Commit graph

708,694 commits

Author SHA1 Message Date
Alex Deucher
d5dc36a45e drm/amdgpu/gfx8: enable cp/rlc ints after we disable clockgating
Even if we disable clockgating, we still need to make sure the
cp/rlc interrupts are enabled for powergating which might still
be enabled.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:55 -04:00
Alex Deucher
dfa6c82ee5 drm/amdgpu/gfx7: enable cp/rlc ints after we disable clockgating
Even if we disable clockgating, we still need to make sure the
cp/rlc interrupts are enabled for powergating which might still
be enabled.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:54 -04:00
Alex Deucher
d17c0faf1d drm/amdgpu/gfx8: move MEC doorbell range setting
It's global, not queue specific, so move it out of the
kiq register init function.

Tested-and-Reviewed-by:  Rex Zhu <Rex.Zhu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:53 -04:00
Alex Deucher
a545e491bb drm/amdgpu/gfx8: fix resume of KIQ and KCQs
No need to reset the wptr and clear the rings.  The UNMAP_QUEUES
packet writes the current MQD state back the MQD on suspend,
so there is no need to reset it as well.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:52 -04:00
Alex Deucher
9d11ca9c09 drm/amdgpu/gfx8: properly disable the KCQs in hw_fini
Use the UNMAP_QUEUES packet to have the KIQ properly
disable them.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:51 -04:00
Alex Deucher
3d7e30b381 drm/amdgpu/gfx8: use new KIQ packet defines
Rather than open coding them.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:51 -04:00
Alex Deucher
346586d567 drm/amdgpu/gfx8: move SET_RESOURCES into the same command stream
As the KCQ setup.  This way we only have to wait once for the
entire MEC.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:50 -04:00
Alex Deucher
c3a49ab54b drm/amdgpu/gfx8: wait once for all KCQs to be created
Rather than waiting for each queue.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:49 -04:00
Alex Deucher
3930011594 drm/amdgpu: split gfx_v8_0_kiq_init_queue into two
One for KIQ and one for the KCQ. This simplifies the logic and
allows for future optimizations.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:48 -04:00
Alex Deucher
f776952b76 drm/amdgpu/gfx8: wait for completion in KIQ init
We need to make sure the various init sequences submitted
to KIQ complete before testing the rings.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:47 -04:00
Alex Deucher
4fdca894bb Revert "drm/amd/amdgpu: Disable GFX_PG on Carrizo until compute issues solved"
Re-enable GFX PG.  It's working properly with MEC now that KIQ is
enabled.

Reviewed-by: Samuel  Li <samuel.li@amd.com>

This reverts commit e9ef19aa1bdeac380662a112f1d03a7c3477527f.
2017-05-24 17:39:47 -04:00
David Panariti
b4e40676e4 drm/amdgpu: Switch baremetal to use KIQ for compute ring management. (v3)
KIQ is the Kernel Interface Queue for managing the MEC.  Rather than setting
up rings via direct MMIO of ring registers, the rings are configured via
special packets sent to the KIQ.  The allows the MEC to better manage shared
resources and certain power events.

v2: squash in s3/s4 fix from Rex
v3: further fixes from Rex

Signed-off-by: David Panariti <David.Panariti@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Acked-by: Tom St Denis <tom.stdenis@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:46 -04:00
Alex Deucher
a576fe5151 drm/amdgpu/gfx8: set doorbell range for polaris as well
Add missing chips to the doorbell range setup.  These
were missed in the KIQ code.  Fixes power and performance
regressions with KIQ.  Spotted by Rex.

Tested-and-Reviewed-by:  Rex Zhu <Rex.Zhu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:45 -04:00
Alex Deucher
ed6f55d1a9 drm/amdgpu/gfx8: add additional MQD initialization
Need to properly set the MTYPE and ROQ space setting.
This should fix performance regressions with KIQ enabled.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:44 -04:00
Eric Huang
b6dc60cf79 drm/amd/powerplay: fix pcie dpm table for vega10
This resolves pcie low speed problem.

Signed-off-by: Eric Huang <JinHuiEric.Huang@amd.com>
Acked-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:43 -04:00
Rex Zhu
9312d9a6a0 drm/amd/powerplay: update vega10 smu interface version to E.
need update smu firmware to version 0x1c20.

Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Reviewws-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:43 -04:00
Rex Zhu
14641ac4eb drm/amd/powerplay: delete dead code in vega10_thermal.c
Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:42 -04:00
Rex Zhu
fbf66a3c9c drm/amd/powerplay: Add Vega10 Powertune Table v3 support.
Handle the latest powerplay table format; includes Boost
State support.

Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Reviewws-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:41 -04:00
Rex Zhu
676b4087fc drm/amd/powerplay: convert from number of lanes to lane bits on vega10
We need a mask.

Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Reviewws-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:40 -04:00
Rex Zhu
9c2cc3a10c drm/amd/powerplay: fix bug in processing CKS_Enable bit.
Typo in the mask.

Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:39 -04:00
Rex Zhu
ab5cf3a551 drm/amd/powerplay: add avfs fuse overdriver func.
Add a function to look up the AVFS fuse values for vega10
These are used to populate the avfs fuse table in the smu.

Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:38 -04:00
Eric Huang
d6c025d243 drm/amd/powerplay: add power profile support for Vega10 (v2)
This implements the workload specific interface of optimized
compute power profile for Vega10.

v2: squash in fix (Tom)

Signed-off-by: Eric Huang <JinHuiEric.Huang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:38 -04:00
Alex Deucher
ea289b39a6 drm/amdgpu/gfx9: drop duplicate gfx info init (v3)
Taken care of by gpu info firmware now.

v2: rebase
v3: rework based on latest firmware

Reviewed-by: Junwei Zhang <Jerry.Zhang@amd.com>
Tested-by: Junwei Zhang <Jerry.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:37 -04:00
Alex Deucher
e2a75f88c3 drm/amdgpu: parse the gpu_info firmware (v4)
And populate the gfx structures from it.

v2: update the structures updated by the table
v3: rework based on new table structure
v4: simplify things

Reviewed-by: Junwei Zhang <Jerry.Zhang@amd.com>
Tested-by: Junwei Zhang <Jerry.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:36 -04:00
Alex Deucher
8ae1a33648 drm/amdgpu: add gpu_info firmware (v3)
Add a new gpu info firmware to store gpu specific configuration
data.  This allows us to store hw constants in a unified place.

v2: adjust structure and elements
v3: further restructure

Reviewed-by: Junwei Zhang <Jerry.Zhang@amd.com>
Tested-by: Junwei Zhang <Jerry.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:35 -04:00
Dan Carpenter
2f590f8419 drm/amd/powerplay: fix a signedness bugs
Smatch complains about a signedness bug here:

        vega10_hwmgr.c:4202 vega10_force_clock_level()
        warn: always true condition '(i >= 0) => (0-u32max >= 0)'

Fixes: 7b52db39a4 ("drm/amd/powerplay: fix bug sclk/mclk
                     level can't be set on vega10.")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Eric Huang <JinHuiEric.Huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:34 -04:00
Chunming Zhou
ca7962d8cc drm/amdgpu: fix NULL pointer panic of emit_gds_switch
[  338.384770] BUG: unable to handle kernel NULL pointer dereference at           (null)
[  338.384817] IP: [<          (null)>]           (null)
[  338.385505] RIP: 0010:[<0000000000000000>]  [<          (null)>]           (null)
[  338.385950] Call Trace:
[  338.385993]  [<ffffffffa05d2313>] ? amdgpu_vm_flush+0x283/0x400 [amdgpu]
[  338.386025]  [<ffffffff811818d3>] ? printk+0x4d/0x4f
[  338.386074]  [<ffffffffa05d4906>] amdgpu_ib_schedule+0x4a6/0x4d0 [amdgpu]
[  338.386140]  [<ffffffffa0673e54>] amdgpu_job_run+0x64/0x180 [amdgpu]
[  338.386203]  [<ffffffffa0672e09>] amd_sched_main+0x2e9/0x4a0 [amdgpu]
[  338.386232]  [<ffffffff810bfce0>] ? prepare_to_wait_event+0x110/0x110
[  338.386295]  [<ffffffffa0672b20>] ? amd_sched_select_entity+0xe0/0xe0 [amdgpu]
[  338.386327]  [<ffffffff8109b423>] kthread+0xd3/0xf0
[  338.386349]  [<ffffffff8109b350>] ? kthread_park+0x60/0x60
[  338.386376]  [<ffffffff817e1ee5>] ret_from_fork+0x25/0x30
[  338.386401] Code:  Bad RIP value.
[  338.386420] RIP  [<          (null)>]           (null)
[  338.386443]  RSP <ffffc90001bd7d40>
[  338.386458] CR2: 0000000000000000
[  338.398508] ---[ end trace 4c66fcdc74b9a0a2 ]---

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:34 -04:00
Lyude
e12fcff799 drm/radeon: Unbreak HPD handling for r600+
We end up reading the interrupt register for HPD5, and then writing it
to HPD6 which on systems without anything using HPD5 results in
permanently disabling hotplug on one of the display outputs after the
first time we acknowledge a hotplug interrupt from the GPU.

This code is really bad. But for now, let's just fix this. I will
hopefully have a large patch series to refactor all of this soon.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Lyude <lyude@redhat.com>
Cc: stable@vger.kernel.org
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:33 -04:00
Alex Deucher
ae5037dcfe drm/amd/powerplay/smu7: disable mclk switching for high refresh rates
Even if the vblank period would allow it, it still seems to
be problematic on some cards.

bug: https://bugs.freedesktop.org/show_bug.cgi?id=96868

Cc: stable@vger.kernel.org
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:32 -04:00
Alex Deucher
14efcf11e9 drm/amd/powerplay/smu7: add vblank check for mclk switching (v2)
Check to make sure the vblank period is long enough to support
mclk switching.

v2: drop needless initial assignment (Nils)

bug: https://bugs.freedesktop.org/show_bug.cgi?id=96868

Cc: stable@vger.kernel.org
Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Rex Zhu <Rex.Zhu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:32 -04:00
Alex Deucher
ab03d9fe50 drm/radeon/ci: disable mclk switching for high refresh rates (v2)
Even if the vblank period would allow it, it still seems to
be problematic on some cards.

v2: fix logic inversion (Nils)

bug: https://bugs.freedesktop.org/show_bug.cgi?id=96868

Cc: stable@vger.kernel.org
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:31 -04:00
Alex Deucher
7121316b6d drm/amdgpu/ci: disable mclk switching for high refresh rates (v2)
Even if the vblank period would allow it, it still seems to
be problematic on some cards.

v2: fix logic inversion (Nils)

bug: https://bugs.freedesktop.org/show_bug.cgi?id=96868

Cc: stable@vger.kernel.org
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:30 -04:00
Christian König
32601d48d7 drm/amdgpu: fix fundamental suspend/resume issue
Reinitializing the VM manager during suspend/resume is a very very bad
idea since all the VMs are still active and kicking.

This can lead to random VM faults after resume when new processes
become the same client ID assigned.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
2017-05-24 17:39:30 -04:00
Dan Carpenter
3083696a1e drm/amd/powerplay: fix a signedness bugs
Smatch complains about a signedness bug here:

        vega10_hwmgr.c:4202 vega10_force_clock_level()
        warn: always true condition '(i >= 0) => (0-u32max >= 0)'

Fixes: 7b52db39a4 ("drm/amd/powerplay: fix bug sclk/mclk
                     level can't be set on vega10.")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Eric Huang <JinHuiEric.Huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 16:49:34 -04:00
Chunming Zhou
7c4378f452 drm/amdgpu: fix NULL pointer panic of emit_gds_switch
[  338.384770] BUG: unable to handle kernel NULL pointer dereference at           (null)
[  338.384817] IP: [<          (null)>]           (null)
[  338.385505] RIP: 0010:[<0000000000000000>]  [<          (null)>]           (null)
[  338.385950] Call Trace:
[  338.385993]  [<ffffffffa05d2313>] ? amdgpu_vm_flush+0x283/0x400 [amdgpu]
[  338.386025]  [<ffffffff811818d3>] ? printk+0x4d/0x4f
[  338.386074]  [<ffffffffa05d4906>] amdgpu_ib_schedule+0x4a6/0x4d0 [amdgpu]
[  338.386140]  [<ffffffffa0673e54>] amdgpu_job_run+0x64/0x180 [amdgpu]
[  338.386203]  [<ffffffffa0672e09>] amd_sched_main+0x2e9/0x4a0 [amdgpu]
[  338.386232]  [<ffffffff810bfce0>] ? prepare_to_wait_event+0x110/0x110
[  338.386295]  [<ffffffffa0672b20>] ? amd_sched_select_entity+0xe0/0xe0 [amdgpu]
[  338.386327]  [<ffffffff8109b423>] kthread+0xd3/0xf0
[  338.386349]  [<ffffffff8109b350>] ? kthread_park+0x60/0x60
[  338.386376]  [<ffffffff817e1ee5>] ret_from_fork+0x25/0x30
[  338.386401] Code:  Bad RIP value.
[  338.386420] RIP  [<          (null)>]           (null)
[  338.386443]  RSP <ffffc90001bd7d40>
[  338.386458] CR2: 0000000000000000
[  338.398508] ---[ end trace 4c66fcdc74b9a0a2 ]---

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 16:49:09 -04:00
Lyude
3d18e33735 drm/radeon: Unbreak HPD handling for r600+
We end up reading the interrupt register for HPD5, and then writing it
to HPD6 which on systems without anything using HPD5 results in
permanently disabling hotplug on one of the display outputs after the
first time we acknowledge a hotplug interrupt from the GPU.

This code is really bad. But for now, let's just fix this. I will
hopefully have a large patch series to refactor all of this soon.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Lyude <lyude@redhat.com>
Cc: stable@vger.kernel.org
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 16:46:43 -04:00
Alex Deucher
2275a3a2fe drm/amd/powerplay/smu7: disable mclk switching for high refresh rates
Even if the vblank period would allow it, it still seems to
be problematic on some cards.

bug: https://bugs.freedesktop.org/show_bug.cgi?id=96868

Cc: stable@vger.kernel.org
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 16:46:01 -04:00
Alex Deucher
09be4a5219 drm/amd/powerplay/smu7: add vblank check for mclk switching (v2)
Check to make sure the vblank period is long enough to support
mclk switching.

v2: drop needless initial assignment (Nils)

bug: https://bugs.freedesktop.org/show_bug.cgi?id=96868

Cc: stable@vger.kernel.org
Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Rex Zhu <Rex.Zhu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 16:45:28 -04:00
Alex Deucher
58d7e3e427 drm/radeon/ci: disable mclk switching for high refresh rates (v2)
Even if the vblank period would allow it, it still seems to
be problematic on some cards.

v2: fix logic inversion (Nils)

bug: https://bugs.freedesktop.org/show_bug.cgi?id=96868

Cc: stable@vger.kernel.org
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 16:45:03 -04:00
Alex Deucher
0a646f331d drm/amdgpu/ci: disable mclk switching for high refresh rates (v2)
Even if the vblank period would allow it, it still seems to
be problematic on some cards.

v2: fix logic inversion (Nils)

bug: https://bugs.freedesktop.org/show_bug.cgi?id=96868

Cc: stable@vger.kernel.org
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 16:44:26 -04:00
Eric Dumazet
ce682ef6e3 tcp: fix TCP_SYNCNT flakes
After the mentioned commit, some of our packetdrill tests became flaky.

TCP_SYNCNT socket option can limit the number of SYN retransmits.

retransmits_timed_out() has to compare times computations based on
local_clock() while timers are based on jiffies. With NTP adjustments
and roundings we can observe 999 ms delay for 1000 ms timers.
We end up sending one extra SYN packet.

Gimmick added in commit 6fa12c8503 ("Revert Backoff [v3]: Calculate
TCP's connection close threshold as a time value") makes no
real sense for TCP_SYN_SENT sockets where no RTO backoff can happen at
all.

Lets use a simpler logic for TCP_SYN_SENT sockets and remove @syn_set
parameter from retransmits_timed_out()

Fixes: 9a568de481 ("tcp: switch TCP TS option (RFC 7323) to 1ms clock")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2017-05-24 16:29:57 -04:00
Vivien Didelot
64dba236a1 net: dsa: support cross-chip ageing time
Now that the switchdev bridge ageing time attribute is propagated to all
switch chips of the fabric, each switch can check if the requested value
is valid and program itself, so that the whole fabric shares a common
ageing time setting.

This is especially needed for switch chips in between others, containing
no bridge port members but evidently used in the data path.

To achieve that, remove the condition which skips the other switches. We
also don't need to identify the target switch anymore, thus remove the
sw_index member of the dsa_notifier_ageing_time_info notifier structure.

On ZII Dev Rev B (with two 88E6352 and one 88E6185) and ZII Dev Rev C
(with two 88E6390X), we have the following hardware configuration:

    # ip link add name br0 type bridge
    # ip link set master br0 dev lan6
    br0: port 1(lan6) entered blocking state
    br0: port 1(lan6) entered disabled state
    # echo 2000 > /sys/class/net/br0/bridge/ageing_time

Before this patch:

    zii-rev-b# cat /sys/kernel/debug/mv88e6xxx/sw*/age_time
    300000
    300000
    15000

    zii-rev-c# cat /sys/kernel/debug/mv88e6xxx/sw*/age_time
    300000
    18750

After this patch:

    zii-rev-b# cat /sys/kernel/debug/mv88e6xxx/sw*/age_time
    15000
    15000
    15000

    zii-rev-c# cat /sys/kernel/debug/mv88e6xxx/sw*/age_time
    18750
    18750

Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2017-05-24 16:27:47 -04:00
David S. Miller
e6a88e4ba8 Merge branch 'q-in-q-checksums'
Daniel Borkmann says:

====================
BPF pruning follow-up

Follow-up to fix incorrect pruning when alignment tracking is
in use and to properly clear regs after call to not leave stale
data behind. For details, please see individual patches.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2017-05-24 16:27:22 -04:00
Vlad Yasevich
2836b4f224 virtio-net: enable TSO/checksum offloads for Q-in-Q vlans
Since virtio does not provide it's own ndo_features_check handler,
TSO, and now checksum offload, are disabled for stacked vlans.
Re-enable the support and let the host take care of it.  This
restores/improves Guest-to-Guest performance over Q-in-Q vlans.

Acked-by: Jason Wang <jasowang@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Vladislav Yasevich <vyasevic@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2017-05-24 16:27:14 -04:00
Vlad Yasevich
cc6e9de62a be2net: Fix offload features for Q-in-Q packets
At least some of the be2net cards do not seem to be capabled
of performing checksum offload computions on Q-in-Q packets.
In these case, the recevied checksum on the remote is invalid
and TCP syn packets are dropped.

This patch adds a call to check disbled acceleration features
on Q-in-Q tagged traffic.

CC: Sathya Perla <sathya.perla@broadcom.com>
CC: Ajit Khaparde <ajit.khaparde@broadcom.com>
CC: Sriharsha Basavapatna <sriharsha.basavapatna@broadcom.com>
CC: Somnath Kotur <somnath.kotur@broadcom.com>
Signed-off-by: Vladislav Yasevich <vyasevic@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2017-05-24 16:27:14 -04:00
Vlad Yasevich
35d2f80b07 vlan: Fix tcp checksum offloads in Q-in-Q vlans
It appears that TCP checksum offloading has been broken for
Q-in-Q vlans.  The behavior was execerbated by the
series
    commit afb0bc972b ("Merge branch 'stacked_vlan_tso'")
that that enabled accleleration features on stacked vlans.

However, event without that series, it is possible to trigger
this issue.  It just requires a lot more specialized configuration.

The root cause is the interaction between how
netdev_intersect_features() works, the features actually set on
the vlan devices and HW having the ability to run checksum with
longer headers.

The issue starts when netdev_interesect_features() replaces
NETIF_F_HW_CSUM with a combination of NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM,
if the HW advertises IP|IPV6 specific checksums.  This happens
for tagged and multi-tagged packets.   However, HW that enables
IP|IPV6 checksum offloading doesn't gurantee that packets with
arbitrarily long headers can be checksummed.

This patch disables IP|IPV6 checksums on the packet for multi-tagged
packets.

CC: Toshiaki Makita <makita.toshiaki@lab.ntt.co.jp>
CC: Michal Kubecek <mkubecek@suse.cz>
Signed-off-by: Vladislav Yasevich <vyasevic@redhat.com>
Acked-by: Toshiaki Makita <makita.toshiaki@lab.ntt.co.jp>
Signed-off-by: David S. Miller <davem@davemloft.net>
2017-05-24 16:27:14 -04:00
David S. Miller
39a7d52766 Merge branch 'mlxsw-tcpflags'
Jiri Pirko says:

====================
net: add tcp flags match support to flower and offload it

This patch adds support to dissect tcp flags, match on them using
flower classifier and offload such rules to mlxsw Spectrum devices.

v1->v2:
- removed no longer relevant comment from patch 1 as suggested by Or
- sent correct patches this time
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2017-05-24 16:22:13 -04:00
Jiri Pirko
8a41d845c4 mlxsw: spectrum_flower: Add support for tcp flags
Allow to offload rules that contain tcp flags within the mask.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2017-05-24 16:22:11 -04:00
Jiri Pirko
dea2d6457f mlxsw: spectrum: Add acl block containing tcp flags for ipv4
Add acl block called "ipv4" which contains tcp flags.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2017-05-24 16:22:11 -04:00
Jiri Pirko
b4d39b4b54 mlxsw: acl: Add tcp flags acl element
Define new element for tcp flags and place it into scratch area.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2017-05-24 16:22:11 -04:00