[stable] [PATCH 1/2] coresight: etb10: Do not call smp_processor_id from preemptible

[stable] [PATCH 1/2] coresight: etb10: Do not call smp_processor_id from preemptible

[Suzuki K Poulose]

[ Upstream commit 730766bae3280a25d40ea76a53dc6342e84e6513 ]

During a perf session we try to allocate buffers on the "node" associated
with the CPU the event is bound to. If it is not bound to a CPU, we
use the current CPU node, using smp_processor_id(). However this is unsafe
in a pre-emptible context and could generate the splats as below :

 BUG: using smp_processor_id() in preemptible [00000000] code: perf/2544

Use NUMA_NO_NODE hint instead of using the current node for events
not bound to CPUs.

Fixes: 2997aa4063d97fdb39 ("coresight: etb10: implementing AUX API")
Cc: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Cc: stable <stable@vger.kernel.org> # v4.9 to v4.19
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Link: https://lore.kernel.org/r/20190620221237.3536-5-mathieu.poirier@linaro.org
---
 drivers/hwtracing/coresight/coresight-etb10.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/drivers/hwtracing/coresight/coresight-etb10.c b/drivers/hwtracing/coresight/coresight-etb10.c
index 0dad8626bcfb..6cf28b049635 100644
--- a/drivers/hwtracing/coresight/coresight-etb10.c
+++ b/drivers/hwtracing/coresight/coresight-etb10.c
@@ -275,9 +275,7 @@ static void *etb_alloc_buffer(struct coresight_device *csdev, int cpu,
 	int node;
 	struct cs_buffers *buf;
 
-	if (cpu == -1)
-		cpu = smp_processor_id();
-	node = cpu_to_node(cpu);
+	node = (cpu == -1) ? NUMA_NO_NODE : cpu_to_node(cpu);
 
 	buf = kzalloc_node(sizeof(struct cs_buffers), GFP_KERNEL, node);
 	if (!buf)
-- 
2.24.1

[stable] [PATCH 2/2] coresight: tmc-etf: Do not call smp_processor_id from preemptible

[Suzuki K Poulose]

[ Upstream commit 024c1fd9dbcc1d8a847f1311f999d35783921b7f ]

During a perf session we try to allocate buffers on the "node" associated
with the CPU the event is bound to. If it is not bound to a CPU, we
use the current CPU node, using smp_processor_id(). However this is unsafe
in a pre-emptible context and could generate the splats as below :

 BUG: using smp_processor_id() in preemptible [00000000] code: perf/2544
 caller is tmc_alloc_etf_buffer+0x5c/0x60
 CPU: 2 PID: 2544 Comm: perf Not tainted 5.1.0-rc6-147786-g116841e #344
 Hardware name: ARM LTD ARM Juno Development Platform/ARM Juno Development Platform, BIOS EDK II Feb  1 2019
 Call trace:
  dump_backtrace+0x0/0x150
  show_stack+0x14/0x20
  dump_stack+0x9c/0xc4
  debug_smp_processor_id+0x10c/0x110
  tmc_alloc_etf_buffer+0x5c/0x60
  etm_setup_aux+0x1c4/0x230
  rb_alloc_aux+0x1b8/0x2b8
  perf_mmap+0x35c/0x478
  mmap_region+0x34c/0x4f0
  do_mmap+0x2d8/0x418
  vm_mmap_pgoff+0xd0/0xf8
  ksys_mmap_pgoff+0x88/0xf8
  __arm64_sys_mmap+0x28/0x38
  el0_svc_handler+0xd8/0x138
  el0_svc+0x8/0xc

Use NUMA_NO_NODE hint instead of using the current node for events
not bound to CPUs.

Fixes: 2e499bbc1a929ac ("coresight: tmc: implementing TMC-ETF AUX space API")
Cc: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Cc: stable <stable@vger.kernel.org> # v4.9 to v4.19
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Link: https://lore.kernel.org/r/20190620221237.3536-4-mathieu.poirier@linaro.org
---
 drivers/hwtracing/coresight/coresight-tmc-etf.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/drivers/hwtracing/coresight/coresight-tmc-etf.c b/drivers/hwtracing/coresight/coresight-tmc-etf.c
index e31061308e19..e90af39283b1 100644
--- a/drivers/hwtracing/coresight/coresight-tmc-etf.c
+++ b/drivers/hwtracing/coresight/coresight-tmc-etf.c
@@ -304,9 +304,7 @@ static void *tmc_alloc_etf_buffer(struct coresight_device *csdev, int cpu,
 	int node;
 	struct cs_buffers *buf;
 
-	if (cpu == -1)
-		cpu = smp_processor_id();
-	node = cpu_to_node(cpu);
+	node = (cpu == -1) ? NUMA_NO_NODE : cpu_to_node(cpu);
 
 	/* Allocate memory structure for interaction with Perf */
 	buf = kzalloc_node(sizeof(struct cs_buffers), GFP_KERNEL, node);
-- 
2.24.1

Re: [stable] [PATCH 1/2] coresight: etb10: Do not call smp_processor_id from preemptible

[Sasha Levin]

On Wed, Jan 08, 2020 at 11:05:40AM +0000, Suzuki K Poulose wrote:
>[ Upstream commit 730766bae3280a25d40ea76a53dc6342e84e6513 ]
>
>During a perf session we try to allocate buffers on the "node" associated
>with the CPU the event is bound to. If it is not bound to a CPU, we
>use the current CPU node, using smp_processor_id(). However this is unsafe
>in a pre-emptible context and could generate the splats as below :
>
> BUG: using smp_processor_id() in preemptible [00000000] code: perf/2544
>
>Use NUMA_NO_NODE hint instead of using the current node for events
>not bound to CPUs.
>
>Fixes: 2997aa4063d97fdb39 ("coresight: etb10: implementing AUX API")
>Cc: Mathieu Poirier <mathieu.poirier@linaro.org>
>Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
>Cc: stable <stable@vger.kernel.org> # v4.9 to v4.19
>Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
>Link: https://lore.kernel.org/r/20190620221237.3536-5-mathieu.poirier@linaro.org

I've queued this for 4.9-4.19. There was a simple conflict on 4.9 which
also had to be resolved.

-- 
Thanks,
Sasha

Re: [stable] [PATCH 1/2] coresight: etb10: Do not call smp_processor_id from preemptible

[Suzuki Kuruppassery Poulose]

On 09/01/2020 14:35, Sasha Levin wrote:
> On Wed, Jan 08, 2020 at 11:05:40AM +0000, Suzuki K Poulose wrote:
>> [ Upstream commit 730766bae3280a25d40ea76a53dc6342e84e6513 ]
>>
>> During a perf session we try to allocate buffers on the "node" associated
>> with the CPU the event is bound to. If it is not bound to a CPU, we
>> use the current CPU node, using smp_processor_id(). However this is 
>> unsafe
>> in a pre-emptible context and could generate the splats as below :
>>
>> BUG: using smp_processor_id() in preemptible [00000000] code: perf/2544
>>
>> Use NUMA_NO_NODE hint instead of using the current node for events
>> not bound to CPUs.
>>
>> Fixes: 2997aa4063d97fdb39 ("coresight: etb10: implementing AUX API")
>> Cc: Mathieu Poirier <mathieu.poirier@linaro.org>
>> Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
>> Cc: stable <stable@vger.kernel.org> # v4.9 to v4.19
>> Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
>> Link: 
>> https://lore.kernel.org/r/20190620221237.3536-5-mathieu.poirier@linaro.org 
>>
> 
> I've queued this for 4.9-4.19. There was a simple conflict on 4.9 which
> also had to be resolved.
> 


Thanks Sasha !

Suzuki

Re: [stable] [PATCH 1/2] coresight: etb10: Do not call smp_processor_id from preemptible

[Greg KH]

On Thu, Jan 09, 2020 at 02:36:17PM +0000, Suzuki Kuruppassery Poulose wrote:
> On 09/01/2020 14:35, Sasha Levin wrote:
> > On Wed, Jan 08, 2020 at 11:05:40AM +0000, Suzuki K Poulose wrote:
> > > [ Upstream commit 730766bae3280a25d40ea76a53dc6342e84e6513 ]
> > > 
> > > During a perf session we try to allocate buffers on the "node" associated
> > > with the CPU the event is bound to. If it is not bound to a CPU, we
> > > use the current CPU node, using smp_processor_id(). However this is
> > > unsafe
> > > in a pre-emptible context and could generate the splats as below :
> > > 
> > > BUG: using smp_processor_id() in preemptible [00000000] code: perf/2544
> > > 
> > > Use NUMA_NO_NODE hint instead of using the current node for events
> > > not bound to CPUs.
> > > 
> > > Fixes: 2997aa4063d97fdb39 ("coresight: etb10: implementing AUX API")
> > > Cc: Mathieu Poirier <mathieu.poirier@linaro.org>
> > > Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
> > > Cc: stable <stable@vger.kernel.org> # v4.9 to v4.19
> > > Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
> > > Link: https://lore.kernel.org/r/20190620221237.3536-5-mathieu.poirier@linaro.org
> > > 
> > 
> > I've queued this for 4.9-4.19. There was a simple conflict on 4.9 which
> > also had to be resolved.
> > 
> 
> 
> Thanks Sasha !

Note, these had to all be dropped as they broke the build :(

So can you please send us patches that at least build?  :)

thanks,

greg k-h

Re: [stable] [PATCH 1/2] coresight: etb10: Do not call smp_processor_id from preemptible

[Suzuki Kuruppassery Poulose]

Hi Greg,

On 15/01/2020 15:11, Greg KH wrote:
> On Thu, Jan 09, 2020 at 02:36:17PM +0000, Suzuki Kuruppassery Poulose wrote:
>> On 09/01/2020 14:35, Sasha Levin wrote:
>>> On Wed, Jan 08, 2020 at 11:05:40AM +0000, Suzuki K Poulose wrote:
>>>> [ Upstream commit 730766bae3280a25d40ea76a53dc6342e84e6513 ]
>>>>
>>>> During a perf session we try to allocate buffers on the "node" associated
>>>> with the CPU the event is bound to. If it is not bound to a CPU, we
>>>> use the current CPU node, using smp_processor_id(). However this is
>>>> unsafe
>>>> in a pre-emptible context and could generate the splats as below :
>>>>
>>>> BUG: using smp_processor_id() in preemptible [00000000] code: perf/2544
>>>>
>>>> Use NUMA_NO_NODE hint instead of using the current node for events
>>>> not bound to CPUs.
>>>>
>>>> Fixes: 2997aa4063d97fdb39 ("coresight: etb10: implementing AUX API")
>>>> Cc: Mathieu Poirier <mathieu.poirier@linaro.org>
>>>> Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
>>>> Cc: stable <stable@vger.kernel.org> # v4.9 to v4.19
>>>> Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
>>>> Link: https://lore.kernel.org/r/20190620221237.3536-5-mathieu.poirier@linaro.org
>>>>
>>>
>>> I've queued this for 4.9-4.19. There was a simple conflict on 4.9 which
>>> also had to be resolved.
>>>
>>
>>
>> Thanks Sasha !
> 
> Note, these had to all be dropped as they broke the build :(
> 
> So can you please send us patches that at least build?  :)
> 

Do you have a build failure log ? I did build test it before sending it 
over. I tried it again on 4.9, 4.14 and 4.19. I don't hit any build
failures here.

Please could you share the log if you have it handy ?

Suzuki

Re: [stable] [PATCH 1/2] coresight: etb10: Do not call smp_processor_id from preemptible

[Greg KH]

On Wed, Jan 15, 2020 at 04:44:29PM +0000, Suzuki Kuruppassery Poulose wrote:
> 
> Hi Greg,
> 
> On 15/01/2020 15:11, Greg KH wrote:
> > On Thu, Jan 09, 2020 at 02:36:17PM +0000, Suzuki Kuruppassery Poulose wrote:
> > > On 09/01/2020 14:35, Sasha Levin wrote:
> > > > On Wed, Jan 08, 2020 at 11:05:40AM +0000, Suzuki K Poulose wrote:
> > > > > [ Upstream commit 730766bae3280a25d40ea76a53dc6342e84e6513 ]
> > > > > 
> > > > > During a perf session we try to allocate buffers on the "node" associated
> > > > > with the CPU the event is bound to. If it is not bound to a CPU, we
> > > > > use the current CPU node, using smp_processor_id(). However this is
> > > > > unsafe
> > > > > in a pre-emptible context and could generate the splats as below :
> > > > > 
> > > > > BUG: using smp_processor_id() in preemptible [00000000] code: perf/2544
> > > > > 
> > > > > Use NUMA_NO_NODE hint instead of using the current node for events
> > > > > not bound to CPUs.
> > > > > 
> > > > > Fixes: 2997aa4063d97fdb39 ("coresight: etb10: implementing AUX API")
> > > > > Cc: Mathieu Poirier <mathieu.poirier@linaro.org>
> > > > > Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
> > > > > Cc: stable <stable@vger.kernel.org> # v4.9 to v4.19
> > > > > Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
> > > > > Link: https://lore.kernel.org/r/20190620221237.3536-5-mathieu.poirier@linaro.org
> > > > > 
> > > > 
> > > > I've queued this for 4.9-4.19. There was a simple conflict on 4.9 which
> > > > also had to be resolved.
> > > > 
> > > 
> > > 
> > > Thanks Sasha !
> > 
> > Note, these had to all be dropped as they broke the build :(
> > 
> > So can you please send us patches that at least build?  :)
> > 
> 
> Do you have a build failure log ? I did build test it before sending it
> over. I tried it again on 4.9, 4.14 and 4.19. I don't hit any build
> failures here.
> 
> Please could you share the log if you have it handy ?

It was in the stable -rc review emails, I don't have it handy, sorry.

greg k-h

Re: [stable] [PATCH 1/2] coresight: etb10: Do not call smp_processor_id from preemptible

[Suzuki Kuruppassery Poulose]

On 15/01/2020 17:21, Greg KH wrote:
> On Wed, Jan 15, 2020 at 04:44:29PM +0000, Suzuki Kuruppassery Poulose wrote:
>>
>> Hi Greg,
>>
>> On 15/01/2020 15:11, Greg KH wrote:
>>> On Thu, Jan 09, 2020 at 02:36:17PM +0000, Suzuki Kuruppassery Poulose wrote:
>>>> On 09/01/2020 14:35, Sasha Levin wrote:
>>>>> On Wed, Jan 08, 2020 at 11:05:40AM +0000, Suzuki K Poulose wrote:
>>>>>> [ Upstream commit 730766bae3280a25d40ea76a53dc6342e84e6513 ]
>>>>>>
>>>>>> During a perf session we try to allocate buffers on the "node" associated
>>>>>> with the CPU the event is bound to. If it is not bound to a CPU, we
>>>>>> use the current CPU node, using smp_processor_id(). However this is
>>>>>> unsafe
>>>>>> in a pre-emptible context and could generate the splats as below :
>>>>>>
>>>>>> BUG: using smp_processor_id() in preemptible [00000000] code: perf/2544
>>>>>>
>>>>>> Use NUMA_NO_NODE hint instead of using the current node for events
>>>>>> not bound to CPUs.
>>>>>>
>>>>>> Fixes: 2997aa4063d97fdb39 ("coresight: etb10: implementing AUX API")
>>>>>> Cc: Mathieu Poirier <mathieu.poirier@linaro.org>
>>>>>> Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
>>>>>> Cc: stable <stable@vger.kernel.org> # v4.9 to v4.19
>>>>>> Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
>>>>>> Link: https://lore.kernel.org/r/20190620221237.3536-5-mathieu.poirier@linaro.org
>>>>>>
>>>>>
>>>>> I've queued this for 4.9-4.19. There was a simple conflict on 4.9 which
>>>>> also had to be resolved.
>>>>>
>>>>
>>>>
>>>> Thanks Sasha !
>>>
>>> Note, these had to all be dropped as they broke the build :(
>>>
>>> So can you please send us patches that at least build?  :)
>>>
>>
>> Do you have a build failure log ? I did build test it before sending it
>> over. I tried it again on 4.9, 4.14 and 4.19. I don't hit any build
>> failures here.
>>
>> Please could you share the log if you have it handy ?
> 
> It was in the stable -rc review emails, I don't have it handy, sorry.
> 

I think there is a bit of confusion here. If you're referring to

https://lkml.org/lkml/2020/1/11/634

as the build failure report, this is precisely my series fixes.
I sent this series to address the build break reported by Nathan.
The original patches were picked up from the "Fixes" tag automatically
which broke the build due to missing "event" parameter. This series
fixes those build issues and for sure builds fine for the affected
versions. Trust me ;-)

Cheers
Suzuki



> greg k-h
> 

Re: [stable] [PATCH 1/2] coresight: etb10: Do not call smp_processor_id from preemptible

[Greg KH]

On Wed, Jan 15, 2020 at 05:28:19PM +0000, Suzuki Kuruppassery Poulose wrote:
> On 15/01/2020 17:21, Greg KH wrote:
> > On Wed, Jan 15, 2020 at 04:44:29PM +0000, Suzuki Kuruppassery Poulose wrote:
> > > 
> > > Hi Greg,
> > > 
> > > On 15/01/2020 15:11, Greg KH wrote:
> > > > On Thu, Jan 09, 2020 at 02:36:17PM +0000, Suzuki Kuruppassery Poulose wrote:
> > > > > On 09/01/2020 14:35, Sasha Levin wrote:
> > > > > > On Wed, Jan 08, 2020 at 11:05:40AM +0000, Suzuki K Poulose wrote:
> > > > > > > [ Upstream commit 730766bae3280a25d40ea76a53dc6342e84e6513 ]
> > > > > > > 
> > > > > > > During a perf session we try to allocate buffers on the "node" associated
> > > > > > > with the CPU the event is bound to. If it is not bound to a CPU, we
> > > > > > > use the current CPU node, using smp_processor_id(). However this is
> > > > > > > unsafe
> > > > > > > in a pre-emptible context and could generate the splats as below :
> > > > > > > 
> > > > > > > BUG: using smp_processor_id() in preemptible [00000000] code: perf/2544
> > > > > > > 
> > > > > > > Use NUMA_NO_NODE hint instead of using the current node for events
> > > > > > > not bound to CPUs.
> > > > > > > 
> > > > > > > Fixes: 2997aa4063d97fdb39 ("coresight: etb10: implementing AUX API")
> > > > > > > Cc: Mathieu Poirier <mathieu.poirier@linaro.org>
> > > > > > > Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
> > > > > > > Cc: stable <stable@vger.kernel.org> # v4.9 to v4.19
> > > > > > > Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
> > > > > > > Link: https://lore.kernel.org/r/20190620221237.3536-5-mathieu.poirier@linaro.org
> > > > > > > 
> > > > > > 
> > > > > > I've queued this for 4.9-4.19. There was a simple conflict on 4.9 which
> > > > > > also had to be resolved.
> > > > > > 
> > > > > 
> > > > > 
> > > > > Thanks Sasha !
> > > > 
> > > > Note, these had to all be dropped as they broke the build :(
> > > > 
> > > > So can you please send us patches that at least build?  :)
> > > > 
> > > 
> > > Do you have a build failure log ? I did build test it before sending it
> > > over. I tried it again on 4.9, 4.14 and 4.19. I don't hit any build
> > > failures here.
> > > 
> > > Please could you share the log if you have it handy ?
> > 
> > It was in the stable -rc review emails, I don't have it handy, sorry.
> > 
> 
> I think there is a bit of confusion here. If you're referring to
> 
> https://lkml.org/lkml/2020/1/11/634
> 
> as the build failure report, this is precisely my series fixes.
> I sent this series to address the build break reported by Nathan.
> The original patches were picked up from the "Fixes" tag automatically
> which broke the build due to missing "event" parameter. This series
> fixes those build issues and for sure builds fine for the affected
> versions. Trust me ;-)

Ok, for some reason it looked like the "original" commits were added to
the tree, not your backports.  I've queued up your backports now, that
should solve the issue.

thanks,

greg k-h