Bug 97711 - r8712 generate crash message([cut here] stuff) in dmesg since 4.1RC1
Summary: r8712 generate crash message([cut here] stuff) in dmesg since 4.1RC1
Status: CLOSED CODE_FIX
Alias: None
Product: Drivers
Classification: Unclassified
Component: Staging (show other bugs)
Hardware: x86-64 Linux
: P1 normal
Assignee: drivers_staging@kernel-bugs.osdl.org
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2015-05-04 19:13 UTC by Arek Ruśniak
Modified: 2015-06-09 16:09 UTC (History)
1 user (show)

See Also:
Kernel Version: 4.1RC1
Subsystem:
Regression: No
Bisected commit-id:


Attachments
dmesg for linux 4.1rc2 with r8712u problem. (86.04 KB, application/octet-stream)
2015-05-04 19:13 UTC, Arek Ruśniak
Details
dmesg for 4.1rc2 + reverted stuff (65.23 KB, application/octet-stream)
2015-05-04 19:20 UTC, Arek Ruśniak
Details
kernel log for the good one kernel v4.0 (92.71 KB, application/octet-stream)
2015-05-04 19:25 UTC, Arek Ruśniak
Details
patch to test (1.43 KB, patch)
2015-05-14 13:54 UTC, Sudip
Details | Diff
config for rc3 (159.16 KB, application/octet-stream)
2015-05-14 18:20 UTC, Arek Ruśniak
Details
dmesg for rc3+patch (73.25 KB, application/octet-stream)
2015-05-14 20:26 UTC, Arek Ruśniak
Details
dmesg for rc3 without patch "clean" (113.07 KB, application/octet-stream)
2015-05-14 20:33 UTC, Arek Ruśniak
Details
patch to test (21.31 KB, patch)
2015-05-15 05:19 UTC, Sudip
Details | Diff

Description Arek Ruśniak 2015-05-04 19:13:45 UTC
Created attachment 175831 [details]
dmesg for linux 4.1rc2 with r8712u problem.

result of bisecting:

6501c8e7d86cca5fab74f873810cbc573273d1b4 is the first bad commit
commit 6501c8e7d86cca5fab74f873810cbc573273d1b4
Author: Vaishali Thakkar <vthakkar1994@gmail.com>
Date:   Fri Mar 6 16:23:35 2015 +0530

    Staging: rtl8712: Eliminate use of _cancel_timer_ex
    
    Use timer API function del_timer_sync instead of driver
    specific function _cancel_timer_ex as besides deactivating
    a timer, it ensures that the timer is stopped on all
    CPUs before the driver exists. Also, definition of function
    _cancel_timer_ex is removed as it is no longer needed after
    this change.
    
    This is done using Coccinelle and semantic patch used for
    this is as follows:
    
    @@ expression x; @@
    
    - _cancel_timer_ex (&x);
    + del_timer_sync (&x);
    
    Signed-off-by: Vaishali Thakkar <vthakkar1994@gmail.com>
    Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

:040000 040000 eaab34b0be7461ea124eb924a374babf19bb05f9 32fe24820b768fdebf160e003e57a4d2b9cf3505 M	drivers

Everything works really good as before, no wifi networking problem etc.
Comment 1 Arek Ruśniak 2015-05-04 19:20:20 UTC
Created attachment 175841 [details]
dmesg for 4.1rc2 + reverted stuff

kernel log after reverting commits:
382d020f4459cd77237c5463098935fd64afdab3 
"Staging: rtl8712: Eliminate use of _cancel_timer"
and 
6501c8e7d86cca5fab74f873810cbc573273d1b4 
"Staging: rtl8712: Eliminate use of _cancel_timer_ex"
Comment 2 Arek Ruśniak 2015-05-04 19:25:23 UTC
Created attachment 175851 [details]
kernel log for the good one kernel v4.0

nice and clean kernel log, without strange calling CRDA again and again.
Comment 3 Sudip 2015-05-14 12:18:24 UTC
Hi,
Can you please check if you have the same problem with 4.1-rc3 also. If you have please attach your config file also.
Comment 4 Sudip 2015-05-14 13:54:46 UTC
Created attachment 176851 [details]
patch to test

Please test the attached patch. It will apply to 4.1-rc2 and 4.1-rc3 both. Not a formal patch for submission, just for testing.
Comment 5 Arek Ruśniak 2015-05-14 18:09:22 UTC
Thx for reply. In rc3 and next-20150512 problem is still here. I'm building the kernel with patch from above, so i try and call you back asap.
Comment 6 Arek Ruśniak 2015-05-14 18:20:22 UTC
Created attachment 176891 [details]
config for rc3

i'm pretty sure it's the same config as for rc2 & rc1.
Comment 7 Arek Ruśniak 2015-05-14 20:26:20 UTC
Created attachment 176921 [details]
dmesg for rc3+patch

This patch didn't work for me, but now i have no spam at logs. [cut here] staff appears only one time for over 1h.
Comment 8 Arek Ruśniak 2015-05-14 20:33:43 UTC
Created attachment 176931 [details]
dmesg for rc3 without patch "clean"

Uptime about 25 min.
Comment 9 Sudip 2015-05-15 05:19:51 UTC
Created attachment 176961 [details]
patch to test

Hi,
Thanks for your time, can you please test this one as well. This should solve all the problems. As before this will apply to 4.1-rc2 , 4.1-rc3 and also linux-next.
Comment 10 Arek Ruśniak 2015-05-15 08:16:37 UTC
Yes, problems are gone. Tested for rc3 only.

Thanks again.
Comment 11 Sudip 2015-05-15 08:25:14 UTC
Great. Please modify the status of this bug to VERIFIED. I will submit the patch with Reported-by and Tested-by both your name. After that one is accepted then you can mark this as CLOSED.

Thanks for your time.
Comment 12 Arek Ruśniak 2015-05-15 08:40:06 UTC
I can't do this, the olny choice for status is RESOLVED (or NEW) but no CLOSED, VERIFIED ...strange.
Comment 13 Sudip 2015-05-15 08:49:02 UTC
The FAQ page is saying about many different statuses including VERIFIED and CLOSED. maybe we need to raise a bug about bugzilla :)
Then leave it for now. I will post the patch today, later we will see how to close it.
Comment 14 Arek Ruśniak 2015-05-16 08:11:31 UTC
I need set status to [RESOLVED] and wait for QA guy to verify it. But i really don't know who is that guy (probably someone from Quatar? ;).
see this:
https://bugzilla.kernel.org/page.cgi?id=fields.html#bug_status
Comment 15 Sudip 2015-05-16 08:31:24 UTC
patch has been posted. https://lkml.org/lkml/2015/5/15/226

If it is applied then the correct status should be FIXED.
Comment 16 Arek Ruśniak 2015-05-16 09:48:05 UTC
And you are right... it should be and it is :) Now i can set VERIFIED or CLOSED (maybe I am QA in this case). 
When patch will land into mainline (probably rc4) i will close this. 
Thanks again for help & patience :)
Comment 17 Sudip 2015-05-16 09:53:46 UTC
Nope, rc4 not possible. rc4 will be released tomorrow. may be in rc6. First it will be applied to staging-testing then staging-next, from there it will go to linux-next.
Comment 18 Sudip 2015-06-08 11:40:43 UTC
Hi,
You can close the bug now. The fix is in 4.1-rc7 now.
Comment 19 Arek Ruśniak 2015-06-09 16:09:30 UTC
Thanks again
best regards

Note You need to log in before you can comment on or make changes to this bug.