Kernel 5.10 USB issues

Kernel 5.10 USB issues

[Jupiter]

Hi,

I connected a USB bus between an iMX6ULZ and an 4G LTE modem, it was
running well in kernel 4.19.75. I've just upgraded kernel to 5.10.59,
the USB is broken, the USB drivers were disconnected / connected
repeatedly:

[12781.730889] usb 1-1: USB disconnect, device number 84
[12781.808317] option1 ttyUSB0: GSM modem (1-port) converter now
disconnected from ttyUSB0
[12781.839256] option 1-1:1.0: device disconnected
[12781.959687] option1 ttyUSB1: GSM modem (1-port) converter now
disconnected from ttyUSB1
[12782.009476] option 1-1:1.2: device disconnected
[12785.607012] usb 1-1: new high-speed USB device number 85 using ci_hdrc
[12785.825218] usb 1-1: New USB device found, idVendor=05c6,
idProduct=90b2, bcdDevice= 0.00
[12785.833828] usb 1-1: New USB device strings: Mfr=3, Product=2, SerialNumber=4
[12785.841222] usb 1-1: Product: Qualcomm CDMA Technologies MSM
[12785.847049] usb 1-1: Manufacturer: Qualcomm, Incorporated
[12785.852536] usb 1-1: SerialNumber: 674a1bcf
[12785.932542] option 1-1:1.0: GSM modem (1-port) converter detected
[12785.957103] usb 1-1: GSM modem (1-port) converter now attached to ttyUSB0
[12785.975924] option 1-1:1.2: GSM modem (1-port) converter detected
[12786.047708] usb 1-1: GSM modem (1-port) converter now attached to ttyUSB1

What could I be missing in kernel 5?

Thank you.

Kind regards,

- JH

Re: Kernel 5.10 USB issues

[Greg KH]

On Wed, Nov 24, 2021 at 11:46:06AM +1100, Jupiter wrote:
> Hi,
> 
> I connected a USB bus between an iMX6ULZ and an 4G LTE modem, it was
> running well in kernel 4.19.75. I've just upgraded kernel to 5.10.59,
> the USB is broken, the USB drivers were disconnected / connected
> repeatedly:
> 
> [12781.730889] usb 1-1: USB disconnect, device number 84
> [12781.808317] option1 ttyUSB0: GSM modem (1-port) converter now
> disconnected from ttyUSB0

That is usually a hardware problem, the kernel can not disconnect a
device from the bus through software.

But, if 4.19 is working, can you use 'git bisect' between the two
kernels to find what caused the problem?

thanks,

greg k-h

Re: Kernel 5.10 USB issues

[Jupiter]

Thanks Greg.

> That is usually a hardware problem, the kernel can not disconnect a
> device from the bus through software.

Understood, that is why I used the same HW to test both 4.19.75 and
5.10.59, that should rule out the HW problems, right?

> But, if 4.19 is working, can you use 'git bisect' between the two
> kernels to find what caused the problem?

My bad, I have never used git bisect before, a quick google search did
not help, the following command does not make sense at all, what are
right commands to check 'git bisect'  to find what caused the problem?

For 4.19 kernel
$ git bisect start
Already on 'linux-4.19.y'
Your branch is behind 'origin/linux-4.19.y' by 13117 commits, and can
be fast-forwarded.
  (use "git pull" to update your local branch)

For 5.10 kernel
$  git bisect start
Already on 'linux-5.10.y'
Your branch is behind 'origin/linux-5.10.y' by 2311 commits, and can
be fast-forwarded.
  (use "git pull" to update your local branch)

Thank you.

- jh

Re: Kernel 5.10 USB issues

[Greg KH]

On Wed, Nov 24, 2021 at 09:06:50PM +1100, Jupiter wrote:
> Thanks Greg.
> 
> > That is usually a hardware problem, the kernel can not disconnect a
> > device from the bus through software.
> 
> Understood, that is why I used the same HW to test both 4.19.75 and
> 5.10.59, that should rule out the HW problems, right?
> 
> > But, if 4.19 is working, can you use 'git bisect' between the two
> > kernels to find what caused the problem?
> 
> My bad, I have never used git bisect before, a quick google search did
> not help, the following command does not make sense at all, what are
> right commands to check 'git bisect'  to find what caused the problem?
> 
> For 4.19 kernel
> $ git bisect start
> Already on 'linux-4.19.y'
> Your branch is behind 'origin/linux-4.19.y' by 13117 commits, and can
> be fast-forwarded.
>   (use "git pull" to update your local branch)
> 
> For 5.10 kernel
> $  git bisect start
> Already on 'linux-5.10.y'
> Your branch is behind 'origin/linux-5.10.y' by 2311 commits, and can
> be fast-forwarded.
>   (use "git pull" to update your local branch)

Start working on the "master" branch and do:
	git bisect start
	git bisect good v4.19
	git bisect bad v5.10

and then go from there, building and testing the kernels it gives you.

Trying to bisect across the stable branches can be done, but it is very
tricky and not something that if you have not done 'git bisect' before,
I would recommend.  The above should be sufficient.

thanks,

greg k-h

Re: Kernel 5.10 USB issues

[Jupiter]

Hi Greg,

Thanks for your patience, here is what I did. I thought I followed the
instructions, any useful information?

> Start working on the "master" branch and do:

$ git bisect start

$ git bisect good v4.19

$ git bisect bad v5.10

Bisecting: 91917 revisions left to test after this (roughly 17 steps)
[386403a115f95997c2715691226e11a7b5cffcfd] Merge
git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next

$ git bisect good
Bisecting: 46699 revisions left to test after this (roughly 16 steps)
[2e63f6ce7ed2c4ff83ba30ad9ccad422289a6c63] Merge branch
'uaccess.comedi' of
git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs

$ git bisect good
Bisecting: 23349 revisions left to test after this (roughly 15 steps)
[0f5d0a4c01cc1b7b0590baa18809a96d45a1e278] thermal: don't make
THERMAL_NETLINK 'default y'

$ git bisect good
Bisecting: 12085 revisions left to test after this (roughly 14 steps)
[726eb70e0d34dc4bc4dada71f52bba8ed638431e] Merge tag
'char-misc-5.10-rc1' of
git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc

$ git bisect good
Bisecting: 6064 revisions left to test after this (roughly 13 steps)
[c8b5e2600a2cfa1cdfbecf151afd67aee227381d] io_uring: use type
appropriate io_kiocb handler for double poll

$ git bisect good
Bisecting: 2856 revisions left to test after this (roughly 12 steps)
[e533cda12d8f0e7936354bafdc85c81741f805d2] Merge tag 'armsoc-dt' of
git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc

$ git bisect good
Bisecting: 1423 revisions left to test after this (roughly 11 steps)
[64b609d6a60c8a54ff9c43819803e6a7861e7ed1] Merge tag
'perf-urgent-2020-11-15' of
git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

$ git bisect good
Bisecting: 724 revisions left to test after this (roughly 10 steps)
[80e1e1761d1a9eefda4d1545f8b6c0a2e46d4e3f] Merge tag 'scsi-fixes' of
git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi

$ git bisect good
Bisecting: 362 revisions left to test after this (roughly 9 steps)
[d4e904198c5b46c140fdd04492df6ec31f1f03a5] Merge tag
'5.10-rc6-smb3-fixes-part2' of git://git.samba.org/sfrench/cifs-2.6

$ git bisect good
Bisecting: 225 revisions left to test after this (roughly 8 steps)
[9fca90cf28920c6d0723d7efd1eae0b0fb90309c] Merge tag 'for-linus' of
git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma

$ git bisect good
Bisecting: 115 revisions left to test after this (roughly 7 steps)
[ba9c1201beaa86a773e83be5654602a0667e4a4d] mm/hugetlb: clear
compound_nr before freeing gigantic pages

$ git bisect good
Bisecting: 62 revisions left to test after this (roughly 6 steps)
[1de5d12b769017f7354e7151ce1c26eb1d9ba2e4] Merge tag 'mmc-v5.10-rc4-2'
of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/mmc

$ git bisect good
Bisecting: 27 revisions left to test after this (roughly 5 steps)
[7b1b868e1d9156484ccce9bf11122c053de82617] Merge tag 'for-linus' of
git://git.kernel.org/pub/scm/virt/kvm/kvm

$ git bisect good
Bisecting: 9 revisions left to test after this (roughly 4 steps)
[d2360a398f0b68722641c59aeb2623e79bd03e34] Merge tag
'block-5.10-2020-12-12' of git://git.kernel.dk/linux-block

$ git bisect good
Bisecting: 4 revisions left to test after this (roughly 2 steps)
[29ac40cbed2bc06fa218ca25d7f5e280d3d08a25] x86/mm/mem_encrypt: Fix
definition of PMD_FLAGS_DEC_WP

$ git bisect good
Bisecting: 2 revisions left to test after this (roughly 1 step)
[190113b4c6531c8e09b31d5235f9b5175cbb0f72] x86/apic/vector: Fix
ordering in vector assignment

$ git bisect good
Bisecting: 0 revisions left to test after this (roughly 1 step)
[ec6f5e0e5ca0764b4bc522c9f9d5abf876a0e3e3] Merge tag
'x86-urgent-2020-12-13' of
git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

$ git bisect good
2c85ebc57b3e1817b6ce1a6b703928e113a90442 is the first bad commit
commit 2c85ebc57b3e1817b6ce1a6b703928e113a90442
Author: Linus Torvalds <torvalds@linux-foundation.org>
Date:   Sun Dec 13 14:41:30 2020 -0800

    Linux 5.10

:100644 100644 9ec53d947628d4dd919f84f348ab8223ef69b4e3
e30cf02da8b892f3fc95492635f426b8383e5980 M  Makefile

$ git bisect bad
ec6f5e0e5ca0764b4bc522c9f9d5abf876a0e3e3 was both good and bad

Re: Kernel 5.10 USB issues

[Greg KH]

On Thu, Nov 25, 2021 at 05:03:33PM +1100, Jupiter wrote:
> Hi Greg,
> 
> Thanks for your patience, here is what I did. I thought I followed the
> instructions, any useful information?
> 
> > Start working on the "master" branch and do:
> 
> $ git bisect start
> 
> $ git bisect good v4.19
> 
> $ git bisect bad v5.10
> 
> Bisecting: 91917 revisions left to test after this (roughly 17 steps)
> [386403a115f95997c2715691226e11a7b5cffcfd] Merge
> git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next
> 
> $ git bisect good
> Bisecting: 46699 revisions left to test after this (roughly 16 steps)
> [2e63f6ce7ed2c4ff83ba30ad9ccad422289a6c63] Merge branch
> 'uaccess.comedi' of
> git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs
> 
> $ git bisect good
> Bisecting: 23349 revisions left to test after this (roughly 15 steps)
> [0f5d0a4c01cc1b7b0590baa18809a96d45a1e278] thermal: don't make
> THERMAL_NETLINK 'default y'
> 
> $ git bisect good
> Bisecting: 12085 revisions left to test after this (roughly 14 steps)
> [726eb70e0d34dc4bc4dada71f52bba8ed638431e] Merge tag
> 'char-misc-5.10-rc1' of
> git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc
> 
> $ git bisect good
> Bisecting: 6064 revisions left to test after this (roughly 13 steps)
> [c8b5e2600a2cfa1cdfbecf151afd67aee227381d] io_uring: use type
> appropriate io_kiocb handler for double poll
> 
> $ git bisect good
> Bisecting: 2856 revisions left to test after this (roughly 12 steps)
> [e533cda12d8f0e7936354bafdc85c81741f805d2] Merge tag 'armsoc-dt' of
> git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc
> 
> $ git bisect good
> Bisecting: 1423 revisions left to test after this (roughly 11 steps)
> [64b609d6a60c8a54ff9c43819803e6a7861e7ed1] Merge tag
> 'perf-urgent-2020-11-15' of
> git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
> 
> $ git bisect good
> Bisecting: 724 revisions left to test after this (roughly 10 steps)
> [80e1e1761d1a9eefda4d1545f8b6c0a2e46d4e3f] Merge tag 'scsi-fixes' of
> git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
> 
> $ git bisect good
> Bisecting: 362 revisions left to test after this (roughly 9 steps)
> [d4e904198c5b46c140fdd04492df6ec31f1f03a5] Merge tag
> '5.10-rc6-smb3-fixes-part2' of git://git.samba.org/sfrench/cifs-2.6
> 
> $ git bisect good
> Bisecting: 225 revisions left to test after this (roughly 8 steps)
> [9fca90cf28920c6d0723d7efd1eae0b0fb90309c] Merge tag 'for-linus' of
> git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma
> 
> $ git bisect good
> Bisecting: 115 revisions left to test after this (roughly 7 steps)
> [ba9c1201beaa86a773e83be5654602a0667e4a4d] mm/hugetlb: clear
> compound_nr before freeing gigantic pages
> 
> $ git bisect good
> Bisecting: 62 revisions left to test after this (roughly 6 steps)
> [1de5d12b769017f7354e7151ce1c26eb1d9ba2e4] Merge tag 'mmc-v5.10-rc4-2'
> of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/mmc
> 
> $ git bisect good
> Bisecting: 27 revisions left to test after this (roughly 5 steps)
> [7b1b868e1d9156484ccce9bf11122c053de82617] Merge tag 'for-linus' of
> git://git.kernel.org/pub/scm/virt/kvm/kvm
> 
> $ git bisect good
> Bisecting: 9 revisions left to test after this (roughly 4 steps)
> [d2360a398f0b68722641c59aeb2623e79bd03e34] Merge tag
> 'block-5.10-2020-12-12' of git://git.kernel.dk/linux-block
> 
> $ git bisect good
> Bisecting: 4 revisions left to test after this (roughly 2 steps)
> [29ac40cbed2bc06fa218ca25d7f5e280d3d08a25] x86/mm/mem_encrypt: Fix
> definition of PMD_FLAGS_DEC_WP
> 
> $ git bisect good
> Bisecting: 2 revisions left to test after this (roughly 1 step)
> [190113b4c6531c8e09b31d5235f9b5175cbb0f72] x86/apic/vector: Fix
> ordering in vector assignment
> 
> $ git bisect good
> Bisecting: 0 revisions left to test after this (roughly 1 step)
> [ec6f5e0e5ca0764b4bc522c9f9d5abf876a0e3e3] Merge tag
> 'x86-urgent-2020-12-13' of
> git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
> 
> $ git bisect good
> 2c85ebc57b3e1817b6ce1a6b703928e113a90442 is the first bad commit
> commit 2c85ebc57b3e1817b6ce1a6b703928e113a90442
> Author: Linus Torvalds <torvalds@linux-foundation.org>
> Date:   Sun Dec 13 14:41:30 2020 -0800
> 
>     Linux 5.10
> 
> :100644 100644 9ec53d947628d4dd919f84f348ab8223ef69b4e3
> e30cf02da8b892f3fc95492635f426b8383e5980 M  Makefile

This looks odd.  git seems to think that the Makefile change caused the
problem.

You did build/boot these kernels properly, right?

I don't know what to suggest anymore, sorry.

greg k-h

Re: Kernel 5.10 USB issues

[Andy Shevchenko]

On Fri, Nov 26, 2021 at 1:04 PM Jupiter <jupiter.hce@gmail.com> wrote:
>
> Hi Greg,
>
> Thanks for your patience, here is what I did. I thought I followed the
> instructions, any useful information?
>
> > Start working on the "master" branch and do:

But do you actually _build and run_ kernel each time after running
`git bisect {good,bad}`? It looks like you simply typed it in a row.
Also there is a subcommand to `git bisect` which prints log in
reproducible format (someone else mey run it on their side and
compare).

-- 
With Best Regards,
Andy Shevchenko

Re: Kernel 5.10 USB issues

[Jupiter]

> But do you actually _build and run_ kernel each time after running
> `git bisect {good,bad}`? It looks like you simply typed it in a row.
> Also there is a subcommand to `git bisect` which prints log in
> reproducible format (someone else mey run it on their side and
> compare).

I was running `git bisect`, it could be the build issues, I am still
digging it further.

Thank you very much.

- jh