FSL P5020/P5040: DPAA Ethernet issue with the latest Git kernel

5 messages, 2 authors, 2020-07-07 · open the first message on its own page

FSL P5020/P5040: DPAA Ethernet issue with the latest Git kernel

From: Christian Zigotzky <hidden>
Date: 2020-06-21 06:54:55

Hello Alexander,

The DPAA Ethernet doesn’t work anymore on our FSL P5020/P5040 boards [1] since the RC1 of kernel 5.8 [2].
We bisected last days [3] and found the problematic commit [4]. I was able to revert it [5]. After that the DPAA Ethernet works again. I created a patch for reverting the commit [5]. This patch works and I will use it for the RC2.
Could you please check your commit? [4]

Thanks,
Christian

[1] http://wiki.amiga.org/index.php?title=X5000
[2] https://forum.hyperion-entertainment.com/viewtopic.php?p=50885#p50885
[3] https://forum.hyperion-entertainment.com/viewtopic.php?p=50892#p50892
[4] lib: fix bitmap_parse() on 64-bit big endian archs: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=81c4f4d924d5d009b5ed785a3e22b18d0f7b831f
[5] https://forum.hyperion-entertainment.com/viewtopic.php?p=50982#p50982

Re: FSL P5020/P5040: DPAA Ethernet issue with the latest Git kernel

From: Alexander Gordeev <agordeev@linux.ibm.com>
Date: 2020-06-24 21:17:31

On Sun, Jun 21, 2020 at 08:40:14AM +0200, Christian Zigotzky wrote:
Hello Alexander,

The DPAA Ethernet doesn’t work anymore on our FSL P5020/P5040 boards [1] since the RC1 of kernel 5.8 [2].
We bisected last days [3] and found the problematic commit [4]. I was able to revert it [5]. After that the DPAA Ethernet works again. I created a patch for reverting the commit [5]. This patch works and I will use it for the RC2.
Could you please check your commit? [4]
Hi Christian,

Could you please check if the kernel passes CONFIG_TEST_BITMAP self-test?

Thanks!
Thanks,
Christian

[1] http://wiki.amiga.org/index.php?title=X5000
[2] https://forum.hyperion-entertainment.com/viewtopic.php?p=50885#p50885
[3] https://forum.hyperion-entertainment.com/viewtopic.php?p=50892#p50892
[4] lib: fix bitmap_parse() on 64-bit big endian archs: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=81c4f4d924d5d009b5ed785a3e22b18d0f7b831f
[5] https://forum.hyperion-entertainment.com/viewtopic.php?p=50982#p50982

Re: FSL P5020/P5040: DPAA Ethernet issue with the latest Git kernel

From: Christian Zigotzky <hidden>
Date: 2020-06-24 23:03:54

On 24 June 2020, at 10:23 am, Alexander Gordeev wrote:
On Sun, Jun 21, 2020 at 08:40:14AM +0200, Christian Zigotzky wrote:
quoted
Hello Alexander,

The DPAA Ethernet doesn’t work anymore on our FSL P5020/P5040 boards [1] since the RC1 of kernel 5.8 [2].
We bisected last days [3] and found the problematic commit [4]. I was able to revert it [5]. After that the DPAA Ethernet works again. I created a patch for reverting the commit [5]. This patch works and I will use it for the RC2.
Could you please check your commit? [4]
Hi Christian,

Could you please check if the kernel passes CONFIG_TEST_BITMAP self-test?

Thanks!
quoted
Thanks,
Christian

[1] http://wiki.amiga.org/index.php?title=X5000
[2] https://forum.hyperion-entertainment.com/viewtopic.php?p=50885#p50885
[3] https://forum.hyperion-entertainment.com/viewtopic.php?p=50892#p50892
[4] lib: fix bitmap_parse() on 64-bit big endian archs: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=81c4f4d924d5d009b5ed785a3e22b18d0f7b831f
[5] https://forum.hyperion-entertainment.com/viewtopic.php?p=50982#p50982
Hi Alexander,

Thanks for your reply!

I compiled a test kernel with the option "CONFIG_TEST_BITMAP=y" 
yesterday. After that Skateman and I booted it and looked for the bitmap 
tests with "dmesg | grep -i bitmap".

Results:

FSL P5020:

[    0.297756] test_bitmap: loaded.
[    0.298113] test_bitmap: parselist: 14: input is '0-2047:128/256' OK, 
Time: 562
[    0.298142] test_bitmap: parselist_user: 14: input is 
'0-2047:128/256' OK, Time: 761
[    0.301553] test_bitmap: all 1663 tests passed

FSL P5040:

[    0.296563] test_bitmap: loaded.
[    0.296894] test_bitmap: parselist: 14: input is '0-2047:128/256' OK, 
Time: 540
[    0.296920] test_bitmap: parselist_user: 14: input is 
'0-2047:128/256' OK, Time: 680
[    0.299994] test_bitmap: all 1663 tests passed

Thanks,
Christian

Re: FSL P5020/P5040: DPAA Ethernet issue with the latest Git kernel

From: Alexander Gordeev <agordeev@linux.ibm.com>
Date: 2020-06-25 10:24:26

On Thu, Jun 25, 2020 at 01:01:52AM +0200, Christian Zigotzky wrote:
[...]
I compiled a test kernel with the option "CONFIG_TEST_BITMAP=y"
yesterday. After that Skateman and I booted it and looked for the
bitmap tests with "dmesg | grep -i bitmap".

Results:

FSL P5020:

[    0.297756] test_bitmap: loaded.
[    0.298113] test_bitmap: parselist: 14: input is '0-2047:128/256'
OK, Time: 562
[    0.298142] test_bitmap: parselist_user: 14: input is
'0-2047:128/256' OK, Time: 761
[    0.301553] test_bitmap: all 1663 tests passed

FSL P5040:

[    0.296563] test_bitmap: loaded.
[    0.296894] test_bitmap: parselist: 14: input is '0-2047:128/256'
OK, Time: 540
[    0.296920] test_bitmap: parselist_user: 14: input is
'0-2047:128/256' OK, Time: 680
[    0.299994] test_bitmap: all 1663 tests passed
Thanks for the test! So it works as expected.

I would suggest to compare what is going on on the device probing
with and without the bisected commit.

There seems to be MAC and PHY mode initialization issue that might
resulted from the bitmap format change.

I put Madalin and Sascha on CC as they have done some works on
this part recently.

Thanks!

Thanks,
Christian

Re: FSL P5020/P5040: DPAA Ethernet issue with the latest Git kernel

From: Christian Zigotzky <hidden>
Date: 2020-07-07 07:54:25

On 25 June 2020 at 12:22 pm, Alexander Gordeev wrote:
On Thu, Jun 25, 2020 at 01:01:52AM +0200, Christian Zigotzky wrote:
[...]
quoted
I compiled a test kernel with the option "CONFIG_TEST_BITMAP=y"
yesterday. After that Skateman and I booted it and looked for the
bitmap tests with "dmesg | grep -i bitmap".

Results:

FSL P5020:

[    0.297756] test_bitmap: loaded.
[    0.298113] test_bitmap: parselist: 14: input is '0-2047:128/256'
OK, Time: 562
[    0.298142] test_bitmap: parselist_user: 14: input is
'0-2047:128/256' OK, Time: 761
[    0.301553] test_bitmap: all 1663 tests passed

FSL P5040:

[    0.296563] test_bitmap: loaded.
[    0.296894] test_bitmap: parselist: 14: input is '0-2047:128/256'
OK, Time: 540
[    0.296920] test_bitmap: parselist_user: 14: input is
'0-2047:128/256' OK, Time: 680
[    0.299994] test_bitmap: all 1663 tests passed
Thanks for the test! So it works as expected.

I would suggest to compare what is going on on the device probing
with and without the bisected commit.

There seems to be MAC and PHY mode initialization issue that might
resulted from the bitmap format change.

I put Madalin and Sascha on CC as they have done some works on
this part recently.

Thanks!
Hi All,

The issue still exists [1] so we still need the dpaa patch [2]. Could 
you please check the problematic commit [3]?

Thanks,
Christian

[1] https://forum.hyperion-entertainment.com/viewtopic.php?p=50885#p50885
[2] https://forum.hyperion-entertainment.com/viewtopic.php?p=50982#p50982
[3] https://forum.hyperion-entertainment.com/viewtopic.php?p=50980#p50980
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help