Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.debian.kernel > #72028 > unrolled thread

[Question] Backporting ACPI fix to Buster kernel?

Back to article view | Back to linux.debian.kernel

Contents

  [Question] Backporting ACPI fix to Buster kernel? Laurențiu Păncescu <lpancescu@gmail.com> - 2021-05-30 08:30 +0200
    Re: [Question] Backporting ACPI fix to Buster kernel? Georgi Naplatanov <gosho@oles.biz> - 2021-05-30 09:50 +0200
      Re: [Question] Backporting ACPI fix to Buster kernel? Laurențiu Păncescu <lpancescu@gmail.com> - 2021-05-30 20:40 +0200
    Re: [Question] Backporting ACPI fix to Buster kernel? Salvatore Bonaccorso <carnil@debian.org> - 2021-06-02 17:50 +0200
      Re: [Question] Backporting ACPI fix to Buster kernel? Laurențiu Păncescu <lpancescu@gmail.com> - 2021-06-03 20:10 +0200

#72028 — [Question] Backporting ACPI fix to Buster kernel?

Hello Debian Kernel Team,

I discovered an older Asus netbook always reports its battery as 100% 
charged even after being unplugged for several hours and with its CPU 
intentionally maxed out. This seems to be upstream bug #199981 in 
kernels 4.17-4.20 and affected Asus-made laptops (I see the same 
behavior with and without acpi_osi=Linux). [1]

Is there any chance of backporting the fix to Buster? Should I file a 
bug about that?

If not, do the kernels on backports get quick security updates, 
comparable with kernels in stable? Even now, during hard and soon full 
freeze?

Thanks,
Laurentiu

[1] https://bugzilla.kernel.org/show_bug.cgi?id=199981

[toc] | [next] | [standalone]

#72030

On 5/30/21 9:25 AM, Laurențiu Păncescu wrote:
> Hello Debian Kernel Team,
> 
> I discovered an older Asus netbook always reports its battery as 100%
> charged even after being unplugged for several hours and with its CPU
> intentionally maxed out. This seems to be upstream bug #199981 in
> kernels 4.17-4.20 and affected Asus-made laptops (I see the same
> behavior with and without acpi_osi=Linux). [1]
> 
> Is there any chance of backporting the fix to Buster? Should I file a
> bug about that?
> 
> If not, do the kernels on backports get quick security updates,
> comparable with kernels in stable? Even now, during hard and soon full
> freeze?
> 
> Thanks,
> Laurentiu
> 
> [1] https://bugzilla.kernel.org/show_bug.cgi?id=199981
> 

Hi Laurențiu,

Debian unstable and testing have no security support. As Linus Torvalds
has said, security issues are bugs so the bugs will be fixed but
unstable and testing have no official security support in Debian.

On the other hand Debian 11 (testing) (AMD64) is quite stable and
current stable (Debian 10) has more bugs than testing (Debian 11) so you
can check existing bugs for Debian 11 and if you decide then you can
install and use Debian testing (Debian 11).

https://bugs.debian.org/release-critical/

HTH

Kind regards
Georgi

PS: I'm Debian user and I'm not involved in Debian project development

[toc] | [prev] | [next] | [standalone]

#72035

Hi Georgi,

On 5/30/21 9:44 AM, Georgi Naplatanov wrote:
> Debian unstable and testing have no security support. As Linus Torvalds
> has said, security issues are bugs so the bugs will be fixed but
> unstable and testing have no official security support in Debian.

I know, I remembered security support for testing was planned or at 
least discussed several years ago. I first started to run Debian in 2001 
(2.2r3), but ran mostly stable over the years. The packages are old, but 
not as old as on RHEL, and I like not having surprises and not getting 
delays and bugs I wouldn't have hit on stable. Stable is very well 
tested and generally works pretty well, except for minor annoyances. I 
ran testing for a few months, many years ago, and returned to stable.

> On the other hand Debian 11 (testing) (AMD64) is quite stable and
> current stable (Debian 10) has more bugs than testing (Debian 11) so you
> can check existing bugs for Debian 11 and if you decide then you can
> install and use Debian testing (Debian 11).

I think running testing now would be ok due to the freeze before the 
release. For future reference, I was told today in #debian-kernel on 
OFTC that "kernels with important security fixes usually go into testing 
within a week during the freeze", and I was also pointed to #988422 - 
kernel 5.10.40 will probably get pushed to testing soon (hopefully also 
on bpo), so running a backported kernel would probably be ok if the fix 
doesn't get backported to 4.19 for the next point release.

Thanks,
Laurențiu

[toc] | [prev] | [next] | [standalone]

#72072

Hi,

On Sun, May 30, 2021 at 08:25:05AM +0200, Laurențiu Păncescu wrote:
> Hello Debian Kernel Team,
> 
> I discovered an older Asus netbook always reports its battery as 100%
> charged even after being unplugged for several hours and with its CPU
> intentionally maxed out. This seems to be upstream bug #199981 in kernels
> 4.17-4.20 and affected Asus-made laptops (I see the same behavior with and
> without acpi_osi=Linux). [1]
> 
> Is there any chance of backporting the fix to Buster? Should I file a bug
> about that?

In general the best way to reach that is to make sure it reaches the
upstream 4.19.y stable branches by asking upstrema to backport the fix
to the series. If it is applicably this is not a problem in most
cases. As in the Debian stable release we follow upstream 4.19.y then
the fix will be included automatically at some point.

Regards,
Salvatore

[toc] | [prev] | [next] | [standalone]

#72075

Hi,

On 6/2/21 5:46 PM, Salvatore Bonaccorso wrote:
> In general the best way to reach that is to make sure it reaches the
> upstream 4.19.y stable branches by asking upstrema to backport the fix
> to the series. If it is applicably this is not a problem in most
> cases. As in the Debian stable release we follow upstream 4.19.y then
> the fix will be included automatically at some point.

Thanks, Salvatore. I reached out on upstream's stable mailing list, and 
Greg KH replied that the patch doesn't apply cleanly and I can backport 
it and sent the patch. So I git-cloned the stable tree and checked out 
branch linux-4.19.y, but I'm unable to compile it (without making any 
changes).

I first tried the approach described in the Debian Kernel Handbook:
$ cp /boot/config-4.19.0-16-amd64 .config
$ scripts/config --disable MODULE_SIG
$ scripts/config --disable DEBUG_INFO
$ make oldconfig
$ make clean
$ make-deb-pkg
...
make[4]: *** No rule to make target 
'debian/certs/debian-uefi-certs.pem', needed by 
'certs/x509_certificate_list'.  Stop.

I then found the instructions for compiling the kernel from the Debian 
Handbook for Buster[1], but it stops before even compiling anything:

$ cp /boot/config-4.19.0-16-amd64 .config
$ make oldconfig
$ vim .config # add CONFIG_TRUSTED_KEYS=""
$ make clean
$ make-deb-pkg LOCALVERSION=-plaur KDEB_PKGVERSION=$(make kernelversion)-1
...
dpkg-source: info: use the '3.0 (quilt)' format to have separate and 
documented changes to upstream files, see dpkg-source(1)
dpkg-source: error: unrepresentable changes to source
dpkg-buildpackage: error: dpkg-source -i.git -b . subprocess returned 
exit status 1
make[1]: *** [scripts/package/Makefile:75: deb-pkg] Error 1
make: *** [Makefile:1390: deb-pkg] Error 2

What is the proper way to compile the upstream 4.19.y kernel, can 
anybody help?

Thanks,
Laurențiu


[1] https://debian-handbook.info/browse/stable/sect.kernel-compilation.html

[toc] | [prev] | [standalone]

Back to top | Article view | linux.debian.kernel

csiph-web