Age | Commit message (Collapse) | Author |
|
Added the following new sysfs files:
*.fman/*.port/fm_port_regs
*.fman/*.port/fm_port_bmi_regs
*.fman/*.port/fm_port_qmi_regs
*.fman/fm_regs
*.fman/fm_ctrl_code_ver
*.fman/fm_muram_free_size
*.fman/fm_plcr_regs
*.fman/fm_kg_pe_regs
*.fman/fm_kg_regs
*.fman/cls_plans/cls_plan_*
*.fman/schemes/scheme_*
*.fman/profiles/profile_*
*.fman/fm_prs_regs
*.fman/fm_fpm_regs
*.fman/tnums_dbg/tnum_dbg_*
Signed-off-by: Mandy Lavi <mandy.lavi@freescale.com>
Change-Id: If625ef3550bf66da5fb6fb282975a70df78438b1
Reviewed-on: http://git.am.freescale.net:8181/5843
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Madalin-Cristian Bucur <madalin.bucur@freescale.com>
Reviewed-by: Thomas Trefny <Tom.Trefny@freescale.com>
|
|
Add an optional .dts parameter to the Offline Port node, that
would specify a set of TX queues. The Offline Port probe routine
will take care of creating those queues. This makes the Offline
Port an independent structure with respect to its TX queues
creation.
Signed-off-by: Bogdan Purcareata <bogdan.purcareata@freescale.com>
Change-Id: I8f1c045285b373360355b9462c502ce0e745fce9
Reviewed-on: http://git.am.freescale.net:8181/5204
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Cristian-Constantin Sovaiala <Cristian.Sovaiala@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
Automatically set DTC bit in fd.cmd for all traffic originating
from macless interfaces. Update dpa_shared_tx code to bypass
calling dpa_enable_tx_csum when priv->mac_dev == NULL.
Signed-off-by: Bogdan Purcareata <bogdan.purcareata@freescale.com>
Change-Id: I66c6662559c3c560cba1b55ab082f8390a30988d
Reviewed-on: http://git.am.freescale.net:8181/5205
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Bucur Madalin-Cristian-B32716 <madalin.bucur@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
There might be CPUs which are not enabled at boottime, therefore
the pool channel is not currently set to the SDQCR of the affined
portals of those CPUs.
This means that the portals affined to the offline CPUs and which
are migrated to the boot CPU, will not issue interrupts to the
boot CPU in first instance and to the affined CPU (when this comes
online) in the second instance.
Therefore we set the pool channel to the SDQCR of each affined portal.
Signed-off-by: Cristian Sovaiala <cristian.sovaiala@freescale.com>
Change-Id: I63f5649e0c2169617dff991f08dd872a435578a8
Reviewed-on: http://git.am.freescale.net:8181/5629
Reviewed-by: Bucur Madalin-Cristian-B32716 <madalin.bucur@freescale.com>
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Ruxandra Ioana Radulescu <ruxandra.radulescu@freescale.com>
Reviewed-by: Bogdan Hamciuc <bogdan.hamciuc@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
In case an error is encountered during probe, don't actually free
the structure containing global buffer pool information unless its
reference count has reached zero.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@freescale.com>
Change-Id: Ica7b995ecf31a878a9e7685ec905119f1c38f17b
Reviewed-on: http://git.am.freescale.net:8181/5186
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Bogdan Hamciuc <bogdan.hamciuc@freescale.com>
Reviewed-by: Cristian-Constantin Sovaiala <Cristian.Sovaiala@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
fsl_fm_rx_extra headroom must have a value of at least 16 bytes.
The Kconfig option has this lower limit, but when configuring the
parameter through bootargs no check was made.
So enforce the lower limit for all input methods, to avoid invalid
values.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@freescale.com>
Change-Id: Ibf6dfe82d7497745b04eb46e0e7b0a1497c3aa7f
Reviewed-on: http://git.am.freescale.net:8181/5172
Reviewed-by: Bogdan Hamciuc <bogdan.hamciuc@freescale.com>
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
Make sure the support for buffer recycling on FMan v3 platforms
gets selected for the S/G driver as well.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@freescale.com>
Change-Id: I39cfaa9482f6204ddc2b788f635de39de6b4c8f3
Reviewed-on: http://git.am.freescale.net:8181/5146
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Bogdan Hamciuc <bogdan.hamciuc@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
At bootime, we currently initialize a NAPI instance per each
online core, per interface. Booting with a subset of cores and
bringing up more cores afterwards means that no NAPI instances
are created for the newly added cores.
Therefore we create a NAPI instance for each possible core in
the system such that there will always be <total_no_of_cores>
NAPI instances for each interface.
This also gurantees that we do enable/disable NAPI on all possible
cores at ifconfig up/down thus eliminating the need to keep track
of the cores that were online at the previous ifconfig operation.
Even though we do create a NAPI instance and enable it for all
possible cores (including offline cores), the scheduler ensures
that NAPI will not be scheduled on offline cores.
Signed-off-by: Cristian Sovaiala <cristian.sovaiala@freescale.com>
Change-Id: I666bbb870fb8645495a80343cdbefb047c086f4c
Reviewed-on: http://git.am.freescale.net:8181/5118
Reviewed-by: Bogdan Hamciuc <bogdan.hamciuc@freescale.com>
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
Signed-off-by: Shengzhou Liu <Shengzhou.Liu@freescale.com>
Change-Id: I841731e258b147b425aec91224e1fff6ed4d4239
Reviewed-on: http://git.am.freescale.net:8181/4363
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Zang Tiefei-R61911 <tie-fei.zang@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
The current sysfs code shows only the default buffer pool. With this
patch, sysfs code will show all created buffer pools, including the
default one.
Signed-off-by: Marian Rotariu <marian.rotariu@freescale.com>
Change-Id: Ia3ff0e72f2806f7a0d7e94a520b12829af8a6a23
Reviewed-on: http://git.am.freescale.net:8181/4847
Reviewed-by: Cristian-Constantin Sovaiala <Cristian.Sovaiala@freescale.com>
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@freescale.com>
Change-Id: I3cc85c746ac1449e1d8a4192ba680c3083adebb8
Reviewed-on: http://git.am.freescale.net:8181/4842
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Cristian-Constantin Sovaiala <Cristian.Sovaiala@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
Recent changes in the driver moved the allocation of the skb on Rx
past the point where the timestamp/1588 code used the skb.
Reorder a bit the code to avoid NULL pointer dereferencing.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@freescale.com>
Change-Id: I65acb8f1a055d925434a5361ca4570032bae9047
Reviewed-on: http://git.am.freescale.net:8181/4841
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Cristian-Constantin Sovaiala <Cristian.Sovaiala@freescale.com>
Reviewed-by: Bucur Madalin-Cristian-B32716 <madalin.bucur@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
Organize all Ethernet driver options under CONFIG_FSL_DPAA_ETH
(as opposed to the current Kconfig flat structure).
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@freescale.com>
Change-Id: I5285ff7fa935b933b439581fbee078a3518ea986
Reviewed-on: http://git.am.freescale.net:8181/4840
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Cristian-Constantin Sovaiala <Cristian.Sovaiala@freescale.com>
Reviewed-by: Bucur Madalin-Cristian-B32716 <madalin.bucur@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
Add a Kconfig option that enables debug support in the driver.
For a start, move (most of) the BUG_ON()'s in the driver's fastpath
under this debug option.
New debug information (such as trace events, frame content dumps, etc)
can be added later under the same umbrella.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@freescale.com>
Change-Id: I8023a75751cdad9e80d96957b9071f707a03e2b2
Reviewed-on: http://git.am.freescale.net:8181/4839
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Cristian-Constantin Sovaiala <Cristian.Sovaiala@freescale.com>
Reviewed-by: Bucur Madalin-Cristian-B32716 <madalin.bucur@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
The timestamping code relies on Tx confirmation routine to get the
Tx timestamp of a frame. So if timestamp support is activated, don't
recycle frames at all.
Also fix a couple of compilation warnings when building with
timestamp support on.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@freescale.com>
Change-Id: I5ee3a5b727e28ee337dc9da5c16ad0d1dabe4317
Reviewed-on: http://git.am.freescale.net:8181/4838
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Cristian-Constantin Sovaiala <Cristian.Sovaiala@freescale.com>
Reviewed-by: Bucur Madalin-Cristian-B32716 <madalin.bucur@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
A bit of cleanup in function skb_to_contig_fd().
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@freescale.com>
Change-Id: Ifbf32b298d1588574b9eabf90d0bd1607b733d2c
Reviewed-on: http://git.am.freescale.net:8181/4837
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Bucur Madalin-Cristian-B32716 <madalin.bucur@freescale.com>
Reviewed-by: Cristian-Constantin Sovaiala <Cristian.Sovaiala@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
Function dpa_bpid2pool() returns NULL if there's no initialized buffer
pool associated with the given bpid.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@freescale.com>
Change-Id: I42b286c1050f273774d645fabee49c105d7ab6f1
Reviewed-on: http://git.am.freescale.net:8181/4836
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Bucur Madalin-Cristian-B32716 <madalin.bucur@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
The unit test code was rendered obsolete by the addition of the SG
codepath. Removing the unit test code for now.
Signed-off-by: Madalin Bucur <madalin.bucur@freescale.com>
Change-Id: I9cfe97e83a91e0418523ff8ec756105f1b435079
Reviewed-on: http://git.am.freescale.net:8181/4821
Reviewed-by: Bogdan Hamciuc <bogdan.hamciuc@freescale.com>
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Cristian-Constantin Sovaiala <Cristian.Sovaiala@freescale.com>
Reviewed-by: Radulescu Ruxandra Ioana-B05472 <ruxandra.radulescu@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
|
|
|
|
Signed-off-by: Mandy Lavi <mandy.lavi@freescale.com>
Change-Id: I4ce7f181158f5df8191eea7ad75c0e84cfd01f1b
Reviewed-on: http://git.am.freescale.net:8181/4573
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Zanoschi Aurelian-B43522 <Aurelian.Zanoschi@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
git://git.am.freescale.net/gitolite/users/b32716/dpaa-eth.git
|
|
The wrong return type of dpa_bp_priv_seed() caused a crash
on the termination optimized driver flavor.
Signed-off-by: Madalin Bucur <madalin.bucur@freescale.com>
Reviewed-on: http://git.am.freescale.net:8181/3696
Reviewed-by: Sovaiala Cristian-Constantin-B39531 <Cristian.Sovaiala@freescale.com>
Reviewed-by: Radulescu Ruxandra Ioana-B05472 <ruxandra.radulescu@freescale.com>
Reviewed-by: Hamciuc Bogdan-BHAMCIU1 <bogdan.hamciuc@freescale.com>
Change-Id: I538a1529f1c0e27ff5b199e9675a56029f448e64
Reviewed-on: http://git.am.freescale.net:8181/4643
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Bogdan Hamciuc <bogdan.hamciuc@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
When draining buffers in batches of eight, we must take into account
the fact that BMan will return an error code if there are less than
eight buffers left in the pool; when this happens, drain the remaining
buffers one by one.
And since we're at buffer pool cleanup, make sure allocated bpools get
freed in case of errors inside the probe function.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@freescale.com>
Reviewed-on: http://git.am.freescale.net:8181/4191
Reviewed-by: Bucur Madalin-Cristian-B32716 <madalin.bucur@freescale.com>
Change-Id: If264ca3dd4a8287bc5e5ee4e1c3a10ffe18663b7
Reviewed-on: http://git.am.freescale.net:8181/4657
Reviewed-by: Bogdan Hamciuc <bogdan.hamciuc@freescale.com>
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
In case of a Tx error at FMan level or an ERN, if the frame was
marked as recycleable we need to put the buffers in the target pools,
otherwise we mess up the pool counters. This was done correctly for
the ERN callback, now fixing the Tx error one as well.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@freescale.com>
Reviewed-on: http://git.am.freescale.net:8181/4192
Reviewed-by: Bucur Madalin-Cristian-B32716 <madalin.bucur@freescale.com>
Change-Id: Iafc436208fe560fd098728e421b105cf568017d8
Reviewed-on: http://git.am.freescale.net:8181/4656
Reviewed-by: Bogdan Hamciuc <bogdan.hamciuc@freescale.com>
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
We randomly used either __this_cpu_ptr() or per_cpu_ptr() on current
processor.
Use only __this_cpu_ptr() in ordr to make the code more uniform.
No functional or performance impact.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@freescale.com>
Reviewed-on: http://git.am.freescale.net:8181/4193
Reviewed-by: Bucur Madalin-Cristian-B32716 <madalin.bucur@freescale.com>
Change-Id: Ifd63f277fe35b3a44f7b7dc7beacf6b86765e900
Reviewed-on: http://git.am.freescale.net:8181/4654
Reviewed-by: Bogdan Hamciuc <bogdan.hamciuc@freescale.com>
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
Fix a series of checkpatch complaints in the dpaa_eth code.
Signed-off-by: Madalin Bucur <madalin.bucur@freescale.com>
Reviewed-on: http://git.am.freescale.net:8181/3811
Change-Id: I3674db6613d311431b13c0eae65a977263957647
Reviewed-on: http://git.am.freescale.net:8181/4646
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Bogdan Hamciuc <bogdan.hamciuc@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
Assign boolean constants to boolean variables instead of integer
constants.
Signed-off-by: Cristian Bercaru <cristian.bercaru@freescale.com>
Reviewed-on: http://git.am.freescale.net:8181/4021
Reviewed-by: Bucur Madalin-Cristian-B32716 <madalin.bucur@freescale.com>
Reviewed-by: Radulescu Ruxandra Ioana-B05472 <ruxandra.radulescu@freescale.com>
Change-Id: I7f3550f1af032349dbedfe59ae39a337195b2aff
Reviewed-on: http://git.am.freescale.net:8181/4645
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Bogdan Hamciuc <bogdan.hamciuc@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
The common code contains some parts that are not used in the
private interface code. Moved that code to a new file.
Signed-off-by: Madalin Bucur <madalin.bucur@freescale.com>
Reviewed-on: http://git.am.freescale.net:8181/3810
Change-Id: I0e5e74b5eb531d20c6de1b1c74a5c1049439c890
Reviewed-on: http://git.am.freescale.net:8181/4644
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Bogdan Hamciuc <bogdan.hamciuc@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
Undoing the merge conflict artifacts, out of order and incomplete
patch application code changes. This reverts the sdk-kernel-3.8
branch to the state at 70c2c840d8ff41359aafd7f510144c235a32fce1.
Signed-off-by: Madalin Bucur <madalin.bucur@freescale.com>
Change-Id: Ib7282b356eeb38884fce716e247620c431bd3c72
Reviewed-on: http://git.am.freescale.net:8181/4642
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Bogdan Hamciuc <bogdan.hamciuc@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
Recent changes in the driver moved the allocation of the skb on Rx
past the point where the timestamp/1588 code used the skb.
Reorder a bit the code to avoid NULL pointer dereferencing.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@freescale.com>
Change-Id: I09929c13bc88048be77c8678c1fb9956fec41039
Reviewed-on: http://git.am.freescale.net:8181/4443
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Bucur Madalin-Cristian-B32716 <madalin.bucur@freescale.com>
Reviewed-by: Sovaiala Cristian-Constantin-B39531 <Cristian.Sovaiala@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
Recent changes in the driver moved the allocation of the skb on Rx
past the point where the timestamp/1588 code used the skb.
Reorder a bit the code to avoid NULL pointer dereferencing.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@freescale.com>
Change-Id: I09929c13bc88048be77c8678c1fb9956fec41039
Reviewed-on: http://git.am.freescale.net:8181/4443
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Bucur Madalin-Cristian-B32716 <madalin.bucur@freescale.com>
Reviewed-by: Sovaiala Cristian-Constantin-B39531 <Cristian.Sovaiala@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
In case of a Tx error at FMan level or an ERN, if the frame was
marked as recycleable we need to put the buffers in the target pools,
otherwise we mess up the pool counters. This was done correctly for
the ERN callback, now fixing the Tx error one as well.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@freescale.com>
Change-Id: I0faa77a31af9086806713e593da72b03c2bd3223
Reviewed-on: http://git.am.freescale.net:8181/4192
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Bucur Madalin-Cristian-B32716 <madalin.bucur@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
Reviewed-on: http://git.am.freescale.net:8181/4470
Reviewed-by: Sovaiala Cristian-Constantin-B39531 <Cristian.Sovaiala@freescale.com>
|
|
The reference to the buffer pool can be just as easily obtained from
the netdev private structure, so do that and reduce the size of the percpu
private struct.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@freescale.com>
Change-Id: I78aea782014b863994585beccac52567b242a72e
Reviewed-on: http://git.am.freescale.net:8181/4194
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Bucur Madalin-Cristian-B32716 <madalin.bucur@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
Reviewed-on: http://git.am.freescale.net:8181/4469
Reviewed-by: Sovaiala Cristian-Constantin-B39531 <Cristian.Sovaiala@freescale.com>
|
|
When draining buffers in batches of eight, we must take into account
the fact that BMan will return an error code if there are less than
eight buffers left in the pool; when this happens, drain the remaining
buffers one by one.
And since we're at buffer pool cleanup, make sure allocated bpools get
freed in case of errors inside the probe function.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@freescale.com>
Change-Id: I9b2516ea76a19487a2f75412722ea21e67247ed2
Reviewed-on: http://git.am.freescale.net:8181/4191
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Bucur Madalin-Cristian-B32716 <madalin.bucur@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
Reviewed-on: http://git.am.freescale.net:8181/4471
Reviewed-by: Sovaiala Cristian-Constantin-B39531 <Cristian.Sovaiala@freescale.com>
|
|
We randomly used either __this_cpu_ptr() or per_cpu_ptr() on current
processor.
Use only __this_cpu_ptr() in ordr to make the code more uniform.
No functional or performance impact.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@freescale.com>
Change-Id: I1d449ec799e8d920935325d283f0ba510ee56e87
Reviewed-on: http://git.am.freescale.net:8181/4193
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Bucur Madalin-Cristian-B32716 <madalin.bucur@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
Reviewed-on: http://git.am.freescale.net:8181/4468
Reviewed-by: Sovaiala Cristian-Constantin-B39531 <Cristian.Sovaiala@freescale.com>
|
|
Instead use a callback for freeing individual buffers acquired from
the pool. This solution is more flexible and allows for simpler code
when new pools will be added to the driver.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@freescale.com>
Change-Id: I52376b5dab53957e497b112df7c42ef18d7a9347
Reviewed-on: http://git.am.freescale.net:8181/4190
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-on: http://git.am.freescale.net:8181/4467
Reviewed-by: Sovaiala Cristian-Constantin-B39531 <Cristian.Sovaiala@freescale.com>
|
|
Interfaces only get their PHYs initialized when they are first
brought up, therefore running ethtool on an interface that is down
prints an error message and returns with error code.
Lower the print to debug mode and return non-error code.
Signed-off-by: Cristian Sovaiala <cristian.sovaiala@freescale.com>
Change-Id: Ic4e9304cf9e6152868af74a31132e92d95713b62
Reviewed-on: http://git.am.freescale.net:8181/4259
Reviewed-by: Radulescu Ruxandra Ioana-B05472 <ruxandra.radulescu@freescale.com>
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Bucur Madalin-Cristian-B32716 <madalin.bucur@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
Reviewed-on: http://git.am.freescale.net:8181/4466
Reviewed-by: Sovaiala Cristian-Constantin-B39531 <Cristian.Sovaiala@freescale.com>
|
|
Don't refill the Rx global pool every time it drops below its target
count, instead wait for the number of buffers to drop below a certain
threshold.
This was not so important until now, but is necessary in view of the
recycling mechanism.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@freescale.com>
Change-Id: I341f6bf7bd746465cf6d17644a4edbd1b3d20aba
Reviewed-on: http://git.am.freescale.net:8181/4189
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Bucur Madalin-Cristian-B32716 <madalin.bucur@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
Reviewed-on: http://git.am.freescale.net:8181/4464
Reviewed-by: Sovaiala Cristian-Constantin-B39531 <Cristian.Sovaiala@freescale.com>
|
|
Reorganize a bit the recycling code without any actual changes in
the logic.
Reordering of conditions based on which is most likely to fail first
brings a bit of performance improvement in case no recycling is
actually done.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@freescale.com>
Change-Id: I6cf3995576b756d9aec678f92f679ee1fa461587
Reviewed-on: http://git.am.freescale.net:8181/4188
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Bucur Madalin-Cristian-B32716 <madalin.bucur@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
Reviewed-on: http://git.am.freescale.net:8181/4465
Reviewed-by: Sovaiala Cristian-Constantin-B39531 <Cristian.Sovaiala@freescale.com>
|
|
When draining buffers in batches of eight, we must take into account
the fact that BMan will return an error code if there are less than
eight buffers left in the pool; when this happens, drain the remaining
buffers one by one.
And since we're at buffer pool cleanup, make sure allocated bpools get
freed in case of errors inside the probe function.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@freescale.com>
Change-Id: Id0ae7ed369ded0fff7359ae20b6bf5df64e77fa1
Reviewed-on: http://git.am.freescale.net:8181/4442
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Bucur Madalin-Cristian-B32716 <madalin.bucur@freescale.com>
Reviewed-by: Sovaiala Cristian-Constantin-B39531 <Cristian.Sovaiala@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
The platform device was not released.
Signed-off-by: Madalin Bucur <madalin.bucur@freescale.com>
Change-Id: I44c134aba26f5eca5cad60fba3d0051ae9f99c8f
Reviewed-on: http://git.am.freescale.net:8181/3911
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Radulescu Ruxandra Ioana-B05472 <ruxandra.radulescu@freescale.com>
Reviewed-by: Rivera Jose-B46482 <Jose.G.Rivera@freescale.com>
Reviewed-on: http://git.am.freescale.net:8181/4462
Reviewed-by: Sovaiala Cristian-Constantin-B39531 <Cristian.Sovaiala@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
Cleanup includes and fix non-smp build, removed unnecessary info.
Signed-off-by: Madalin Bucur <madalin.bucur@freescale.com>
Change-Id: I32596b8eb0c4b674de1f3684993d1ba09f20947e
Reviewed-on: http://git.am.freescale.net:8181/3856
Reviewed-by: Bucur Madalin-Cristian-B32716 <madalin.bucur@freescale.com>
Reviewed-by: Rivera Jose-B46482 <Jose.G.Rivera@freescale.com>
Tested-by: Rivera Jose-B46482 <Jose.G.Rivera@freescale.com>
Reviewed-on: http://git.am.freescale.net:8181/4461
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Sovaiala Cristian-Constantin-B39531 <Cristian.Sovaiala@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
Signed-off-by: Sunil Kumar Kori <Sunil.Kori@freescale.com>
Earlier to enable ASF hooks in Linux, we are required to enable ASF support
using "make menuconfig" before compiling Linux. This patch enables
ASF support as default Kernel configuration and also will enable required
configuratuin flags by ASF.
Change-Id: I3467513f782cecef0bbcdf0c9aa4bb9be4f9f195
Reviewed-on: http://git.am.freescale.net:8181/4488
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Bucur Madalin-Cristian-B32716 <madalin.bucur@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
Fix a series of checkpatch complaints in the dpaa_eth code.
Signed-off-by: Madalin Bucur <madalin.bucur@freescale.com>
Change-Id: I341d5835bb0a4491456947b8e1ae714e095cd580
Reviewed-on: http://git.am.freescale.net:8181/3811
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Rivera Jose-B46482 <Jose.G.Rivera@freescale.com>
Reviewed-on: http://git.am.freescale.net:8181/4460
Reviewed-by: Sovaiala Cristian-Constantin-B39531 <Cristian.Sovaiala@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
The common code contains some parts that are not used in the
private interface code. Moved that code to a new file.
Signed-off-by: Madalin Bucur <madalin.bucur@freescale.com>
Change-Id: I5095bc0e7503c413d39e78cd5dff7bf84eebaca9
Reviewed-on: http://git.am.freescale.net:8181/3810
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Rivera Jose-B46482 <Jose.G.Rivera@freescale.com>
Reviewed-on: http://git.am.freescale.net:8181/4458
Reviewed-by: Sovaiala Cristian-Constantin-B39531 <Cristian.Sovaiala@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
The wrong return type of dpa_bp_priv_seed() caused a crash
on the termination optimized driver flavor.
Signed-off-by: Madalin Bucur <madalin.bucur@freescale.com>
Change-Id: I22cec61c063a901a4d8fd40653fb4fe2cba31435
Reviewed-on: http://git.am.freescale.net:8181/3696
Reviewed-by: Sovaiala Cristian-Constantin-B39531 <Cristian.Sovaiala@freescale.com>
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Radulescu Ruxandra Ioana-B05472 <ruxandra.radulescu@freescale.com>
Reviewed-by: Hamciuc Bogdan-BHAMCIU1 <bogdan.hamciuc@freescale.com>
Reviewed-by: Rivera Jose-B46482 <Jose.G.Rivera@freescale.com>
Reviewed-on: http://git.am.freescale.net:8181/4457
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
Conflicts temporarily resolved for:
drivers/net/ethernet/freescale/dpa/dpaa_eth.c
Permanent solution needs to be provided by the author of
commit ec6aa025.
|
|
Assign boolean constants to boolean variables instead of integer
constants.
Signed-off-by: Cristian Bercaru <cristian.bercaru@freescale.com>
Change-Id: Ib7ecb86273355d0dd6c4f06fe410d572e613c570
Reviewed-on: http://git.am.freescale.net:8181/4021
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Bucur Madalin-Cristian-B32716 <madalin.bucur@freescale.com>
Reviewed-by: Radulescu Ruxandra Ioana-B05472 <ruxandra.radulescu@freescale.com>
Reviewed-by: Rivera Jose-B46482 <Jose.G.Rivera@freescale.com>
Reviewed-on: http://git.am.freescale.net:8181/4459
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|
|
When the stack drops an incoming frame it also increments the
rx_dropped counter in the netdevice stats. We were also updating
this counter inside the driver, resulting in incorrect statistics
for dropped frames.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@freescale.com>
Change-Id: Ieeef79a277f62cf935c779e0e93b018d4e2b284c
Reviewed-on: http://git.am.freescale.net:8181/4441
Tested-by: Review Code-CDREVIEW <CDREVIEW@freescale.com>
Reviewed-by: Bucur Madalin-Cristian-B32716 <madalin.bucur@freescale.com>
Reviewed-by: Rivera Jose-B46482 <German.Rivera@freescale.com>
|