Beowulf keyboard missing keystrokes

bobemoe · 2020-07-04 13:01:19

I've just upgraded from ascii to beowulf and there seems to be a strange issue with the keyboard.

There seems to be a laggy feeling to the text coming onto the screen, lots of characters are missing and occasionally they come out in the wrong order!

If I type slowly this is not such an issue but I usually type quite fast.

Any thoughts?

As an example I have type this quickly without correcting anything and include the original copy below:

Thanks And great work all

----

I've jsut upgade rom aci to beowulf and tehre seem to be astrnaeg ise with teh jkeyboard.

There em so be aig laggyfeelino the text comin onto the ce, lots of charaters are missin and ocasionalyl they ome out in the wrong orde!

If I type loly this is not sch an isus but I usualy type rpet fast.

An thoughts?

As an exaple I hve type this quickly without corectin aytihng and inclue the rignal copy belo:

Thanks And geat wrok al :

Last edited by bobemoe (2020-07-04 13:01:54)

bobemoe · 2020-07-04 13:13:54

This is not limited to X but also an issue in a real tty1.

bobemoe · 2020-07-04 13:25:50

So after watching `top` and digging through some logs and then a reboot the issue seems to have vanished! I'm happy, if not a little confused. I've never seen anything like that before, I've been using Linux for years and Devuan since the start.

Thanks again all and sorry to bother you

I'll let you know if it reoccurs.

Thoughts still welcome for curiosity sake

bobemoe · 2020-07-04 14:54:10

The issue is back!! Immediately after resuming from suspend it is happening again. I will reboot and suspend again too see if its consistent...

Head_on_a_Stick · 2020-07-04 15:05:57

An intermittent problem would suggest a hardware issue. Does it also occur in the live environment?

bobemoe wrote:

I've just upgraded from ascii to beowulf

You should probably explain your method, the more detail you include the better the chance we will spot any potential mistakes.

bobemoe · 2020-07-04 15:24:03

So it does seem pretty consistent. After 2 suspends the issue returns. I've confirmed this three times. An IRQ16 issue is seen in dmesg upon the issue occurring.

[  155.169229] irq 16: nobody cared (try booting with the "irqpoll" option)
[  155.169238] CPU: 0 PID: 4465 Comm: pm-powersave Not tainted 4.19.0-9-amd64 #1 Debian 4.19.118-2+deb10u1
[  155.169239] Hardware name: Hewlett-Packard HP xw4550 Workstation/0AC8h, BIOS 786F7 v01.07 04/23/2012
[  155.169240] Call Trace:
[  155.169244]  <IRQ>
[  155.169252]  dump_stack+0x66/0x90
[  155.169257]  __report_bad_irq+0x3a/0xb4
[  155.169260]  note_interrupt.cold.9+0xa/0x64
[  155.169263]  handle_irq_event_percpu+0x6d/0x80
[  155.169266]  handle_irq_event+0x3c/0x5c
[  155.169268]  handle_fasteoi_irq+0xa3/0x160
[  155.169271]  handle_irq+0x1f/0x30
[  155.169274]  do_IRQ+0x49/0xe0
[  155.169277]  common_interrupt+0xf/0xf
[  155.169278]  </IRQ>
[  155.169282] RIP: 0010:_cond_resched+0x5/0x30
[  155.169285] Code: 01 eb ba 66 90 bf 01 00 00 00 e8 26 f5 ff ff 65 48 8b 04 25 40 5c 01 00 48 8b 00 a8 08 75 e6 c3 0f 1f 44 00 00 66 66 66 66 90 <65> 8b 05 d4 b9 4e 7a 85 c0 74 08 e8 db 87 9c ff 31 c0 c3 e8 c3 ff
[  155.169286] RSP: 0000:ffffb54b449d7ee0 EFLAGS: 00000202 ORIG_RAX: ffffffffffffffda
[  155.169288] RAX: 0000000000000001 RBX: 0000000000000004 RCX: 0000000000000000
[  155.169289] RDX: 0000000000000001 RSI: 0000000000000004 RDI: ffff945d3503aaf0
[  155.169290] RBP: 00007fe763a9c4ad R08: 0000000000000000 R09: 0000000000000000
[  155.169291] R10: 0000000000000000 R11: 0000000000000000 R12: ffffb54b449d7f58
[  155.169292] R13: ffff945d3503aa80 R14: 0000000000000054 R15: ffff945c7eb21b00
[  155.169297]  __do_page_fault+0x14c/0x4f0
[  155.169300]  ? page_fault+0x8/0x30
[  155.169302]  page_fault+0x1e/0x30
[  155.169304] RIP: 0033:0x55d66372d76a
[  155.169306] Code: 00 00 e8 d9 fe ff ff 66 0f 1f 84 00 00 00 00 00 41 54 49 89 fc 55 53 0f b6 1f 80 fb 5f 74 12 e8 0c 4e ff ff 4c 89 e5 48 8b 00 <f6> 44 58 01 04 74 26 4c 89 e5 0f 1f 40 00 48 83 c5 01 0f b6 5d 00
[  155.169307] RSP: 002b:00007ffd0ba1fac0 EFLAGS: 00010207
[  155.169308] RAX: 00007fe763a9c3e0 RBX: 0000000000000066 RCX: 0000000000000075
[  155.169309] RDX: 0000000000000000 RSI: 000055d6647e7c00 RDI: 000055d6647e93ee
[  155.169310] RBP: 000055d6647e93ee R08: 0000000000000000 R09: 0000000000000010
[  155.169311] R10: fffffffffffffe28 R11: 0000000000000202 R12: 000055d6647e93ee
[  155.169312] R13: 000055d6647e7c00 R14: 000055d6647e7e68 R15: 000055d6647e8aa0
[  155.169314] handlers:
[  155.169341] [<0000000021e5ee98>] usb_hcd_irq [usbcore]
[  155.169366] [<000000002666a161>] ata_bmdma_interrupt [libata]
[  155.169384] [<00000000c654a7e1>] azx_interrupt [snd_hda_codec]
[  155.169385] Disabling IRQ #16

This has enabled me to search further, looks upstream kernel issue, lots of good info here: https://bugzilla.redhat.com/show_bug.cgi?id=474624

bobemoe · 2020-07-04 16:04:52

Looks like adding noirqdebug to kernel cmdline has fixed the issue. I've suspended 5 times now with no issue.

I also tried acpi=routeirq and pci=msi which did not help.

Strange I found the answer in a 2009 thread about kernel 2.x! Something regressed?

I'll report back if it occurs again.

Software freedom FTW

Last edited by bobemoe (2020-07-04 16:10:28)

bobemoe · 2020-07-05 05:35:04

Since noirqdebug the USB/keyboard input is perfect, but I'm hearing glitches on the sound card now and the system seems a bit sluggish, presumably because the IRQ's are being handled differently.

@Head_on_a_Stick I've not tried live yet, I'll give it a go.

bobemoe · 2021-08-17 09:48:16

Time for an update!...

The issue still appears to be present on Chimaera. Exactly the same. Upon resuming from the second suspend a single system beep is heard, and an IRQ error is seen in dmesg, as per post #6

After that the USB keyboard is slow to respond, and even mixes up letters, as demonstrated in second half of post #1

This persists until a reboot, and then all is normal until the second suspend again.

Previously I noted noirqdebug prevented the issue, but introduced another, audio/video glitches and overall sluggishness. I tried to live with it for a while but soon downgraded (reinstalled) back to Ascii.

I've been on Ascii for a year since then, suspending daily for months, with no issue whatsoever.

Recently upgraded to Chimaera only to be quickly reminded of this problem!

To rather belatedly answer @Head_on_a_Stick: Yes, this does occur on the Beowulf live environment too, but not Ascii.

I guess next step would be to try on Debian live?

Head_on_a_Stick · 2021-08-17 12:52:40

bobemoe wrote:

I guess next step would be to try on Debian live?

Yes. The new bullseye images should be available now.

You can use the journal to check for errors:

journalctl --follow

bobemoe · 2021-09-06 11:22:03

So the results are in, all using Debian Live Desktop xfce4.

* Debian 9 (Stretch): Suspend & resume 3 times, perfect.
* Debian 10 (Buster): Suspend & resume 2 times and the bug manifests, exactly as described above.
* Debian 11 (Bullseye): Suspend doesn't even work. The screen goes no signal, capslock stops responding, but the power light does not flash to indicate the machine is suspended. Power or keyboard buttons don't resume. Only way forward is a hard reset. Not sure how I'd get any logs for that.

And to recap, the Devuan versions, xfce4:
* ASCII (Stretch): Suspend & resume, perfect.
* Beowulf (Buster): Suspend & resume 2 times and the bug manifests.
* Chimaera (Bullseye): Suspend & resume 2 times and the bug manifests.

So it seems that ASCII and Beowulf mirror the behaviour of Stretch and Buster, being the last working and introduction of the bug respectively.

For Devuan the next release nothing changes; the bug still remains. But for Debian, the next release things get even worst! Strange!

Guess I need to report this upstream somewhere?

Last edited by bobemoe (2021-09-06 11:22:22)

The officially official Devuan Forum!

#1 2020-07-04 13:01:19

Beowulf keyboard missing keystrokes

#2 2020-07-04 13:13:54

Re: Beowulf keyboard missing keystrokes

#3 2020-07-04 13:25:50

Re: Beowulf keyboard missing keystrokes

#4 2020-07-04 14:54:10

Re: Beowulf keyboard missing keystrokes

#5 2020-07-04 15:05:57

Re: Beowulf keyboard missing keystrokes

#6 2020-07-04 15:24:03

Re: Beowulf keyboard missing keystrokes

#7 2020-07-04 16:04:52

Re: Beowulf keyboard missing keystrokes

#8 2020-07-05 05:35:04

Re: Beowulf keyboard missing keystrokes

#9 2021-08-17 09:48:16

Re: Beowulf keyboard missing keystrokes

#10 2021-08-17 12:52:40

Re: Beowulf keyboard missing keystrokes

#11 2021-09-06 11:22:03

Re: Beowulf keyboard missing keystrokes

Board footer