summaryrefslogtreecommitdiff
path: root/drivers/firewire/core-transaction.c
AgeCommit message (Collapse)Author
2011-01-04firewire: use split transaction timeout only for split transactionsClemens Ladisch
Instead of starting the split transaction timeout timer when any request is submitted, start it only when the destination's ACK_PENDING has been received. This prevents us from using a timeout that is too short, and, if the controller's AT queue is emptying very slowly, from cancelling a packet that has not yet been sent. Signed-off-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-12-13firewire: make PHY packet header format consistentClemens Ladisch
Change the header of PHY packets to be sent to include a pseudo transaction code. This makes the header consistent with that of received PHY packets, and allows at_context_queue_packet() and log_ar_at_event() to see the packet type directly instead of having to deduce it from the header length or even from the header contents. Signed-off-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-12-12firewire: core: Update WARN usesJoe Perches
Add missing newlines. Signed-off-by: Joe Perches <joe@perches.com> Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-08-19firewire: core: do not use del_timer_sync() in interrupt contextClemens Ladisch
Because we might be in interrupt context, replace del_timer_sync() with del_timer(). If the timer is already running, we know that it will clean up the transaction, so we do not need to do any further processing in the normal transaction handler. Many thanks to Yong Zhang for diagnosing this. Reported-by: Stefan Richter <stefanr@s5r6.in-berlin.de> Signed-off-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-07-23firewire: core: fix upper bound of possible CSR allocationsStefan Richter
region->end is defined as an upper bound of the requested address range, exclusive --- i.e. as an address outside of the range in which the requested CSR is to be placed. Hence 0x0001,0000,0000,0000 is the biggest valid region->end, not 0x0000,ffff,ffff,fffc like the current check asserted. For simplicity, the fix drops the region->end & 3 test because there is no actual problem with these bits set in region->end. The allocated address range will be quadlet aligned and of a size of multiple quadlets due to the checks for region->start & 3 and handler->length & 3 alone. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-07-23firewire: cdev: add PHY packet receptionStefan Richter
Add an FW_CDEV_IOC_RECEIVE_PHY_PACKETS ioctl() and FW_CDEV_EVENT_PHY_PACKET_RECEIVED poll()/read() event for /dev/fw*. This can be used to get information from remote PHYs by remote access PHY packets. This is also the 2nd half of the functionality (the receive part) to support a userspace implementation of a VersaPHY transaction layer. Safety considerations: - PHY packets are generally broadcasts, hence some kind of elevated privileges should be required of a process to be able to listen in on PHY packets. This implementation assumes that a process that is allowed to open the /dev/fw* of a local node does have this privilege. There was an inconclusive discussion about introducing POSIX capabilities as a means to check for user privileges for these kinds of operations. Other limitations: - PHY packet reception may be switched on by ioctl() but cannot be switched off again. It would be trivial to provide an off switch, but this is not worth the code. The client should simply close() the fd then, or just ignore further events. - For sake of simplicity of API and kernel-side implementation, no filter per packet content is provided. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-07-23firewire: normalize status values in packet callbacksStefan Richter
core-transaction.c transmit_complete_callback() and close_transaction() expect packet callback status to be an ACK or RCODE, and ACKs get translated to RCODEs for transaction callbacks. An old comment on the packet callback API (been there from the initial submission of the stack) and the dummy_driver implementation of send_request/send_response deviated from this as they also included -ERRNO in the range of status values. Let's narrow status values down to ACK and RCODE to prevent surprises. RCODE_CANCELLED is chosen as the dummy_driver's RCODE as its meaning of "transaction timed out" comes closest to what happens when a transaction coincides with card removal. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-07-13firewire: core: integrate software-forced bus resets with bus managementStefan Richter
Bus resets which are triggered - by the kernel drivers after updates of the local nodes' config ROM, - by userspace software via ioctl shall be deferred until after >=2 seconds after the last bus reset. If multiple modifications of the local nodes' config ROM happen in a row, only a single bus reset should happen after them. When the local node's link goes from inactive to active or vice versa, and at the two occasions of bus resets mentioned above --- and if the current gap count differs from 63 --- the bus reset should be preceded by a PHY configuration packet that reaffirms the gap count. Otherwise a bus manager would have to reset the bus again right after that. This is necessary to promote bus stability, e.g. leave grace periods for allocations and reallocations of isochronous channels and bandwidth, SBP-2 reconnections etc.; see IEEE 1394 clause 8.2.1. This change implements all of the above by moving bus reset initiation into a delayed work (except for bus resets which are triggered by the bus manager workqueue job and are performed there immediately). It comes with a necessary addition to the card driver methods that allows to get the current gap count from PHY registers. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-07-13firewire: core: fix fw_send_request kerneldoc commentStefan Richter
The present inline documentation of the fw_send_request() in-kernel API refers to userland code that is not applicable to kernel drivers at all. Reported-by: Ben Gamari <bgamari.foss@gmail.com> While we are at fixing the whole documentation of fw_send_request(), also improve the rest of firewire-core's kerneldoc comments: - Add a bit of text concerning fw_run_transaction()'s call parameters. - Append () to function names and tab-align parameter descriptions as suggested by the example in Documentation/kernel-doc-nano-HOWTO.txt. - Remove kerneldoc markers from comments on static functions. - Remove outdated parameter descriptions at build_tree(). Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-06-20firewire: expose extended tcode of incoming lock requests to (userspace) driversJay Fenlason
When a remote device does a LOCK_REQUEST, the core does not pass the extended tcode to userspace. This patch makes it use the juju-specific tcodes listed in firewire-constants.h for incoming requests. Signed-off-by: Jay Fenlason <fenlason@redhat.com> This matches how tcode in the API for outbound requests is treated. Affects kernelspace and userspace drivers alike, but at the moment there are no kernespace drivers that receive lock requests. Split out from a combo patch, slightly reordered, changelog reworded. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-06-20firewire: remove an unused function argumentStefan Richter
void (*fw_address_callback_t)(..., int speed, ...) is the speed that a remote node chose to transmit a request to us. In case of split transactions, firewire-core will transmit the response at that speed. Upper layer drivers on the other hand (firewire-net, -sbp2, firedtv, and userspace drivers) cannot do anything useful with that speed datum, except log it for debug purposes. But data that is merely potentially (not even actually) used for debug purposes does not belong into the API. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-06-19firewire: core: remove unused variableStefan Richter
which caused gcc 4.6 to warn about variable 'destination' set but not used. Since the hardware ensures that we receive only response packets with proper destination node ID (in a given bus generation), we have no use for destination here in the core as well as in upper layers. (This is different with request packets. There we pass destination node ID to upper layers because they may for example need to check whether this was an unicast or broadcast request.) Reported-and-Tested-By: Justin P. Mattock <justinmattock@gmail.com> Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-06-19firewire: rename CSR access driver methodsStefan Richter
Rather than "read a Control and Status Registers (CSR) Architecture register" I prefer to say "read a Control and Status Register". Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-06-19firewire: core: combine some repeated codeStefan Richter
All of these CSRs have the same read/ write/ aynthing-else handling, except for CSR_PRIORITY_BUDGET which might not be implemented. The CSR_CYCLE_TIME read handler implementation accepted 4-byte-sized block write requests before this change but this is just silly; the register is only required to support quadlet read and write requests like the other r/w CSR core and Serial-Bus-dependent registers. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-06-19firewire: normalize STATE_CLEAR/SET CSR access interfaceStefan Richter
Push the maintenance of STATE_CLEAR/SET.abdicate down into the card driver. This way, the read/write_csr_reg driver method works uniformly across all CSR offsets. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-06-19firewire: replace get_features card driver hookStefan Richter
by feature variables in the fw_card struct. The hook appeared to be an unnecessary abstraction in the card driver interface. Cleaner would be to pass those feature flags as arguments to fw_card_initialize() or fw_card_add(), but the FairnessControl register is in the SCLK domain and may therefore not be accessible while Link Power Status is off, i.e. before the card->driver->enable call from fw_card_add(). Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-06-19firewire: 'add CSR_... support' addendumStefan Richter
Add a comment on which of the conflicting NODE_IDS specifications we implement. Reduce a comment on rather irrelevant register bits that can all be looked up in the spec (or from now on in the code history). Directly include the required indirectly included bug.h. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-06-10firewire: core: add CSR abdicate supportClemens Ladisch
Implement the abdicate bit, which is required for bus manager capable nodes and tested by the Base 1394 Test Suite. Finally, something to do at a command reset! :-) Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
2010-06-10firewire: add CSR cmstr supportClemens Ladisch
Implement the cmstr bit, which is required for cycle master capable nodes and tested for by the Base 1394 Test Suite. This bit allows the bus master to disable cycle start packets; there are bus master implementations that actually do this. Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
2010-06-10firewire: core: add CSR MAINT_UTILITY supportClemens Ladisch
Implement the MAIN_UTILITY register, which is utterly optional but useful as a safe target for diagnostic read/write/broadcast transactions. Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
2010-06-10firewire: add CSR PRIORITY_BUDGET supportClemens Ladisch
If supported by the OHCI controller, implement the PRIORITY_BUDGET register, which is required for nodes that can use asynchronous priority arbitration. To allow the core to determine what features the lowlevel device supports, add a new card driver callback. Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
2010-06-10firewire: add CSR BUSY_TIMEOUT supportClemens Ladisch
Implement the BUSY_TIMEOUT register, which is required for nodes that support retries. Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
2010-06-10firewire: add CSR BUS_TIME supportClemens Ladisch
Implement the BUS_TIME register, which is required for cycle master capable nodes and tested for by the Base 1393 Test Suite. Even when there is not yet bus master initialization support, this register allows us to work together with other bus masters. Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
2010-06-10firewire: add CSR CYCLE_TIME write supportClemens Ladisch
The specification requires that CYCLE_TIME is writable so that it can be initialized, so we better implement it. Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
2010-06-10firewire: core: add CSR SPLIT_TIMEOUT supportClemens Ladisch
Implement the SPLIT_TIMEOUT registers. Besides being required by the spec, this is desirable for some IIDC devices and necessary for many audio devices to be able to increase the timeout from userspace. Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
2010-06-10firewire: core: add CSR RESET_START supportClemens Ladisch
This implements the RESET_START register (as a dummy) to make the Base 1394 Test Suite happy. Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
2010-06-10firewire: add CSR NODE_IDS supportClemens Ladisch
The NODE_IDS register, and especially its bus_id field, is quite useless because 1394.1 requires that the bus_id field always stays 0x3ff. However, the 1394 specification requires this register on all transaction capable nodes, and the Base 1394 Test Suite tests for it, so we better implement it. Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
2010-06-10firewire: add read_csr_reg driver callbackClemens Ladisch
To prepare for the following additions of more OHCI-implemented CSR registers, replace the get_cycle_time driver callback with a generic CSR register callback. Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
2010-06-10firewire: core: add CSR STATE_CLEAR/STATE_SET supportClemens Ladisch
The state registers are zero and read-only in this implementation, so they are not of much use. However, the specification requires that they are present for transaction capable nodes, and the Base 1394 Test Suite tests for them, so we better implement them. Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
2010-06-09firewire: core: trivial fix for warning stringsStefan Richter
WARN's format string argument should not carry a printk level prefix. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-06-09firewire: check cdev response lengthClemens Ladisch
Add a check that the data length in the SEND_RESPONSE ioctl is correct. Incidentally, this also fixes the previously wrong response length of software-handled lock requests. Signed-off-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-05-18firewire: core: use separate timeout for each transactionClemens Ladisch
Using a single timeout for all transaction that need to be flushed does not work if the submission of new transactions can defer the timeout indefinitely into the future. We need to have timeouts that do not change due to other transactions; the simplest way to do this is with a separate timer for each transaction. Signed-off-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de> (+ one lockdep annotation)
2010-05-18firewire: core: Fix tlabel exhaustion problemPeter Hurley
fw_core_handle_response() was not properly clearing tlabel_mask. This was resulting in premature tlabel exhaustion. Signed-off-by: Peter Hurley <phurley@charter.net> This fixes an omission in 2.6.31-rc1 commit 1e626fdc "firewire: core: use more outbound tlabels" which prevented to really use 64 instead of 32 transaction labels, as soon as split transactions occurred that had their AR-resp tasklet run after the AT-req tasklet. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-04-19firewire: core: make transaction label allocation more robustClemens Ladisch
If one request is so long-lived that it does not get a response before the following 63 requests, its bit in tlabel_mask is still set when the next request tries to allocate a transaction label for that number. In this state, while the first request is not completed or timed out, no new requests can be submitted. To fix this, skip over any label still in use, and do not error out unless we have entirely run out of labels. Signed-off-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-02-20firewire: remove incomplete Bus_Time CSR supportStefan Richter
The current implementation of Bus_Time read access was buggy since it did not ensure that Bus_Time.second_count_hi and second_count_lo came from the same 128 seconds period. Reported-by: HÃ¥kan Johansson <f96hajo@chalmers.se> Instead of a fix, remove Bus_Time register support altogether. The spec requires all cycle master capable nodes to implement this (all Linux nodes are cycle master capable) while it also says that it "may" be initialized by the bus manager or by the IRM standing in for a bus manager. (Neither Linux' firewire-core nor ieee1394 nodemgr implement this.) Since we cannot rely on Bus_Time having been initialized by a bus manager, it is better to return an error instead of a nonsensical value on a read request to Bus_Time. Alternatively, we could fix the Bus_Time read integrity bug _and_ implement (a) cycle master's write support of the register as well as (b) bus manager's Bus_Time initialization service, i.e. preservation of the Bus_Time when the cycle master node of a bus changes. However, that would be quite some code for a feature that is unreliable to begin with and very likely unused in practice. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2009-12-29firewire: fix use of multiple AV/C devices, allow multiple FCP listenersClemens Ladisch
Control of more than one AV/C device at once --- e.g. camcorders, tape decks, audio devices, TV tuners --- failed or worked only unreliably, depending on driver implementation. This affected kernelspace and userspace drivers alike and was caused by firewire-core's inability to accept multiple registrations of FCP listeners. The fix allows multiple address handlers to be registered for the FCP command and response registers. When a request for these registers is received, all handlers are invoked, and the Firewire response is generated by the core and not by any handler. The cdev API does not change, i.e., userspace is still expected to send a response for FCP requests; this response is silently ignored. Signed-off-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de> (changelog, rebased, whitespace)
2009-10-31firewire: ohci: 0 may be a valid DMA addressStefan Richter
I was told that there are obscure architectures with non-coherent DMA which may DMA-map to bus address 0. We shall not use 0 as a magic number of uninitialized bus address variables. The packet->payload_length > 0 test cannot be used either (except in at_context_queue_packet) because local requests are not DMA-mapped regardless of payload_length. Hence add a state flag to struct fw_packet. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2009-10-31firewire: core: WARN on wrong usage of core transaction functionsStefan Richter
In the code path which creates request packets, clearly mark a switch branch which must never be reached with a WARN. In the code path which creates response packets, replace a BUG by a friendlier to debug WARN. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2009-10-14firewire: core: optimize Topology Map creationStefan Richter
The Topology Map of the local node was created in CPU byte order, then a temporary big endian copy was created to compute the CRC, and when a read request to the Topology Map arrived it had to be converted to big endian byte order again. We now generate it in big endian byte order in the first place. This also rids us of 1000 bytes stack usage in tasklet context. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2009-09-12firewire: core: fix topology map response handlerStefan Richter
This register is 1 kBytes large. Adjust topology_map.length to prevent registration of other response handlers in this region and to make sure that we respond to requests to the upper half of the register. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2009-06-14firewire: core: use more outbound tlabelsStefan Richter
Tlabel is a 6 bits wide datum. Wrap it after 63 rather than 31 for more safety against transaction label exhaustion and potential responders' transaction layer bugs. (As noted by Guus Sliepen, this change requires an expansion of tlabel_mask to 64 bits.) Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2009-06-05firewire: rename source filesStefan Richter
The source files of firewire-core, firewire-ohci, firewire-sbp2, i.e. "drivers/firewire/fw-*.c" are renamed to "drivers/firewire/core-*.c", "drivers/firewire/ohci.c", "drivers/firewire/sbp2.c". The old fw- prefix was redundant to the directory name. The new core- prefix distinguishes the files according to which driver they belong to. This change comes a little late, but still before further firewire drivers are added as anticipated RSN. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>