Re: 3.9 merge window kernels

Re: 3.9 merge window kernels

[poma]

On 02/25/13 01:22, Antti Palosaari wrote:
[…]
>>> Poma, you should probably just start filing bugs for things you hit.
>>> In this case, the skge backtrace is just a warning but it can be fixed
>>> up relatively easily.
>>>
>>
>> https://bugzilla.redhat.com/show_bug.cgi?id=914994
>> Josh, Mauro thanks for the overview. :)
>> Antti, cheers. ;)
>>
>> poma
> 
> I cannot see these warnings at all. What is Kernel option to enable
> those debug(?) warnings? From which menu it could be located when make
> menuconfig ?
> 

http://kojipkgs.fedoraproject.org/packages/kernel/3.9.0/0.rc0.git7.1.fc19/data/logs/x86_64/build.log
Even after reapplying Stephen's "skge: check for PCI dma mapping
errors"[1] from David's 'net-next' tree, no luck.
…
WARNING: at lib/dma-debug.c:933 check_unmap+0x47b/0x960()
skge 0000:01:09.0: DMA-API: device driver failed to check map
error[device address=0x000000010287094a] [size=90 bytes] [mapped as single]
…


Cheers,
poma


[1]
http://git.kernel.org/?p=linux/kernel/git/davem/net-next.git;a=commit;h=d0249e44432aa0ffcf710b64449b8eaa3722547e



_______________________________________________
kernel mailing list
kernel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/kernel

Re: 3.9 merge window kernels

[Josh Boyer]

On Tue, Feb 26, 2013 at 01:05:05AM +0100, poma wrote:
> On 02/25/13 01:22, Antti Palosaari wrote:
> […]
> >>> Poma, you should probably just start filing bugs for things you hit.
> >>> In this case, the skge backtrace is just a warning but it can be fixed
> >>> up relatively easily.
> >>>
> >>
> >> https://bugzilla.redhat.com/show_bug.cgi?id=914994
> >> Josh, Mauro thanks for the overview. :)
> >> Antti, cheers. ;)
> >>
> >> poma
> > 
> > I cannot see these warnings at all. What is Kernel option to enable
> > those debug(?) warnings? From which menu it could be located when make
> > menuconfig ?
> > 
> 
> http://kojipkgs.fedoraproject.org/packages/kernel/3.9.0/0.rc0.git7.1.fc19/data/logs/x86_64/build.log
> Even after reapplying Stephen's "skge: check for PCI dma mapping
> errors"[1] from David's 'net-next' tree, no luck.
> …
> WARNING: at lib/dma-debug.c:933 check_unmap+0x47b/0x960()
> skge 0000:01:09.0: DMA-API: device driver failed to check map
> error[device address=0x000000010287094a] [size=90 bytes] [mapped as single]
> …

I'm confused.  You pointed to the build I did this morning, which
definitely doesn't include the commit you mentioned.  Are you saying you
took this morning's build and applied the patch yourself?  If so, did it
really get applied?  Do you have logs?  Do you have more than just that
tiny snippet of error message?

josh

Re: 3.9 merge window kernels

[poma]

On 02/26/13 02:12, Josh Boyer wrote:
> On Tue, Feb 26, 2013 at 01:05:05AM +0100, poma wrote:
>> On 02/25/13 01:22, Antti Palosaari wrote:
>> […]
>>>>> Poma, you should probably just start filing bugs for things you hit.
>>>>> In this case, the skge backtrace is just a warning but it can be fixed
>>>>> up relatively easily.
>>>>>
>>>>
>>>> https://bugzilla.redhat.com/show_bug.cgi?id=914994
>>>> Josh, Mauro thanks for the overview. :)
>>>> Antti, cheers. ;)
>>>>
>>>> poma
>>>
>>> I cannot see these warnings at all. What is Kernel option to enable
>>> those debug(?) warnings? From which menu it could be located when make
>>> menuconfig ?
>>>
>>
>> http://kojipkgs.fedoraproject.org/packages/kernel/3.9.0/0.rc0.git7.1.fc19/data/logs/x86_64/build.log
>> Even after reapplying Stephen's "skge: check for PCI dma mapping
>> errors"[1] from David's 'net-next' tree, no luck.
>> …
>> WARNING: at lib/dma-debug.c:933 check_unmap+0x47b/0x960()
>> skge 0000:01:09.0: DMA-API: device driver failed to check map
>> error[device address=0x000000010287094a] [size=90 bytes] [mapped as single]
>> …
> 
> I'm confused.  You pointed to the build I did this morning, which
> definitely doesn't include the commit you mentioned.  Are you saying you
> took this morning's build and applied the patch yourself?  If so, did it
> really get applied?  Do you have logs?  Do you have more than just that
> tiny snippet of error message?
> 

build.log is for Antti's eyes only. :)
Logs are virtually the same, with or without that *old* commit, which is
actually replaced by this one[1].
So of course it isn't in 3.9.0-0.rc0.git7.1.fc19.x86_64. ;)
Building particular module groups(skge & Co.) rather than build an
entire kernel tree isn't big deal, likewise.
You know that better than me. ;)
The real question is whether someone will help squeeze this bug.


Cheers,
poma


[1]
http://git.kernel.org/?p=linux/kernel/git/davem/net-next.git;a=commit;h=aadf1f0fc851d71095d2773bcda73e1423e6b730

Re: 3.9 merge window kernels

[Antti Palosaari]

On 02/26/2013 04:18 AM, poma wrote:
> On 02/26/13 02:12, Josh Boyer wrote:
>> On Tue, Feb 26, 2013 at 01:05:05AM +0100, poma wrote:
>>> On 02/25/13 01:22, Antti Palosaari wrote:
>>> […]
>>>>>> Poma, you should probably just start filing bugs for things you hit.
>>>>>> In this case, the skge backtrace is just a warning but it can be fixed
>>>>>> up relatively easily.
>>>>>>
>>>>>
>>>>> https://bugzilla.redhat.com/show_bug.cgi?id=914994
>>>>> Josh, Mauro thanks for the overview. :)
>>>>> Antti, cheers. ;)
>>>>>
>>>>> poma
>>>>
>>>> I cannot see these warnings at all. What is Kernel option to enable
>>>> those debug(?) warnings? From which menu it could be located when make
>>>> menuconfig ?
>>>>
>>>
>>> http://kojipkgs.fedoraproject.org/packages/kernel/3.9.0/0.rc0.git7.1.fc19/data/logs/x86_64/build.log
>>> Even after reapplying Stephen's "skge: check for PCI dma mapping
>>> errors"[1] from David's 'net-next' tree, no luck.
>>> …
>>> WARNING: at lib/dma-debug.c:933 check_unmap+0x47b/0x960()
>>> skge 0000:01:09.0: DMA-API: device driver failed to check map
>>> error[device address=0x000000010287094a] [size=90 bytes] [mapped as single]
>>> …
>>
>> I'm confused.  You pointed to the build I did this morning, which
>> definitely doesn't include the commit you mentioned.  Are you saying you
>> took this morning's build and applied the patch yourself?  If so, did it
>> really get applied?  Do you have logs?  Do you have more than just that
>> tiny snippet of error message?
>>
>
> build.log is for Antti's eyes only. :)
> Logs are virtually the same, with or without that *old* commit, which is
> actually replaced by this one[1].
> So of course it isn't in 3.9.0-0.rc0.git7.1.fc19.x86_64. ;)
> Building particular module groups(skge & Co.) rather than build an
> entire kernel tree isn't big deal, likewise.
> You know that better than me. ;)
> The real question is whether someone will help squeeze this bug.

I want just know if I could reproduce that AF9015 error message or was 
it just warning. Is it something I have to fix for Kernel 3.10 or 
earlier. I am running Fedora 17 AMD64 and could surely compile any 
Kernel needed.

regards
Antti

-- 
http://palosaari.fi/

Re: 3.9 merge window kernels

[Antti Palosaari]

On 02/26/2013 09:51 AM, Antti Palosaari wrote:
> On 02/26/2013 04:18 AM, poma wrote:
>> On 02/26/13 02:12, Josh Boyer wrote:
>>> On Tue, Feb 26, 2013 at 01:05:05AM +0100, poma wrote:
>>>> On 02/25/13 01:22, Antti Palosaari wrote:
>>>> […]
>>>>>>> Poma, you should probably just start filing bugs for things you hit.
>>>>>>> In this case, the skge backtrace is just a warning but it can be
>>>>>>> fixed
>>>>>>> up relatively easily.
>>>>>>>
>>>>>>
>>>>>> https://bugzilla.redhat.com/show_bug.cgi?id=914994
>>>>>> Josh, Mauro thanks for the overview. :)
>>>>>> Antti, cheers. ;)
>>>>>>
>>>>>> poma
>>>>>
>>>>> I cannot see these warnings at all. What is Kernel option to enable
>>>>> those debug(?) warnings? From which menu it could be located when make
>>>>> menuconfig ?
>>>>>
>>>>
>>>> http://kojipkgs.fedoraproject.org/packages/kernel/3.9.0/0.rc0.git7.1.fc19/data/logs/x86_64/build.log
>>>>
>>>> Even after reapplying Stephen's "skge: check for PCI dma mapping
>>>> errors"[1] from David's 'net-next' tree, no luck.
>>>> …
>>>> WARNING: at lib/dma-debug.c:933 check_unmap+0x47b/0x960()
>>>> skge 0000:01:09.0: DMA-API: device driver failed to check map
>>>> error[device address=0x000000010287094a] [size=90 bytes] [mapped as
>>>> single]
>>>> …
>>>
>>> I'm confused.  You pointed to the build I did this morning, which
>>> definitely doesn't include the commit you mentioned.  Are you saying you
>>> took this morning's build and applied the patch yourself?  If so, did it
>>> really get applied?  Do you have logs?  Do you have more than just that
>>> tiny snippet of error message?
>>>
>>
>> build.log is for Antti's eyes only. :)
>> Logs are virtually the same, with or without that *old* commit, which is
>> actually replaced by this one[1].
>> So of course it isn't in 3.9.0-0.rc0.git7.1.fc19.x86_64. ;)
>> Building particular module groups(skge & Co.) rather than build an
>> entire kernel tree isn't big deal, likewise.
>> You know that better than me. ;)
>> The real question is whether someone will help squeeze this bug.
>
> I want just know if I could reproduce that AF9015 error message or was
> it just warning. Is it something I have to fix for Kernel 3.10 or
> earlier. I am running Fedora 17 AMD64 and could surely compile any
> Kernel needed.

OK, found it finally.
Kernel hacking  --->  Enable debugging of DMA-API usage

There was rather many DVB USB drivers using USB bulk buffers from the 
stack. I will fix at least some of those, lets say for 3.10.

regards
Antti
-- 
http://palosaari.fi/