Core126 udevd problem

Core126 udevd problem

[Tapani Tarvainen]

[-- Attachment #1: Type: text/plain, Size: 881 bytes --]

Core126 upgrade killed a machine: boot freezes after udevd start.

Experimenting with fresh (flash) installation image (and new CF card)
gave same result.

Up to and including core125 it works like charm (also tested with a
fresh image).

This is a bit difficult to debug: the last message displayed is

"Starting udev daemon...   [ OK ]"

After that even keyboard is dead (caps lock doesn't work, nor alt-ctrl-del).

And as this happens before disks are mounted, there's no log to read.

This is a rather old machine, VIA C3 CPU, 512MB RAM, booting of CF card,
but there's nothing obviously wrong with the hardware, and as noted it
worked just fine up to and including core125.

I could leave it running core125, but a firewall that can't be upgraded
is not a workable long-term proposition.

Any suggestions as to how I could try to track the problem down?

-- 
Tapani Tarvainen

Re: Core126 udevd problem

[Michael Tremer]

[-- Attachment #1: Type: text/plain, Size: 1441 bytes --]

Hello,

I think you might be running into this:

https://git.ipfire.org/?p=ipfire-2.x.git;a=commitdiff;h=4c76d08b2a1ef5ac9ff8b546c0d887e342adec1c

The framebuffer driver is no longer blacklisted and that causes the kernel to crash on my VIA devices.

You can mount the hard drive in a different computer and edit the file manually if that is an option for you.

Best,
-Michael

> On 2 Jan 2019, at 17:13, Tapani Tarvainen <ipfire(a)tapanitarvainen.fi> wrote:
> 
> Core126 upgrade killed a machine: boot freezes after udevd start.
> 
> Experimenting with fresh (flash) installation image (and new CF card)
> gave same result.
> 
> Up to and including core125 it works like charm (also tested with a
> fresh image).
> 
> This is a bit difficult to debug: the last message displayed is
> 
> "Starting udev daemon...   [ OK ]"
> 
> After that even keyboard is dead (caps lock doesn't work, nor alt-ctrl-del).
> 
> And as this happens before disks are mounted, there's no log to read.
> 
> This is a rather old machine, VIA C3 CPU, 512MB RAM, booting of CF card,
> but there's nothing obviously wrong with the hardware, and as noted it
> worked just fine up to and including core125.
> 
> I could leave it running core125, but a firewall that can't be upgraded
> is not a workable long-term proposition.
> 
> Any suggestions as to how I could try to track the problem down?
> 
> -- 
> Tapani Tarvainen

Re: Core126 udevd problem

[Tapani Tarvainen]

[-- Attachment #1: Type: text/plain, Size: 2128 bytes --]

Indeed, restoring /etc/modprobe.d/framebuffer.conf fixed it - thank you!

(As noted the machine boots off a CF card and I have another machine with
card reader so editing it is easy.)

That leads to a followup question, however: how can I prevent this
from happening again with next upgrade, and with other machines that
are likely to have the same issue? Guess I could add a custom boot
script that runs before udev start and blacklists it again - any
better ideas?

Regards,

-- 
Tapani

On Wed, Jan 02, 2019 at 05:24:01PM +0000, Michael Tremer (michael.tremer(a)ipfire.org) wrote:
> 
> Hello,
> 
> I think you might be running into this:
> 
> https://git.ipfire.org/?p=ipfire-2.x.git;a=commitdiff;h=4c76d08b2a1ef5ac9ff8b546c0d887e342adec1c
> 
> The framebuffer driver is no longer blacklisted and that causes the kernel to crash on my VIA devices.
> 
> You can mount the hard drive in a different computer and edit the file manually if that is an option for you.
> 
> Best,
> -Michael
> 
> > On 2 Jan 2019, at 17:13, Tapani Tarvainen <ipfire(a)tapanitarvainen.fi> wrote:
> > 
> > Core126 upgrade killed a machine: boot freezes after udevd start.
> > 
> > Experimenting with fresh (flash) installation image (and new CF card)
> > gave same result.
> > 
> > Up to and including core125 it works like charm (also tested with a
> > fresh image).
> > 
> > This is a bit difficult to debug: the last message displayed is
> > 
> > "Starting udev daemon...   [ OK ]"
> > 
> > After that even keyboard is dead (caps lock doesn't work, nor alt-ctrl-del).
> > 
> > And as this happens before disks are mounted, there's no log to read.
> > 
> > This is a rather old machine, VIA C3 CPU, 512MB RAM, booting of CF card,
> > but there's nothing obviously wrong with the hardware, and as noted it
> > worked just fine up to and including core125.
> > 
> > I could leave it running core125, but a firewall that can't be upgraded
> > is not a workable long-term proposition.
> > 
> > Any suggestions as to how I could try to track the problem down?
> > 
> > -- 
> > Tapani Tarvainen

Re: Core126 udevd problem

[Michael Tremer]

[-- Attachment #1: Type: text/plain, Size: 2362 bytes --]

Hi,

This file will be shipped with the next update. This was an unintended bug.

-Michael

> On 2 Jan 2019, at 17:39, Tapani Tarvainen <ipfire(a)tapanitarvainen.fi> wrote:
> 
> Indeed, restoring /etc/modprobe.d/framebuffer.conf fixed it - thank you!
> 
> (As noted the machine boots off a CF card and I have another machine with
> card reader so editing it is easy.)
> 
> That leads to a followup question, however: how can I prevent this
> from happening again with next upgrade, and with other machines that
> are likely to have the same issue? Guess I could add a custom boot
> script that runs before udev start and blacklists it again - any
> better ideas?
> 
> Regards,
> 
> -- 
> Tapani
> 
> On Wed, Jan 02, 2019 at 05:24:01PM +0000, Michael Tremer (michael.tremer(a)ipfire.org) wrote:
>> 
>> Hello,
>> 
>> I think you might be running into this:
>> 
>> https://git.ipfire.org/?p=ipfire-2.x.git;a=commitdiff;h=4c76d08b2a1ef5ac9ff8b546c0d887e342adec1c
>> 
>> The framebuffer driver is no longer blacklisted and that causes the kernel to crash on my VIA devices.
>> 
>> You can mount the hard drive in a different computer and edit the file manually if that is an option for you.
>> 
>> Best,
>> -Michael
>> 
>>> On 2 Jan 2019, at 17:13, Tapani Tarvainen <ipfire(a)tapanitarvainen.fi> wrote:
>>> 
>>> Core126 upgrade killed a machine: boot freezes after udevd start.
>>> 
>>> Experimenting with fresh (flash) installation image (and new CF card)
>>> gave same result.
>>> 
>>> Up to and including core125 it works like charm (also tested with a
>>> fresh image).
>>> 
>>> This is a bit difficult to debug: the last message displayed is
>>> 
>>> "Starting udev daemon...   [ OK ]"
>>> 
>>> After that even keyboard is dead (caps lock doesn't work, nor alt-ctrl-del).
>>> 
>>> And as this happens before disks are mounted, there's no log to read.
>>> 
>>> This is a rather old machine, VIA C3 CPU, 512MB RAM, booting of CF card,
>>> but there's nothing obviously wrong with the hardware, and as noted it
>>> worked just fine up to and including core125.
>>> 
>>> I could leave it running core125, but a firewall that can't be upgraded
>>> is not a workable long-term proposition.
>>> 
>>> Any suggestions as to how I could try to track the problem down?
>>> 
>>> -- 
>>> Tapani Tarvainen

Re: Core126 udevd problem

[Tapani Tarvainen]

[-- Attachment #1: Type: text/plain, Size: 2697 bytes --]

Ah, OK then. Thank you for very quick and helpful response!

Tapani

On Wed, Jan 02, 2019 at 05:46:44PM +0000, Michael Tremer (michael.tremer(a)ipfire.org) wrote:
> 
> Hi,
> 
> This file will be shipped with the next update. This was an unintended bug.
> 
> -Michael
> 
> > On 2 Jan 2019, at 17:39, Tapani Tarvainen <ipfire(a)tapanitarvainen.fi> wrote:
> > 
> > Indeed, restoring /etc/modprobe.d/framebuffer.conf fixed it - thank you!
> > 
> > (As noted the machine boots off a CF card and I have another machine with
> > card reader so editing it is easy.)
> > 
> > That leads to a followup question, however: how can I prevent this
> > from happening again with next upgrade, and with other machines that
> > are likely to have the same issue? Guess I could add a custom boot
> > script that runs before udev start and blacklists it again - any
> > better ideas?
> > 
> > Regards,
> > 
> > -- 
> > Tapani
> > 
> > On Wed, Jan 02, 2019 at 05:24:01PM +0000, Michael Tremer (michael.tremer(a)ipfire.org) wrote:
> >> 
> >> Hello,
> >> 
> >> I think you might be running into this:
> >> 
> >> https://git.ipfire.org/?p=ipfire-2.x.git;a=commitdiff;h=4c76d08b2a1ef5ac9ff8b546c0d887e342adec1c
> >> 
> >> The framebuffer driver is no longer blacklisted and that causes the kernel to crash on my VIA devices.
> >> 
> >> You can mount the hard drive in a different computer and edit the file manually if that is an option for you.
> >> 
> >> Best,
> >> -Michael
> >> 
> >>> On 2 Jan 2019, at 17:13, Tapani Tarvainen <ipfire(a)tapanitarvainen.fi> wrote:
> >>> 
> >>> Core126 upgrade killed a machine: boot freezes after udevd start.
> >>> 
> >>> Experimenting with fresh (flash) installation image (and new CF card)
> >>> gave same result.
> >>> 
> >>> Up to and including core125 it works like charm (also tested with a
> >>> fresh image).
> >>> 
> >>> This is a bit difficult to debug: the last message displayed is
> >>> 
> >>> "Starting udev daemon...   [ OK ]"
> >>> 
> >>> After that even keyboard is dead (caps lock doesn't work, nor alt-ctrl-del).
> >>> 
> >>> And as this happens before disks are mounted, there's no log to read.
> >>> 
> >>> This is a rather old machine, VIA C3 CPU, 512MB RAM, booting of CF card,
> >>> but there's nothing obviously wrong with the hardware, and as noted it
> >>> worked just fine up to and including core125.
> >>> 
> >>> I could leave it running core125, but a firewall that can't be upgraded
> >>> is not a workable long-term proposition.
> >>> 
> >>> Any suggestions as to how I could try to track the problem down?
> >>> 
> >>> -- 
> >>> Tapani Tarvainen
> 

-- 
Tapani Tarvainen