Bug 1323420 - keyboard not working on 4.5.0 kernel
Summary: keyboard not working on 4.5.0 kernel
Keywords:
Status: CLOSED CURRENTRELEASE
Alias: None
Product: Fedora
Classification: Fedora
Component: kernel
Version: 24
Hardware: x86_64
OS: Linux
unspecified
high
Target Milestone: ---
Assignee: Kernel Maintainer List
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2016-04-02 19:19 UTC by Herbert Carl Meyer
Modified: 2016-10-03 12:15 UTC (History)
8 users (show)

Fixed In Version:
Clone Of:
Environment:
Last Closed: 2016-10-03 12:15:28 UTC
Type: Bug
Embargoed:


Attachments (Terms of Use)
dmesg output 4.4.6 kernel (68.11 KB, text/plain)
2016-04-11 22:21 UTC, Herbert Carl Meyer
no flags Details
lsmod output 4.4.6 kernel (5.09 KB, text/plain)
2016-04-11 22:22 UTC, Herbert Carl Meyer
no flags Details
dmesg output, 4.5.0 kernel, 100 line sample (10.72 KB, text/plain)
2016-04-12 00:17 UTC, Herbert Carl Meyer
no flags Details
lsmod output 4.5.0 kernel (5.02 KB, text/plain)
2016-04-12 00:18 UTC, Herbert Carl Meyer
no flags Details
kernel messages for 4.4.6 kernel, successful boot (101.17 KB, text/plain)
2016-04-30 01:43 UTC, Herbert Carl Meyer
no flags Details
kernel messages for 4.5.2 kernel, with sp5100_tco blacklisted (86.32 KB, text/plain)
2016-04-30 01:51 UTC, Herbert Carl Meyer
no flags Details

Description Herbert Carl Meyer 2016-04-02 19:19:34 UTC
Description of problem:

My aspire v5 keyboard does not work on kernel versions 4.5.0

Version-Release number of selected component (if applicable):

4.5.0

How reproducible:

100 %

Steps to Reproduce:

1. upgrade to fedora 24 alpha
 
2. complete update, boot to 4.5.0-rc7 kernel

3. try to login

Actual results:

Unable to login, keyboard does not work, mouse (touchpad) does
 
Expected results:

normal keyboard login

Additional info:

both initial rc7 and released 4.5.0 kernel fail.

Older f23 kernels still boot

Hardware is aspire v5 122p with touchscreen

Comment 1 Josh Boyer 2016-04-11 14:20:48 UTC
On the older kernel, what driver is used for your keyboard?  Please provide the output of 'lsmod' and 'dmesg' from the working kernel.  If you can ssh into the machine or use a different keyboard, do the same for the non-working kernel.

Comment 2 Herbert Carl Meyer 2016-04-11 22:21:58 UTC
Created attachment 1146152 [details]
dmesg output 4.4.6 kernel

Comment 3 Herbert Carl Meyer 2016-04-11 22:22:48 UTC
Created attachment 1146153 [details]
lsmod output 4.4.6 kernel

Comment 4 Herbert Carl Meyer 2016-04-12 00:17:26 UTC
Created attachment 1146170 [details]
dmesg output, 4.5.0 kernel, 100 line sample

Comment 5 Herbert Carl Meyer 2016-04-12 00:18:20 UTC
Created attachment 1146171 [details]
lsmod output 4.5.0 kernel

Comment 6 Herbert Carl Meyer 2016-04-16 00:53:02 UTC
4.5.1-300 kernel does not work

Comment 7 Herbert Carl Meyer 2016-04-23 00:05:59 UTC
4.5.2 kernel does not work

Comment 8 Herbert Carl Meyer 2016-04-26 01:26:24 UTC
4.5.2-301 kernel does not work

Comment 9 Herbert Carl Meyer 2016-04-26 01:44:46 UTC
I think it might be related to:

https://bugzilla.kernel.org/show_bug.cgi?id=114201

Which is for similar AMD hardware. I will drink another cup of coffee and research.

Gee, I wish I could remember x86 interrupt hardware. 8259 really sucked, as I remember. 

I think it has something to do will the interrupt mask not being set to disable on gp02, whatever that is.

Comment 10 Herbert Carl Meyer 2016-04-26 03:04:11 UTC
lsmod output for the 4.5.0 kernel shows the sp5100_tco module, lsmod for the 4.4 kernel does not.

I will blacklist the sp5100_tco, and report back. Tomorrow.

Comment 11 Herbert Carl Meyer 2016-04-26 23:42:58 UTC
Blacklisting the sp5100_tco device on the boot line does not help.

Comment 12 Justin M. Forbes 2016-04-27 13:38:58 UTC
Unfortunately, that module is only a part of the problem that was introduced. I was hoping for proper fix there. Either way, can I get the dmesg from 4.5 with the module blacklisted, so we can actually see what is going on?

Comment 13 Herbert Carl Meyer 2016-04-29 11:50:33 UTC
I do not think getting dmesg is feasible, it is quickly filled and the ring overwritten by unhandled interrupt messages. My attempt using ssh login actually produced 88 meg of data from dmesg. Clearly an overrun.

I have been reviewing how logs/journal works, and think I might capture the requested info in the logs if I change MaxLevelKMsg=notice to =debug in /etc/systemd/journald.conf. 

I don't know if journald starts up early enough in boot to catch the messages. I will try it later today, Fri 4/29.

Comment 14 Herbert Carl Meyer 2016-04-30 01:43:44 UTC
Created attachment 1152471 [details]
kernel messages for 4.4.6 kernel, successful boot

Comment 15 Herbert Carl Meyer 2016-04-30 01:49:31 UTC
1) 4.5.2-302 kernel does not work

2) I was able to change the kernel message log level, and have output that may be diagnostic. I am having a temporary failure attaching them. I tried to blacklist the sp5100_tco device on the 4.5.2 kernel.

Does anyone on this side of the Pacific have an In-Circuit-Emulator for AMD hardware ?

Comment 16 Herbert Carl Meyer 2016-04-30 01:51:53 UTC
Created attachment 1152472 [details]
kernel messages for 4.5.2 kernel, with sp5100_tco blacklisted

Comment 17 Herbert Carl Meyer 2016-05-05 12:33:54 UTC
Lucas Stach has posted a patch, he says:

Commits 190aa4304de6 (Add AMD Mullins platform support) and
cca118fa2a0a94 (Add AMD Carrizo platform support) enabled the
driver on a lot more devices, but the following commit missed
a single location in the code when checking if the SB800 register
offsets should be used. This leads to the wrong register being
written which in turn causes ACPI to go haywire.

Fix this by introducing a helper function to check for the new
register layout and use this consistently.

See patch at:

http://www.spinics.net/lists/linux-watchdog/msg09165.html

Comment 18 Herbert Carl Meyer 2016-05-06 23:27:19 UTC
Kernel version 4.5.3-300 does not work

Comment 19 Herbert Carl Meyer 2016-05-14 08:18:03 UTC
Kernel version 4.5.4-300 does work, on my Acer V5 hardware. Thank you.

I have two other AMD systems, I will install F24 beta on at least one, as time allows.

I like AMD hardware because it is cheap. Cheap is a virtue.

Comment 20 Herbert Carl Meyer 2016-05-14 21:58:18 UTC
I have installed the F24 beta on another system, a Samsung s3 laptop, with AMD hardware. Kernel 4.5.3-300.fc24.x86_64 works on it. 

No, I do not wanna explain it. 4.5.4 kernel is in the first batch of updates.

Comment 21 Laura Abbott 2016-09-23 19:16:13 UTC
*********** MASS BUG UPDATE **************
 
We apologize for the inconvenience.  There is a large number of bugs to go through and several of them have gone stale.  Due to this, we are doing a mass bug update across all of the Fedora 24 kernel bugs.
 
Fedora 24 has now been rebased to 4.7.4-200.fc24.  Please test this kernel update (or newer) and let us know if you issue has been resolved or if it is still present with the newer kernel.
 
If you have moved on to Fedora 25, and are still experiencing this issue, please change the version to Fedora 25.
 
If you experience different issues, please open a new bug report for those.

Comment 22 Herbert Carl Meyer 2016-10-03 12:15:28 UTC
Kernel fixed upstream. Involves lenovo hardware watchdog.
Please close.


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