irq, void *dev_instance) if ((status & EpicNapiEvent) && !ep->reschedule_in_poll) { spin_lock(&ep->napi_lock); - if (netif_rx_schedule_prep(dev)) { + if (netif_rx_schedule_prep(dev, &ep->napi)) { epic_napi_irq_off(dev, ep); - __netif_rx_schedule(dev); + __netif_rx_schedule(dev, &ep->napi); } else ep->reschedule_in_poll++; spin_unlock(&ep->napi_lock); @@ -1257,26 +1260,22 @@ static void epic_rx_err(struct net_device *dev, struct epic_private *ep) outw(RxQueued, ioaddr + COMMAND); } -static int epic_poll(struct net_device *dev, int *budget) +static int epic_poll(struct napi_struct *napi, int budget) { - struct epic_private *ep = dev->priv; - int work_done = 0, orig_budget; + struct epic_private *ep = container_of(napi, struct epic_private, napi); + struct net_device *dev = ep->mii.dev; + int work_done = 0; long ioaddr = dev->base_addr; - orig_budget = (*budget > dev->quota) ? dev->quota : *budget; - rx_action: epic_tx(dev, ep); - work_done += epic_rx(dev, *budget); + work_done += epic_rx(dev, budget); epic_rx_err(dev, ep); - *budget -= work_done; - dev->quota -= work_done; - - if (netif_running(dev) && (work_done < orig_budget)) { + if (netif_running(dev) && (work_done < budget)) { unsigned long flags; int more; @@ -1286,7 +1285,7 @@ rx_action: more = ep->reschedule_in_poll; if (!more) { - __netif_rx_complete(dev); + __netif_rx_complete(dev, napi); outl(EpicNapiEvent, ioaddr + INTSTAT); epic_napi_irq_on(dev, ep); } else @@ -1298,7 +1297,7 @@ rx_action: goto rx_action; } - return (work_done >= orig_budget); + return work_done; } static int epic_close(struct net_device *dev) @@ -1309,6 +1308,7 @@ static int epic_close(struct net_device *dev) int i; netif_stop_queue(dev); + napi_disable(&ep->napi); if (debug > 1) printk(KERN_DEBUG "%s: Shutting down ethercard, status was %2.2x.\n", diff --git a/drivers/net/fec_8xx/fec_8xx.h b/drivers/net/fec_8xx/fec_8xx.h index 5af60b0..f3b1c6f 100644 --- a/drivers/net/fec_8xx/fec_8xx.h +++ b/drivers/net/fec_8xx/fec_8xx.h @@ -105,6 +105,8 @@ struct fec; struct fec_enet_private { spinlock_t lock; /* during all ops except TX pckt processing */ spinlock_t tx_lock; /* during fec_start_xmit and fec_tx */ + struct net_device *dev; + struct napi_struct napi; int fecno; struct fec *fecp; const struct fec_platform_info *fpi; diff --git a/drivers/net/fec_8xx/fec_main.c b/drivers/net/fec_8xx/fec_main.c index e5502af..6348fb9 100644 --- a/drivers/net/fec_8xx/fec_main.c +++ b/drivers/net/fec_8xx/fec_main.c @@ -465,9 +465,9 @@ void fec_stop(struct net_device *dev) } /* common receive function */ -static int fec_enet_rx_common(struct net_device *dev, int *budget) +static int fec_enet_rx_common(struct fec_enet_private *ep, + struct net_device *dev, int budget) { - struct fec_enet_private *fep = netdev_priv(dev); fec_t *fecp = fep->fecp; const struct fec_platform_info *fpi = fep->fpi; cbd_t *bdp; @@ -475,11 +475,8 @@ static int fec_enet_rx_common(struct net_device *dev, int *budget) int received = 0; __u16 pkt_len, sc; int curidx; - int rx_work_limit; if (fpi->use_napi) { - rx_work_limit = min(dev->quota, *budget); - if (!netif_running(dev)) return 0; } @@ -530,11 +527,6 @@ static int fec_enet_rx_common(struct net_device *dev, int *budget) BUG_ON(skbn == NULL); } else { - - /* napi, got packet but no quota */ - if (fpi->use_napi && --rx_work_limit < 0) - break; - skb = fep->rx_skbuff[curidx]; BUG_ON(skb == NULL); @@ -599,25 +591,24 @@ static int fec_enet_rx_common(struct net_device *dev, int *budget) * able to keep up at the expense of system resources. */ FW(fecp, r_des_active, 0x01000000); + + if (received >= budget) + break; + } fep->cur_rx = bdp; if (fpi->use_napi) { - dev->quota -= received; - *budget -= received; - - if (rx_work_limit < 0) - return 1; /* not done */ + if (received < budget) { + netif_rx_complete(dev, &fep->napi); - /* done */ - netif_rx_complete(dev); - - /* enable RX interrupt bits */ - FS(fecp, imask, FEC_ENET_RXF | FEC_ENET_RXB); + /* enable RX interrupt bits */ + FS(fecp, imask, FEC_ENET_RXF | FEC_ENET_RXB); + } } - return 0; + return received; } static void fec_enet_tx(struct net_device *dev) @@ -743,12 +734,12 @@ fec_enet_interrupt(int irq, void *dev_id) if ((int_events & FEC_ENET_RXF) != 0) { if (!fpi->use_napi) - fec_enet_rx_common(dev, NULL); + fec_enet_rx_common(fep, dev, ~0); else { - if (netif_rx_schedule_prep(dev)) { + if (netif_rx_schedule_prep(dev, &fep->napi)) { /* disable rx interrupts */ FC(fecp, imask, FEC_ENET_RXF | FEC_ENET_RXB); - __netif_rx_schedule(dev); + __netif_rx_schedule(dev, &fep->napi); } else { printk(KERN_ERR DRV_MODULE_NAME ": %s driver bug! interrupt while in poll!\n", @@ -893,10 +884,13 @@ static int fec_enet_open(struct net_device *dev) const struct fec_platform_info *fpi = fep->fpi; unsigned long flags; + napi_enable(&fep->napi); + /* Install our interrupt handler. */ if (request_irq(fpi->fec_irq, fec_enet_interrupt, 0, "fec", dev) != 0) { printk(KERN_ERR DRV_MODULE_NAME ": %s Could not allocate FEC IRQ!", dev->name); + napi_disable(&fep->napi); return -EINVAL; } @@ -907,6 +901,7 @@ static int fec_enet_open(struct net_device *dev) printk(KERN_ERR DRV_MODULE_NAME ": %s Could not allocate PHY IRQ!", dev->name); free_irq(fpi->fec_irq, dev); + napi_disable(&fep->napi); return -EINVAL; } @@ -932,6 +927,7 @@ static int fec_enet_close(struct net_device *dev) unsigned long flags; netif_stop_queue(dev); + napi_disable(&fep->napi); netif_carrier_off(dev); if (fpi->use_mdio) @@ -955,9 +951,12 @@ static struct net_device_stats *fec_enet_get_stats(struct net_device *dev) return &fep->stats; } -static int fec_enet_poll(struct net_device *dev, int *budget) +static int fec_enet_poll(struct napi_struct *napi, int budget) { - return fec_enet_rx_common(dev, budget); + struct fec_enet_private *fep = container_of(napi, struct fec_enet_private, napi); + struct net_device *dev = fep->dev; + + return fec_enet_rx_common(fep, dev, budget); } /*************************************************************************/ @@ -1107,6 +1106,7 @@ int fec_8xx_init_one(const struct fec_platform_info *fpi, SET_MODULE_OWNER(dev); fep = netdev_priv(dev); + fep->dev = dev; /* partial reset of FEC */ fec_whack_reset(fecp); @@ -1172,10 +1172,9 @@ int fec_8xx_init_one(const struct fec_platform_info *fpi, dev->get_stats = fec_enet_get_stats; dev->set_multicast_list = fec_set_multicast_list; dev->set_mac_address = fec_set_mac_address; - if (fpi->use_napi) { - dev->poll = fec_enet_poll; - dev->weight = fpi->napi_weight; - } + netif_napi_add(dev, &fec->napi, + fec_enet_poll, fpi->napi_weight); + dev->ethtool_ops = &fec_ethtool_ops; dev->do_ioctl = fec_ioctl; diff --git a/drivers/net/forcedeth.c b/drivers/net/forcedeth.c index 1938d6d..24c1294 100644 --- a/drivers/net/forcedeth.c +++ b/drivers/net/forcedeth.c @@ -159,6 +159,8 @@ #define dprintk(x...) do { } while (0) #endif +#define TX_WORK_PER_LOOP 64 +#define RX_WORK_PER_LOOP 64 /* * Hardware access: @@ -745,6 +747,9 @@ struct nv_skb_map { struct fe_priv { spinlock_t lock; + struct net_device *dev; + struct napi_struct napi; + /* General data: * Locking: spin_lock(&np->lock); */ struct net_device_stats stats; @@ -1586,9 +1591,10 @@ static int nv_alloc_rx_optimized(struct net_device *dev) static void nv_do_rx_refill(unsigned long data) { struct net_device *dev = (struct net_device *) data; + struct fe_priv *np = netdev_priv(dev); /* Just reschedule NAPI rx processing */ - netif_rx_schedule(dev); + netif_rx_schedule(dev, &np->napi); } #else static void nv_do_rx_refill(unsigned long data) @@ -2997,7 +3003,7 @@ static irqreturn_t nv_nic_irq(int foo, void *data) #ifdef CONFIG_FORCEDETH_NAPI if (events & NVREG_IRQ_RX_ALL) { - netif_rx_schedule(dev); + netif_rx_schedule(dev, &np->napi); /* Disable furthur receive irq's */ spin_lock(&np->lock); @@ -3010,7 +3016,7 @@ static irqreturn_t nv_nic_irq(int foo, void *data) spin_unlock(&np->lock); } #else - if (nv_rx_process(dev, dev->weight)) { + if (nv_rx_process(dev, RX_WORK_PER_LOOP)) { if (unlikely(nv_alloc_rx(dev))) { spin_lock(&np->lock); if (!np->in_shutdown) @@ -3079,8 +3085,6 @@ static irqreturn_t nv_nic_irq(int foo, void *data) return IRQ_RETVAL(i); } -#define TX_WORK_PER_LOOP 64 -#define RX_WORK_PER_LOOP 64 /** * All _optimized functions are used to help increase performance * (reduce CPU and increase throughput). They use descripter version 3, @@ -3114,7 +3118,7 @@ static irqreturn_t nv_nic_irq_optimized(int foo, void *data) #ifdef CONFIG_FORCEDETH_NAPI if (events & NVREG_IRQ_RX_ALL) { - netif_rx_schedule(dev); + netif_rx_schedule(dev, &np->napi); /* Disable furthur receive irq's */ spin_lock(&np->lock); @@ -3127,7 +3131,7 @@ static irqreturn_t nv_nic_irq_optimized(int foo, void *data) spin_unlock(&np->lock); } #else - if (nv_rx_process_optimized(dev, dev->weight)) { + if (nv_rx_process_optimized(dev, RX_WORK_PER_LOOP)) { if (unlikely(nv_alloc_rx_optimized(dev))) { spin_lock(&np->lock); if (!np->in_shutdown) @@ -3245,19 +3249,19 @@ static irqreturn_t nv_nic_irq_tx(int foo, void *data) } #ifdef CONFIG_FORCEDETH_NAPI -static int nv_napi_poll(struct net_device *dev, int *budget) +static int nv_napi_poll(struct napi_struct *napi, int budget) { - int pkts, limit = min(*budget, dev->quota); - struct fe_priv *np = netdev_priv(dev); + struct fe_priv *np = container_of(napi, struct fe_priv, napi); + struct net_device *dev = np->dev; u8 __iomem *base = get_hwbase(dev); unsigned long flags; - int retcode; + int pkts, retcode; if (np->desc_ver == DESC_VER_1 || np->desc_ver == DESC_VER_2) { - pkts = nv_rx_process(dev, limit); + pkts = nv_rx_process(dev, budget); retcode = nv_alloc_rx(dev); } else { - pkts = nv_rx_process_optimized(dev, limit); + pkts = nv_rx_process_optimized(dev, budget); retcode = nv_alloc_rx_optimized(dev); } @@ -3268,13 +3272,12 @@ static int nv_napi_poll(struct net_device *dev, int *budget) spin_unlock_irqrestore(&np->lock, flags); } - if (pkts < limit) { - /* all done, no more packets present */ - netif_rx_complete(dev); - + if (pkts < budget) { /* re-enable receive interrupts */ spin_lock_irqsave(&np->lock, flags); + __netif_rx_complete(dev, napi); + np->irqmask |= NVREG_IRQ_RX_ALL; if (np->msi_flags & NV_MSI_X_ENABLED) writel(NVREG_IRQ_RX_ALL, base + NvRegIrqMask); @@ -3282,13 +3285,8 @@ static int nv_napi_poll(struct net_device *dev, int *budget) writel(np->irqmask, base + NvRegIrqMask); spin_unlock_irqrestore(&np->lock, flags); - return 0; - } else { - /* used up our quantum, so reschedule */ - dev->quota -= pkts; - *budget -= pkts; - return 1; } + return pkts; } #endif @@ -3296,6 +3294,7 @@ static int nv_napi_poll(struct net_device *dev, int *budget) static irqreturn_t nv_nic_irq_rx(int foo, void *data) { struct net_device *dev = (struct net_device *) data; + struct fe_priv *np = netdev_priv(dev); u8 __iomem *base = get_hwbase(dev); u32 events; @@ -3303,7 +3302,7 @@ static irqreturn_t nv_nic_irq_rx(int foo, void *data) writel(NVREG_IRQ_RX_ALL, base + NvRegMSIXIrqStatus); if (events) { - netif_rx_schedule(dev); + netif_rx_schedule(dev, &np->napi); /* disable receive interrupts on the nic */ writel(NVREG_IRQ_RX_ALL, base + NvRegIrqMask); pci_push(base); @@ -3329,7 +3328,7 @@ static irqreturn_t nv_nic_irq_rx(int foo, void *data) if (!(events & np->irqmask)) break; - if (nv_rx_process_optimized(dev, dev->weight)) { + if (nv_rx_process_optimized(dev, RX_WORK_PER_LOOP)) { if (unlikely(nv_alloc_rx_optimized(dev))) { spin_lock_irqsave(&np->lock, flags); if (!np->in_shutdown) @@ -4620,7 +4619,9 @@ static void nv_self_test(struct net_device *dev, struct ethtool_test *test, u64 if (test->flags & ETH_TEST_FL_OFFLINE) { if (netif_running(dev)) { netif_stop_queue(dev); - netif_poll_disable(dev); +#ifdef CONFIG_FORCEDETH_NAPI + napi_disable(&np->napi); +#endif netif_tx_lock_bh(dev); spin_lock_irq(&np->lock); nv_disable_hw_interrupts(dev, np->irqmask); @@ -4679,7 +4680,9 @@ static void nv_self_test(struct net_device *dev, struct ethtool_test *test, u64 nv_start_rx(dev); nv_start_tx(dev); netif_start_queue(dev); - netif_poll_enable(dev); +#ifdef CONFIG_FORCEDETH_NAPI + napi_enable(&np->napi); +#endif nv_enable_hw_interrupts(dev, np->irqmask); } } @@ -4911,7 +4914,9 @@ static int nv_open(struct net_device *dev) nv_start_rx(dev); nv_start_tx(dev); netif_start_queue(dev); - netif_poll_enable(dev); +#ifdef CONFIG_FORCEDETH_NAPI + napi_enable(&np->napi); +#endif if (ret) { netif_carrier_on(dev); @@ -4942,7 +4947,9 @@ static int nv_close(struct net_device *dev) spin_lock_irq(&np->lock); np->in_shutdown = 1; spin_unlock_irq(&np->lock); - netif_poll_disable(dev); +#ifdef CONFIG_FORCEDETH_NAPI + napi_disable(&np->napi); +#endif synchronize_irq(dev->irq); del_timer_sync(&np->oom_kick); @@ -4994,6 +5001,7 @@ static int __devinit nv_probe(struct pci_dev *pci_dev, const struct pci_device_i goto out; np = netdev_priv(dev); + np->dev = dev; np->pci_dev = pci_dev; spin_lock_init(&np->lock); SET_MODULE_OWNER(dev); @@ -5155,9 +5163,8 @@ static int __devinit nv_probe(struct pci_dev *pci_dev, const struct pci_device_i #ifdef CONFIG_NET_POLL_CONTROLLER dev->poll_controller = nv_poll_controller; #endif - dev->weight = RX_WORK_PER_LOOP; #ifdef CONFIG_FORCEDETH_NAPI - dev->poll = nv_napi_poll; + netif_napi_add(dev, &np->napi, nv_napi_poll, RX_WORK_PER_LOOP); #endif SET_ETHTOOL_OPS(dev, &ops); dev->tx_timeout = nv_tx_timeout; diff --git a/drivers/net/fs_enet/fs_enet-main.c b/drivers/net/fs_enet/fs_enet-main.c index a4a2a0e..c509cb1 100644 --- a/drivers/net/fs_enet/fs_enet-main.c +++ b/drivers/net/fs_enet/fs_enet-main.c @@ -70,18 +70,16 @@ static void fs_set_multicast_list(struct net_device *dev) } /* NAPI receive function */ -static int fs_enet_rx_napi(struct net_device *dev, int *budget) +static int fs_enet_rx_napi(struct napi_struct *napi, int budget) { - struct fs_enet_private *fep = netdev_priv(dev); + struct fs_enet_private *fep = container_of(napi, struct fs_enet_private, napi); + struct net_device *dev = to_net_dev(fep->dev); const struct fs_platform_info *fpi = fep->fpi; cbd_t *bdp; struct sk_buff *skb, *skbn, *skbt; int received = 0; u16 pkt_len, sc; int curidx; - int rx_work_limit = 0; /* pacify gcc */ - - rx_work_limit = min(dev->quota, *budget); if (!netif_running(dev)) return 0; @@ -96,7 +94,6 @@ static int fs_enet_rx_napi(struct net_device *dev, int *budget) (*fep->ops->napi_clear_rx_event)(dev); while (((sc = CBDR_SC(bdp)) & BD_ENET_RX_EMPTY) == 0) { - curidx = bdp - fep->rx_bd_base; /* @@ -136,11 +133,6 @@ static int fs_enet_rx_napi(struct net_device *dev, int *budget) skbn = skb; } else { - - /* napi, got packet but no quota */ - if (--rx_work_limit < 0) - break; - skb = fep->rx_skbuff[curidx]; dma_unmap_single(fep->dev, CBDR_BUFADDR(bdp), @@ -199,22 +191,19 @@ static int fs_enet_rx_napi(struct net_device *dev, int *budget) bdp = fep->rx_bd_base; (*fep->ops->rx_bd_done)(dev); + + if (received >= budget) + break; } fep->cur_rx = bdp; - dev->quota -= received; - *budget -= received; - - if (rx_work_limit < 0) - return 1; /* not done */ - - /* done */ - netif_rx_complete(dev); - - (*fep->ops->napi_enable_rx)(dev); - - return 0; + if (received >= budget) { + /* done */ + netif_rx_complete(dev, napi); + (*fep->ops->napi_enable_rx)(dev); + } + return received; } /* non NAPI receive function */ @@ -470,7 +459,7 @@ fs_enet_interrupt(int irq, void *dev_id) if (!fpi->use_napi) fs_enet_rx_non_napi(dev); else { - napi_ok = netif_rx_schedule_prep(dev); + napi_ok = napi_schedule_prep(&fep->napi); (*fep->ops->napi_disable_rx)(dev); (*fep->ops->clear_int_events)(dev, fep->ev_napi_rx); @@ -478,7 +467,7 @@ fs_enet_interrupt(int irq, void *dev_id) /* NOTE: it is possible for FCCs in NAPI mode */ /* to submit a spurious interrupt while in poll */ if (napi_ok) - __netif_rx_schedule(dev); + __netif_rx_schedule(dev, &fep->napi); } } @@ -799,18 +788,22 @@ static int fs_enet_open(struct net_device *dev) int r; int err; + napi_enable(&fep->napi); + /* Install our interrupt handler. */ r = fs_request_irq(dev, fep->interrupt, "fs_enet-mac", fs_enet_interrupt); if (r != 0) { printk(KERN_ERR DRV_MODULE_NAME ": %s Could not allocate FS_ENET IRQ!", dev->name); + napi_disable(&fep->napi); return -EINVAL; } err = fs_init_phy(dev); - if(err) + if(err) { + napi_disable(&fep->napi); return err; - + } phy_start(fep->phydev); return 0; @@ -823,6 +816,7 @@ static int fs_enet_close(struct net_device *dev) netif_stop_queue(dev); netif_carrier_off(dev); + napi_disable(&fep->napi); phy_stop(fep->phydev); spin_lock_irqsave(&fep->lock, flags); @@ -1047,10 +1041,9 @@ static struct net_device *fs_init_instance(struct device *dev, ndev->stop = fs_enet_close; ndev->get_stats = fs_enet_get_stats; ndev->set_multicast_list = fs_set_multicast_list; - if (fpi->use_napi) { - ndev->poll = fs_enet_rx_napi; - ndev->weight = fpi->napi_weight; - } + netif_napi_add(ndev, &fep->napi, + fs_enet_rx_napi, fpi->napi_weight); + ndev->ethtool_ops = &fs_ethtool_ops; ndev->do_ioctl = fs_ioctl; diff --git a/drivers/net/fs_enet/fs_enet.h b/drivers/net/fs_enet/fs_enet.h index 569be22..46d0606 100644 --- a/drivers/net/fs_enet/fs_enet.h +++ b/drivers/net/fs_enet/fs_enet.h @@ -121,6 +121,7 @@ struct fs_enet_mii_bus { }; struct fs_enet_private { + struct napi_struct napi; struct device *dev; /* pointer back to the device (must be initialized first) */ spinlock_t lock; /* during all ops except TX pckt processing */ spinlock_t tx_lock; /* during fs_start_xmit and fs_tx */ diff --git a/drivers/net/gianfar.c b/drivers/net/gianfar.c index f926905..bd2de32 100644 --- a/drivers/net/gianfar.c +++ b/drivers/net/gianfar.c @@ -134,7 +134,7 @@ static void gfar_configure_serdes(struct net_device *dev); extern int gfar_local_mdio_write(struct gfar_mii *regs, int mii_id, int regnum, u16 value); extern int gfar_local_mdio_read(struct gfar_mii *regs, int mii_id, int regnum); #ifdef CONFIG_GFAR_NAPI -static int gfar_poll(struct net_device *dev, int *budget); +static int gfar_poll(struct napi_struct *napi, int budget); #endif #ifdef CONFIG_NET_POLL_CONTROLLER static void gfar_netpoll(struct net_device *dev); @@ -188,6 +188,7 @@ static int gfar_probe(struct platform_device *pdev) return -ENOMEM; priv = netdev_priv(dev); + priv->dev = dev; /* Set the info in the priv to the current info */ priv->einfo = einfo; @@ -261,10 +262,7 @@ static int gfar_probe(struct platform_device *pdev) dev->hard_start_xmit = gfar_start_xmit; dev->tx_timeout = gfar_timeout; dev->watchdog_timeo = TX_TIMEOUT; -#ifdef CONFIG_GFAR_NAPI - dev->poll = gfar_poll; - dev->weight = GFAR_DEV_WEIGHT; -#endif + netif_napi_add(dev, &priv->napi, gfar_poll, GFAR_DEV_WEIGHT); #ifdef CONFIG_NET_POLL_CONTROLLER dev->poll_controller = gfar_netpoll; #endif @@ -939,6 +937,8 @@ static int gfar_enet_open(struct net_device *dev) { int err; + napi_enable(&priv->napi); + /* Initialize a bunch of registers */ init_registers(dev); @@ -946,10 +946,14 @@ static int gfar_enet_open(struct net_device *dev) err = init_phy(dev); - if(err) + if(err) { + napi_disable(&priv->napi); return err; + } err = startup_gfar(dev); + if (err) + napi_disable(&priv->napi); netif_start_queue(dev); @@ -1102,6 +1106,9 @@ static int gfar_start_xmit(struct sk_buff *skb, struct net_device *dev) static int gfar_close(struct net_device *dev) { struct gfar_private *priv = netdev_priv(dev); + + napi_disable(&priv->napi); + stop_gfar(dev); /* Disconnect from the PHY */ @@ -1318,7 +1325,7 @@ struct sk_buff * gfar_new_skb(struct net_device *dev, struct rxbd8 *bdp) return NULL; alignamount = RXBUF_ALIGNMENT - - (((unsigned) skb->data) & (RXBUF_ALIGNMENT - 1)); + (((unsigned long) skb->data) & (RXBUF_ALIGNMENT - 1)); /* We need the data buffer to be aligned properly. We will reserve * as many bytes as needed to align the data properly @@ -1390,12 +1397,12 @@ irqreturn_t gfar_receive(int irq, void *dev_id) /* support NAPI */ #ifdef CONFIG_GFAR_NAPI - if (netif_rx_schedule_prep(dev)) { + if (netif_rx_schedule_prep(dev, &priv->napi)) { tempval = gfar_read(&priv->regs->imask); tempval &= IMASK_RX_DISABLED; gfar_write(&priv->regs->imask, tempval); - __netif_rx_schedule(dev); + __netif_rx_schedule(dev, &priv->napi); } else { if (netif_msg_rx_err(priv)) printk(KERN_DEBUG "%s: receive called twice (%x)[%x]\n", @@ -1569,23 +1576,16 @@ int gfar_clean_rx_ring(struct net_device *dev, int rx_work_limit) } #ifdef CONFIG_GFAR_NAPI -static int gfar_poll(struct net_device *dev, int *budget) +static int gfar_poll(struct napi_struct *napi, int budget) { + struct gfar_private *priv = container_of(napi, struct gfar_private, napi); + struct net_device *dev = priv->dev; int howmany; - struct gfar_private *priv = netdev_priv(dev); - int rx_work_limit = *budget; - - if (rx_work_limit > dev->quota) - rx_work_limit = dev->quota; - howmany = gfar_clean_rx_ring(dev, rx_work_limit); + howmany = gfar_clean_rx_ring(dev, budget); - dev->quota -= howmany; - rx_work_limit -= howmany; - *budget -= howmany; - - if (rx_work_limit > 0) { - netif_rx_complete(dev); + if (howmany < budget) { + netif_rx_complete(dev, napi); /* Clear the halt bit in RSTAT */ gfar_write(&priv->regs->rstat, RSTAT_CLEAR_RHALT); @@ -1601,8 +1601,7 @@ static int gfar_poll(struct net_device *dev, int *budget) gfar_write(&priv->regs->rxic, 0); } - /* Return 1 if there's more work to do */ - return (rx_work_limit > 0) ? 0 : 1; + return howmany; } #endif diff --git a/drivers/net/gianfar.h b/drivers/net/gianfar.h index d8e779c..b8714e0 100644 --- a/drivers/net/gianfar.h +++ b/drivers/net/gianfar.h @@ -691,6 +691,9 @@ struct gfar_private { /* RX Locked fields */ spinlock_t rxlock; + struct net_device *dev; + struct napi_struct napi; + /* skb array and index */ struct sk_buff ** rx_skbuff; u16 skb_currx; diff --git a/drivers/net/ibmveth.c b/drivers/net/ibmveth.c index acba90f..78e28ad 100644 --- a/drivers/net/ibmveth.c +++ b/drivers/net/ibmveth.c @@ -83,7 +83,7 @@ static int ibmveth_open(struct net_device *dev); static int ibmveth_close(struct net_device *dev); static int ibmveth_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd); -static int ibmveth_poll(struct net_device *dev, int *budget); +static int ibmveth_poll(struct napi_struct *napi, int budget); static int ibmveth_start_xmit(struct sk_buff *skb, struct net_device *dev); static struct net_device_stats *ibmveth_get_stats(struct net_device *dev); static void ibmveth_set_multicast_list(struct net_device *dev); @@ -480,6 +480,8 @@ static int ibmveth_open(struct net_device *netdev) ibmveth_debug_printk("open starting\n"); + napi_enable(&adapter->napi); + for(i = 0; i<IbmVethNumBufferPools; i++) rxq_entries += adapter->rx_buff_pool[i].size; @@ -489,6 +491,7 @@ static int ibmveth_open(struct net_device *netdev) if(!adapter->buffer_list_addr || !adapter->filter_list_addr) { ibmveth_error_printk("unable to allocate filter or buffer list pages\n"); ibmveth_cleanup(adapter); + napi_disable(&adapter->napi); return -ENOMEM; } @@ -498,6 +501,7 @@ static int ibmveth_open(struct net_device *netdev) if(!adapter->rx_queue.queue_addr) { ibmveth_error_printk("unable to allocate rx queue pages\n"); ibmveth_cleanup(adapter); + napi_disable(&adapter->napi); return -ENOMEM; } @@ -514,6 +518,7 @@ static int ibmveth_open(struct net_device *netdev) (dma_mapping_error(adapter->rx_queue.queue_dma))) { ibmveth_error_printk("unable to map filter or buffer list pages\n"); ibmveth_cleanup(adapter); + napi_disable(&adapter->napi); return -ENOMEM; } @@ -545,6 +550,7 @@ static int ibmveth_open(struct net_device *netdev) rxq_desc.desc, mac_address); ibmveth_cleanup(adapter); + napi_disable(&adapter->napi); return -ENONET; } @@ -555,6 +561,7 @@ static int ibmveth_open(struct net_device *netdev) ibmveth_error_printk("unable to alloc pool\n"); adapter->rx_buff_pool[i].active = 0; ibmveth_cleanup(adapter); + napi_disable(&adapter->napi); return -ENOMEM ; } } @@ -567,6 +574,7 @@ static int ibmveth_open(struct net_device *netdev) } while (H_IS_LONG_BUSY(rc) || (rc == H_BUSY)); ibmveth_cleanup(adapter); + napi_disable(&adapter->napi); return rc; } @@ -587,6 +595,8 @@ static int ibmveth_close(struct net_device *netdev) ibmveth_debug_printk("close starting\n"); + napi_disable(&adapter->napi); + if (!adapter->pool_config) netif_stop_queue(netdev); @@ -767,80 +777,68 @@ out: spin_lock_irqsave(&adapter->stats_lock, flags); return 0; } -static int ibmveth_poll(struct net_device *netdev, int *budget) +static int ibmveth_poll(struct napi_struct *napi, int budget) { - struct ibmveth_adapter *adapter = netdev->priv; - int max_frames_to_process = netdev->quota; + struct ibmveth_adapter *adapter = container_of(napi, struct ibmveth_adapter, napi); + struct net_device *netdev = adapter->netdev; int frames_processed = 0; - int more_work = 1; unsigned long lpar_rc; restart_poll: do { - struct net_device *netdev = adapter->netdev; - - if(ibmveth_rxq_pending_buffer(adapter)) { - struct sk_buff *skb; + struct sk_buff *skb; - rmb(); + if (!ibmveth_rxq_pending_buffer(adapter)) + break; - if(!ibmveth_rxq_buffer_valid(adapter)) { - wmb(); /* suggested by larson1 */ - adapter->rx_invalid_buffer++; - ibmveth_debug_printk("recycling invalid buffer\n"); - ibmveth_rxq_recycle_buffer(adapter); - } else { - int length = ibmveth_rxq_frame_length(adapter); - int offset = ibmveth_rxq_frame_offset(adapter); - skb = ibmveth_rxq_get_buffer(adapter); + rmb(); + if (!ibmveth_rxq_buffer_valid(adapter)) { + wmb(); /* suggested by larson1 */ + adapter->rx_invalid_buffer++; + ibmveth_debug_printk("recycling invalid buffer\n"); + ibmveth_rxq_recycle_buffer(adapter); + } else { + int length = ibmveth_rxq_frame_length(adapter); + int offset = ibmveth_rxq_frame_offset(adapter); + skb = ibmveth_rxq_get_buffer(adapter); - ibmveth_rxq_harvest_buffer(adapter); + ibmveth_rxq_harvest_buffer(adapter); - skb_reserve(skb, offset); - skb_put(skb, length); - skb->protocol = eth_type_trans(skb, netdev); + skb_reserve(skb, offset); + skb_put(skb, length); + skb->protocol = eth_type_trans(skb, netdev); - netif_receive_skb(skb); /* send it up */ + netif_receive_skb(skb); /* send it up */ - adapter->stats.rx_packets++; - adapter->stats.rx_bytes += length; - frames_processed++; - netdev->last_rx = jiffies; - } - } else { - more_work = 0; + adapter->stats.rx_packets++; + adapter->stats.rx_bytes += length; + frames_processed++; + netdev->last_rx = jiffies; } - } while(more_work && (frames_processed < max_frames_to_process)); + } while (frames_processed < budget); ibmveth_replenish_task(adapter); - if(more_work) { - /* more work to do - return that we are not done yet */ - netdev->quota -= frames_processed; - *budget -= frames_processed; - return 1; - } - - /* we think we are done - reenable interrupts, then check once more to make sure we are done */ - lpar_rc = h_vio_signal(adapter->vdev->unit_address, VIO_IRQ_ENABLE); + if (frames_processed < budget) { + /* We think we are done - reenable interrupts, + * then check once more to make sure we are done. + */ + lpar_rc = h_vio_signal(adapter->vdev->unit_address, + VIO_IRQ_ENABLE); - ibmveth_assert(lpar_rc == H_SUCCESS); + ibmveth_assert(lpar_rc == H_SUCCESS); - netif_rx_complete(netdev); + netif_rx_complete(netdev, napi); - if(ibmveth_rxq_pending_buffer(adapter) && netif_rx_reschedule(netdev, frames_processed)) - { - lpar_rc = h_vio_signal(adapter->vdev->unit_address, VIO_IRQ_DISABLE); - ibmveth_assert(lpar_rc == H_SUCCESS); - more_work = 1; - goto restart_poll; + if (ibmveth_rxq_pending_buffer(adapter) && + netif_rx_reschedule(netdev, napi)) { + lpar_rc = h_vio_signal(adapter->vdev->unit_address, + VIO_IRQ_DISABLE); + goto restart_poll; + } } - netdev->quota -= frames_processed; - *budget -= frames_processed; - - /* we really are done */ - return 0; + return frames_processed; } static irqreturn_t ibmveth_interrupt(int irq, void *dev_instance) @@ -849,10 +847,11 @@ static irqreturn_t ibmveth_interrupt(int irq, void *dev_instance) struct ibmveth_adapter *adapter = netdev->priv; unsigned long lpar_rc; - if(netif_rx_schedule_prep(netdev)) { - lpar_rc = h_vio_signal(adapter->vdev->unit_address, VIO_IRQ_DISABLE); + if (netif_rx_schedule_prep(netdev, &adapter->napi)) { + lpar_rc = h_vio_signal(adapter->vdev->unit_address, + VIO_IRQ_DISABLE); ibmveth_assert(lpar_rc == H_SUCCESS); - __netif_rx_schedule(netdev); + __netif_rx_schedule(netdev, &adapter->napi); } return IRQ_HANDLED; } @@ -1004,6 +1003,8 @@ static int __devinit ibmveth_probe(struct vio_dev *dev, const struct vio_device_ adapter->mcastFilterSize= *mcastFilterSize_p; adapter->pool_config = 0; + netif_napi_add(netdev, &adapter->napi, ibmveth_poll, 16); + /* Some older boxes running PHYP non-natively have an OF that returns a 8-byte local-mac-address field (and the first 2 bytes have to be ignored) while newer boxes' OF return @@ -1020,8 +1021,6 @@ static int __devinit ibmveth_probe(struct vio_dev *dev, const struct vio_device_ netdev->irq = dev->irq; netdev->open = ibmveth_open; - netdev->poll = ibmveth_poll; - netdev->weight = 16; netdev->stop = ibmveth_close; netdev->hard_start_xmit = ibmveth_start_xmit; netdev->get_stats = ibmveth_get_stats; diff --git a/drivers/net/ibmveth.h b/drivers/net/ibmveth.h index 72cc15a..e056941 100644 --- a/drivers/net/ibmveth.h +++ b/drivers/net/ibmveth.h @@ -112,6 +112,7 @@ struct ibmveth_rx_q { struct ibmveth_adapter { struct vio_dev *vdev; struct net_device *netdev; + struct napi_struct napi; struct net_device_stats stats; unsigned int mcastFilterSize; unsigned long mac_addr; diff --git a/drivers/net/ixgb/ixgb.h b/drivers/net/ixgb/ixgb.h index 3569d5b..1eee889 100644 --- a/drivers/net/ixgb/ixgb.h +++ b/drivers/net/ixgb/ixgb.h @@ -184,6 +184,7 @@ struct ixgb_adapter { boolean_t rx_csum; /* OS defined structs */ + struct napi_struct napi; struct net_device *netdev; struct pci_dev *pdev; struct net_device_stats net_stats; diff --git a/drivers/net/ixgb/ixgb_main.c b/drivers/net/ixgb/ixgb_main.c index 991c883..e3f27c6 100644 --- a/drivers/net/ixgb/ixgb_main.c +++ b/drivers/net/ixgb/ixgb_main.c @@ -97,7 +97,7 @@ static irqreturn_t ixgb_intr(int irq, void *data); static boolean_t ixgb_clean_tx_irq(struct ixgb_adapter *adapter); #ifdef CONFIG_IXGB_NAPI -static int ixgb_clean(struct net_device *netdev, int *budget); +static int ixgb_clean(struct napi_struct *napi, int budget); static boolean_t ixgb_clean_rx_irq(struct ixgb_adapter *adapter, int *work_done, int work_to_do); #else @@ -288,7 +288,7 @@ ixgb_up(struct ixgb_adapter *adapter) mod_timer(&adapter->watchdog_timer, jiffies); #ifdef CONFIG_IXGB_NAPI - netif_poll_enable(netdev); + napi_enable(&adapter->napi); #endif ixgb_irq_enable(adapter); @@ -309,7 +309,7 @@ ixgb_down(struct ixgb_adapter *adapter, boolean_t kill_watchdog) if(kill_watchdog) del_timer_sync(&adapter->watchdog_timer); #ifdef CONFIG_IXGB_NAPI - netif_poll_disable(netdev); + napi_disable(&adapter->napi); #endif adapter->link_speed = 0; adapter->link_duplex = 0; @@ -421,8 +421,7 @@ ixgb_probe(struct pci_dev *pdev, netdev->tx_timeout = &ixgb_tx_timeout; netdev->watchdog_timeo = 5 * HZ; #ifdef CONFIG_IXGB_NAPI - netdev->poll = &ixgb_clean; - netdev->weight = 64; + netif_napi_add(netdev, &adapter->napi, ixgb_clean, 64); #endif netdev->vlan_rx_register = ixgb_vlan_rx_register; netdev->vlan_rx_add_vid = ixgb_vlan_rx_add_vid; @@ -1746,7 +1745,7 @@ ixgb_intr(int irq, void *data) } #ifdef CONFIG_IXGB_NAPI - if(netif_rx_schedule_prep(netdev)) { + if (netif_rx_schedule_prep(netdev, &adapter->napi)) { /* Disable interrupts and register for poll. The flush of the posted write is intentionally left out. @@ -1754,7 +1753,7 @@ ixgb_intr(int irq, void *data) atomic_inc(&adapter->irq_sem); IXGB_WRITE_REG(&adapter->hw, IMC, ~0); - __netif_rx_schedule(netdev); + __netif_rx_schedule(netdev, &adapter->napi); } #else /* yes, that is actually a & and it is meant to make sure that @@ -1776,27 +1775,23 @@ ixgb_intr(int irq, void *data) **/ static int -ixgb_clean(struct net_device *netdev, int *budget) +ixgb_clean(struct napi_struct *napi, int budget) { - struct ixgb_adapter *adapter = netdev_priv(netdev); - int work_to_do = min(*budget, netdev->quota); + struct ixgb_adapter *adapter = container_of(napi, struct ixgb_adapter, napi); + struct net_device *netdev = adapter->netdev; int tx_cleaned; int work_done = 0; tx_cleaned = ixgb_clean_tx_irq(adapter); - ixgb_clean_rx_irq(adapter, &work_done, work_to_do); - - *budget -= work_done; - netdev->quota -= work_done; + ixgb_clean_rx_irq(adapter, &work_done, budget); /* if no Tx and not enough Rx work done, exit the polling mode */ if((!tx_cleaned && (work_done == 0)) || !netif_running(netdev)) { - netif_rx_complete(netdev); + netif_rx_complete(netdev, napi); ixgb_irq_enable(adapter); - return 0; } - return 1; + return work_done; } #endif diff --git a/drivers/net/ixp2000/ixpdev.c b/drivers/net/ixp2000/ixpdev.c index d9ce1ae..6c0dd49 100644 --- a/drivers/net/ixp2000/ixpdev.c +++ b/drivers/net/ixp2000/ixpdev.c @@ -74,9 +74,9 @@ static int ixpdev_xmit(struct sk_buff *skb, struct net_device *dev) } -static int ixpdev_rx(struct net_device *dev, int *budget) +static int ixpdev_rx(struct net_device *dev, int processed, int budget) { - while (*budget > 0) { + while (processed < budget) { struct ixpdev_rx_desc *desc; struct sk_buff *skb; void *buf; @@ -122,29 +122,34 @@ static int ixpdev_rx(struct net_device *dev, int *budget) err: ixp2000_reg_write(RING_RX_PENDING, _desc); - dev->quota--; - (*budget)--; + processed++; } - return 1; + return processed; } /* dev always points to nds[0]. */ -static int ixpdev_poll(struct net_device *dev, int *budget) +static int ixpdev_poll(struct napi_struct *napi, int budget) { + struct ixpdev_priv *ip = container_of(napi, struct ixpdev_priv, napi); + struct net_device *dev = ip->dev; + int rx; + /* @@@ Have to stop polling when nds[0] is administratively * downed while we are polling. */ + rx = 0; do { ixp2000_reg_write(IXP2000_IRQ_THD_RAW_STATUS_A_0, 0x00ff); - if (ixpdev_rx(dev, budget)) - return 1; + rx = ixpdev_rx(dev, rx, budget); + if (rx >= budget) + break; } while (ixp2000_reg_read(IXP2000_IRQ_THD_RAW_STATUS_A_0) & 0x00ff); - netif_rx_complete(dev); + netif_rx_complete(dev, napi); ixp2000_reg_write(IXP2000_IRQ_THD_ENABLE_SET_A_0, 0x00ff); - return 0; + return rx; } static void ixpdev_tx_complete(void) @@ -199,9 +204,12 @@ static irqreturn_t ixpdev_interrupt(int irq, void *dev_id) * Any of the eight receive units signaled RX? */ if (status & 0x00ff) { + struct net_device *dev = nds[0]; + struct ixpdev_priv *ip = netdev_priv(dev); + ixp2000_reg_wrb(IXP2000_IRQ_THD_ENABLE_CLEAR_A_0, 0x00ff); - if (likely(__netif_rx_schedule_prep(nds[0]))) { - __netif_rx_schedule(nds[0]); + if (likely(napi_schedule_prep(&ip->napi))) { + __netif_rx_schedule(dev, &ip->napi); } else { printk(KERN_CRIT "ixp2000: irq while polling!!\n"); } @@ -232,11 +240,13 @@ static int ixpdev_open(struct net_device *dev) struct ixpdev_priv *ip = netdev_priv(dev); int err; + napi_enable(&ip->napi); if (!nds_open++) { err = request_irq(IRQ_IXP2000_THDA0, ixpdev_interrupt, IRQF_SHARED, "ixp2000_eth", nds); if (err) { nds_open--; + napi_disable(&ip->napi); return err; } @@ -254,6 +264,7 @@ static int ixpdev_close(struct net_device *dev) struct ixpdev_priv *ip = netdev_priv(dev); netif_stop_queue(dev); + napi_disable(&ip->napi); set_port_admin_status(ip->channel, 0); if (!--nds_open) { @@ -274,7 +285,6 @@ struct net_device *ixpdev_alloc(int channel, int sizeof_priv) return NULL; dev->hard_start_xmit = ixpdev_xmit; - dev->poll = ixpdev_poll; dev->open = ixpdev_open; dev->stop = ixpdev_close; #ifdef CONFIG_NET_POLL_CONTROLLER @@ -282,9 +292,10 @@ struct net_device *ixpdev_alloc(int channel, int sizeof_priv) #endif dev->features |= NETIF_F_SG | NETIF_F_HW_CSUM; - dev->weight = 64; ip = netdev_priv(dev); + ip->dev = dev; + netif_napi_add(dev, &ip->napi, ixpdev_poll, 64); ip->channel = channel; ip->tx_queue_entries = 0; diff --git a/drivers/net/ixp2000/ixpdev.h b/drivers/net/ixp2000/ixpdev.h index bd686cb..391ece6 100644 --- a/drivers/net/ixp2000/ixpdev.h +++ b/drivers/net/ixp2000/ixpdev.h @@ -14,6 +14,8 @@ struct ixpdev_priv { + struct net_device *dev; + struct napi_struct napi; int channel; int tx_queue_entries; }; diff --git a/drivers/net/macb.c b/drivers/net/macb.c index a4bb026..74c3f7a 100644 --- a/drivers/net/macb.c +++ b/drivers/net/macb.c @@ -470,47 +470,41 @@ static int macb_rx(struct macb *bp, int budget) return received; } -static int macb_poll(struct net_device *dev, int *budget) +static int macb_poll(struct napi_struct *napi, int budget) { - struct macb *bp = netdev_priv(dev); - int orig_budget, work_done, retval = 0; + struct macb *bp = container_of(napi, struct macb, napi); + struct net_device *dev = bp->dev; + int work_done; u32 status; status = macb_readl(bp, RSR); macb_writel(bp, RSR, status); + work_done = 0; if (!status) { /* * This may happen if an interrupt was pending before * this function was called last time, and no packets * have been received since. */ - netif_rx_complete(dev); + netif_rx_complete(dev, napi); goto out; } dev_dbg(&bp->pdev->dev, "poll: status = %08lx, budget = %d\n", - (unsigned long)status, *budget); + (unsigned long)status, budget); if (!(status & MACB_BIT(REC))) { dev_warn(&bp->pdev->dev, "No RX buffers complete, status = %02lx\n", (unsigned long)status); - netif_rx_complete(dev); + netif_rx_complete(dev, napi); goto out; } - orig_budget = *budget; - if (orig_budget > dev->quota) - orig_budget = dev->quota; - - work_done = macb_rx(bp, orig_budget); - if (work_done < orig_budget) { - netif_rx_complete(dev); - retval = 0; - } else { - retval = 1; - } + work_done = macb_rx(bp, budget); + if (work_done < budget) + netif_rx_complete(dev, napi); /* * We've done what we can to clean the buffers. Make sure we @@ -521,7 +515,7 @@ out: /* TODO: Handle errors */ - return retval; + return work_done; } static irqreturn_t macb_interrupt(int irq, void *dev_id) @@ -545,7 +539,7 @@ static irqreturn_t macb_interrupt(int irq, void *dev_id) } if (status & MACB_RX_INT_FLAGS) { - if (netif_rx_schedule_prep(dev)) { + if (netif_rx_schedule_prep(dev, &bp->napi)) { /* * There's no point taking any more interrupts * until we have processed the buffers @@ -553,7 +547,7 @@ static irqreturn_t macb_interrupt(int irq, void *dev_id) macb_writel(bp, IDR, MACB_RX_INT_FLAGS); dev_dbg(&bp->pdev->dev, "scheduling RX softirq\n"); - __netif_rx_schedule(dev); + __netif_rx_schedule(dev, &bp->napi); } } @@ -937,6 +931,8 @@ static int macb_open(struct net_device *dev) return err; } + napi_enable(&bp->napi); + macb_init_rings(bp); macb_init_hw(bp); @@ -954,6 +950,7 @@ static int macb_close(struct net_device *dev) unsigned long flags; netif_stop_queue(dev); + napi_disable(&bp->napi); if (bp->phy_dev) phy_stop(bp->phy_dev); @@ -1146,8 +1143,7 @@ static int __devinit macb_probe(struct platform_device *pdev) dev->get_stats = macb_get_stats; dev->set_multicast_list = macb_set_rx_mode; dev->do_ioctl = macb_ioctl; - dev->poll = macb_poll; - dev->weight = 64; + netif_napi_add(dev, &bp->napi, macb_poll, 64); dev->ethtool_ops = &macb_ethtool_ops; dev->base_addr = regs->start; diff --git a/drivers/net/macb.h b/drivers/net/macb.h index 4e3283e..57b85ac 100644 --- a/drivers/net/macb.h +++ b/drivers/net/macb.h @@ -374,6 +374,7 @@ struct macb { struct clk *pclk; struct clk *hclk; struct net_device *dev; + struct napi_struct napi; struct net_device_stats stats; struct macb_stats hw_stats; diff --git a/drivers/net/mv643xx_eth.c b/drivers/net/mv643xx_eth.c index 3153356..702eba5 100644 --- a/drivers/net/mv643xx_eth.c +++ b/drivers/net/mv643xx_eth.c @@ -66,7 +66,7 @@ static int mv643xx_eth_change_mtu(struct net_device *, int); static struct net_device_stats *mv643xx_eth_get_stats(struct net_device *); static void eth_port_init_mac_tables(unsigned int eth_port_num); #ifdef MV643XX_NAPI -static int mv643xx_poll(struct net_device *dev, int *budget); +static int mv643xx_poll(struct napi_struct *napi, int budget); #endif static int ethernet_phy_get(unsigned int eth_port_num); static void ethernet_phy_set(unsigned int eth_port_num, int phy_addr); @@ -562,7 +562,7 @@ static irqreturn_t mv643xx_eth_int_handler(int irq, void *dev_id) /* wait for previous write to complete */ mv_read(MV643XX_ETH_INTERRUPT_MASK_REG(port_num)); - netif_rx_schedule(dev); + netif_rx_schedule(dev, &mp->napi); } #else if (eth_int_cause & ETH_INT_CAUSE_RX) @@ -880,6 +880,10 @@ static int mv643xx_eth_open(struct net_device *dev) mv643xx_eth_rx_refill_descs(dev); /* Fill RX ring with skb's */ +#ifdef MV643XX_NAPI + napi_enable(&mp->napi); +#endif + eth_port_start(dev); /* Interrupt Coalescing */ @@ -982,7 +986,7 @@ static int mv643xx_eth_stop(struct net_device *dev) mv_read(MV643XX_ETH_INTERRUPT_MASK_REG(port_num)); #ifdef MV643XX_NAPI - netif_poll_disable(dev); + napi_disable(&mp->napi); #endif netif_carrier_off(dev); netif_stop_queue(dev); @@ -992,10 +996,6 @@ static int mv643xx_eth_stop(struct net_device *dev) mv643xx_eth_free_tx_rings(dev); mv643xx_eth_free_rx_rings(dev); -#ifdef MV643XX_NAPI - netif_poll_enable(dev); -#endif - free_irq(dev->irq, dev); return 0; @@ -1007,11 +1007,12 @@ static int mv643xx_eth_stop(struct net_device *dev) * * This function is used in case of NAPI */ -static int mv643xx_poll(struct net_device *dev, int *budget) +static int mv643xx_poll(struct napi_struct *napi, int budget) { - struct mv643xx_private *mp = netdev_priv(dev); - int done = 1, orig_budget, work_done; + struct mv643xx_private *mp = container_of(napi, struct mv643xx_private, napi); + struct net_device *dev = mp->dev; unsigned int port_num = mp->port_num; + int work_done; #ifdef MV643XX_TX_FAST_REFILL if (++mp->tx_clean_threshold > 5) { @@ -1020,27 +1021,20 @@ static int mv643xx_poll(struct net_device *dev, int *budget) } #endif + work_done = 0; if ((mv_read(MV643XX_ETH_RX_CURRENT_QUEUE_DESC_PTR_0(port_num))) - != (u32) mp->rx_used_desc_q) { - orig_budget = *budget; - if (orig_budget > dev->quota) - orig_budget = dev->quota; - work_done = mv643xx_eth_receive_queue(dev, orig_budget); - *budget -= work_done; - dev->quota -= work_done; - if (work_done >= orig_budget) - done = 0; - } + != (u32) mp->rx_used_desc_q) + work_done = mv643xx_eth_receive_queue(dev, budget); - if (done) { - netif_rx_complete(dev); + if (work_done < budget) { + netif_rx_complete(dev, napi); mv_write(MV643XX_ETH_INTERRUPT_CAUSE_REG(port_num), 0); mv_write(MV643XX_ETH_INTERRUPT_CAUSE_EXTEND_REG(port_num), 0); mv_write(MV643XX_ETH_INTERRUPT_MASK_REG(port_num), ETH_INT_UNMASK_ALL); } - return done ? 0 : 1; + return work_done; } #endif @@ -1333,6 +1327,10 @@ static int mv643xx_eth_probe(struct platform_device *pdev) platform_set_drvdata(pdev, dev); mp = netdev_priv(dev); + mp->dev = dev; +#ifdef MV643XX_NAPI + netif_napi_add(dev, &mp->napi, mv643xx_poll, 64); +#endif res = platform_get_resource(pdev, IORESOURCE_IRQ, 0); BUG_ON(!res); @@ -1347,10 +1345,6 @@ static int mv643xx_eth_probe(struct platform_device *pdev) /* No need to Tx Timeout */ dev->tx_timeout = mv643xx_eth_tx_timeout; -#ifdef MV643XX_NAPI - dev->poll = mv643xx_poll; - dev->weight = 64; -#endif #ifdef CONFIG_NET_POLL_CONTROLLER dev->poll_controller = mv643xx_netpoll; diff --git a/drivers/net/mv643xx_eth.h b/drivers/net/mv643xx_eth.h index 565b966..be669eb 100644 --- a/drivers/net/mv643xx_eth.h +++ b/drivers/net/mv643xx_eth.h @@ -320,6 +320,8 @@ struct mv643xx_private { struct work_struct tx_timeout_task; + struct net_device *dev; + struct napi_struct napi; struct net_device_stats stats; struct mv643xx_mib_counters mib_counters; spinlock_t lock; diff --git a/drivers/net/myri10ge/myri10ge.c b/drivers/net/myri10ge/myri10ge.c index 556962f..a30146e 100644 --- a/drivers/net/myri10ge/myri10ge.c +++ b/drivers/net/myri10ge/myri10ge.c @@ -163,6 +163,7 @@ struct myri10ge_priv { int small_bytes; int big_bytes; struct net_device *dev; + struct napi_struct napi; struct net_device_stats stats; u8 __iomem *sram; int sram_size; @@ -1100,7 +1101,7 @@ static inline void myri10ge_tx_done(struct myri10ge_priv *mgp, int mcp_index) } } -static inline void myri10ge_clean_rx_done(struct myri10ge_priv *mgp, int *limit) +static inline int myri10ge_clean_rx_done(struct myri10ge_priv *mgp, int budget) { struct myri10ge_rx_done *rx_done = &mgp->rx_done; unsigned long rx_bytes = 0; @@ -1109,10 +1110,11 @@ static inline void myri10ge_clean_rx_done(struct myri10ge_priv *mgp, int *limit) int idx = rx_done->idx; int cnt = rx_done->cnt; + int work_done = 0; u16 length; __wsum checksum; - while (rx_done->entry[idx].length != 0 && *limit != 0) { + while (rx_done->entry[idx].length != 0 && work_done++ < budget) { length = ntohs(rx_done->entry[idx].length); rx_done->entry[idx].length = 0; checksum = csum_unfold(rx_done->entry[idx].checksum); @@ -1128,10 +1130,6 @@ static inline void myri10ge_clean_rx_done(struct myri10ge_priv *mgp, int *limit) rx_bytes += rx_ok * (unsigned long)length; cnt++; idx = cnt & (myri10ge_max_intr_slots - 1); - - /* limit potential for livelock by only handling a - * limited number of frames. */ - (*limit)--; } rx_done->idx = idx; rx_done->cnt = cnt; @@ -1145,6 +1143,7 @@ static inline void myri10ge_clean_rx_done(struct myri10ge_priv *mgp, int *limit) if (mgp->rx_big.fill_cnt - mgp->rx_big.cnt < myri10ge_fill_thresh) myri10ge_alloc_rx_pages(mgp, &mgp->rx_big, mgp->big_bytes, 0); + return work_done; } static inline void myri10ge_check_statblock(struct myri10ge_priv *mgp) @@ -1189,26 +1188,21 @@ static inline void myri10ge_check_statblock(struct myri10ge_priv *mgp) } } -static int myri10ge_poll(struct net_device *netdev, int *budget) +static int myri10ge_poll(struct napi_struct *napi, int budget) { - struct myri10ge_priv *mgp = netdev_priv(netdev); + struct myri10ge_priv *mgp = container_of(napi, struct myri10ge_priv, napi); + struct net_device *netdev = mgp->dev; struct myri10ge_rx_done *rx_done = &mgp->rx_done; - int limit, orig_limit, work_done; + int work_done; /* process as many rx events as NAPI will allow */ - limit = min(*budget, netdev->quota); - orig_limit = limit; - myri10ge_clean_rx_done(mgp, &limit); - work_done = orig_limit - limit; - *budget -= work_done; - netdev->quota -= work_done; + work_done = myri10ge_clean_rx_done(mgp, budget); if (rx_done->entry[rx_done->idx].length == 0 || !netif_running(netdev)) { - netif_rx_complete(netdev); + netif_rx_complete(netdev, napi); put_be32(htonl(3), mgp->irq_claim); - return 0; } - return 1; + return work_done; } static irqreturn_t myri10ge_intr(int irq, void *arg) @@ -1226,7 +1220,7 @@ static irqreturn_t myri10ge_intr(int irq, void *arg) /* low bit indicates receives are present, so schedule * napi poll handler */ if (stats->valid & 1) - netif_rx_schedule(mgp->dev); + netif_rx_schedule(mgp->dev, &mgp->napi); if (!mgp->msi_enabled) { put_be32(0, mgp->irq_deassert); @@ -1853,7 +1847,7 @@ static int myri10ge_open(struct net_device *dev) mgp->link_state = htonl(~0U); mgp->rdma_tags_available = 15; - netif_poll_enable(mgp->dev); /* must happen prior to any irq */ + napi_enable(&mgp->napi); /* must happen prior to any irq */ status = myri10ge_send_cmd(mgp, MXGEFW_CMD_ETHERNET_UP, &cmd, 0); if (status) { @@ -1897,7 +1891,7 @@ static int myri10ge_close(struct net_device *dev) del_timer_sync(&mgp->watchdog_timer); mgp->running = MYRI10GE_ETH_STOPPING; - netif_poll_disable(mgp->dev); + napi_disable(&mgp->napi); netif_carrier_off(dev); netif_stop_queue(dev); old_down_cnt = mgp->down_cnt; @@ -2857,6 +2851,8 @@ static int myri10ge_probe(struct pci_dev *pdev, const struct pci_device_id *ent) mgp = netdev_priv(netdev); memset(mgp, 0, sizeof(*mgp)); mgp->dev = netdev; + netif_napi_add(netdev, &mgp->napi, + myri10ge_poll, myri10ge_napi_weight); mgp->pdev = pdev; mgp->csum_flag = MXGEFW_FLAGS_CKSUM; mgp->pause = myri10ge_flow_control; @@ -2981,8 +2977,6 @@ static int myri10ge_probe(struct pci_dev *pdev, const struct pci_device_id *ent) netdev->features = NETIF_F_SG | NETIF_F_HW_CSUM | NETIF_F_TSO; if (dac_enabled) netdev->features |= NETIF_F_HIGHDMA; - netdev->poll = myri10ge_poll; - netdev->weight = myri10ge_napi_weight; /* make sure we can get an irq, and that MSI can be * setup (if available). Also ensure netdev->irq diff --git a/drivers/net/natsemi.c b/drivers/net/natsemi.c index b47a12d..43cfa4b 100644 --- a/drivers/net/natsemi.c +++ b/drivers/net/natsemi.c @@ -560,6 +560,8 @@ struct netdev_private { /* address of a sent-in-place packet/buffer, for later free() */ struct sk_buff *tx_skbuff[TX_RING_SIZE]; dma_addr_t tx_dma[TX_RING_SIZE]; + struct net_device *dev; + struct napi_struct napi; struct net_device_stats stats; /* Media monitoring timer */ struct timer_list timer; @@ -636,7 +638,7 @@ static void init_registers(struct net_device *dev); static int start_tx(struct sk_buff *skb, struct net_device *dev); static irqreturn_t intr_handler(int irq, void *dev_instance); static void netdev_error(struct net_device *dev, int intr_status); -static int natsemi_poll(struct net_device *dev, int *budget); +static int natsemi_poll(struct napi_struct *napi, int budget); static void netdev_rx(struct net_device *dev, int *work_done, int work_to_do); static void netdev_tx_done(struct net_device *dev); static int natsemi_change_mtu(struct net_device *dev, int new_mtu); @@ -861,6 +863,7 @@ static int __devinit natsemi_probe1 (struct pci_dev *pdev, dev->irq = irq; np = netdev_priv(dev); + netif_napi_add(dev, &np->napi, natsemi_poll, 64); np->pci_dev = pdev; pci_set_drvdata(pdev, dev); @@ -931,8 +934,6 @@ static int __devinit natsemi_probe1 (struct pci_dev *pdev, dev->do_ioctl = &netdev_ioctl; dev->tx_timeout = &tx_timeout; dev->watchdog_timeo = TX_TIMEOUT; - dev->poll = natsemi_poll; - dev->weight = 64; #ifdef CONFIG_NET_POLL_CONTROLLER dev->poll_controller = &natsemi_poll_controller; @@ -1554,6 +1555,8 @@ static int netdev_open(struct net_device *dev) free_irq(dev->irq, dev); return i; } + napi_enable(&np->napi); + init_ring(dev); spin_lock_irq(&np->lock); init_registers(dev); @@ -2200,10 +2203,10 @@ static irqreturn_t intr_handler(int irq, void *dev_instance) prefetch(&np->rx_skbuff[np->cur_rx % RX_RING_SIZE]); - if (netif_rx_schedule_prep(dev)) { + if (netif_rx_schedule_prep(dev, &np->napi)) { /* Disable interrupts and register for poll */ natsemi_irq_disable(dev); - __netif_rx_schedule(dev); + __netif_rx_schedule(dev, &np->napi); } else printk(KERN_WARNING "%s: Ignoring interrupt, status %#08x, mask %#08x.\n", @@ -2216,12 +2219,11 @@ static irqreturn_t intr_handler(int irq, void *dev_instance) /* This is the NAPI poll routine. As well as the standard RX handling * it also handles all other interrupts that the chip might raise. */ -static int natsemi_poll(struct net_device *dev, int *budget) +static int natsemi_poll(struct napi_struct *napi, int budget) { - struct netdev_private *np = netdev_priv(dev); + struct netdev_private *np = container_of(napi, struct netdev_private, napi); + struct net_device *dev = np->dev; void __iomem * ioaddr = ns_ioaddr(dev); - - int work_to_do = min(*budget, dev->quota); int work_done = 0; do { @@ -2236,7 +2238,7 @@ static int natsemi_poll(struct net_device *dev, int *budget) if (np->intr_status & (IntrRxDone | IntrRxIntr | RxStatusFIFOOver | IntrRxErr | IntrRxOverrun)) { - netdev_rx(dev, &work_done, work_to_do); + netdev_rx(dev, &work_done, budget); } if (np->intr_status & @@ -2250,16 +2252,13 @@ static int natsemi_poll(struct net_device *dev, int *budget) if (np->intr_status & IntrAbnormalSummary) netdev_error(dev, np->intr_status); - *budget -= work_done; - dev->quota -= work_done; - - if (work_done >= work_to_do) - return 1; + if (work_done >= budget) + return work_done; np->intr_status = readl(ioaddr + IntrStatus); } while (np->intr_status); - netif_rx_complete(dev); + netif_rx_complete(dev, napi); /* Reenable interrupts providing nothing is trying to shut * the chip down. */ @@ -2268,7 +2267,7 @@ static int natsemi_poll(struct net_device *dev, int *budget) natsemi_irq_enable(dev); spin_unlock(&np->lock); - return 0; + return work_done; } /* This routine is logically part of the interrupt handler, but separated @@ -3158,6 +3157,8 @@ static int netdev_close(struct net_device *dev) dev->name, np->cur_tx, np->dirty_tx, np->cur_rx, np->dirty_rx); + napi_disable(&np->napi); + /* * FIXME: what if someone tries to close a device * that is suspended? @@ -3253,7 +3254,7 @@ static void __devexit natsemi_remove1 (struct pci_dev *pdev) * disable_irq() to enforce synchronization. * * natsemi_poll: checks before reenabling interrupts. suspend * sets hands_off, disables interrupts and then waits with - * netif_poll_disable(). + * napi_disable(). * * Interrupts must be disabled, otherwise hands_off can cause irq storms. */ @@ -3279,7 +3280,7 @@ static int natsemi_suspend (struct pci_dev *pdev, pm_message_t state) spin_unlock_irq(&np->lock); enable_irq(dev->irq); - netif_poll_disable(dev); + napi_disable(&np->napi); /* Update the error counts. */ __get_stats(dev); @@ -3320,6 +3321,8 @@ static int natsemi_resume (struct pci_dev *pdev) pci_enable_device(pdev); /* pci_power_on(pdev); */ + napi_enable(&np->napi); + natsemi_reset(dev); init_ring(dev); disable_irq(dev->irq); @@ -3333,7 +3336,6 @@ static int natsemi_resume (struct pci_dev *pdev) mod_timer(&np->timer, jiffies + 1*HZ); } netif_device_attach(dev); - netif_poll_enable(dev); out: rtnl_unlock(); return 0; diff --git a/drivers/net/netxen/netxen_nic.h b/drivers/net/netxen/netxen_nic.h index d4c92cc..aaa3493 100644 --- a/drivers/net/netxen/netxen_nic.h +++ b/drivers/net/netxen/netxen_nic.h @@ -880,6 +880,7 @@ struct netxen_adapter { struct netxen_adapter *master; struct net_device *netdev; struct pci_dev *pdev; + struct napi_struct napi; struct net_device_stats net_stats; unsigned char mac_addr[ETH_ALEN]; int mtu; diff --git a/drivers/net/netxen/netxen_nic_main.c b/drivers/net/netxen/netxen_nic_main.c index 3122d01..a10bbef 100644 --- a/drivers/net/netxen/netxen_nic_main.c +++ b/drivers/net/netxen/netxen_nic_main.c @@ -68,7 +68,7 @@ static void netxen_tx_timeout(struct net_device *netdev); static void netxen_tx_timeout_task(struct work_struct *work); static void netxen_watchdog(unsigned long); static int netxen_handle_int(struct netxen_adapter *, struct net_device *); -static int netxen_nic_poll(struct net_device *dev, int *budget); +static int netxen_nic_poll(struct napi_struct *napi, int budget); #ifdef CONFIG_NET_POLL_CONTROLLER static void netxen_nic_poll_controller(struct net_device *netdev); #endif @@ -402,6 +402,9 @@ netxen_nic_probe(struct pci_dev *pdev, const struct pci_device_id *ent) adapter->netdev = netdev; adapter->pdev = pdev; + netif_napi_add(netdev, &adapter->napi, + netxen_nic_poll, NETXEN_NETDEV_WEIGHT); + /* this will be read from FW later */ adapter->intr_scheme = -1; @@ -422,8 +425,6 @@ netxen_nic_probe(struct pci_dev *pdev, const struct pci_device_id *ent) netxen_nic_change_mtu(netdev, netdev->mtu); SET_ETHTOOL_OPS(netdev, &netxen_nic_ethtool_ops); - netdev->poll = netxen_nic_poll; - netdev->weight = NETXEN_NETDEV_WEIGHT; #ifdef CONFIG_NET_POLL_CONTROLLER netdev->poll_controller = netxen_nic_poll_controller; #endif @@ -885,6 +886,8 @@ static int netxen_nic_open(struct net_device *netdev) if (!adapter->driver_mismatch) mod_timer(&adapter->watchdog_timer, jiffies); + napi_enable(&adapter->napi); + netxen_nic_enable_int(adapter); /* Done here again so that even if phantom sw overwrote it, @@ -894,6 +897,7 @@ static int netxen_nic_open(struct net_device *netdev) del_timer_sync(&adapter->watchdog_timer); printk(KERN_ERR "%s: Failed to initialize port %d\n", netxen_nic_driver_name, adapter->portnum); + napi_disable(&adapter->napi); return -EIO; } if (adapter->macaddr_set) @@ -923,6 +927,7 @@ static int netxen_nic_close(struct net_device *netdev) netif_carrier_off(netdev); netif_stop_queue(netdev); + napi_disable(&adapter->napi); netxen_nic_disable_int(adapter); @@ -1243,11 +1248,11 @@ netxen_handle_int(struct netxen_adapter *adapter, struct net_device *netdev) netxen_nic_disable_int(adapter); if (netxen_nic_rx_has_work(adapter) || netxen_nic_tx_has_work(adapter)) { - if (netif_rx_schedule_prep(netdev)) { + if (netif_rx_schedule_prep(netdev, &adapter->napi)) { /* * Interrupts are already disabled. */ - __netif_rx_schedule(netdev); + __netif_rx_schedule(netdev, &adapter->napi); } else { static unsigned int intcount = 0; if ((++intcount & 0xfff) == 0xfff) @@ -1305,14 +1310,13 @@ irqreturn_t netxen_intr(int irq, void *data) return IRQ_HANDLED; } -static int netxen_nic_poll(struct net_device *netdev, int *budget) +static int netxen_nic_poll(struct napi_struct *napi, int budget) { - struct netxen_adapter *adapter = netdev_priv(netdev); - int work_to_do = min(*budget, netdev->quota); + struct netxen_adapter *adapter = container_of(napi, struct netxen_adapter, napi); + struct net_device *netdev = adapter->netdev; int done = 1; int ctx; - int this_work_done; - int work_done = 0; + int work_done; DPRINTK(INFO, "polling for %d descriptors\n", *budget); @@ -1330,16 +1334,11 @@ static int netxen_nic_poll(struct net_device *netdev, int *budget) * packets are on one context, it gets only half of the quota, * and ends up not processing it. */ - this_work_done = netxen_process_rcv_ring(adapter, ctx, - work_to_do / - MAX_RCV_CTX); - work_done += this_work_done; + work_done += netxen_process_rcv_ring(adapter, ctx, + budget / MAX_RCV_CTX); } - netdev->quota -= work_done; - *budget -= work_done; - - if (work_done >= work_to_do && netxen_nic_rx_has_work(adapter) != 0) + if (work_done >= budget && netxen_nic_rx_has_work(adapter) != 0) done = 0; if (netxen_process_cmd_ring((unsigned long)adapter) == 0) @@ -1348,11 +1347,11 @@ static int netxen_nic_poll(struct net_device *netdev, int *budget) DPRINTK(INFO, "new work_done: %d work_to_do: %d\n", work_done, work_to_do); if (done) { - netif_rx_complete(netdev); + netif_rx_complete(netdev, napi); netxen_nic_enable_int(adapter); } - return !done; + return work_done; } #ifdef CONFIG_NET_POLL_CONTROLLER diff --git a/drivers/net/pasemi_mac.c b/drivers/net/pasemi_mac.c index 0b3066a..e63cc33 100644 --- a/drivers/net/pasemi_mac.c +++ b/drivers/net/pasemi_mac.c @@ -584,7 +584,7 @@ static irqreturn_t pasemi_mac_rx_intr(int irq, void *data) if (*mac->rx_status & PAS_STATUS_TIMER) reg |= PAS_IOB_DMA_RXCH_RESET_TINTC; - netif_rx_schedule(dev); + netif_rx_schedule(dev, &mac->napi); pci_write_config_dword(mac->iob_pdev, PAS_IOB_DMA_RXCH_RESET(mac->dma_rxch), reg); @@ -808,7 +808,7 @@ static int pasemi_mac_open(struct net_device *dev) dev_warn(&mac->pdev->dev, "phy init failed: %d\n", ret); netif_start_queue(dev); - netif_poll_enable(dev); + napi_enable(&mac->napi); /* Interrupts are a bit different for our DMA controller: While * it's got one a regular PCI device header, the interrupt there @@ -845,7 +845,7 @@ static int pasemi_mac_open(struct net_device *dev) out_rx_int: free_irq(mac->tx_irq, dev); out_tx_int: - netif_poll_disable(dev); + napi_disable(&mac->napi); netif_stop_queue(dev); pasemi_mac_free_tx_resources(dev); out_tx_resources: @@ -869,6 +869,7 @@ static int pasemi_mac_close(struct net_device *dev) } netif_stop_queue(dev); + napi_disable(&mac->napi); /* Clean out any pending buffers */ pasemi_mac_clean_tx(mac); @@ -1047,26 +1048,20 @@ static void pasemi_mac_set_rx_mode(struct net_device *dev) } -static int pasemi_mac_poll(struct net_device *dev, int *budget) +static int pasemi_mac_poll(struct napi_struct *napi, int budget) { - int pkts, limit = min(*budget, dev->quota); - struct pasemi_mac *mac = netdev_priv(dev); - - pkts = pasemi_mac_clean_rx(mac, limit); + struct pasemi_mac *mac = container_of(napi, struct pasemi_mac, napi); + struct net_device *dev = mac->netdev; + int pkts; - dev->quota -= pkts; - *budget -= pkts; - - if (pkts < limit) { + pkts = pasemi_mac_clean_rx(mac, budget); + if (pkts < budget) { /* all done, no more packets present */ - netif_rx_complete(dev); + netif_rx_complete(dev, napi); pasemi_mac_restart_rx_intr(mac); - return 0; - } else { - /* used up our quantum, so reschedule */ - return 1; } + return pkts; } static int __devinit @@ -1099,6 +1094,10 @@ pasemi_mac_probe(struct pci_dev *pdev, const struct pci_device_id *ent) mac->netdev = dev; mac->dma_pdev = pci_get_device(PCI_VENDOR_ID_PASEMI, 0xa007, NULL); + netif_napi_add(dev, &mac->napi, pasemi_mac_poll, 64); + + dev->features = NETIF_F_HW_CSUM; + if (!mac->dma_pdev) { dev_err(&pdev->dev, "Can't find DMA Controller\n"); err = -ENODEV; @@ -1150,9 +1149,6 @@ pasemi_mac_probe(struct pci_dev *pdev, const struct pci_device_id *ent) dev->hard_start_xmit = pasemi_mac_start_tx; dev->get_stats = pasemi_mac_get_stats; dev->set_multicast_list = pasemi_mac_set_rx_mode; - dev->weight = 64; - dev->poll = pasemi_mac_poll; - dev->features = NETIF_F_HW_CSUM; /* The dma status structure is located in the I/O bridge, and * is cache coherent. diff --git a/drivers/net/pasemi_mac.h b/drivers/net/pasemi_mac.h index c29ee15..85d3b78 100644 --- a/drivers/net/pasemi_mac.h +++ b/drivers/net/pasemi_mac.h @@ -56,6 +56,7 @@ struct pasemi_mac { struct pci_dev *dma_pdev; struct pci_dev *iob_pdev; struct phy_device *phydev; + struct napi_struct napi; struct net_device_stats stats; /* Pointer to the cacheable per-channel status registers */ diff --git a/drivers/net/pcnet32.c b/drivers/net/pcnet32.c index e6a6753..a997349 100644 --- a/drivers/net/pcnet32.c +++ b/drivers/net/pcnet32.c @@ -280,6 +280,8 @@ struct pcnet32_private { unsigned int dirty_rx, /* ring entries to be freed. */ dirty_tx; + struct net_device *dev; + struct napi_struct napi; struct net_device_stats stats; char tx_full; char phycount; /* number of phys found */ @@ -440,15 +442,21 @@ static struct pcnet32_access pcnet32_dwio = { static void pcnet32_netif_stop(struct net_device *dev) { + struct pcnet32_private *lp = netdev_priv(dev); dev->trans_start = jiffies; - netif_poll_disable(dev); +#ifdef CONFIG_PCNET32_NAPI + napi_disable(&lp->napi); +#endif netif_tx_disable(dev); } static void pcnet32_netif_start(struct net_device *dev) { + struct pcnet32_private *lp = netdev_priv(dev); netif_wake_queue(dev); - netif_poll_enable(dev); +#ifdef CONFIG_PCNET32_NAPI + napi_enable(&lp->napi); +#endif } /* @@ -816,7 +824,7 @@ static int pcnet32_set_ringparam(struct net_device *dev, if ((1 << i) != lp->rx_ring_size) pcnet32_realloc_rx_ring(dev, lp, i); - dev->weight = lp->rx_ring_size / 2; + lp->napi.weight = lp->rx_ring_size / 2; if (netif_running(dev)) { pcnet32_netif_start(dev); @@ -1255,7 +1263,7 @@ static void pcnet32_rx_entry(struct net_device *dev, return; } -static int pcnet32_rx(struct net_device *dev, int quota) +static int pcnet32_rx(struct net_device *dev, int budget) { struct pcnet32_private *lp = netdev_priv(dev); int entry = lp->cur_rx & lp->rx_mod_mask; @@ -1263,7 +1271,7 @@ static int pcnet32_rx(struct net_device *dev, int quota) int npackets = 0; /* If we own the next entry, it's a new packet. Send it up. */ - while (quota > npackets && (short)le16_to_cpu(rxp->status) >= 0) { + while (npackets < budget && (short)le16_to_cpu(rxp->status) >= 0) { pcnet32_rx_entry(dev, lp, rxp, entry); npackets += 1; /* @@ -1379,15 +1387,16 @@ static int pcnet32_tx(struct net_device *dev) } #ifdef CONFIG_PCNET32_NAPI -static int pcnet32_poll(struct net_device *dev, int *budget) +static int pcnet32_poll(struct napi_struct *napi, int budget) { - struct pcnet32_private *lp = netdev_priv(dev); - int quota = min(dev->quota, *budget); + struct pcnet32_private *lp = container_of(napi, struct pcnet32_private, napi); + struct net_device *dev = lp->dev; unsigned long ioaddr = dev->base_addr; unsigned long flags; + int work_done; u16 val; - quota = pcnet32_rx(dev, quota); + work_done = pcnet32_rx(dev, budget); spin_lock_irqsave(&lp->lock, flags); if (pcnet32_tx(dev)) { @@ -1399,28 +1408,22 @@ static int pcnet32_poll(struct net_device *dev, int *budget) } spin_unlock_irqrestore(&lp->lock, flags); - *budget -= quota; - dev->quota -= quota; - - if (dev->quota == 0) { - return 1; - } - - netif_rx_complete(dev); - - spin_lock_irqsave(&lp->lock, flags); + if (work_done < budget) { + spin_lock_irqsave(&lp->lock, flags); - /* clear interrupt masks */ - val = lp->a.read_csr(ioaddr, CSR3); - val &= 0x00ff; - lp->a.write_csr(ioaddr, CSR3, val); + __netif_rx_complete(dev, napi); - /* Set interrupt enable. */ - lp->a.write_csr(ioaddr, CSR0, CSR0_INTEN); - mmiowb(); - spin_unlock_irqrestore(&lp->lock, flags); + /* clear interrupt masks */ + val = lp->a.read_csr(ioaddr, CSR3); + val &= 0x00ff; + lp->a.write_csr(ioaddr, CSR3, val); - return 0; + /* Set interrupt enable. */ + lp->a.write_csr(ioaddr, CSR0, CSR0_INTEN); + mmiowb(); + spin_unlock_irqrestore(&lp->lock, flags); + } + return work_done; } #endif @@ -1815,6 +1818,8 @@ pcnet32_probe1(unsigned long ioaddr, int shared, struct pci_dev *pdev) } lp->pci_dev = pdev; + lp->dev = dev; + spin_lock_init(&lp->lock); SET_MODULE_OWNER(dev); @@ -1843,6 +1848,10 @@ pcnet32_probe1(unsigned long ioaddr, int shared, struct pci_dev *pdev) lp->mii_if.mdio_read = mdio_read; lp->mii_if.mdio_write = mdio_write; +#ifdef CONFIG_PCNET32_NAPI + netif_napi_add(dev, &lp->napi, pcnet32_poll, lp->rx_ring_size / 2); +#endif + if (fdx && !(lp->options & PCNET32_PORT_ASEL) && ((cards_found >= MAX_UNITS) || full_duplex[cards_found])) lp->options |= PCNET32_PORT_FD; @@ -1953,10 +1962,6 @@ pcnet32_probe1(unsigned long ioaddr, int shared, struct pci_dev *pdev) dev->ethtool_ops = &pcnet32_ethtool_ops; dev->tx_timeout = pcnet32_tx_timeout; dev->watchdog_timeo = (5 * HZ); - dev->weight = lp->rx_ring_size / 2; -#ifdef CONFIG_PCNET32_NAPI - dev->poll = pcnet32_poll; -#endif #ifdef CONFIG_NET_POLL_CONTROLLER dev->poll_controller = pcnet32_poll_controller; @@ -2276,6 +2281,10 @@ static int pcnet32_open(struct net_device *dev) goto err_free_ring; } +#ifdef CONFIG_PCNET32_NAPI + napi_enable(&lp->napi); +#endif + /* Re-initialize the PCNET32, and start it when done. */ lp->a.write_csr(ioaddr, 1, (lp->init_dma_addr & 0xffff)); lp->a.write_csr(ioaddr, 2, (lp->init_dma_addr >> 16)); @@ -2599,18 +2608,18 @@ pcnet32_interrupt(int irq, void *dev_id) /* unlike for the lance, there is no restart needed */ } #ifdef CONFIG_PCNET32_NAPI - if (netif_rx_schedule_prep(dev)) { + if (netif_rx_schedule_prep(dev, &lp->napi)) { u16 val; /* set interrupt masks */ val = lp->a.read_csr(ioaddr, CSR3); val |= 0x5f00; lp->a.write_csr(ioaddr, CSR3, val); mmiowb(); - __netif_rx_schedule(dev); + __netif_rx_schedule(dev, &lp->napi); break; } #else - pcnet32_rx(dev, dev->weight); + pcnet32_rx(dev, lp->napi.weight); if (pcnet32_tx(dev)) { /* reset the chip to clear the error condition, then restart */ lp->a.reset(ioaddr); @@ -2645,6 +2654,9 @@ static int pcnet32_close(struct net_device *dev) del_timer_sync(&lp->watchdog_timer); netif_stop_queue(dev); +#ifdef CONFIG_PCNET32_NAPI + napi_disable(&lp->napi); +#endif spin_lock_irqsave(&lp->lock, flags); diff --git a/drivers/net/ps3_gelic_net.c b/drivers/net/ps3_gelic_net.c index e565039..92561c0 100644 --- a/drivers/net/ps3_gelic_net.c +++ b/drivers/net/ps3_gelic_net.c @@ -556,6 +556,7 @@ static int gelic_net_stop(struct net_device *netdev) { struct gelic_net_card *card = netdev_priv(netdev); + napi_disable(&card->napi); netif_stop_queue(netdev); /* turn off DMA, force end */ @@ -987,32 +988,24 @@ refill: * if the quota is exceeded, but the driver has still packets. * */ -static int gelic_net_poll(struct net_device *netdev, int *budget) +static int gelic_net_poll(struct napi_struct *napi, int budget) { - struct gelic_net_card *card = netdev_priv(netdev); - int packets_to_do, packets_done = 0; - int no_more_packets = 0; - - packets_to_do = min(*budget, netdev->quota); + struct gelic_net_card *card = container_of(napi, struct gelic_net_card, napi); + struct net_device *netdev = card->netdev; + int packets_done = 0; - while (packets_to_do) { - if (gelic_net_decode_one_descr(card)) { - packets_done++; - packets_to_do--; - } else { - /* no more packets for the stack */ - no_more_packets = 1; + while (packets_done < budget) { + if (!gelic_net_decode_one_descr(card)) break; - } + + packets_done++; } - netdev->quota -= packets_done; - *budget -= packets_done; - if (no_more_packets) { - netif_rx_complete(netdev); + + if (packets_done < budget) { + netif_rx_complete(netdev, napi); gelic_net_rx_irq_on(card); - return 0; - } else - return 1; + } + return packets_done; } /** * gelic_net_change_mtu - changes the MTU of an interface @@ -1055,7 +1048,7 @@ static irqreturn_t gelic_net_interrupt(int irq, void *ptr) if (status & GELIC_NET_RXINT) { gelic_net_rx_irq_off(card); - netif_rx_schedule(netdev); + netif_rx_schedule(netdev, &card->napi); } if (status & GELIC_NET_TXINT) { @@ -1159,6 +1152,8 @@ static int gelic_net_open(struct net_device *netdev) if (gelic_net_alloc_rx_skbs(card)) goto alloc_skbs_failed; + napi_enable(&card->napi); + card->tx_dma_progress = 0; card->ghiintmask = GELIC_NET_RXINT | GELIC_NET_TXINT; @@ -1360,9 +1355,6 @@ static void gelic_net_setup_netdev_ops(struct net_device *netdev) /* tx watchdog */ netdev->tx_timeout = &gelic_net_tx_timeout; netdev->watchdog_timeo = GELIC_NET_WATCHDOG_TIMEOUT; - /* NAPI */ - netdev->poll = &gelic_net_poll; - netdev->weight = GELIC_NET_NAPI_WEIGHT; netdev->ethtool_ops = &gelic_net_ethtool_ops; } @@ -1390,6 +1382,9 @@ static int gelic_net_setup_netdev(struct gelic_net_card *card) gelic_net_setup_netdev_ops(netdev); + netif_napi_add(netdev, &card->napi, + gelic_net_poll, GELIC_NET_NAPI_WEIGHT); + netdev->features = NETIF_F_IP_CSUM; status = lv1_net_control(bus_id(card), dev_id(card), diff --git a/drivers/net/ps3_gelic_net.h b/drivers/net/ps3_gelic_net.h index a9c4c4f..9685602 100644 --- a/drivers/net/ps3_gelic_net.h +++ b/drivers/net/ps3_gelic_net.h @@ -194,6 +194,7 @@ struct gelic_net_descr_chain { struct gelic_net_card { struct net_device *netdev; + struct napi_struct napi; /* * hypervisor requires irq_status should be * 8 bytes aligned, but u64 member is diff --git a/drivers/net/qla3xxx.c b/drivers/net/qla3xxx.c index ea15131..bf9f8f6 100755 --- a/drivers/net/qla3xxx.c +++ b/drivers/net/qla3xxx.c @@ -2310,10 +2310,10 @@ static int ql_tx_rx_clean(struct ql3_adapter *qdev, return work_done; } -static int ql_poll(struct net_device *ndev, int *budget) +static int ql_poll(struct napi_struct *napi, int budget) { - struct ql3_adapter *qdev = netdev_priv(ndev); - int work_to_do = min(*budget, ndev->quota); + struct ql3_adapter *qdev = container_of(napi, struct ql3_adapter, napi); + struct net_device *ndev = qdev->ndev; int rx_cleaned = 0, tx_cleaned = 0; unsigned long hw_flags; struct ql3xxx_port_registers __iomem *port_regs = qdev->mem_map_registers; @@ -2321,16 +2321,13 @@ static int ql_poll(struct net_device *ndev, int *budget) if (!netif_carrier_ok(ndev)) goto quit_polling; - ql_tx_rx_clean(qdev, &tx_cleaned, &rx_cleaned, work_to_do); - *budget -= rx_cleaned; - ndev->quota -= rx_cleaned; + ql_tx_rx_clean(qdev, &tx_cleaned, &rx_cleaned, budget); - if( tx_cleaned + rx_cleaned != work_to_do || + if (tx_cleaned + rx_cleaned != budget || !netif_running(ndev)) { quit_polling: - netif_rx_complete(ndev); - spin_lock_irqsave(&qdev->hw_lock, hw_flags); + __netif_rx_complete(ndev, napi); ql_update_small_bufq_prod_index(qdev); ql_update_lrg_bufq_prod_index(qdev); writel(qdev->rsp_consumer_index, @@ -2338,9 +2335,8 @@ quit_polling: spin_unlock_irqrestore(&qdev->hw_lock, hw_flags); ql_enable_interrupts(qdev); - return 0; } - return 1; + return tx_cleaned + rx_cleaned; } static irqreturn_t ql3xxx_isr(int irq, void *dev_id) @@ -2390,8 +2386,8 @@ static irqreturn_t ql3xxx_isr(int irq, void *dev_id) spin_unlock(&qdev->adapter_lock); } else if (value & ISP_IMR_DISABLE_CMPL_INT) { ql_disable_interrupts(qdev); - if (likely(netif_rx_schedule_prep(ndev))) { - __netif_rx_schedule(ndev); + if (likely(netif_rx_schedule_prep(ndev, &qdev->napi))) { + __netif_rx_schedule(ndev, &qdev->napi); } } else { return IRQ_NONE; @@ -3617,7 +3613,7 @@ static int ql_adapter_down(struct ql3_adapter *qdev, int do_reset) del_timer_sync(&qdev->adapter_timer); - netif_poll_disable(ndev); + napi_disable(&qdev->napi); if (do_reset) { int soft_reset; @@ -3705,7 +3701,7 @@ static int ql_adapter_up(struct ql3_adapter *qdev) mod_timer(&qdev->adapter_timer, jiffies + HZ * 1); - netif_poll_enable(ndev); + napi_enable(&qdev->napi); ql_enable_interrupts(qdev); return 0; @@ -4061,8 +4057,7 @@ static int __devinit ql3xxx_probe(struct pci_dev *pdev, ndev->tx_timeout = ql3xxx_tx_timeout; ndev->watchdog_timeo = 5 * HZ; - ndev->poll = &ql_poll; - ndev->weight = 64; + netif_napi_add(ndev, &qdev->napi, ql_poll, 64); ndev->irq = pdev->irq; diff --git a/drivers/net/qla3xxx.h b/drivers/net/qla3xxx.h index 4a832c4..aa2216f 100755 --- a/drivers/net/qla3xxx.h +++ b/drivers/net/qla3xxx.h @@ -1175,6 +1175,8 @@ struct ql3_adapter { struct pci_dev *pdev; struct net_device *ndev; /* Parent NET device */ + struct napi_struct napi; + /* Hardware information */ u8 chip_rev_id; u8 pci_slot; diff --git a/drivers/net/r8169.c b/drivers/net/r8169.c index c76dd29..3f2306e 100644 --- a/drivers/net/r8169.c +++ b/drivers/net/r8169.c @@ -384,6 +384,7 @@ struct rtl8169_private { void __iomem *mmio_addr; /* memory map physical address */ struct pci_dev *pci_dev; /* Index of PCI device */ struct net_device *dev; + struct napi_struct napi; struct net_device_stats stats; /* statistics of net device */ spinlock_t lock; /* spin lock flag */ u32 msg_enable; @@ -443,13 +444,13 @@ static void rtl_set_rx_mode(struct net_device *dev); static void rtl8169_tx_timeout(struct net_device *dev); static struct net_device_stats *rtl8169_get_stats(struct net_device *dev); static int rtl8169_rx_interrupt(struct net_device *, struct rtl8169_private *, - void __iomem *); + void __iomem *, u32 budget); static int rtl8169_change_mtu(struct net_device *dev, int new_mtu); static void rtl8169_down(struct net_device *dev); static void rtl8169_rx_clear(struct rtl8169_private *tp); #ifdef CONFIG_R8169_NAPI -static int rtl8169_poll(struct net_device *dev, int *budget); +static int rtl8169_poll(struct napi_struct *napi, int budget); #endif static const unsigned int rtl8169_rx_config = @@ -1656,8 +1657,7 @@ rtl8169_init_one(struct pci_dev *pdev, const struct pci_device_id *ent) dev->set_mac_address = rtl_set_mac_address; #ifdef CONFIG_R8169_NAPI - dev->poll = rtl8169_poll; - dev->weight = R8169_NAPI_WEIGHT; + netif_napi_add(dev, &tp->napi, rtl8169_poll, R8169_NAPI_WEIGHT); #endif #ifdef CONFIG_R8169_VLAN @@ -1777,6 +1777,10 @@ static int rtl8169_open(struct net_device *dev) if (retval < 0) goto err_release_ring_2; +#ifdef CONFIG_R8169_NAPI + napi_enable(&tp->napi); +#endif + rtl_hw_start(dev); rtl8169_request_timer(dev); @@ -2082,7 +2086,9 @@ static int rtl8169_change_mtu(struct net_device *dev, int new_mtu) if (ret < 0) goto out; - netif_poll_enable(dev); +#ifdef CONFIG_R8169_NAPI + napi_enable(&tp->napi); +#endif rtl_hw_start(dev); @@ -2274,11 +2280,15 @@ static void rtl8169_wait_for_quiescence(struct net_device *dev) synchronize_irq(dev->irq); /* Wait for any pending NAPI task to complete */ - netif_poll_disable(dev); +#ifdef CONFIG_R8169_NAPI + napi_disable(&tp->napi); +#endif rtl8169_irq_mask_and_ack(ioaddr); - netif_poll_enable(dev); +#ifdef CONFIG_R8169_NAPI + napi_enable(&tp->napi); +#endif } static void rtl8169_reinit_task(struct work_struct *work) @@ -2322,7 +2332,7 @@ static void rtl8169_reset_task(struct work_struct *work) rtl8169_wait_for_quiescence(dev); - rtl8169_rx_interrupt(dev, tp, tp->mmio_addr); + rtl8169_rx_interrupt(dev, tp, tp->mmio_addr, ~(u32)0); rtl8169_tx_clear(tp); if (tp->dirty_rx == tp->cur_rx) { @@ -2636,14 +2646,14 @@ out: static int rtl8169_rx_interrupt(struct net_device *dev, struct rtl8169_private *tp, - void __iomem *ioaddr) + void __iomem *ioaddr, u32 budget) { unsigned int cur_rx, rx_left; unsigned int delta, count; cur_rx = tp->cur_rx; rx_left = NUM_RX_DESC + tp->dirty_rx - cur_rx; - rx_left = rtl8169_rx_quota(rx_left, (u32) dev->quota); + rx_left = rtl8169_rx_quota(rx_left, budget); for (; rx_left > 0; rx_left--, cur_rx++) { unsigned int entry = cur_rx % NUM_RX_DESC; @@ -2792,8 +2802,8 @@ static irqreturn_t rtl8169_interrupt(int irq, void *dev_instance) RTL_W16(IntrMask, tp->intr_event & ~tp->napi_event); tp->intr_mask = ~tp->napi_event; - if (likely(netif_rx_schedule_prep(dev))) - __netif_rx_schedule(dev); + if (likely(netif_rx_schedule_prep(dev, &tp->napi))) + __netif_rx_schedule(dev, &tp->napi); else if (netif_msg_intr(tp)) { printk(KERN_INFO "%s: interrupt %04x in poll\n", dev->name, status); @@ -2803,7 +2813,7 @@ static irqreturn_t rtl8169_interrupt(int irq, void *dev_instance) #else /* Rx interrupt */ if (status & (RxOK | RxOverflow | RxFIFOOver)) - rtl8169_rx_interrupt(dev, tp, ioaddr); + rtl8169_rx_interrupt(dev, tp, ioaddr, ~(u32)0); /* Tx interrupt */ if (status & (TxOK | TxErr)) @@ -2826,20 +2836,18 @@ out: } #ifdef CONFIG_R8169_NAPI -static int rtl8169_poll(struct net_device *dev, int *budget) +static int rtl8169_poll(struct napi_struct *napi, int budget) { - unsigned int work_done, work_to_do = min(*budget, dev->quota); - struct rtl8169_private *tp = netdev_priv(dev); + struct rtl8169_private *tp = container_of(napi, struct rtl8169_private, napi); + struct net_device *dev = tp->dev; void __iomem *ioaddr = tp->mmio_addr; + int work_done; - work_done = rtl8169_rx_interrupt(dev, tp, ioaddr); + work_done = rtl8169_rx_interrupt(dev, tp, ioaddr, (u32) budget); rtl8169_tx_interrupt(dev, tp, ioaddr); - *budget -= work_done; - dev->quota -= work_done; - - if (work_done < work_to_do) { - netif_rx_complete(dev); + if (work_done < budget) { + netif_rx_complete(dev, napi); tp->intr_mask = 0xffff; /* * 20040426: the barrier is not strictly required but the @@ -2851,7 +2859,7 @@ static int rtl8169_poll(struct net_device *dev, int *budget) RTL_W16(IntrMask, tp->intr_event); } - return (work_done >= work_to_do); + return work_done; } #endif @@ -2880,7 +2888,7 @@ core_down: synchronize_irq(dev->irq); if (!poll_locked) { - netif_poll_disable(dev); + napi_disable(&tp->napi); poll_locked++; } @@ -2918,8 +2926,6 @@ static int rtl8169_close(struct net_device *dev) free_irq(dev->irq, dev); - netif_poll_enable(dev); - pci_free_consistent(pdev, R8169_RX_RING_BYTES, tp->RxDescArray, tp->RxPhyAddr); pci_free_consistent(pdev, R8169_TX_RING_BYTES, tp->TxDescArray, diff --git a/drivers/net/s2io.c b/drivers/net/s2io.c index 24feb00..dd01232 100644 --- a/drivers/net/s2io.c +++ b/drivers/net/s2io.c @@ -2568,7 +2568,7 @@ static void free_rx_buffers(struct s2io_nic *sp) /** * s2io_poll - Rx interrupt handler for NAPI support - * @dev : pointer to the device structure. + * @napi : pointer to the napi structure. * @budget : The number of packets that were budgeted to be processed * during one pass through the 'Poll" function. * Description: @@ -2579,9 +2579,10 @@ static void free_rx_buffers(struct s2io_nic *sp) * 0 on success and 1 if there are No Rx packets to be processed. */ -static int s2io_poll(struct net_device *dev, int *budget) +static int s2io_poll(struct napi_struct *napi, int budget) { - struct s2io_nic *nic = dev->priv; + struct s2io_nic *nic = container_of(napi, struct s2io_nic, napi); + struct net_device *dev = nic->dev; int pkt_cnt = 0, org_pkts_to_process; struct mac_info *mac_control; struct config_param *config; @@ -2592,9 +2593,7 @@ static int s2io_poll(struct net_device *dev, int *budget) mac_control = &nic->mac_control; config = &nic->config; - nic->pkts_to_process = *budget; - if (nic->pkts_to_process > dev->quota) - nic->pkts_to_process = dev->quota; + nic->pkts_to_process = budget; org_pkts_to_process = nic->pkts_to_process; writeq(S2IO_MINUS_ONE, &bar0->rx_traffic_int); @@ -2608,12 +2607,8 @@ static int s2io_poll(struct net_device *dev, int *budget) goto no_rx; } } - if (!pkt_cnt) - pkt_cnt = 1; - dev->quota -= pkt_cnt; - *budget -= pkt_cnt; - netif_rx_complete(dev); + netif_rx_complete(dev, napi); for (i = 0; i < config->rx_ring_num; i++) { if (fill_rx_buffers(nic, i) == -ENOMEM) { @@ -2626,12 +2621,9 @@ static int s2io_poll(struct net_device *dev, int *budget) writeq(0x0, &bar0->rx_traffic_mask); readl(&bar0->rx_traffic_mask); atomic_dec(&nic->isr_cnt); - return 0; + return pkt_cnt; no_rx: - dev->quota -= pkt_cnt; - *budget -= pkt_cnt; - for (i = 0; i < config->rx_ring_num; i++) { if (fill_rx_buffers(nic, i) == -ENOMEM) { DBG_PRINT(INFO_DBG, "%s:Out of memory", dev->name); @@ -2640,7 +2632,7 @@ no_rx: } } atomic_dec(&nic->isr_cnt); - return 1; + return pkt_cnt; } #ifdef CONFIG_NET_POLL_CONTROLLER @@ -3809,6 +3801,8 @@ static int s2io_open(struct net_device *dev) netif_carrier_off(dev); sp->last_link_state = 0; + napi_enable(&sp->napi); + /* Initialize H/W and enable interrupts */ err = s2io_card_up(sp); if (err) { @@ -3828,6 +3822,7 @@ static int s2io_open(struct net_device *dev) return 0; hw_init_failed: + napi_disable(&sp->napi); if (sp->intr_type == MSI_X) { if (sp->entries) { kfree(sp->entries); @@ -3861,6 +3856,7 @@ static int s2io_close(struct net_device *dev) struct s2io_nic *sp = dev->priv; netif_stop_queue(dev); + napi_disable(&sp->napi); /* Reset card, kill tasklet and free Tx and Rx buffers. */ s2io_card_down(sp); @@ -4232,8 +4228,8 @@ static irqreturn_t s2io_isr(int irq, void *dev_id) if (napi) { if (reason & GEN_INTR_RXTRAFFIC) { - if ( likely ( netif_rx_schedule_prep(dev)) ) { - __netif_rx_schedule(dev); + if (likely (netif_rx_schedule_prep(dev, &sp->napi))) { + __netif_rx_schedule(dev, &sp->napi); writeq(S2IO_MINUS_ONE, &bar0->rx_traffic_mask); } else @@ -7215,8 +7211,7 @@ s2io_init_nic(struct pci_dev *pdev, const struct pci_device_id *pre) * will use eth_mac_addr() for dev->set_mac_address * mac address will be set every time dev->open() is called */ - dev->poll = s2io_poll; - dev->weight = 32; + netif_napi_add(dev, &sp->napi, s2io_poll, 32); #ifdef CONFIG_NET_POLL_CONTROLLER dev->poll_controller = s2io_netpoll; diff --git a/drivers/net/s2io.h b/drivers/net/s2io.h index 92983ee..420fefb 100644 --- a/drivers/net/s2io.h +++ b/drivers/net/s2io.h @@ -786,6 +786,7 @@ struct s2io_nic { */ int pkts_to_process; struct net_device *dev; + struct napi_struct napi; struct mac_info mac_control; struct config_param config; struct pci_dev *pdev; @@ -1019,7 +1020,7 @@ static void s2io_set_multicast(struct net_device *dev); static int rx_osm_handler(struct ring_info *ring_data, struct RxD_t * rxdp); static void s2io_link(struct s2io_nic * sp, int link); static void s2io_reset(struct s2io_nic * sp); -static int s2io_poll(struct net_device *dev, int *budget); +static int s2io_poll(struct napi_struct *napi, int budget); static void s2io_init_pci(struct s2io_nic * sp); static int s2io_set_mac_addr(struct net_device *dev, u8 * addr); static void s2io_alarm_handle(unsigned long data); diff --git a/drivers/net/sb1250-mac.c b/drivers/net/sb1250-mac.c index e7fdcf1..53845eb 100644 --- a/drivers/net/sb1250-mac.c +++ b/drivers/net/sb1250-mac.c @@ -238,6 +238,7 @@ struct sbmac_softc { */ struct net_device *sbm_dev; /* pointer to linux device */ + struct napi_struct napi; spinlock_t sbm_lock; /* spin lock */ struct timer_list sbm_timer; /* for monitoring MII */ struct net_device_stats sbm_stats; @@ -320,7 +321,7 @@ static struct net_device_stats *sbmac_get_stats(struct net_device *dev); static void sbmac_set_rx_mode(struct net_device *dev); static int sbmac_mii_ioctl(struct net_device *dev, struct ifreq *rq, int cmd); static int sbmac_close(struct net_device *dev); -static int sbmac_poll(struct net_device *poll_dev, int *budget); +static int sbmac_poll(struct napi_struct *napi, int budget); static int sbmac_mii_poll(struct sbmac_softc *s,int noisy); static int sbmac_mii_probe(struct net_device *dev); @@ -2154,20 +2155,13 @@ static irqreturn_t sbmac_intr(int irq,void *dev_instance) * Transmits on channel 0 */ - if (isr & (M_MAC_INT_CHANNEL << S_MAC_TX_CH0)) { + if (isr & (M_MAC_INT_CHANNEL << S_MAC_TX_CH0)) sbdma_tx_process(sc,&(sc->sbm_txdma), 0); -#ifdef CONFIG_NETPOLL_TRAP - if (netpoll_trap()) { - if (test_and_clear_bit(__LINK_STATE_XOFF, &dev->state)) - __netif_schedule(dev); - } -#endif - } if (isr & (M_MAC_INT_CHANNEL << S_MAC_RX_CH0)) { - if (netif_rx_schedule_prep(dev)) { + if (netif_rx_schedule_prep(dev, &sc->napi)) { __raw_writeq(0, sc->sbm_imr); - __netif_rx_schedule(dev); + __netif_rx_schedule(dev, &sc->napi); /* Depend on the exit from poll to reenable intr */ } else { @@ -2470,8 +2464,8 @@ static int sbmac_init(struct net_device *dev, int idx) dev->do_ioctl = sbmac_mii_ioctl; dev->tx_timeout = sbmac_tx_timeout; dev->watchdog_timeo = TX_TIMEOUT; - dev->poll = sbmac_poll; - dev->weight = 16; + + netif_napi_add(dev, &sc->napi, sbmac_poll, 16); dev->change_mtu = sb1250_change_mtu; #ifdef CONFIG_NET_POLL_CONTROLLER @@ -2537,6 +2531,8 @@ static int sbmac_open(struct net_device *dev) return -EINVAL; } + napi_enable(&sc->napi); + /* * Configure default speed */ @@ -2850,6 +2846,8 @@ static int sbmac_close(struct net_device *dev) unsigned long flags; int irq; + napi_disable(&sc->napi); + sbmac_set_channel_state(sc,sbmac_state_off); del_timer_sync(&sc->sbm_timer); @@ -2874,26 +2872,17 @@ static int sbmac_close(struct net_device *dev) return 0; } -static int sbmac_poll(struct net_device *dev, int *budget) +static int sbmac_poll(struct napi_struct *napi, int budget) { - int work_to_do; + struct sbmac_softc *sc = container_of(napi, struct sbmac_softc, napi); + struct net_device *dev = sc->sbm_dev; int work_done; - struct sbmac_softc *sc = netdev_priv(dev); - - work_to_do = min(*budget, dev->quota); - work_done = sbdma_rx_process(sc, &(sc->sbm_rxdma), work_to_do, 1); - - if (work_done > work_to_do) - printk(KERN_ERR "%s exceeded work_to_do budget=%d quota=%d work-done=%d\n", - sc->sbm_dev->name, *budget, dev->quota, work_done); + work_done = sbdma_rx_process(sc, &(sc->sbm_rxdma), budget, 1); sbdma_tx_process(sc, &(sc->sbm_txdma), 1); - *budget -= work_done; - dev->quota -= work_done; - - if (work_done < work_to_do) { - netif_rx_complete(dev); + if (work_done < budget) { + netif_rx_complete(dev, napi); #ifdef CONFIG_SBMAC_COALESCE __raw_writeq(((M_MAC_INT_EOP_COUNT | M_MAC_INT_EOP_TIMER) << S_MAC_TX_CH0) | @@ -2905,7 +2894,7 @@ static int sbmac_poll(struct net_device *dev, int *budget) #endif } - return (work_done >= work_to_do); + return work_done; } #if defined(SBMAC_ETH0_HWADDR) || defined(SBMAC_ETH1_HWADDR) || defined(SBMAC_ETH2_HWADDR) || defined(SBMAC_ETH3_HWADDR) diff --git a/drivers/net/sis190.c b/drivers/net/sis190.c index d470b19..038ccfb 100644 --- a/drivers/net/sis190.c +++ b/drivers/net/sis190.c @@ -47,24 +47,13 @@ #define PHY_ID_ANY 0x1f #define MII_REG_ANY 0x1f -#ifdef CONFIG_SIS190_NAPI -#define NAPI_SUFFIX "-NAPI" -#else -#define NAPI_SUFFIX "" -#endif - -#define DRV_VERSION "1.2" NAPI_SUFFIX +#define DRV_VERSION "1.2" #define DRV_NAME "sis190" #define SIS190_DRIVER_NAME DRV_NAME " Gigabit Ethernet driver " DRV_VERSION #define PFX DRV_NAME ": " -#ifdef CONFIG_SIS190_NAPI -#define sis190_rx_skb netif_receive_skb -#define sis190_rx_quota(count, quota) min(count, quota) -#else #define sis190_rx_skb netif_rx #define sis190_rx_quota(count, quota) count -#endif #define MAC_ADDR_LEN 6 @@ -1115,10 +1104,8 @@ static void sis190_down(struct net_device *dev) synchronize_irq(dev->irq); - if (!poll_locked) { - netif_poll_disable(dev); + if (!poll_locked) poll_locked++; - } synchronize_sched(); @@ -1137,8 +1124,6 @@ static int sis190_close(struct net_device *dev) free_irq(dev->irq, dev); - netif_poll_enable(dev); - pci_free_consistent(pdev, TX_RING_BYTES, tp->TxDescRing, tp->tx_dma); pci_free_consistent(pdev, RX_RING_BYTES, tp->RxDescRing, tp->rx_dma); diff --git a/drivers/net/skge.c b/drivers/net/skge.c index e3d8520..0bf46ed 100644 --- a/drivers/net/skge.c +++ b/drivers/net/skge.c @@ -2528,7 +2528,7 @@ static int skge_up(struct net_device *dev) skge_write32(hw, B0_IMSK, hw->intr_mask); spin_unlock_irq(&hw->hw_lock); - netif_poll_enable(dev); + napi_enable(&skge->napi); return 0; free_rx_ring: @@ -2558,7 +2558,7 @@ static int skge_down(struct net_device *dev) if (hw->chip_id == CHIP_ID_GENESIS && hw->phy_type == SK_PHY_XMAC) del_timer_sync(&skge->link_timer); - netif_poll_disable(dev); + napi_disable(&skge->napi); netif_carrier_off(dev); spin_lock_irq(&hw->hw_lock); @@ -3044,14 +3044,13 @@ static void skge_tx_done(struct net_device *dev) } } -static int skge_poll(struct net_device *dev, int *budget) +static int skge_poll(struct napi_struct *napi, int to_do) { - struct skge_port *skge = netdev_priv(dev); + struct skge_port *skge = container_of(napi, struct skge_port, napi); + struct net_device *dev = skge->netdev; struct skge_hw *hw = skge->hw; struct skge_ring *ring = &skge->rx_ring; struct skge_element *e; - unsigned long flags; - int to_do = min(dev->quota, *budget); int work_done = 0; skge_tx_done(dev); @@ -3082,20 +3081,16 @@ static int skge_poll(struct net_device *dev, int *budget) wmb(); skge_write8(hw, Q_ADDR(rxqaddr[skge->port], Q_CSR), CSR_START); - *budget -= work_done; - dev->quota -= work_done; - - if (work_done >= to_do) - return 1; /* not done */ - - spin_lock_irqsave(&hw->hw_lock, flags); - __netif_rx_complete(dev); - hw->intr_mask |= napimask[skge->port]; - skge_write32(hw, B0_IMSK, hw->intr_mask); - skge_read32(hw, B0_IMSK); - spin_unlock_irqrestore(&hw->hw_lock, flags); + if (work_done < to_do) { + spin_lock_irq(&hw->hw_lock); + __netif_rx_complete(dev, napi); + hw->intr_mask |= napimask[skge->port]; + skge_write32(hw, B0_IMSK, hw->intr_mask); + skge_read32(hw, B0_IMSK); + spin_unlock_irq(&hw->hw_lock); + } - return 0; + return work_done; } /* Parity errors seem to happen when Genesis is connected to a switch @@ -3252,8 +3247,9 @@ static irqreturn_t skge_intr(int irq, void *dev_id) } if (status & (IS_XA1_F|IS_R1_F)) { + struct skge_port *skge = netdev_priv(hw->dev[0]); hw->intr_mask &= ~(IS_XA1_F|IS_R1_F); - netif_rx_schedule(hw->dev[0]); + netif_rx_schedule(hw->dev[0], &skge->napi); } if (status & IS_PA_TO_TX1) @@ -3271,13 +3267,14 @@ static irqreturn_t skge_intr(int irq, void *dev_id) skge_mac_intr(hw, 0); if (hw->dev[1]) { + struct skge_port *skge = netdev_priv(hw->dev[1]); + if (status & (IS_XA2_F|IS_R2_F)) { hw->intr_mask &= ~(IS_XA2_F|IS_R2_F); - netif_rx_schedule(hw->dev[1]); + netif_rx_schedule(hw->dev[1], &skge->napi); } if (status & IS_PA_TO_RX2) { - struct skge_port *skge = netdev_priv(hw->dev[1]); ++skge->net_stats.rx_over_errors; skge_write16(hw, B3_PA_CTRL, PA_CLR_TO_RX2); } @@ -3569,8 +3566,6 @@ static struct net_device *skge_devinit(struct skge_hw *hw, int port, SET_ETHTOOL_OPS(dev, &skge_ethtool_ops); dev->tx_timeout = skge_tx_timeout; dev->watchdog_timeo = TX_WATCHDOG; - dev->poll = skge_poll; - dev->weight = NAPI_WEIGHT; #ifdef CONFIG_NET_POLL_CONTROLLER dev->poll_controller = skge_netpoll; #endif @@ -3580,6 +3575,7 @@ static struct net_device *skge_devinit(struct skge_hw *hw, int port, dev->features |= NETIF_F_HIGHDMA; skge = netdev_priv(dev); + netif_napi_add(dev, &skge->napi, skge_poll, NAPI_WEIGHT); skge->netdev = dev; skge->hw = hw; skge->msg_enable = netif_msg_init(debug, default_msg); diff --git a/drivers/net/skge.h b/drivers/net/skge.h index edd7146..dd0fd45 100644 --- a/drivers/net/skge.h +++ b/drivers/net/skge.h @@ -2448,6 +2448,7 @@ enum pause_status { struct skge_port { struct skge_hw *hw; struct net_device *netdev; + struct napi_struct napi; int port; u32 msg_enable; diff --git a/drivers/net/sky2.c b/drivers/net/sky2.c index ea117fc..a0d75b0 100644 --- a/drivers/net/sky2.c +++ b/drivers/net/sky2.c @@ -1130,7 +1130,7 @@ static void sky2_vlan_rx_register(struct net_device *dev, struct vlan_group *grp u16 port = sky2->port; netif_tx_lock_bh(dev); - netif_poll_disable(sky2->hw->dev[0]); + napi_disable(&hw->napi); sky2->vlgrp = grp; if (grp) { @@ -1145,7 +1145,7 @@ static void sky2_vlan_rx_register(struct net_device *dev, struct vlan_group *grp TX_VLAN_TAG_OFF); } - netif_poll_enable(sky2->hw->dev[0]); + napi_enable(&hw->napi); netif_tx_unlock_bh(dev); } #endif @@ -1385,9 +1385,13 @@ static int sky2_up(struct net_device *dev) sky2_prefetch_init(hw, txqaddr[port], sky2->tx_le_map, TX_RING_SIZE - 1); + napi_enable(&hw->napi); + err = sky2_rx_start(sky2); - if (err) + if (err) { + napi_disable(&hw->napi); goto err_out; + } /* Enable interrupts from phy/mac for port */ imask = sky2_read32(hw, B0_IMSK); @@ -1676,6 +1680,8 @@ static int sky2_down(struct net_device *dev) /* Stop more packets from being queued */ netif_stop_queue(dev); + napi_disable(&hw->napi); + /* Disable port IRQ */ imask = sky2_read32(hw, B0_IMSK); imask &= ~portirq_msk[port]; @@ -2016,7 +2022,7 @@ static int sky2_change_mtu(struct net_device *dev, int new_mtu) dev->trans_start = jiffies; /* prevent tx timeout */ netif_stop_queue(dev); - netif_poll_disable(hw->dev[0]); + napi_disable(&hw->napi); synchronize_irq(hw->pdev->irq); @@ -2043,12 +2049,16 @@ static int sky2_change_mtu(struct net_device *dev, int new_mtu) err = sky2_rx_start(sky2); sky2_write32(hw, B0_IMSK, imask); + /* Unconditionally re-enable NAPI because even if we + * call dev_close() that will do a napi_disable(). + */ + napi_enable(&hw->napi); + if (err) dev_close(dev); else { gma_write16(hw, port, GM_GP_CTRL, ctl); - netif_poll_enable(hw->dev[0]); netif_wake_queue(dev); } @@ -2544,18 +2554,15 @@ static int sky2_rx_hung(struct net_device *dev) static void sky2_watchdog(unsigned long arg) { struct sky2_hw *hw = (struct sky2_hw *) arg; - struct net_device *dev; /* Check for lost IRQ once a second */ if (sky2_read32(hw, B0_ISRC)) { - dev = hw->dev[0]; - if (__netif_rx_schedule_prep(dev)) - __netif_rx_schedule(dev); + napi_schedule(&hw->napi); } else { int i, active = 0; for (i = 0; i < hw->ports; i++) { - dev = hw->dev[i]; + struct net_device *dev = hw->dev[i]; if (!netif_running(dev)) continue; ++active; @@ -2605,11 +2612,11 @@ static void sky2_err_intr(struct sky2_hw *hw, u32 status) sky2_le_error(hw, 1, Q_XA2, TX_RING_SIZE); } -static int sky2_poll(struct net_device *dev0, int *budget) +static int sky2_poll(struct napi_struct *napi, int work_limit) { - struct sky2_hw *hw = ((struct sky2_port *) netdev_priv(dev0))->hw; - int work_done; + struct sky2_hw *hw = container_of(napi, struct sky2_hw, napi); u32 status = sky2_read32(hw, B0_Y2_SP_EISR); + int work_done; if (unlikely(status & Y2_IS_ERROR)) sky2_err_intr(hw, status); @@ -2620,31 +2627,27 @@ static int sky2_poll(struct net_device *dev0, int *budget) if (status & Y2_IS_IRQ_PHY2) sky2_phy_intr(hw, 1); - work_done = sky2_status_intr(hw, min(dev0->quota, *budget)); - *budget -= work_done; - dev0->quota -= work_done; + work_done = sky2_status_intr(hw, work_limit); /* More work? */ - if (hw->st_idx != sky2_read16(hw, STAT_PUT_IDX)) - return 1; + if (hw->st_idx == sky2_read16(hw, STAT_PUT_IDX)) { + /* Bug/Errata workaround? + * Need to kick the TX irq moderation timer. + */ + if (sky2_read8(hw, STAT_TX_TIMER_CTRL) == TIM_START) { + sky2_write8(hw, STAT_TX_TIMER_CTRL, TIM_STOP); + sky2_write8(hw, STAT_TX_TIMER_CTRL, TIM_START); + } - /* Bug/Errata workaround? - * Need to kick the TX irq moderation timer. - */ - if (sky2_read8(hw, STAT_TX_TIMER_CTRL) == TIM_START) { - sky2_write8(hw, STAT_TX_TIMER_CTRL, TIM_STOP); - sky2_write8(hw, STAT_TX_TIMER_CTRL, TIM_START); + napi_complete(napi); + sky2_read32(hw, B0_Y2_SP_LISR); } - netif_rx_complete(dev0); - - sky2_read32(hw, B0_Y2_SP_LISR); - return 0; + return work_done; } static irqreturn_t sky2_intr(int irq, void *dev_id) { struct sky2_hw *hw = dev_id; - struct net_device *dev0 = hw->dev[0]; u32 status; /* Reading this mask interrupts as side effect */ @@ -2653,8 +2656,8 @@ static irqreturn_t sky2_intr(int irq, void *dev_id) return IRQ_NONE; prefetch(&hw->st_le[hw->st_idx]); - if (likely(__netif_rx_schedule_prep(dev0))) - __netif_rx_schedule(dev0); + + napi_schedule(&hw->napi); return IRQ_HANDLED; } @@ -2663,10 +2666,8 @@ static irqreturn_t sky2_intr(int irq, void *dev_id) static void sky2_netpoll(struct net_device *dev) { struct sky2_port *sky2 = netdev_priv(dev); - struct net_device *dev0 = sky2->hw->dev[0]; - if (netif_running(dev) && __netif_rx_schedule_prep(dev0)) - __netif_rx_schedule(dev0); + napi_schedule(&sky2->hw->napi); } #endif @@ -2914,8 +2915,6 @@ static void sky2_restart(struct work_struct *work) sky2_write32(hw, B0_IMSK, 0); sky2_read32(hw, B0_IMSK); - netif_poll_disable(hw->dev[0]); - for (i = 0; i < hw->ports; i++) { dev = hw->dev[i]; if (netif_running(dev)) @@ -2924,7 +2923,6 @@ static void sky2_restart(struct work_struct *work) sky2_reset(hw); sky2_write32(hw, B0_IMSK, Y2_IS_BASE); - netif_poll_enable(hw->dev[0]); for (i = 0; i < hw->ports; i++) { dev = hw->dev[i]; @@ -3735,7 +3733,7 @@ static int sky2_debug_show(struct seq_file *seq, void *v) { struct net_device *dev = seq->private; const struct sky2_port *sky2 = netdev_priv(dev); - const struct sky2_hw *hw = sky2->hw; + struct sky2_hw *hw = sky2->hw; unsigned port = sky2->port; unsigned idx, last; int sop; @@ -3748,7 +3746,7 @@ static int sky2_debug_show(struct seq_file *seq, void *v) sky2_read32(hw, B0_IMSK), sky2_read32(hw, B0_Y2_SP_ICR)); - netif_poll_disable(hw->dev[0]); + napi_disable(&hw->napi); last = sky2_read16(hw, STAT_PUT_IDX); if (hw->st_idx == last) @@ -3818,7 +3816,7 @@ static int sky2_debug_show(struct seq_file *seq, void *v) last = sky2_read16(hw, Y2_QADDR(rxqaddr[port], PREF_UNIT_PUT_IDX)), sky2_read16(hw, Y2_QADDR(rxqaddr[port], PREF_UNIT_LAST_IDX))); - netif_poll_enable(hw->dev[0]); + napi_enable(&hw->napi); return 0; } @@ -3943,15 +3941,8 @@ static __devinit struct net_device *sky2_init_netdev(struct sky2_hw *hw, SET_ETHTOOL_OPS(dev, &sky2_ethtool_ops); dev->tx_timeout = sky2_tx_timeout; dev->watchdog_timeo = TX_WATCHDOG; - if (port == 0) - dev->poll = sky2_poll; - dev->weight = NAPI_WEIGHT; #ifdef CONFIG_NET_POLL_CONTROLLER - /* Network console (only works on port 0) - * because netpoll makes assumptions about NAPI - */ - if (port == 0) - dev->poll_controller = sky2_netpoll; + dev->poll_controller = sky2_netpoll; #endif sky2 = netdev_priv(dev); @@ -4166,6 +4157,7 @@ static int __devinit sky2_probe(struct pci_dev *pdev, err = -ENOMEM; goto err_out_free_pci; } + netif_napi_add(dev, &hw->napi, sky2_poll, NAPI_WEIGHT); if (!disable_msi && pci_enable_msi(pdev) == 0) { err = sky2_test_msi(hw); @@ -4288,8 +4280,6 @@ static int sky2_suspend(struct pci_dev *pdev, pm_message_t state) if (!hw) return 0; - netif_poll_disable(hw->dev[0]); - for (i = 0; i < hw->ports; i++) { struct net_device *dev = hw->dev[i]; struct sky2_port *sky2 = netdev_priv(dev); @@ -4356,8 +4346,6 @@ static int sky2_resume(struct pci_dev *pdev) } } - netif_poll_enable(hw->dev[0]); - return 0; out: dev_err(&pdev->dev, "resume failed (%d)\n", err); @@ -4374,7 +4362,7 @@ static void sky2_shutdown(struct pci_dev *pdev) if (!hw) return; - netif_poll_disable(hw->dev[0]); + napi_disable(&hw->napi); for (i = 0; i < hw->ports; i++) { struct net_device *dev = hw->dev[i]; diff --git a/drivers/net/sky2.h b/drivers/net/sky2.h index 8bc5c54..f18f875 100644 --- a/drivers/net/sky2.h +++ b/drivers/net/sky2.h @@ -2057,6 +2057,7 @@ struct sky2_port { struct sky2_hw { void __iomem *regs; struct pci_dev *pdev; + struct napi_struct napi; struct net_device *dev[2]; unsigned long flags; #define SKY2_HW_USE_MSI 0x00000001 diff --git a/drivers/net/spider_net.c b/drivers/net/spider_net.c index 82d837a..6d8f2bb 100644 --- a/drivers/net/spider_net.c +++ b/drivers/net/spider_net.c @@ -1278,34 +1278,26 @@ bad_desc: * (using netif_receive_skb). If all/enough packets are up, the driver * reenables interrupts and returns 0. If not, 1 is returned. */ -static int -spider_net_poll(struct net_device *netdev, int *budget) +static int spider_net_poll(struct napi_struct *napi, int budget) { - struct spider_net_card *card = netdev_priv(netdev); - int packets_to_do, packets_done = 0; - int no_more_packets = 0; - - packets_to_do = min(*budget, netdev->quota); - - while (packets_to_do) { - if (spider_net_decode_one_descr(card)) { - packets_done++; - packets_to_do--; - } else { - /* no more packets for the stack */ - no_more_packets = 1; + struct spider_net_card *card = container_of(napi, struct spider_net_card, napi); + struct net_device *netdev = card->netdev; + int packets_done = 0; + + while (packets_done < budget) { + if (!spider_net_decode_one_descr(card)) break; - } + + packets_done++; } if ((packets_done == 0) && (card->num_rx_ints != 0)) { - no_more_packets = spider_net_resync_tail_ptr(card); + if (!spider_net_resync_tail_ptr(card)) + packets_done = budget; spider_net_resync_head_ptr(card); } card->num_rx_ints = 0; - netdev->quota -= packets_done; - *budget -= packets_done; spider_net_refill_rx_chain(card); spider_net_enable_rxdmac(card); @@ -1313,14 +1305,13 @@ spider_net_poll(struct net_device *netdev, int *budget) /* if all packets are in the stack, enable interrupts and return 0 */ /* if not, return 1 */ - if (no_more_packets) { - netif_rx_complete(netdev); + if (packets_done < budget) { + netif_rx_complete(netdev, napi); spider_net_rx_irq_on(card); card->ignore_rx_ramfull = 0; - return 0; } - return 1; + return packets_done; } /** @@ -1560,7 +1551,8 @@ spider_net_handle_error_irq(struct spider_net_card *card, u32 status_reg) spider_net_refill_rx_chain(card); spider_net_enable_rxdmac(card); card->num_rx_ints ++; - netif_rx_schedule(card->netdev); + netif_rx_schedule(card->netdev, + &card->napi); } show_error = 0; break; @@ -1580,7 +1572,8 @@ spider_net_handle_error_irq(struct spider_net_card *card, u32 status_reg) spider_net_refill_rx_chain(card); spider_net_enable_rxdmac(card); card->num_rx_ints ++; - netif_rx_schedule(card->netdev); + netif_rx_schedule(card->netdev, + &card->napi); show_error = 0; break; @@ -1594,7 +1587,8 @@ spider_net_handle_error_irq(struct spider_net_card *card, u32 status_reg) spider_net_refill_rx_chain(card); spider_net_enable_rxdmac(card); card->num_rx_ints ++; - netif_rx_schedule(card->netdev); + netif_rx_schedule(card->netdev, + &card->napi); show_error = 0; break; @@ -1686,11 +1680,11 @@ spider_net_interrupt(int irq, void *ptr) if (status_reg & SPIDER_NET_RXINT ) { spider_net_rx_irq_off(card); - netif_rx_schedule(netdev); + netif_rx_schedule(netdev, &card->napi); card->num_rx_ints ++; } if (status_reg & SPIDER_NET_TXINT) - netif_rx_schedule(netdev); + netif_rx_schedule(netdev, &card->napi); if (status_reg & SPIDER_NET_LINKINT) spider_net_link_reset(netdev); @@ -2034,7 +2028,7 @@ spider_net_open(struct net_device *netdev) netif_start_queue(netdev); netif_carrier_on(netdev); - netif_poll_enable(netdev); + napi_enable(&card->napi); spider_net_enable_interrupts(card); @@ -2204,7 +2198,7 @@ spider_net_stop(struct net_device *netdev) { struct spider_net_card *card = netdev_priv(netdev); - netif_poll_disable(netdev); + napi_disable(&card->napi); netif_carrier_off(netdev); netif_stop_queue(netdev); del_timer_sync(&card->tx_timer); @@ -2304,9 +2298,6 @@ spider_net_setup_netdev_ops(struct net_device *netdev) /* tx watchdog */ netdev->tx_timeout = &spider_net_tx_timeout; netdev->watchdog_timeo = SPIDER_NET_WATCHDOG_TIMEOUT; - /* NAPI */ - netdev->poll = &spider_net_poll; - netdev->weight = SPIDER_NET_NAPI_WEIGHT; /* HW VLAN */ #ifdef CONFIG_NET_POLL_CONTROLLER /* poll controller */ @@ -2351,6 +2342,9 @@ spider_net_setup_netdev(struct spider_net_card *card) card->options.rx_csum = SPIDER_NET_RX_CSUM_DEFAULT; + netif_napi_add(netdev, &card->napi, + spider_net_poll, SPIDER_NET_NAPI_WEIGHT); + spider_net_setup_netdev_ops(netdev); netdev->features = NETIF_F_IP_CSUM | NETIF_F_LLTX; diff --git a/drivers/net/spider_net.h b/drivers/net/spider_net.h index dbbdb8c..a2fcdeb 100644 --- a/drivers/net/spider_net.h +++ b/drivers/net/spider_net.h @@ -466,6 +466,8 @@ struct spider_net_card { struct pci_dev *pdev; struct mii_phy phy; + struct napi_struct napi; + int medium; void __iomem *regs; diff --git a/drivers/net/starfire.c b/drivers/net/starfire.c index 8b64786..3b9336c 100644 --- a/drivers/net/starfire.c +++ b/drivers/net/starfire.c @@ -178,16 +178,13 @@ static int full_duplex[MAX_UNITS] = {0, }; #define skb_num_frags(skb) (skb_shinfo(skb)->nr_frags + 1) #ifdef HAVE_NETDEV_POLL -#define init_poll(dev) \ -do { \ - dev->poll = &netdev_poll; \ - dev->weight = max_interrupt_work; \ -} while (0) -#define netdev_rx(dev, ioaddr) \ +#define init_poll(dev, np) \ + netif_napi_add(dev, &np->napi, netdev_poll, max_interrupt_work) +#define netdev_rx(dev, np, ioaddr) \ do { \ u32 intr_enable; \ - if (netif_rx_schedule_prep(dev)) { \ - __netif_rx_schedule(dev); \ + if (netif_rx_schedule_prep(dev, &np->napi)) { \ + __netif_rx_schedule(dev, &np->napi); \ intr_enable = readl(ioaddr + IntrEnable); \ intr_enable &= ~(IntrRxDone | IntrRxEmpty); \ writel(intr_enable, ioaddr + IntrEnable); \ @@ -204,12 +201,12 @@ do { \ } while (0) #define netdev_receive_skb(skb) netif_receive_skb(skb) #define vlan_netdev_receive_skb(skb, vlgrp, vlid) vlan_hwaccel_receive_skb(skb, vlgrp, vlid) -static int netdev_poll(struct net_device *dev, int *budget); +static int netdev_poll(struct napi_struct *napi, int budget); #else /* not HAVE_NETDEV_POLL */ -#define init_poll(dev) +#define init_poll(dev, np) #define netdev_receive_skb(skb) netif_rx(skb) #define vlan_netdev_receive_skb(skb, vlgrp, vlid) vlan_hwaccel_rx(skb, vlgrp, vlid) -#define netdev_rx(dev, ioaddr) \ +#define netdev_rx(dev, np, ioaddr) \ do { \ int quota = np->dirty_rx + RX_RING_SIZE - np->cur_rx; \ __netdev_rx(dev, "a);\ @@ -599,6 +596,8 @@ struct netdev_private { struct tx_done_desc *tx_done_q; dma_addr_t tx_done_q_dma; unsigned int tx_done; + struct napi_struct napi; + struct net_device *dev; struct net_device_stats stats; struct pci_dev *pci_dev; #ifdef VLAN_SUPPORT @@ -791,6 +790,7 @@ static int __devinit starfire_init_one(struct pci_dev *pdev, dev->irq = irq; np = netdev_priv(dev); + np->dev = dev; np->base = base; spin_lock_init(&np->lock); pci_set_drvdata(pdev, dev); @@ -851,7 +851,7 @@ static int __devinit starfire_init_one(struct pci_dev *pdev, dev->hard_start_xmit = &start_tx; dev->tx_timeout = tx_timeout; dev->watchdog_timeo = TX_TIMEOUT; - init_poll(dev); + init_poll(dev, np); dev->stop = &netdev_close; dev->get_stats = &get_stats; dev->set_multicast_list = &set_rx_mode; @@ -1056,6 +1056,9 @@ static int netdev_open(struct net_device *dev) writel(np->intr_timer_ctrl, ioaddr + IntrTimerCtrl); +#ifdef HAVE_NETDEV_POLL + napi_enable(&np->napi); +#endif netif_start_queue(dev); if (debug > 1) @@ -1330,7 +1333,7 @@ static irqreturn_t intr_handler(int irq, void *dev_instance) handled = 1; if (intr_status & (IntrRxDone | IntrRxEmpty)) - netdev_rx(dev, ioaddr); + netdev_rx(dev, np, ioaddr); /* Scavenge the skbuff list based on the Tx-done queue. There are redundant checks here that may be cleaned up @@ -1531,36 +1534,35 @@ static int __netdev_rx(struct net_device *dev, int *quota) #ifdef HAVE_NETDEV_POLL -static int netdev_poll(struct net_device *dev, int *budget) +static int netdev_poll(struct napi_struct *napi, int budget) { + struct netdev_private *np = container_of(napi, struct netdev_private, napi); + struct net_device *dev = np->dev; u32 intr_status; - struct netdev_private *np = netdev_priv(dev); void __iomem *ioaddr = np->base; - int retcode = 0, quota = dev->quota; + int quota = budget; do { writel(IntrRxDone | IntrRxEmpty, ioaddr + IntrClear); - retcode = __netdev_rx(dev, "a); - *budget -= (dev->quota - quota); - dev->quota = quota; - if (retcode) + if (__netdev_rx(dev, "a)) goto out; intr_status = readl(ioaddr + IntrStatus); } while (intr_status & (IntrRxDone | IntrRxEmpty)); - netif_rx_complete(dev); + netif_rx_complete(dev, napi); intr_status = readl(ioaddr + IntrEnable); intr_status |= IntrRxDone | IntrRxEmpty; writel(intr_status, ioaddr + IntrEnable); out: if (debug > 5) - printk(KERN_DEBUG " exiting netdev_poll(): %d.\n", retcode); + printk(KERN_DEBUG " exiting netdev_poll(): %d.\n", + budget - quota); /* Restart Rx engine if stopped. */ - return retcode; + return budget - quota; } #endif /* HAVE_NETDEV_POLL */ @@ -1904,6 +1906,9 @@ static int netdev_close(struct net_device *dev) int i; netif_stop_queue(dev); +#ifdef HAVE_NETDEV_POLL + napi_disable(&np->napi); +#endif if (debug > 1) { printk(KERN_DEBUG "%s: Shutting down ethercard, Intr status %#8.8x.\n", diff --git a/drivers/net/sungem.c b/drivers/net/sungem.c index 4328038..bf821e9 100644 --- a/drivers/net/sungem.c +++ b/drivers/net/sungem.c @@ -19,7 +19,7 @@ * * gem_change_mtu() and gem_set_multicast() are called with a read_lock() * help by net/core/dev.c, thus they can't schedule. That means they can't - * call netif_poll_disable() neither, thus force gem_poll() to keep a spinlock + * call napi_disable() neither, thus force gem_poll() to keep a spinlock * where it could have been dropped. change_mtu especially would love also to * be able to msleep instead of horrid locked delays when resetting the HW, * but that read_lock() makes it impossible, unless I defer it's action to @@ -878,19 +878,20 @@ static int gem_rx(struct gem *gp, int work_to_do) return work_done; } -static int gem_poll(struct net_device *dev, int *budget) +static int gem_poll(struct napi_struct *napi, int budget) { - struct gem *gp = dev->priv; + struct gem *gp = container_of(napi, struct gem, napi); + struct net_device *dev = gp->dev; unsigned long flags; + int work_done; /* * NAPI locking nightmare: See comment at head of driver */ spin_lock_irqsave(&gp->lock, flags); + work_done = 0; do { - int work_to_do, work_done; - /* Handle anomalies */ if (gp->status & GREG_STAT_ABNORMAL) { if (gem_abnormal_irq(dev, gp, gp->status)) @@ -906,29 +907,25 @@ static int gem_poll(struct net_device *dev, int *budget) /* Run RX thread. We don't use any locking here, * code willing to do bad things - like cleaning the - * rx ring - must call netif_poll_disable(), which + * rx ring - must call napi_disable(), which * schedule_timeout()'s if polling is already disabled. */ - work_to_do = min(*budget, dev->quota); - - work_done = gem_rx(gp, work_to_do); - - *budget -= work_done; - dev->quota -= work_done; + work_done += gem_rx(gp, budget); - if (work_done >= work_to_do) - return 1; + if (work_done >= budget) + return work_done; spin_lock_irqsave(&gp->lock, flags); gp->status = readl(gp->regs + GREG_STAT); } while (gp->status & GREG_STAT_NAPI); - __netif_rx_complete(dev); + __netif_rx_complete(dev, napi); gem_enable_ints(gp); spin_unlock_irqrestore(&gp->lock, flags); - return 0; + + return work_done; } static irqreturn_t gem_interrupt(int irq, void *dev_id) @@ -946,17 +943,17 @@ static irqreturn_t gem_interrupt(int irq, void *dev_id) spin_lock_irqsave(&gp->lock, flags); - if (netif_rx_schedule_prep(dev)) { + if (netif_rx_schedule_prep(dev, &gp->napi)) { u32 gem_status = readl(gp->regs + GREG_STAT); if (gem_status == 0) { - netif_poll_enable(dev); + napi_enable(&gp->napi); spin_unlock_irqrestore(&gp->lock, flags); return IRQ_NONE; } gp->status = gem_status; gem_disable_ints(gp); - __netif_rx_schedule(dev); + __netif_rx_schedule(dev, &gp->napi); } spin_unlock_irqrestore(&gp->lock, flags); @@ -2284,7 +2281,7 @@ static void gem_reset_task(struct work_struct *work) mutex_lock(&gp->pm_mutex); - netif_poll_disable(gp->dev); + napi_disable(&gp->napi); spin_lock_irq(&gp->lock); spin_lock(&gp->tx_lock); @@ -2307,7 +2304,7 @@ static void gem_reset_task(struct work_struct *work) spin_unlock(&gp->tx_lock); spin_unlock_irq(&gp->lock); - netif_poll_enable(gp->dev); + napi_enable(&gp->napi); mutex_unlock(&gp->pm_mutex); } @@ -2324,6 +2321,8 @@ static int gem_open(struct net_device *dev) if (!gp->asleep) rc = gem_do_start(dev); gp->opened = (rc == 0); + if (gp->opened) + napi_enable(&gp->napi); mutex_unlock(&gp->pm_mutex); @@ -2334,9 +2333,7 @@ static int gem_close(struct net_device *dev) { struct gem *gp = dev->priv; - /* Note: we don't need to call netif_poll_disable() here because - * our caller (dev_close) already did it for us - */ + napi_disable(&gp->napi); mutex_lock(&gp->pm_mutex); @@ -2358,7 +2355,7 @@ static int gem_suspend(struct pci_dev *pdev, pm_message_t state) mutex_lock(&gp->pm_mutex); - netif_poll_disable(dev); + napi_disable(&gp->napi); printk(KERN_INFO "%s: suspending, WakeOnLan %s\n", dev->name, @@ -2482,7 +2479,7 @@ static int gem_resume(struct pci_dev *pdev) spin_unlock(&gp->tx_lock); spin_unlock_irqrestore(&gp->lock, flags); - netif_poll_enable(dev); + napi_enable(&gp->napi); mutex_unlock(&gp->pm_mutex); @@ -3121,8 +3118,7 @@ static int __devinit gem_init_one(struct pci_dev *pdev, dev->get_stats = gem_get_stats; dev->set_multicast_list = gem_set_multicast; dev->do_ioctl = gem_ioctl; - dev->poll = gem_poll; - dev->weight = 64; + netif_napi_add(dev, &gp->napi, gem_poll, 64); dev->ethtool_ops = &gem_ethtool_ops; dev->tx_timeout = gem_tx_timeout; dev->watchdog_timeo = 5 * HZ; diff --git a/drivers/net/sungem.h b/drivers/net/sungem.h index 58cf87c..76d760a 100644 --- a/drivers/net/sungem.h +++ b/drivers/net/sungem.h @@ -993,6 +993,7 @@ struct gem { u32 msg_enable; u32 status; + struct napi_struct napi; struct net_device_stats net_stats; int tx_fifo_sz; diff --git a/drivers/net/tc35815.c b/drivers/net/tc35815.c index ec41469..b5e0dff 100644 --- a/drivers/net/tc35815.c +++ b/drivers/net/tc35815.c @@ -414,6 +414,9 @@ enum tc35815_timer_state { struct tc35815_local { struct pci_dev *pci_dev; + struct net_device *dev; + struct napi_struct napi; + /* statistics */ struct net_device_stats stats; struct { @@ -566,7 +569,7 @@ static int tc35815_send_packet(struct sk_buff *skb, struct net_device *dev); static irqreturn_t tc35815_interrupt(int irq, void *dev_id); #ifdef TC35815_NAPI static int tc35815_rx(struct net_device *dev, int limit); -static int tc35815_poll(struct net_device *dev, int *budget); +static int tc35815_poll(struct napi_struct *napi, int budget); #else static void tc35815_rx(struct net_device *dev); #endif @@ -685,6 +688,7 @@ static int __devinit tc35815_init_one (struct pci_dev *pdev, SET_MODULE_OWNER(dev); SET_NETDEV_DEV(dev, &pdev->dev); lp = dev->priv; + lp->dev = dev; /* enable device (incl. PCI PM wakeup), and bus-mastering */ rc = pci_enable_device (pdev); @@ -738,8 +742,7 @@ static int __devinit tc35815_init_one (struct pci_dev *pdev, dev->tx_timeout = tc35815_tx_timeout; dev->watchdog_timeo = TC35815_TX_TIMEOUT; #ifdef TC35815_NAPI - dev->poll = tc35815_poll; - dev->weight = NAPI_WEIGHT; + netif_napi_add(dev, &lp->napi, tc35815_poll, NAPI_WEIGHT); #endif #ifdef CONFIG_NET_POLL_CONTROLLER dev->poll_controller = tc35815_poll_controller; @@ -748,8 +751,6 @@ static int __devinit tc35815_init_one (struct pci_dev *pdev, dev->irq = pdev->irq; dev->base_addr = (unsigned long) ioaddr; - /* dev->priv/lp zeroed and aligned in alloc_etherdev */ - lp = dev->priv; spin_lock_init(&lp->lock); lp->pci_dev = pdev; lp->boardtype = ent->driver_data; @@ -1237,6 +1238,10 @@ tc35815_open(struct net_device *dev) return -EAGAIN; } +#ifdef TC35815_NAPI + napi_enable(&lp->napi); +#endif + /* Reset the hardware here. Don't forget to set the station address. */ spin_lock_irq(&lp->lock); tc35815_chip_init(dev); @@ -1436,6 +1441,7 @@ static int tc35815_do_interrupt(struct net_device *dev, u32 status) static irqreturn_t tc35815_interrupt(int irq, void *dev_id) { struct net_device *dev = dev_id; + struct tc35815_local *lp = netdev_priv(dev); struct tc35815_regs __iomem *tr = (struct tc35815_regs __iomem *)dev->base_addr; #ifdef TC35815_NAPI @@ -1444,8 +1450,8 @@ static irqreturn_t tc35815_interrupt(int irq, void *dev_id) if (!(dmactl & DMA_IntMask)) { /* disable interrupts */ tc_writel(dmactl | DMA_IntMask, &tr->DMA_Ctl); - if (netif_rx_schedule_prep(dev)) - __netif_rx_schedule(dev); + if (netif_rx_schedule_prep(dev, &lp->napi)) + __netif_rx_schedule(dev, &lp->napi); else { printk(KERN_ERR "%s: interrupt taken in poll\n", dev->name); @@ -1726,13 +1732,12 @@ tc35815_rx(struct net_device *dev) } #ifdef TC35815_NAPI -static int -tc35815_poll(struct net_device *dev, int *budget) +static int tc35815_poll(struct napi_struct *napi, int budget) { - struct tc35815_local *lp = dev->priv; + struct tc35815_local *lp = container_of(napi, struct tc35815_local, napi); + struct net_device *dev = lp->dev; struct tc35815_regs __iomem *tr = (struct tc35815_regs __iomem *)dev->base_addr; - int limit = min(*budget, dev->quota); int received = 0, handled; u32 status; @@ -1744,23 +1749,19 @@ tc35815_poll(struct net_device *dev, int *budget) handled = tc35815_do_interrupt(dev, status, limit); if (handled >= 0) { received += handled; - limit -= handled; - if (limit <= 0) + if (received >= budget) break; } status = tc_readl(&tr->Int_Src); } while (status); spin_unlock(&lp->lock); - dev->quota -= received; - *budget -= received; - if (limit <= 0) - return 1; - - netif_rx_complete(dev); - /* enable interrupts */ - tc_writel(tc_readl(&tr->DMA_Ctl) & ~DMA_IntMask, &tr->DMA_Ctl); - return 0; + if (received < budget) { + netif_rx_complete(dev, napi); + /* enable interrupts */ + tc_writel(tc_readl(&tr->DMA_Ctl) & ~DMA_IntMask, &tr->DMA_Ctl); + } + return received; } #endif @@ -1949,7 +1950,11 @@ static int tc35815_close(struct net_device *dev) { struct tc35815_local *lp = dev->priv; + netif_stop_queue(dev); +#ifdef TC35815_NAPI + napi_disable(&lp->napi); +#endif /* Flush the Tx and disable Rx here. */ diff --git a/drivers/net/tg3.c b/drivers/net/tg3.c index 9034a05..ef1e3d1 100644 --- a/drivers/net/tg3.c +++ b/drivers/net/tg3.c @@ -574,7 +574,7 @@ static void tg3_restart_ints(struct tg3 *tp) static inline void tg3_netif_stop(struct tg3 *tp) { tp->dev->trans_start = jiffies; /* prevent tx timeout */ - netif_poll_disable(tp->dev); + napi_disable(&tp->napi); netif_tx_disable(tp->dev); } @@ -585,7 +585,7 @@ static inline void tg3_netif_start(struct tg3 *tp) * so long as all callers are assured to have free tx slots * (such as after tg3_init_hw) */ - netif_poll_enable(tp->dev); + napi_enable(&tp->napi); tp->hw_status->status |= SD_STATUS_UPDATED; tg3_enable_ints(tp); } @@ -3471,11 +3471,12 @@ next_pkt_nopost: return received; } -static int tg3_poll(struct net_device *netdev, int *budget) +static int tg3_poll(struct napi_struct *napi, int budget) { - struct tg3 *tp = netdev_priv(netdev); + struct tg3 *tp = container_of(napi, struct tg3, napi); + struct net_device *netdev = tp->dev; struct tg3_hw_status *sblk = tp->hw_status; - int done; + int work_done = 0; /* handle link change and other phy events */ if (!(tp->tg3_flags & @@ -3494,7 +3495,7 @@ static int tg3_poll(struct net_device *netdev, int *budget) if (sblk->idx[0].tx_consumer != tp->tx_cons) { tg3_tx(tp); if (unlikely(tp->tg3_flags & TG3_FLAG_TX_RECOVERY_PENDING)) { - netif_rx_complete(netdev); + netif_rx_complete(netdev, napi); schedule_work(&tp->reset_task); return 0; } @@ -3502,20 +3503,10 @@ static int tg3_poll(struct net_device *netdev, int *budget) /* run RX thread, within the bounds set by NAPI. * All RX "locking" is done by ensuring outside - * code synchronizes with dev->poll() + * code synchronizes with tg3->napi.poll() */ - if (sblk->idx[0].rx_producer != tp->rx_rcb_ptr) { - int orig_budget = *budget; - int work_done; - - if (orig_budget > netdev->quota) - orig_budget = netdev->quota; - - work_done = tg3_rx(tp, orig_budget); - - *budget -= work_done; - netdev->quota -= work_done; - } + if (sblk->idx[0].rx_producer != tp->rx_rcb_ptr) + work_done = tg3_rx(tp, budget); if (tp->tg3_flags & TG3_FLAG_TAGGED_STATUS) { tp->last_tag = sblk->status_tag; @@ -3524,13 +3515,12 @@ static int tg3_poll(struct net_device *netdev, int *budget) sblk->status &= ~SD_STATUS_UPDATED; /* if no more work, tell net stack and NIC we're done */ - done = !tg3_has_work(tp); - if (done) { - netif_rx_complete(netdev); + if (!tg3_has_work(tp)) { + netif_rx_complete(netdev, napi); tg3_restart_ints(tp); } - return (done ? 0 : 1); + return work_done; } static void tg3_irq_quiesce(struct tg3 *tp) @@ -3577,7 +3567,7 @@ static irqreturn_t tg3_msi_1shot(int irq, void *dev_id) prefetch(&tp->rx_rcb[tp->rx_rcb_ptr]); if (likely(!tg3_irq_sync(tp))) - netif_rx_schedule(dev); /* schedule NAPI poll */ + netif_rx_schedule(dev, &tp->napi); return IRQ_HANDLED; } @@ -3602,7 +3592,7 @@ static irqreturn_t tg3_msi(int irq, void *dev_id) */ tw32_mailbox(MAILBOX_INTERRUPT_0 + TG3_64BIT_REG_LOW, 0x00000001); if (likely(!tg3_irq_sync(tp))) - netif_rx_schedule(dev); /* schedule NAPI poll */ + netif_rx_schedule(dev, &tp->napi); return IRQ_RETVAL(1); } @@ -3644,7 +3634,7 @@ static irqreturn_t tg3_interrupt(int irq, void *dev_id) sblk->status &= ~SD_STATUS_UPDATED; if (likely(tg3_has_work(tp))) { prefetch(&tp->rx_rcb[tp->rx_rcb_ptr]); - netif_rx_schedule(dev); /* schedule NAPI poll */ + netif_rx_schedule(dev, &tp->napi); } else { /* No work, shared interrupt perhaps? re-enable * interrupts, and flush that PCI write @@ -3690,7 +3680,7 @@ static irqreturn_t tg3_interrupt_tagged(int irq, void *dev_id) tw32_mailbox_f(MAILBOX_INTERRUPT_0 + TG3_64BIT_REG_LOW, 0x00000001); if (tg3_irq_sync(tp)) goto out; - if (netif_rx_schedule_prep(dev)) { + if (netif_rx_schedule_prep(dev, &tp->napi)) { prefetch(&tp->rx_rcb[tp->rx_rcb_ptr]); /* Update last_tag to mark that this status has been * seen. Because interrupt may be shared, we may be @@ -3698,7 +3688,7 @@ static irqreturn_t tg3_interrupt_tagged(int irq, void *dev_id) * if tg3_poll() is not scheduled. */ tp->last_tag = sblk->status_tag; - __netif_rx_schedule(dev); + __netif_rx_schedule(dev, &tp->napi); } out: return IRQ_RETVAL(handled); @@ -3737,7 +3727,7 @@ static int tg3_restart_hw(struct tg3 *tp, int reset_phy) tg3_full_unlock(tp); del_timer_sync(&tp->timer); tp->irq_sync = 0; - netif_poll_enable(tp->dev); + napi_enable(&tp->napi); dev_close(tp->dev); tg3_full_lock(tp, 0); } @@ -3932,7 +3922,7 @@ static int tg3_start_xmit(struct sk_buff *skb, struct net_device *dev) len = skb_headlen(skb); /* We are running in BH disabled context with netif_tx_lock - * and TX reclaim runs via tp->poll inside of a software + * and TX reclaim runs via tp->napi.poll inside of a software * interrupt. Furthermore, IRQ processing runs lockless so we have * no IRQ context deadlocks to worry about either. Rejoice! */ @@ -4087,7 +4077,7 @@ static int tg3_start_xmit_dma_bug(struct sk_buff *skb, struct net_device *dev) len = skb_headlen(skb); /* We are running in BH disabled context with netif_tx_lock - * and TX reclaim runs via tp->poll inside of a software + * and TX reclaim runs via tp->napi.poll inside of a software * interrupt. Furthermore, IRQ processing runs lockless so we have * no IRQ context deadlocks to worry about either. Rejoice! */ @@ -7147,6 +7137,8 @@ static int tg3_open(struct net_device *dev) return err; } + napi_enable(&tp->napi); + tg3_full_lock(tp, 0); err = tg3_init_hw(tp, 1); @@ -7174,6 +7166,7 @@ static int tg3_open(struct net_device *dev) tg3_full_unlock(tp); if (err) { + napi_disable(&tp->napi); free_irq(tp->pdev->irq, dev); if (tp->tg3_flags2 & TG3_FLG2_USING_MSI) { pci_disable_msi(tp->pdev); @@ -7199,6 +7192,8 @@ static int tg3_open(struct net_device *dev) tg3_full_unlock(tp); + napi_disable(&tp->napi); + return err; } @@ -7460,6 +7455,7 @@ static int tg3_close(struct net_device *dev) { struct tg3 *tp = netdev_priv(dev); + napi_disable(&tp->napi); cancel_work_sync(&tp->reset_task); netif_stop_queue(dev); @@ -11900,9 +11896,8 @@ static int __devinit tg3_init_one(struct pci_dev *pdev, dev->set_mac_address = tg3_set_mac_addr; dev->do_ioctl = tg3_ioctl; dev->tx_timeout = tg3_tx_timeout; - dev->poll = tg3_poll; + netif_napi_add(dev, &tp->napi, tg3_poll, 64); dev->ethtool_ops = &tg3_ethtool_ops; - dev->weight = 64; dev->watchdog_timeo = TG3_TX_TIMEOUT; dev->change_mtu = tg3_change_mtu; dev->irq = pdev->irq; diff --git a/drivers/net/tg3.h b/drivers/net/tg3.h index 5c21f49..a6a23bb 100644 --- a/drivers/net/tg3.h +++ b/drivers/net/tg3.h @@ -2176,6 +2176,7 @@ struct tg3 { dma_addr_t tx_desc_mapping; /* begin "rx thread" cacheline section */ + struct napi_struct napi; void (*write32_rx_mbox) (struct tg3 *, u32, u32); u32 rx_rcb_ptr; diff --git a/drivers/net/tsi108_eth.c b/drivers/net/tsi108_eth.c index 1aabc91..b3069ee 100644 --- a/drivers/net/tsi108_eth.c +++ b/drivers/net/tsi108_eth.c @@ -79,6 +79,9 @@ struct tsi108_prv_data { void __iomem *regs; /* Base of normal regs */ void __iomem *phyregs; /* Base of register bank used for PHY access */ + struct net_device *dev; + struct napi_struct napi; + unsigned int phy; /* Index of PHY for this interface */ unsigned int irq_num; unsigned int id; @@ -837,13 +840,13 @@ static int tsi108_refill_rx(struct net_device *dev, int budget) return done; } -static int tsi108_poll(struct net_device *dev, int *budget) +static int tsi108_poll(struct napi_struct *napi, int budget) { - struct tsi108_prv_data *data = netdev_priv(dev); + struct tsi108_prv_data *data = container_of(napi, struct tsi108_prv_data, napi); + struct net_device *dev = data->dev; u32 estat = TSI_READ(TSI108_EC_RXESTAT); u32 intstat = TSI_READ(TSI108_EC_INTSTAT); - int total_budget = min(*budget, dev->quota); - int num_received = 0, num_filled = 0, budget_used; + int num_received = 0, num_filled = 0; intstat &= TSI108_INT_RXQUEUE0 | TSI108_INT_RXTHRESH | TSI108_INT_RXOVERRUN | TSI108_INT_RXERROR | TSI108_INT_RXWAIT; @@ -852,7 +855,7 @@ static int tsi108_poll(struct net_device *dev, int *budget) TSI_WRITE(TSI108_EC_INTSTAT, intstat); if (data->rxpending || (estat & TSI108_EC_RXESTAT_Q0_DESCINT)) - num_received = tsi108_complete_rx(dev, total_budget); + num_received = tsi108_complete_rx(dev, budget); /* This should normally fill no more slots than the number of * packets received in tsi108_complete_rx(). The exception @@ -867,7 +870,7 @@ static int tsi108_poll(struct net_device *dev, int *budget) */ if (data->rxfree < TSI108_RXRING_LEN) - num_filled = tsi108_refill_rx(dev, total_budget * 2); + num_filled = tsi108_refill_rx(dev, budget * 2); if (intstat & TSI108_INT_RXERROR) { u32 err = TSI_READ(TSI108_EC_RXERR); @@ -890,14 +893,9 @@ static int tsi108_poll(struct net_device *dev, int *budget) spin_unlock_irq(&data->misclock); } - budget_used = max(num_received, num_filled / 2); - - *budget -= budget_used; - dev->quota -= budget_used; - - if (budget_used != total_budget) { + if (num_received < budget) { data->rxpending = 0; - netif_rx_complete(dev); + netif_rx_complete(dev, napi); TSI_WRITE(TSI108_EC_INTMASK, TSI_READ(TSI108_EC_INTMASK) @@ -906,14 +904,11 @@ static int tsi108_poll(struct net_device *dev, int *budget) TSI108_INT_RXOVERRUN | TSI108_INT_RXERROR | TSI108_INT_RXWAIT)); - - /* IRQs are level-triggered, so no need to re-check */ - return 0; } else { data->rxpending = 1; } - return 1; + return num_received; } static void tsi108_rx_int(struct net_device *dev) @@ -931,7 +926,7 @@ static void tsi108_rx_int(struct net_device *dev) * from tsi108_check_rxring(). */ - if (netif_rx_schedule_prep(dev)) { + if (netif_rx_schedule_prep(dev, &data->napi)) { /* Mask, rather than ack, the receive interrupts. The ack * will happen in tsi108_poll(). */ @@ -942,7 +937,7 @@ static void tsi108_rx_int(struct net_device *dev) | TSI108_INT_RXTHRESH | TSI108_INT_RXOVERRUN | TSI108_INT_RXERROR | TSI108_INT_RXWAIT); - __netif_rx_schedule(dev); + __netif_rx_schedule(dev, &data->napi); } else { if (!netif_running(dev)) { /* This can happen if an interrupt occurs while the @@ -1401,6 +1396,8 @@ static int tsi108_open(struct net_device *dev) TSI_WRITE(TSI108_EC_TXQ_PTRLOW, data->txdma); tsi108_init_phy(dev); + napi_enable(&data->napi); + setup_timer(&data->timer, tsi108_timed_checker, (unsigned long)dev); mod_timer(&data->timer, jiffies + 1); @@ -1425,6 +1422,7 @@ static int tsi108_close(struct net_device *dev) struct tsi108_prv_data *data = netdev_priv(dev); netif_stop_queue(dev); + napi_disable(&data->napi); del_timer_sync(&data->timer); @@ -1562,6 +1560,7 @@ tsi108_init_one(struct platform_device *pdev) printk("tsi108_eth%d: probe...\n", pdev->id); data = netdev_priv(dev); + data->dev = dev; pr_debug("tsi108_eth%d:regs:phyresgs:phy:irq_num=0x%x:0x%x:0x%x:0x%x\n", pdev->id, einfo->regs, einfo->phyregs, @@ -1597,9 +1596,8 @@ tsi108_init_one(struct platform_device *pdev) dev->set_mac_address = tsi108_set_mac; dev->set_multicast_list = tsi108_set_rx_mode; dev->get_stats = tsi108_get_stats; - dev->poll = tsi108_poll; + netif_napi_add(dev, &data->napi, tsi108_poll, 64); dev->do_ioctl = tsi108_do_ioctl; - dev->weight = 64; /* 64 is more suitable for GigE interface - klai */ /* Apparently, the Linux networking code won't use scatter-gather * if the hardware doesn't do checksums. However, it's faster diff --git a/drivers/net/tulip/interrupt.c b/drivers/net/tulip/interrupt.c index 53efd66..3653314 100644 --- a/drivers/net/tulip/interrupt.c +++ b/drivers/net/tulip/interrupt.c @@ -103,28 +103,29 @@ int tulip_refill_rx(struct net_device *dev) void oom_timer(unsigned long data) { struct net_device *dev = (struct net_device *)data; - netif_rx_schedule(dev); + struct tulip_private *tp = netdev_priv(dev); + netif_rx_schedule(dev, &tp->napi); } -int tulip_poll(struct net_device *dev, int *budget) +int tulip_poll(struct napi_struct *napi, int budget) { - struct tulip_private *tp = netdev_priv(dev); + struct tulip_private *tp = container_of(napi, struct tulip_private, napi); + struct net_device *dev = tp->dev; int entry = tp->cur_rx % RX_RING_SIZE; - int rx_work_limit = *budget; + int work_done = 0; +#ifdef CONFIG_TULIP_NAPI_HW_MITIGATION int received = 0; +#endif if (!netif_running(dev)) goto done; - if (rx_work_limit > dev->quota) - rx_work_limit = dev->quota; - #ifdef CONFIG_TULIP_NAPI_HW_MITIGATION /* that one buffer is needed for mit activation; or might be a bug in the ring buffer code; check later -- JHS*/ - if (rx_work_limit >=RX_RING_SIZE) rx_work_limit--; + if (budget >=RX_RING_SIZE) budget--; #endif if (tulip_debug > 4) @@ -144,14 +145,13 @@ int tulip_poll(struct net_device *dev, int *budget) while ( ! (tp->rx_ring[entry].status & cpu_to_le32(DescOwned))) { s32 status = le32_to_cpu(tp->rx_ring[entry].status); - if (tp->dirty_rx + RX_RING_SIZE == tp->cur_rx) break; if (tulip_debug > 5) printk(KERN_DEBUG "%s: In tulip_rx(), entry %d %8.8x.\n", dev->name, entry, status); - if (--rx_work_limit < 0) + if (work_done++ >= budget) goto not_done; if ((status & 0x38008300) != 0x0300) { @@ -238,7 +238,9 @@ int tulip_poll(struct net_device *dev, int *budget) tp->stats.rx_packets++; tp->stats.rx_bytes += pkt_len; } - received++; +#ifdef CONFIG_TULIP_NAPI_HW_MITIGATION + received++; +#endif entry = (++tp->cur_rx) % RX_RING_SIZE; if (tp->cur_rx - tp->dirty_rx > RX_RING_SIZE/4) @@ -296,17 +298,15 @@ done: #endif /* CONFIG_TULIP_NAPI_HW_MITIGATION */ - dev->quota -= received; - *budget -= received; - tulip_refill_rx(dev); /* If RX ring is not full we are out of memory. */ - if (tp->rx_buffers[tp->dirty_rx % RX_RING_SIZE].skb == NULL) goto oom; + if (tp->rx_buffers[tp->dirty_rx % RX_RING_SIZE].skb == NULL) + goto oom; /* Remove us from polling list and enable RX intr. */ - netif_rx_complete(dev); + netif_rx_complete(dev, napi); iowrite32(tulip_tbl[tp->chip_id].valid_intrs, tp->base_addr+CSR7); /* The last op happens after poll completion. Which means the following: @@ -320,28 +320,20 @@ done: * processed irqs. But it must not result in losing events. */ - return 0; + return work_done; not_done: - if (!received) { - - received = dev->quota; /* Not to happen */ - } - dev->quota -= received; - *budget -= received; - if (tp->cur_rx - tp->dirty_rx > RX_RING_SIZE/2 || tp->rx_buffers[tp->dirty_rx % RX_RING_SIZE].skb == NULL) tulip_refill_rx(dev); - if (tp->rx_buffers[tp->dirty_rx % RX_RING_SIZE].skb == NULL) goto oom; - - return 1; + if (tp->rx_buffers[tp->dirty_rx % RX_RING_SIZE].skb == NULL) + goto oom; + return work_done; oom: /* Executed with RX ints disabled */ - /* Start timer, stop polling, but do not enable rx interrupts. */ mod_timer(&tp->oom_timer, jiffies+1); @@ -350,9 +342,9 @@ done: * before we did netif_rx_complete(). See? We would lose it. */ /* remove ourselves from the polling list */ - netif_rx_complete(dev); + netif_rx_complete(dev, napi); - return 0; + return work_done; } #else /* CONFIG_TULIP_NAPI */ @@ -534,7 +526,7 @@ irqreturn_t tulip_interrupt(int irq, void *dev_instance) rxd++; /* Mask RX intrs and add the device to poll list. */ iowrite32(tulip_tbl[tp->chip_id].valid_intrs&~RxPollInt, ioaddr + CSR7); - netif_rx_schedule(dev); + netif_rx_schedule(dev, &tp->napi); if (!(csr5&~(AbnormalIntr|NormalIntr|RxPollInt|TPLnkPass))) break; diff --git a/drivers/net/tulip/tulip.h b/drivers/net/tulip/tulip.h index 16f26a8..5a4d727 100644 --- a/drivers/net/tulip/tulip.h +++ b/drivers/net/tulip/tulip.h @@ -353,6 +353,7 @@ struct tulip_private { int chip_id; int revision; int flags; + struct napi_struct napi; struct net_device_stats stats; struct timer_list timer; /* Media selection timer. */ struct timer_list oom_timer; /* Out of memory timer. */ @@ -429,7 +430,7 @@ extern int tulip_rx_copybreak; irqreturn_t tulip_interrupt(int irq, void *dev_instance); int tulip_refill_rx(struct net_device *dev); #ifdef CONFIG_TULIP_NAPI -int tulip_poll(struct net_device *dev, int *budget); +int tulip_poll(struct napi_struct *napi, int budget); #endif diff --git a/drivers/net/tulip/tulip_core.c b/drivers/net/tulip/tulip_core.c index eca984f..7040a59 100644 --- a/drivers/net/tulip/tulip_core.c +++ b/drivers/net/tulip/tulip_core.c @@ -294,6 +294,10 @@ static void tulip_up(struct net_device *dev) int next_tick = 3*HZ; int i; +#ifdef CONFIG_TULIP_NAPI + napi_enable(&tp->napi); +#endif + /* Wake the chip from sleep/snooze mode. */ tulip_set_power_state (tp, 0, 0); @@ -728,6 +732,10 @@ static void tulip_down (struct net_device *dev) flush_scheduled_work(); +#ifdef CONFIG_TULIP_NAPI + napi_disable(&tp->napi); +#endif + del_timer_sync (&tp->timer); #ifdef CONFIG_TULIP_NAPI del_timer_sync (&tp->oom_timer); @@ -1606,8 +1614,7 @@ static int __devinit tulip_init_one (struct pci_dev *pdev, dev->tx_timeout = tulip_tx_timeout; dev->watchdog_timeo = TX_TIMEOUT; #ifdef CONFIG_TULIP_NAPI - dev->poll = tulip_poll; - dev->weight = 16; + netif_napi_add(dev, &tp->napi, tulip_poll, 16); #endif dev->stop = tulip_close; dev->get_stats = tulip_get_stats; diff --git a/drivers/net/typhoon.c b/drivers/net/typhoon.c index 0358720..0377b8b 100644 --- a/drivers/net/typhoon.c +++ b/drivers/net/typhoon.c @@ -284,6 +284,7 @@ struct typhoon { struct basic_ring rxLoRing; struct pci_dev * pdev; struct net_device * dev; + struct napi_struct napi; spinlock_t state_lock; struct vlan_group * vlgrp; struct basic_ring rxHiRing; @@ -1759,12 +1760,12 @@ typhoon_fill_free_ring(struct typhoon *tp) } static int -typhoon_poll(struct net_device *dev, int *total_budget) +typhoon_poll(struct napi_struct *napi, int budget) { - struct typhoon *tp = netdev_priv(dev); + struct typhoon *tp = container_of(napi, struct typhoon, napi); + struct net_device *dev = tp->dev; struct typhoon_indexes *indexes = tp->indexes; - int orig_budget = *total_budget; - int budget, work_done, done; + int work_done; rmb(); if(!tp->awaiting_resp && indexes->respReady != indexes->respCleared) @@ -1773,30 +1774,16 @@ typhoon_poll(struct net_device *dev, int *total_budget) if(le32_to_cpu(indexes->txLoCleared) != tp->txLoRing.lastRead) typhoon_tx_complete(tp, &tp->txLoRing, &indexes->txLoCleared); - if(orig_budget > dev->quota) - orig_budget = dev->quota; - - budget = orig_budget; work_done = 0; - done = 1; if(indexes->rxHiCleared != indexes->rxHiReady) { - work_done = typhoon_rx(tp, &tp->rxHiRing, &indexes->rxHiReady, + work_done += typhoon_rx(tp, &tp->rxHiRing, &indexes->rxHiReady, &indexes->rxHiCleared, budget); - budget -= work_done; } if(indexes->rxLoCleared != indexes->rxLoReady) { work_done += typhoon_rx(tp, &tp->rxLoRing, &indexes->rxLoReady, - &indexes->rxLoCleared, budget); - } - - if(work_done) { - *total_budget -= work_done; - dev->quota -= work_done; - - if(work_done >= orig_budget) - done = 0; + &indexes->rxLoCleared, budget - work_done); } if(le32_to_cpu(indexes->rxBuffCleared) == tp->rxBuffRing.lastWrite) { @@ -1804,14 +1791,14 @@ typhoon_poll(struct net_device *dev, int *total_budget) typhoon_fill_free_ring(tp); } - if(done) { - netif_rx_complete(dev); + if (work_done < budget) { + netif_rx_complete(dev, napi); iowrite32(TYPHOON_INTR_NONE, tp->ioaddr + TYPHOON_REG_INTR_MASK); typhoon_post_pci_writes(tp->ioaddr); } - return (done ? 0 : 1); + return work_done; } static irqreturn_t @@ -1828,10 +1815,10 @@ typhoon_interrupt(int irq, void *dev_instance) iowrite32(intr_status, ioaddr + TYPHOON_REG_INTR_STATUS); - if(netif_rx_schedule_prep(dev)) { + if (netif_rx_schedule_prep(dev, &tp->napi)) { iowrite32(TYPHOON_INTR_ALL, ioaddr + TYPHOON_REG_INTR_MASK); typhoon_post_pci_writes(ioaddr); - __netif_rx_schedule(dev); + __netif_rx_schedule(dev, &tp->napi); } else { printk(KERN_ERR "%s: Error, poll already scheduled\n", dev->name); @@ -2119,9 +2106,13 @@ typhoon_open(struct net_device *dev) if(err < 0) goto out_sleep; + napi_enable(&tp->napi); + err = typhoon_start_runtime(tp); - if(err < 0) + if(err < 0) { + napi_disable(&tp->napi); goto out_irq; + } netif_start_queue(dev); return 0; @@ -2150,6 +2141,7 @@ typhoon_close(struct net_device *dev) struct typhoon *tp = netdev_priv(dev); netif_stop_queue(dev); + napi_disable(&tp->napi); if(typhoon_stop_runtime(tp, WaitSleep) < 0) printk(KERN_ERR "%s: unable to stop runtime\n", dev->name); @@ -2521,8 +2513,7 @@ typhoon_init_one(struct pci_dev *pdev, const struct pci_device_id *ent) dev->stop = typhoon_close; dev->set_multicast_list = typhoon_set_rx_mode; dev->tx_timeout = typhoon_tx_timeout; - dev->poll = typhoon_poll; - dev->weight = 16; + netif_napi_add(dev, &tp->napi, typhoon_poll, 16); dev->watchdog_timeo = TX_TIMEOUT; dev->get_stats = typhoon_get_stats; dev->set_mac_address = typhoon_set_mac_address; diff --git a/drivers/net/ucc_geth.c b/drivers/net/ucc_geth.c index 9a38dfe..72f617b 100644 --- a/drivers/net/ucc_geth.c +++ b/drivers/net/ucc_geth.c @@ -3582,41 +3582,31 @@ static int ucc_geth_tx(struct net_device *dev, u8 txQ) } #ifdef CONFIG_UGETH_NAPI -static int ucc_geth_poll(struct net_device *dev, int *budget) +static int ucc_geth_poll(struct napi_struct *napi, int budget) { - struct ucc_geth_private *ugeth = netdev_priv(dev); + struct ucc_geth_private *ugeth = container_of(napi, struct ucc_geth_private, napi); + struct net_device *dev = ugeth->dev; struct ucc_geth_info *ug_info; - struct ucc_fast_private *uccf; - int howmany; - u8 i; - int rx_work_limit; - register u32 uccm; + int howmany, i; ug_info = ugeth->ug_info; - rx_work_limit = *budget; - if (rx_work_limit > dev->quota) - rx_work_limit = dev->quota; - howmany = 0; + for (i = 0; i < ug_info->numQueuesRx; i++) + howmany += ucc_geth_rx(ugeth, i, budget - howmany); - for (i = 0; i < ug_info->numQueuesRx; i++) { - howmany += ucc_geth_rx(ugeth, i, rx_work_limit); - } - - dev->quota -= howmany; - rx_work_limit -= howmany; - *budget -= howmany; + if (howmany < budget) { + struct ucc_fast_private *uccf; + u32 uccm; - if (rx_work_limit > 0) { - netif_rx_complete(dev); + netif_rx_complete(dev, napi); uccf = ugeth->uccf; uccm = in_be32(uccf->p_uccm); uccm |= UCCE_RX_EVENTS; out_be32(uccf->p_uccm, uccm); } - return (rx_work_limit > 0) ? 0 : 1; + return howmany; } #endif /* CONFIG_UGETH_NAPI */ @@ -3651,10 +3641,10 @@ static irqreturn_t ucc_geth_irq_handler(int irq, void *info) /* check for receive events that require processing */ if (ucce & UCCE_RX_EVENTS) { #ifdef CONFIG_UGETH_NAPI - if (netif_rx_schedule_prep(dev)) { - uccm &= ~UCCE_RX_EVENTS; + if (netif_rx_schedule_prep(dev, &ugeth->napi)) { + uccm &= ~UCCE_RX_EVENTS; out_be32(uccf->p_uccm, uccm); - __netif_rx_schedule(dev); + __netif_rx_schedule(dev, &ugeth->napi); } #else rx_mask = UCCE_RXBF_SINGLE_MASK; @@ -3717,12 +3707,15 @@ static int ucc_geth_open(struct net_device *dev) return err; } +#ifdef CONFIG_UGETH_NAPI + napi_enable(&ugeth->napi); +#endif err = ucc_geth_startup(ugeth); if (err) { if (netif_msg_ifup(ugeth)) ugeth_err("%s: Cannot configure net device, aborting.", dev->name); - return err; + goto out_err; } err = adjust_enet_interface(ugeth); @@ -3730,7 +3723,7 @@ static int ucc_geth_open(struct net_device *dev) if (netif_msg_ifup(ugeth)) ugeth_err("%s: Cannot configure net device, aborting.", dev->name); - return err; + goto out_err; } /* Set MACSTNADDR1, MACSTNADDR2 */ @@ -3748,7 +3741,7 @@ static int ucc_geth_open(struct net_device *dev) if (err) { if (netif_msg_ifup(ugeth)) ugeth_err("%s: Cannot initialize PHY, aborting.", dev->name); - return err; + goto out_err; } phy_start(ugeth->phydev); @@ -3761,7 +3754,7 @@ static int ucc_geth_open(struct net_device *dev) ugeth_err("%s: Cannot get IRQ for net device, aborting.", dev->name); ucc_geth_stop(ugeth); - return err; + goto out_err; } err = ugeth_enable(ugeth, COMM_DIR_RX_AND_TX); @@ -3769,12 +3762,18 @@ static int ucc_geth_open(struct net_device *dev) if (netif_msg_ifup(ugeth)) ugeth_err("%s: Cannot enable net device, aborting.", dev->name); ucc_geth_stop(ugeth); - return err; + goto out_err; } netif_start_queue(dev); return err; + +out_err: +#ifdef CONFIG_UGETH_NAPI + napi_disable(&ugeth->napi); +#endif + return err; } /* Stops the kernel queue, and halts the controller */ @@ -3784,6 +3783,10 @@ static int ucc_geth_close(struct net_device *dev) ugeth_vdbg("%s: IN", __FUNCTION__); +#ifdef CONFIG_UGETH_NAPI + napi_disable(&ugeth->napi); +#endif + ucc_geth_stop(ugeth); phy_disconnect(ugeth->phydev); @@ -3964,8 +3967,7 @@ static int ucc_geth_probe(struct of_device* ofdev, const struct of_device_id *ma dev->tx_timeout = ucc_geth_timeout; dev->watchdog_timeo = TX_TIMEOUT; #ifdef CONFIG_UGETH_NAPI - dev->poll = ucc_geth_poll; - dev->weight = UCC_GETH_DEV_WEIGHT; + netif_napi_add(dev, &ugeth->napi, ucc_geth_poll, UCC_GETH_DEV_WEIGHT); #endif /* CONFIG_UGETH_NAPI */ dev->stop = ucc_geth_close; dev->get_stats = ucc_geth_get_stats; diff --git a/drivers/net/ucc_geth.h b/drivers/net/ucc_geth.h index bb4dac8..0579ba0 100644 --- a/drivers/net/ucc_geth.h +++ b/drivers/net/ucc_geth.h @@ -1184,6 +1184,7 @@ struct ucc_geth_private { struct ucc_geth_info *ug_info; struct ucc_fast_private *uccf; struct net_device *dev; + struct napi_struct napi; struct net_device_stats stats; /* linux network statistics */ struct ucc_geth *ug_regs; struct ucc_geth_init_pram *p_init_enet_param_shadow; diff --git a/drivers/net/via-rhine.c b/drivers/net/via-rhine.c index b56dff2..7a58990 100644 --- a/drivers/net/via-rhine.c +++ b/drivers/net/via-rhine.c @@ -389,6 +389,8 @@ struct rhine_private { struct pci_dev *pdev; long pioaddr; + struct net_device *dev; + struct napi_struct napi; struct net_device_stats stats; spinlock_t lock; @@ -582,28 +584,25 @@ static void rhine_poll(struct net_device *dev) #endif #ifdef CONFIG_VIA_RHINE_NAPI -static int rhine_napipoll(struct net_device *dev, int *budget) +static int rhine_napipoll(struct napi_struct *napi, int budget) { - struct rhine_private *rp = netdev_priv(dev); + struct rhine_private *rp = container_of(napi, struct rhine_private, napi); + struct net_device *dev = rp->dev; void __iomem *ioaddr = rp->base; - int done, limit = min(dev->quota, *budget); + int work_done; - done = rhine_rx(dev, limit); - *budget -= done; - dev->quota -= done; + work_done = rhine_rx(dev, budget); - if (done < limit) { - netif_rx_complete(dev); + if (work_done < budget) { + netif_rx_complete(dev, napi); iowrite16(IntrRxDone | IntrRxErr | IntrRxEmpty| IntrRxOverflow | IntrRxDropped | IntrRxNoBuf | IntrTxAborted | IntrTxDone | IntrTxError | IntrTxUnderrun | IntrPCIErr | IntrStatsMax | IntrLinkChange, ioaddr + IntrEnable); - return 0; } - else - return 1; + return work_done; } #endif @@ -707,6 +706,7 @@ static int __devinit rhine_init_one(struct pci_dev *pdev, SET_NETDEV_DEV(dev, &pdev->dev); rp = netdev_priv(dev); + rp->dev = dev; rp->quirks = quirks; rp->pioaddr = pioaddr; rp->pdev = pdev; @@ -785,8 +785,7 @@ static int __devinit rhine_init_one(struct pci_dev *pdev, dev->poll_controller = rhine_poll; #endif #ifdef CONFIG_VIA_RHINE_NAPI - dev->poll = rhine_napipoll; - dev->weight = 64; + netif_napi_add(dev, &rp->napi, rhine_napipoll, 64); #endif if (rp->quirks & rqRhineI) dev->features |= NETIF_F_SG|NETIF_F_HW_CSUM; @@ -1061,7 +1060,9 @@ static void init_registers(struct net_device *dev) rhine_set_rx_mode(dev); - netif_poll_enable(dev); +#ifdef CONFIG_VIA_RHINE_NAPI + napi_enable(&rp->napi); +#endif /* Enable interrupts by setting the interrupt mask. */ iowrite16(IntrRxDone | IntrRxErr | IntrRxEmpty| IntrRxOverflow | @@ -1196,6 +1197,10 @@ static void rhine_tx_timeout(struct net_device *dev) /* protect against concurrent rx interrupts */ disable_irq(rp->pdev->irq); +#ifdef CONFIG_VIA_RHINE_NAPI + napi_disable(&rp->napi); +#endif + spin_lock(&rp->lock); /* clear all descriptors */ @@ -1324,7 +1329,7 @@ static irqreturn_t rhine_interrupt(int irq, void *dev_instance) IntrPCIErr | IntrStatsMax | IntrLinkChange, ioaddr + IntrEnable); - netif_rx_schedule(dev); + netif_rx_schedule(dev, &rp->napi); #else rhine_rx(dev, RX_RING_SIZE); #endif @@ -1837,7 +1842,9 @@ static int rhine_close(struct net_device *dev) spin_lock_irq(&rp->lock); netif_stop_queue(dev); - netif_poll_disable(dev); +#ifdef CONFIG_VIA_RHINE_NAPI + napi_disable(&rp->napi); +#endif if (debug > 1) printk(KERN_DEBUG "%s: Shutting down ethercard, " @@ -1936,6 +1943,9 @@ static int rhine_suspend(struct pci_dev *pdev, pm_message_t state) if (!netif_running(dev)) return 0; +#ifdef CONFIG_VIA_RHINE_NAPI + napi_disable(&rp->napi); +#endif netif_device_detach(dev); pci_save_state(pdev); diff --git a/drivers/net/xen-netfront.c b/drivers/net/xen-netfront.c index 4445810..70e551c 100644 --- a/drivers/net/xen-netfront.c +++ b/drivers/net/xen-netfront.c @@ -72,6 +72,7 @@ struct netfront_info { struct list_head list; struct net_device *netdev; + struct napi_struct napi; struct net_device_stats stats; struct xen_netif_tx_front_ring tx; @@ -185,7 +186,8 @@ static int xennet_can_sg(struct net_device *dev) static void rx_refill_timeout(unsigned long data) { struct net_device *dev = (struct net_device *)data; - netif_rx_schedule(dev); + struct netfront_info *np = netdev_priv(dev); + netif_rx_schedule(dev, &np->napi); } static int netfront_tx_slot_available(struct netfront_info *np) @@ -342,12 +344,14 @@ static int xennet_open(struct net_device *dev) memset(&np->stats, 0, sizeof(np->stats)); + napi_enable(&np->napi); + spin_lock_bh(&np->rx_lock); if (netif_carrier_ok(dev)) { xennet_alloc_rx_buffers(dev); np->rx.sring->rsp_event = np->rx.rsp_cons + 1; if (RING_HAS_UNCONSUMED_RESPONSES(&np->rx)) - netif_rx_schedule(dev); + netif_rx_schedule(dev, &np->napi); } spin_unlock_bh(&np->rx_lock); @@ -589,6 +593,7 @@ static int xennet_close(struct net_device *dev) { struct netfront_info *np = netdev_priv(dev); netif_stop_queue(np->netdev); + napi_disable(&np->napi); return 0; } @@ -872,15 +877,16 @@ static int handle_incoming_queue(struct net_device *dev, return packets_dropped; } -static int xennet_poll(struct net_device *dev, int *pbudget) +static int xennet_poll(struct napi_struct *napi, int budget) { - struct netfront_info *np = netdev_priv(dev); + struct netfront_info *np = container_of(napi, struct netfront_info, napi); + struct net_device *dev = np->netdev; struct sk_buff *skb; struct netfront_rx_info rinfo; struct xen_netif_rx_response *rx = &rinfo.rx; struct xen_netif_extra_info *extras = rinfo.extras; RING_IDX i, rp; - int work_done, budget, more_to_do = 1; + int work_done; struct sk_buff_head rxq; struct sk_buff_head errq; struct sk_buff_head tmpq; @@ -899,9 +905,6 @@ static int xennet_poll(struct net_device *dev, int *pbudget) skb_queue_head_init(&errq); skb_queue_head_init(&tmpq); - budget = *pbudget; - if (budget > dev->quota) - budget = dev->quota; rp = np->rx.sring->rsp_prod; rmb(); /* Ensure we see queued responses up to 'rp'. */ @@ -1006,22 +1009,21 @@ err: xennet_alloc_rx_buffers(dev); - *pbudget -= work_done; - dev->quota -= work_done; - if (work_done < budget) { + int more_to_do = 0; + local_irq_save(flags); RING_FINAL_CHECK_FOR_RESPONSES(&np->rx, more_to_do); if (!more_to_do) - __netif_rx_complete(dev); + __netif_rx_complete(dev, napi); local_irq_restore(flags); } spin_unlock(&np->rx_lock); - return more_to_do; + return work_done; } static int xennet_change_mtu(struct net_device *dev, int mtu) @@ -1201,10 +1203,9 @@ static struct net_device * __devinit xennet_create_dev(struct xenbus_device *dev netdev->hard_start_xmit = xennet_start_xmit; netdev->stop = xennet_close; netdev->get_stats = xennet_get_stats; - netdev->poll = xennet_poll; + netif_napi_add(netdev, &np->napi, xennet_poll, 64); netdev->uninit = xennet_uninit; netdev->change_mtu = xennet_change_mtu; - netdev->weight = 64; netdev->features = NETIF_F_IP_CSUM; SET_ETHTOOL_OPS(netdev, &xennet_ethtool_ops); @@ -1349,7 +1350,7 @@ static irqreturn_t xennet_interrupt(int irq, void *dev_id) xennet_tx_buf_gc(dev); /* Under tx_lock: protects access to rx shared-ring indexes. */ if (RING_HAS_UNCONSUMED_RESPONSES(&np->rx)) - netif_rx_schedule(dev); + netif_rx_schedule(dev, &np->napi); } spin_unlock_irqrestore(&np->tx_lock, flags); diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index e679b27..b93575d 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -31,6 +31,7 @@ #ifdef __KERNEL__ #include <linux/timer.h> +#include <linux/delay.h> #include <asm/atomic.h> #include <asm/cache.h> #include <asm/byteorder.h> @@ -38,6 +39,7 @@ #include <linux/device.h> #include <linux/percpu.h> #include <linux/dmaengine.h> +#include <linux/workqueue.h> struct vlan_group; struct ethtool_ops; @@ -258,7 +260,6 @@ enum netdev_state_t __LINK_STATE_PRESENT, __LINK_STATE_SCHED, __LINK_STATE_NOCARRIER, - __LINK_STATE_RX_SCHED, __LINK_STATE_LINKWATCH_PENDING, __LINK_STATE_DORMANT, __LINK_STATE_QDISC_RUNNING, @@ -278,6 +279,110 @@ struct netdev_boot_setup { extern int __init netdev_boot_setup(char *str); /* + * Structure for NAPI scheduling similar to tasklet but with weighting + */ +struct napi_struct { + /* The poll_list must only be managed by the entity which + * changes the state of the NAPI_STATE_SCHED bit. This means + * whoever atomically sets that bit can add this napi_struct + * to the per-cpu poll_list, and whoever clears that bit + * can remove from the list right before clearing the bit. + */ + struct list_head poll_list; + + unsigned long state; + int weight; + int (*poll)(struct napi_struct *, int); +#ifdef CONFIG_NETPOLL + spinlock_t poll_lock; + int poll_owner; + struct net_device *dev; + struct list_head dev_list; +#endif +}; + +enum +{ + NAPI_STATE_SCHED, /* Poll is scheduled */ +}; + +extern void FASTCALL(__napi_schedule(struct napi_struct *n)); + +/** + * napi_schedule_prep - check if napi can be scheduled + * @n: napi context + * + * Test if NAPI routine is already running, and if not mark + * it as running. This is used as a condition variable + * insure only one NAPI poll instance runs + */ +static inline int napi_schedule_prep(struct napi_struct *n) +{ + return !test_and_set_bit(NAPI_STATE_SCHED, &n->state); +} + +/** + * napi_schedule - schedule NAPI poll + * @n: napi context + * + * Schedule NAPI poll routine to be called if it is not already + * running. + */ +static inline void napi_schedule(struct napi_struct *n) +{ + if (napi_schedule_prep(n)) + __napi_schedule(n); +} + +/** + * napi_complete - NAPI processing complete + * @n: napi context + * + * Mark NAPI processing as complete. + */ +static inline void __napi_complete(struct napi_struct *n) +{ + BUG_ON(!test_bit(NAPI_STATE_SCHED, &n->state)); + list_del(&n->poll_list); + smp_mb__before_clear_bit(); + clear_bit(NAPI_STATE_SCHED, &n->state); +} + +static inline void napi_complete(struct napi_struct *n) +{ + local_irq_disable(); + __napi_complete(n); + local_irq_enable(); +} + +/** + * napi_disable - prevent NAPI from scheduling + * @n: napi context + * + * Stop NAPI from being scheduled on this context. + * Waits till any outstanding processing completes. + */ +static inline void napi_disable(struct napi_struct *n) +{ + while (test_and_set_bit(NAPI_STATE_SCHED, &n->state)) + msleep_interruptible(1); +} + +/** + * napi_enable - enable NAPI scheduling + * @n: napi context + * + * Resume NAPI from being scheduled on this context. + * Must be paired with napi_disable. + */ +static inline void napi_enable(struct napi_struct *n) +{ + BUG_ON(!test_bit(NAPI_STATE_SCHED, &n->state)); + smp_mb__before_clear_bit(); + clear_bit(NAPI_STATE_SCHED, &n->state); +} + +/* * The DEVICE structure. * Actually, this whole structure is a big mistake. It mixes I/O * data with strictly "high-level" data, and it has to know about @@ -319,6 +424,9 @@ struct net_device unsigned long state; struct list_head dev_list; +#ifdef CONFIG_NETPOLL + struct list_head napi_list; +#endif /* The device initialization function. Called only once. */ int (*init)(struct net_device *dev); @@ -430,12 +538,6 @@ struct net_device /* * Cache line mostly used on receive path (including eth_type_trans()) */ - struct list_head poll_list ____cacheline_aligned_in_smp; - /* Link to poll list */ - - int (*poll) (struct net_device *dev, int *quota); - int quota; - int weight; unsigned long last_rx; /* Time of last Rx */ /* Interface address info used in eth_type_trans() */ unsigned char dev_addr[MAX_ADDR_LEN]; /* hw address, (before bcast @@ -582,6 +684,12 @@ struct net_device #define NETDEV_ALIGN 32 #define NETDEV_ALIGN_CONST (NETDEV_ALIGN - 1) +/** + * netdev_priv - access network device private data + * @dev: network device + * + * Get network device private data + */ static inline void *netdev_priv(const struct net_device *dev) { return dev->priv; @@ -593,6 +701,23 @@ static inline void *netdev_priv(const struct net_device *dev) */ #define SET_NETDEV_DEV(net, pdev) ((net)->dev.parent = (pdev)) +static inline void netif_napi_add(struct net_device *dev, + struct napi_struct *napi, + int (*poll)(struct napi_struct *, int), + int weight) +{ + INIT_LIST_HEAD(&napi->poll_list); + napi->poll = poll; + napi->weight = weight; +#ifdef CONFIG_NETPOLL + napi->dev = dev; + list_add(&napi->dev_list, &dev->napi_list); + spin_lock_init(&napi->poll_lock); + napi->poll_owner = -1; +#endif + set_bit(NAPI_STATE_SCHED, &napi->state); +} + struct packet_type { __be16 type; /* This is really htons(ether_type). */ struct net_device *dev; /* NULL is wildcarded here */ @@ -678,7 +803,6 @@ static inline int unregister_gifconf(unsigned int family) * Incoming packets are placed on per-cpu queues so that * no locking is needed. */ - struct softnet_data { struct net_device *output_queue; @@ -686,7 +810,7 @@ struct softnet_data struct list_head poll_list; struct sk_buff *completion_queue; - struct net_device backlog_dev; /* Sorry. 8) */ + struct napi_struct backlog; #ifdef CONFIG_NET_DMA struct dma_chan *net_dma; #endif @@ -704,11 +828,24 @@ static inline void netif_schedule(struct net_device *dev) __netif_schedule(dev); } +/** + * netif_start_queue - allow transmit + * @dev: network device + * + * Allow upper layers to call the device hard_start_xmit routine. + */ static inline void netif_start_queue(struct net_device *dev) { clear_bit(__LINK_STATE_XOFF, &dev->state); } +/** + * netif_wake_queue - restart transmit + * @dev: network device + * + * Allow upper layers to call the device hard_start_xmit routine. + * Used for flow control when transmit resources are available. + */ static inline void netif_wake_queue(struct net_device *dev) { #ifdef CONFIG_NETPOLL_TRAP @@ -721,16 +858,35 @@ static inline void netif_wake_queue(struct net_device *dev) __netif_schedule(dev); } +/** + * netif_stop_queue - stop transmitted packets + * @dev: network device + * + * Stop upper layers calling the device hard_start_xmit routine. + * Used for flow control when transmit resources are unavailable. + */ static inline void netif_stop_queue(struct net_device *dev) { set_bit(__LINK_STATE_XOFF, &dev->state); } +/** + * netif_queue_stopped - test if transmit queue is flowblocked + * @dev: network device + * + * Test if transmit queue on device is currently unable to send. + */ static inline int netif_queue_stopped(const struct net_device *dev) { return test_bit(__LINK_STATE_XOFF, &dev->state); } +/** + * netif_running - test if up + * @dev: network device + * + * Test if the device has been brought up. + */ static inline int netif_running(const struct net_device *dev) { return test_bit(__LINK_STATE_START, &dev->state); @@ -742,6 +898,14 @@ static inline int netif_running(const struct net_device *dev) * done at the overall netdevice level. * Also test the device if we're multiqueue. */ + +/** + * netif_start_subqueue - allow sending packets on subqueue + * @dev: network device + * @queue_index: sub queue index + * + * Start individual transmit queue of a device with multiple transmit queues. + */ static inline void netif_start_subqueue(struct net_device *dev, u16 queue_index) { #ifdef CONFIG_NETDEVICES_MULTIQUEUE @@ -749,6 +913,13 @@ static inline void netif_start_subqueue(struct net_device *dev, u16 queue_index) #endif } +/** + * netif_stop_subqueue - stop sending packets on subqueue + * @dev: network device + * @queue_index: sub queue index + * + * Stop individual transmit queue of a device with multiple transmit queues. + */ static inline void netif_stop_subqueue(struct net_device *dev, u16 queue_index) { #ifdef CONFIG_NETDEVICES_MULTIQUEUE @@ -760,6 +931,13 @@ static inline void netif_stop_subqueue(struct net_device *dev, u16 queue_index) #endif } +/** + * netif_subqueue_stopped - test status of subqueue + * @dev: network device + * @queue_index: sub queue index + * + * Check individual transmit queue of a device with multiple transmit queues. + */ static inline int netif_subqueue_stopped(const struct net_device *dev, u16 queue_index) { @@ -771,6 +949,14 @@ static inline int netif_subqueue_stopped(const struct net_device *dev, #endif } + +/** + * netif_wake_subqueue - allow sending packets on subqueue + * @dev: network device + * @queue_index: sub queue index + * + * Resume individual transmit queue of a device with multiple transmit queues. + */ static inline void netif_wake_subqueue(struct net_device *dev, u16 queue_index) { #ifdef CONFIG_NETDEVICES_MULTIQUEUE @@ -784,6 +970,13 @@ static inline void netif_wake_subqueue(struct net_device *dev, u16 queue_index) #endif } +/** + * netif_is_multiqueue - test if device has multiple transmit queues + * @dev: network device + * + * Check if device has multiple transmit queues + * Always falls if NETDEVICE_MULTIQUEUE is not configured + */ static inline int netif_is_multiqueue(const struct net_device *dev) { #ifdef CONFIG_NETDEVICES_MULTIQUEUE @@ -796,20 +989,7 @@ static inline int netif_is_multiqueue(const struct net_device *dev) /* Use this variant when it is known for sure that it * is executing from interrupt context. */ -static inline void dev_kfree_skb_irq(struct sk_buff *skb) -{ - if (atomic_dec_and_test(&skb->users)) { - struct softnet_data *sd; - unsigned long flags; - - local_irq_save(flags); - sd = &__get_cpu_var(softnet_data); - skb->next = sd->completion_queue; - sd->completion_queue = skb; - raise_softirq_irqoff(NET_TX_SOFTIRQ); - local_irq_restore(flags); - } -} +extern void dev_kfree_skb_irq(struct sk_buff *skb); /* Use this variant in places where it could be invoked * either from interrupt or non-interrupt context. @@ -833,18 +1013,28 @@ extern int dev_set_mac_address(struct net_device *, extern int dev_hard_start_xmit(struct sk_buff *skb, struct net_device *dev); -extern void dev_init(void); - extern int netdev_budget; /* Called by rtnetlink.c:rtnl_unlock() */ extern void netdev_run_todo(void); +/** + * dev_put - release reference to device + * @dev: network device + * + * Hold reference to device to keep it from being freed. + */ static inline void dev_put(struct net_device *dev) { atomic_dec(&dev->refcnt); } +/** + * dev_hold - get reference to device + * @dev: network device + * + * Release reference to device to allow it to be freed. + */ static inline void dev_hold(struct net_device *dev) { atomic_inc(&dev->refcnt); @@ -861,6 +1051,12 @@ static inline void dev_hold(struct net_device *dev) extern void linkwatch_fire_event(struct net_device *dev); +/** + * netif_carrier_ok - test if carrier present + * @dev: network device + * + * Check if carrier is present on device + */ static inline int netif_carrier_ok(const struct net_device *dev) { return !test_bit(__LINK_STATE_NOCARRIER, &dev->state); @@ -872,30 +1068,66 @@ extern void netif_carrier_on(struct net_device *dev); extern void netif_carrier_off(struct net_device *dev); +/** + * netif_dormant_on - mark device as dormant. + * @dev: network device + * + * Mark device as dormant (as per RFC2863). + * + * The dormant state indicates that the relevant interface is not + * actually in a condition to pass packets (i.e., it is not 'up') but is + * in a "pending" state, waiting for some external event. For "on- + * demand" interfaces, this new state identifies the situation where the + * interface is waiting for events to place it in the up state. + * + */ static inline void netif_dormant_on(struct net_device *dev) { if (!test_and_set_bit(__LINK_STATE_DORMANT, &dev->state)) linkwatch_fire_event(dev); } +/** + * netif_dormant_off - set device as not dormant. + * @dev: network device + * + * Device is not in dormant state. + */ static inline void netif_dormant_off(struct net_device *dev) { if (test_and_clear_bit(__LINK_STATE_DORMANT, &dev->state)) linkwatch_fire_event(dev); } +/** + * netif_dormant - test if carrier present + * @dev: network device + * + * Check if carrier is present on device + */ static inline int netif_dormant(const struct net_device *dev) { return test_bit(__LINK_STATE_DORMANT, &dev->state); } +/** + * netif_oper_up - test if device is operational + * @dev: network device + * + * Check if carrier is operational + */ static inline int netif_oper_up(const struct net_device *dev) { return (dev->operstate == IF_OPER_UP || dev->operstate == IF_OPER_UNKNOWN /* backward compat */); } -/* Hot-plugging. */ +/** + * netif_device_present - is device available or removed + * @dev: network device + * + * Check if device has not been removed from system. + */ static inline int netif_device_present(struct net_device *dev) { return test_bit(__LINK_STATE_PRESENT, &dev->state); @@ -955,46 +1187,38 @@ static inline u32 netif_msg_init(int debug_value, int default_msg_enable_bits) return (1 << debug_value) - 1; } -/* Test if receive needs to be scheduled */ -static inline int __netif_rx_schedule_prep(struct net_device *dev) -{ - return !test_and_set_bit(__LINK_STATE_RX_SCHED, &dev->state); -} - /* Test if receive needs to be scheduled but only if up */ -static inline int netif_rx_schedule_prep(struct net_device *dev) +static inline int netif_rx_schedule_prep(struct net_device *dev, + struct napi_struct *napi) { - return netif_running(dev) && __netif_rx_schedule_prep(dev); + return netif_running(dev) && napi_schedule_prep(napi); } /* Add interface to tail of rx poll list. This assumes that _prep has * already been called and returned 1. */ - -extern void __netif_rx_schedule(struct net_device *dev); +static inline void __netif_rx_schedule(struct net_device *dev, + struct napi_struct *napi) +{ + dev_hold(dev); + __napi_schedule(napi); +} /* Try to reschedule poll. Called by irq handler. */ -static inline void netif_rx_schedule(struct net_device *dev) +static inline void netif_rx_schedule(struct net_device *dev, + struct napi_struct *napi) { - if (netif_rx_schedule_prep(dev)) - __netif_rx_schedule(dev); + if (netif_rx_schedule_prep(dev, napi)) + __netif_rx_schedule(dev, napi); } -/* Try to reschedule poll. Called by dev->poll() after netif_rx_complete(). - * Do not inline this? - */ -static inline int netif_rx_reschedule(struct net_device *dev, int undo) +/* Try to reschedule poll. Called by dev->poll() after netif_rx_complete(). */ +static inline int netif_rx_reschedule(struct net_device *dev, + struct napi_struct *napi) { - if (netif_rx_schedule_prep(dev)) { - unsigned long flags; - - dev->quota += undo; - - local_irq_save(flags); - list_add_tail(&dev->poll_list, &__get_cpu_var(softnet_data).poll_list); - __raise_softirq_irqoff(NET_RX_SOFTIRQ); - local_irq_restore(flags); + if (napi_schedule_prep(napi)) { + __netif_rx_schedule(dev, napi); return 1; } return 0; @@ -1003,12 +1227,11 @@ static inline int netif_rx_reschedule(struct net_device *dev, int undo) /* same as netif_rx_complete, except that local_irq_save(flags) * has already been issued */ -static inline void __netif_rx_complete(struct net_device *dev) +static inline void __netif_rx_complete(struct net_device *dev, + struct napi_struct *napi) { - BUG_ON(!test_bit(__LINK_STATE_RX_SCHED, &dev->state)); - list_del(&dev->poll_list); - smp_mb__before_clear_bit(); - clear_bit(__LINK_STATE_RX_SCHED, &dev->state); + __napi_complete(napi); + dev_put(dev); } /* Remove interface from poll list: it must be in the poll list @@ -1016,28 +1239,22 @@ static inline void __netif_rx_complete(struct net_device *dev) * it completes the work. The device cannot be out of poll list at this * moment, it is BUG(). */ -static inline void netif_rx_complete(struct net_device *dev) +static inline void netif_rx_complete(struct net_device *dev, + struct napi_struct *napi) { unsigned long flags; local_irq_save(flags); - __netif_rx_complete(dev); + __netif_rx_complete(dev, napi); local_irq_restore(flags); } -static inline void netif_poll_disable(struct net_device *dev) -{ - while (test_and_set_bit(__LINK_STATE_RX_SCHED, &dev->state)) - /* No hurry. */ - schedule_timeout_interruptible(1); -} - -static inline void netif_poll_enable(struct net_device *dev) -{ - smp_mb__before_clear_bit(); - clear_bit(__LINK_STATE_RX_SCHED, &dev->state); -} - +/** + * netif_tx_lock - grab network device transmit lock + * @dev: network device + * + * Get network device transmit lock + */ static inline void netif_tx_lock(struct net_device *dev) { spin_lock(&dev->_xmit_lock); diff --git a/include/linux/netpoll.h b/include/linux/netpoll.h index 29930b71..08dcc39 100644 --- a/include/linux/netpoll.h +++ b/include/linux/netpoll.h @@ -25,8 +25,6 @@ struct netpoll { struct netpoll_info { atomic_t refcnt; - spinlock_t poll_lock; - int poll_owner; int rx_flags; spinlock_t rx_lock; struct netpoll *rx_np; /* netpoll that registered an rx_hook */ @@ -64,32 +62,61 @@ static inline int netpoll_rx(struct sk_buff *skb) return ret; } -static inline void *netpoll_poll_lock(struct net_device *dev) +static inline int netpoll_receive_skb(struct sk_buff *skb) { + if (!list_empty(&skb->dev->napi_list)) + return netpoll_rx(skb); + return 0; +} + +static inline void *netpoll_poll_lock(struct napi_struct *napi) +{ + struct net_device *dev = napi->dev; + rcu_read_lock(); /* deal with race on ->npinfo */ - if (dev->npinfo) { - spin_lock(&dev->npinfo->poll_lock); - dev->npinfo->poll_owner = smp_processor_id(); - return dev->npinfo; + if (dev && dev->npinfo) { + spin_lock(&napi->poll_lock); + napi->poll_owner = smp_processor_id(); + return napi; } return NULL; } static inline void netpoll_poll_unlock(void *have) { - struct netpoll_info *npi = have; + struct napi_struct *napi = have; - if (npi) { - npi->poll_owner = -1; - spin_unlock(&npi->poll_lock); + if (napi) { + napi->poll_owner = -1; + spin_unlock(&napi->poll_lock); } rcu_read_unlock(); } +static inline void netpoll_netdev_init(struct net_device *dev) +{ + INIT_LIST_HEAD(&dev->napi_list); +} + #else -#define netpoll_rx(a) 0 -#define netpoll_poll_lock(a) NULL -#define netpoll_poll_unlock(a) +static inline int netpoll_rx(struct sk_buff *skb) +{ + return 0; +} +static inline int netpoll_receive_skb(struct sk_buff *skb) +{ + return 0; +} +static inline void *netpoll_poll_lock(struct napi_struct *napi) +{ + return NULL; +} +static inline void netpoll_poll_unlock(void *have) +{ +} +static inline void netpoll_netdev_init(struct net_device *dev) +{ +} #endif #endif diff --git a/net/core/dev.c b/net/core/dev.c index a76021c..29cf00c 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -220,7 +220,8 @@ static RAW_NOTIFIER_HEAD(netdev_chain); * Device drivers call our routines to queue packets here. We empty the * queue in the local softnet handler. */ -DEFINE_PER_CPU(struct softnet_data, softnet_data) = { NULL }; + +DEFINE_PER_CPU(struct softnet_data, softnet_data); #ifdef CONFIG_SYSFS extern int netdev_sysfs_init(void); @@ -1018,16 +1019,12 @@ int dev_close(struct net_device *dev) clear_bit(__LINK_STATE_START, &dev->state); /* Synchronize to scheduled poll. We cannot touch poll list, - * it can be even on different cpu. So just clear netif_running(), - * and wait when poll really will happen. Actually, the best place - * for this is inside dev->stop() after device stopped its irq - * engine, but this requires more changes in devices. */ - + * it can be even on different cpu. So just clear netif_running(). + * + * dev->stop() will invoke napi_disable() on all of it's + * napi_struct instances on this device. + */ smp_mb__after_clear_bit(); /* Commit netif_running(). */ - while (test_bit(__LINK_STATE_RX_SCHED, &dev->state)) { - /* No hurry. */ - msleep(1); - } /* * Call the device specific close. This cannot fail. @@ -1233,21 +1230,21 @@ void __netif_schedule(struct net_device *dev) } EXPORT_SYMBOL(__netif_schedule); -void __netif_rx_schedule(struct net_device *dev) +void dev_kfree_skb_irq(struct sk_buff *skb) { - unsigned long flags; + if (atomic_dec_and_test(&skb->users)) { + struct softnet_data *sd; + unsigned long flags; - local_irq_save(flags); - dev_hold(dev); - list_add_tail(&dev->poll_list, &__get_cpu_var(softnet_data).poll_list); - if (dev->quota < 0) - dev->quota += dev->weight; - else - dev->quota = dev->weight; - __raise_softirq_irqoff(NET_RX_SOFTIRQ); - local_irq_restore(flags); + local_irq_save(flags); + sd = &__get_cpu_var(softnet_data); + skb->next = sd->completion_queue; + sd->completion_queue = skb; + raise_softirq_irqoff(NET_TX_SOFTIRQ); + local_irq_restore(flags); + } } -EXPORT_SYMBOL(__netif_rx_schedule); +EXPORT_SYMBOL(dev_kfree_skb_irq); void dev_kfree_skb_any(struct sk_buff *skb) { @@ -1259,7 +1256,12 @@ void dev_kfree_skb_any(struct sk_buff *skb) EXPORT_SYMBOL(dev_kfree_skb_any); -/* Hot-plugging. */ +/** + * netif_device_detach - mark device as removed + * @dev: network device + * + * Mark device as removed from system and therefore no longer available. + */ void netif_device_detach(struct net_device *dev) { if (test_and_clear_bit(__LINK_STATE_PRESENT, &dev->state) && @@ -1269,6 +1271,12 @@ void netif_device_detach(struct net_device *dev) } EXPORT_SYMBOL(netif_device_detach); +/** + * netif_device_attach - mark device as attached + * @dev: network device + * + * Mark device as attached from system and restart if needed. + */ void netif_device_attach(struct net_device *dev) { if (!test_and_set_bit(__LINK_STATE_PRESENT, &dev->state) && @@ -1730,7 +1738,7 @@ enqueue: return NET_RX_SUCCESS; } - netif_rx_schedule(&queue->backlog_dev); + napi_schedule(&queue->backlog); goto enqueue; } @@ -1771,6 +1779,7 @@ static inline struct net_device *skb_bond(struct sk_buff *skb) return dev; } + static void net_tx_action(struct softirq_action *h) { struct softnet_data *sd = &__get_cpu_var(softnet_data); @@ -1927,7 +1936,7 @@ int netif_receive_skb(struct sk_buff *skb) __be16 type; /* if we've gotten here through NAPI, check netpoll */ - if (skb->dev->poll && netpoll_rx(skb)) + if (netpoll_receive_skb(skb)) return NET_RX_DROP; if (!skb->tstamp.tv64) @@ -2017,22 +2026,25 @@ out: return ret; } -static int process_backlog(struct net_device *backlog_dev, int *budget) +static int process_backlog(struct napi_struct *napi, int quota) { int work = 0; - int quota = min(backlog_dev->quota, *budget); struct softnet_data *queue = &__get_cpu_var(softnet_data); unsigned long start_time = jiffies; - backlog_dev->weight = weight_p; - for (;;) { + napi->weight = weight_p; + do { struct sk_buff *skb; struct net_device *dev; local_irq_disable(); skb = __skb_dequeue(&queue->input_pkt_queue); - if (!skb) - goto job_done; + if (!skb) { + __napi_complete(napi); + local_irq_enable(); + break; + } + local_irq_enable(); dev = skb->dev; @@ -2040,67 +2052,86 @@ static int process_backlog(struct net_device *backlog_dev, int *budget) netif_receive_skb(skb); dev_put(dev); + } while (++work < quota && jiffies == start_time); - work++; - - if (work >= quota || jiffies - start_time > 1) - break; - - } - - backlog_dev->quota -= work; - *budget -= work; - return -1; - -job_done: - backlog_dev->quota -= work; - *budget -= work; + return work; +} - list_del(&backlog_dev->poll_list); - smp_mb__before_clear_bit(); - netif_poll_enable(backlog_dev); +/** + * __napi_schedule - schedule for receive + * @napi: entry to schedule + * + * The entry's receive function will be scheduled to run + */ +void fastcall __napi_schedule(struct napi_struct *n) +{ + unsigned long flags; - local_irq_enable(); - return 0; + local_irq_save(flags); + list_add_tail(&n->poll_list, &__get_cpu_var(softnet_data).poll_list); + __raise_softirq_irqoff(NET_RX_SOFTIRQ); + local_irq_restore(flags); } +EXPORT_SYMBOL(__napi_schedule); + static void net_rx_action(struct softirq_action *h) { - struct softnet_data *queue = &__get_cpu_var(softnet_data); + struct list_head *list = &__get_cpu_var(softnet_data).poll_list; unsigned long start_time = jiffies; int budget = netdev_budget; void *have; local_irq_disable(); - while (!list_empty(&queue->poll_list)) { - struct net_device *dev; + while (!list_empty(list)) { + struct napi_struct *n; + int work, weight; - if (budget <= 0 || jiffies - start_time > 1) + /* If softirq window is exhuasted then punt. + * + * Note that this is a slight policy change from the + * previous NAPI code, which would allow up to 2 + * jiffies to pass before breaking out. The test + * used to be "jiffies - start_time > 1". + */ + if (unlikely(budget <= 0 || jiffies != start_time)) goto softnet_break; local_irq_enable(); - dev = list_entry(queue->poll_list.next, - struct net_device, poll_list); - have = netpoll_poll_lock(dev); + /* Even though interrupts have been re-enabled, this + * access is safe because interrupts can only add new + * entries to the tail of this list, and only ->poll() + * calls can remove this head entry from the list. + */ + n = list_entry(list->next, struct napi_struct, poll_list); - if (dev->quota <= 0 || dev->poll(dev, &budget)) { - netpoll_poll_unlock(have); - local_irq_disable(); - list_move_tail(&dev->poll_list, &queue->poll_list); - if (dev->quota < 0) - dev->quota += dev->weight; - else - dev->quota = dev->weight; - } else { - netpoll_poll_unlock(have); - dev_put(dev); - local_irq_disable(); - } + have = netpoll_poll_lock(n); + + weight = n->weight; + + work = n->poll(n, weight); + + WARN_ON_ONCE(work > weight); + + budget -= work; + + local_irq_disable(); + + /* Drivers must not modify the NAPI state if they + * consume the entire weight. In such cases this code + * still "owns" the NAPI instance and therefore can + * move the instance around on the list at-will. + */ + if (unlikely(work == weight)) + list_move_tail(&n->poll_list, list); + + netpoll_poll_unlock(have); } out: local_irq_enable(); + #ifdef CONFIG_NET_DMA /* * There may not be any more sk_buffs coming right now, so push @@ -2115,6 +2146,7 @@ out: } } #endif + return; softnet_break: @@ -3704,6 +3736,7 @@ struct net_device *alloc_netdev_mq(int sizeof_priv, const char *name, dev->egress_subqueue_count = queue_count; dev->get_stats = internal_stats; + netpoll_netdev_init(dev); setup(dev); strcpy(dev->name, name); return dev; @@ -4076,10 +4109,9 @@ static int __init net_dev_init(void) skb_queue_head_init(&queue->input_pkt_queue); queue->completion_queue = NULL; INIT_LIST_HEAD(&queue->poll_list); - set_bit(__LINK_STATE_START, &queue->backlog_dev.state); - queue->backlog_dev.weight = weight_p; - queue->backlog_dev.poll = process_backlog; - atomic_set(&queue->backlog_dev.refcnt, 1); + + queue->backlog.poll = process_backlog; + queue->backlog.weight = weight_p; } netdev_dma_register(); diff --git a/net/core/net-sysfs.c b/net/core/net-sysfs.c index 5c19b06..79159db 100644 --- a/net/core/net-sysfs.c +++ b/net/core/net-sysfs.c @@ -216,20 +216,6 @@ static ssize_t store_tx_queue_len(struct device *dev, return netdev_store(dev, attr, buf, len, change_tx_queue_len); } -NETDEVICE_SHOW(weight, fmt_dec); - -static int change_weight(struct net_device *net, unsigned long new_weight) -{ - net->weight = new_weight; - return 0; -} - -static ssize_t store_weight(struct device *dev, struct device_attribute *attr, - const char *buf, size_t len) -{ - return netdev_store(dev, attr, buf, len, change_weight); -} - static struct device_attribute net_class_attributes[] = { __ATTR(addr_len, S_IRUGO, show_addr_len, NULL), __ATTR(iflink, S_IRUGO, show_iflink, NULL), @@ -246,7 +232,6 @@ static struct device_attribute net_class_attributes[] = { __ATTR(flags, S_IRUGO | S_IWUSR, show_flags, store_flags), __ATTR(tx_queue_len, S_IRUGO | S_IWUSR, show_tx_queue_len, store_tx_queue_len), - __ATTR(weight, S_IRUGO | S_IWUSR, show_weight, store_weight), {} }; diff --git a/net/core/netpoll.c b/net/core/netpoll.c index de1b26a..abe6e3a 100644 --- a/net/core/netpoll.c +++ b/net/core/netpoll.c @@ -119,19 +119,22 @@ static __sum16 checksum_udp(struct sk_buff *skb, struct udphdr *uh, static void poll_napi(struct netpoll *np) { struct netpoll_info *npinfo = np->dev->npinfo; + struct napi_struct *napi; int budget = 16; - if (test_bit(__LINK_STATE_RX_SCHED, &np->dev->state) && - npinfo->poll_owner != smp_processor_id() && - spin_trylock(&npinfo->poll_lock)) { - npinfo->rx_flags |= NETPOLL_RX_DROP; - atomic_inc(&trapped); + list_for_each_entry(napi, &np->dev->napi_list, dev_list) { + if (test_bit(NAPI_STATE_SCHED, &napi->state) && + napi->poll_owner != smp_processor_id() && + spin_trylock(&napi->poll_lock)) { + npinfo->rx_flags |= NETPOLL_RX_DROP; + atomic_inc(&trapped); - np->dev->poll(np->dev, &budget); + napi->poll(napi, budget); - atomic_dec(&trapped); - npinfo->rx_flags &= ~NETPOLL_RX_DROP; - spin_unlock(&npinfo->poll_lock); + atomic_dec(&trapped); + npinfo->rx_flags &= ~NETPOLL_RX_DROP; + spin_unlock(&napi->poll_lock); + } } } @@ -157,7 +160,7 @@ void netpoll_poll(struct netpoll *np) /* Process pending work on NIC */ np->dev->poll_controller(np->dev); - if (np->dev->poll) + if (!list_empty(&np->dev->napi_list)) poll_napi(np); service_arp_queue(np->dev->npinfo); @@ -233,6 +236,17 @@ repeat: return skb; } +static int netpoll_owner_active(struct net_device *dev) +{ + struct napi_struct *napi; + + list_for_each_entry(napi, &dev->napi_list, dev_list) { + if (napi->poll_owner == smp_processor_id()) + return 1; + } + return 0; +} + static void netpoll_send_skb(struct netpoll *np, struct sk_buff *skb) { int status = NETDEV_TX_BUSY; @@ -246,8 +260,7 @@ static void netpoll_send_skb(struct netpoll *np, struct sk_buff *skb) } /* don't get messages out of order, and no recursion */ - if (skb_queue_len(&npinfo->txq) == 0 && - npinfo->poll_owner != smp_processor_id()) { + if (skb_queue_len(&npinfo->txq) == 0 && !netpoll_owner_active(dev)) { unsigned long flags; local_irq_save(flags); @@ -652,8 +665,6 @@ int netpoll_setup(struct netpoll *np) npinfo->rx_flags = 0; npinfo->rx_np = NULL; - spin_lock_init(&npinfo->poll_lock); - npinfo->poll_owner = -1; spin_lock_init(&npinfo->rx_lock); skb_queue_head_init(&npinfo->arp_tx); diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c index 4756d58..2b0b6fa 100644 --- a/net/core/rtnetlink.c +++ b/net/core/rtnetlink.c @@ -634,7 +634,6 @@ static int rtnl_fill_ifinfo(struct sk_buff *skb, struct net_device *dev, NLA_PUT_STRING(skb, IFLA_IFNAME, dev->name); NLA_PUT_U32(skb, IFLA_TXQLEN, dev->tx_queue_len); - NLA_PUT_U32(skb, IFLA_WEIGHT, dev->weight); NLA_PUT_U8(skb, IFLA_OPERSTATE, netif_running(dev) ? dev->operstate : IF_OPER_DOWN); NLA_PUT_U8(skb, IFLA_LINKMODE, dev->link_mode); @@ -834,9 +833,6 @@ static int do_setlink(struct net_device *dev, struct ifinfomsg *ifm, if (tb[IFLA_TXQLEN]) dev->tx_queue_len = nla_get_u32(tb[IFLA_TXQLEN]); - if (tb[IFLA_WEIGHT]) - dev->weight = nla_get_u32(tb[IFLA_WEIGHT]); - if (tb[IFLA_OPERSTATE]) set_operstate(dev, nla_get_u8(tb[IFLA_OPERSTATE])); @@ -1074,8 +1070,6 @@ replay: nla_len(tb[IFLA_BROADCAST])); if (tb[IFLA_TXQLEN]) dev->tx_queue_len = nla_get_u32(tb[IFLA_TXQLEN]); - if (tb[IFLA_WEIGHT]) - dev->weight = nla_get_u32(tb[IFLA_WEIGHT]); if (tb[IFLA_OPERSTATE]) set_operstate(dev, nla_get_u8(tb[IFLA_OPERSTATE])); if (tb[IFLA_LINKMODE]) diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index c81649c..e970e8e 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -256,6 +256,12 @@ static void dev_watchdog_down(struct net_device *dev) netif_tx_unlock_bh(dev); } +/** + * netif_carrier_on - set carrier + * @dev: network device + * + * Device has detected that carrier. + */ void netif_carrier_on(struct net_device *dev) { if (test_and_clear_bit(__LINK_STATE_NOCARRIER, &dev->state)) @@ -264,6 +270,12 @@ void netif_carrier_on(struct net_device *dev) __netdev_watchdog_up(dev); } +/** + * netif_carrier_off - clear carrier + * @dev: network device + * + * Device has detected loss of carrier. + */ void netif_carrier_off(struct net_device *dev) { if (!test_and_set_bit(__LINK_STATE_NOCARRIER, &dev->state)) |