Changes
Page history
cleanups, mostly typo fixes
authored
Apr 10, 2025
by
Uwe Kleine-König
Show whitespace changes
Inline
Side-by-side
20250409.md
View page @
042f7ada
...
...
@@ -63,9 +63,9 @@ Chair: @carnil
*
[
\#1088522
](
https://bugs.debian.org/1088522
)
:
Situation:
RasperryPi, OOT lkrg module, requires config setting enabled, amd64 is enabled.
Situation: Rasp
b
erryPi, OOT lkrg module, requires config setting enabled, amd64 is enabled.
S
urprised that i
s
is not consistent enabled.
Users are s
urprised that i
t
is not consistent
ly
enabled.
Concerns about enabling it, since kconfig help explicitly tells that if concerned about memory, increased size of image.
...
...
@@ -73,19 +73,19 @@ Concerns about enabling it, since kconfig help explicitly tells that if concerne
Remark: Config is enabled on amd64 because we enabled livepatch on amd64.
Is livepatch for all architctures? Just going for arm64 does not sound right. (ppc64el, s390x, ... major architectures)
Is livepatch
available
for all archit
e
ctures? Just going for arm64 does not sound right. (ppc64el, s390x, ... major architectures)
@benh: We could ask people
ask people
(@santiago, @eamanu) working on livepatch implementation.
@benh: We could ask people (@santiago, @eamanu) working on livepatch implementation.
@ema: livepatch only supported on amd64, but they will need as well the option on arm64 when its ready.
livepatch: is patching arbitrary code and needs know where it is located.
@ema: kprobes as well
as
argument
s
to enable the option.
@ema: kprobes as well
is an
argument to enable the option.
@ukleinek: Is it already tech
i
ncally possible that depending bootparameter to get all symbols or not?
@ukleinek: Is it already techn
i
cally possible that depending
on a
boot
parameter to get all symbols or not?
@benh: possibly there is not
not
(yet) an option for that. Don't know if there is a mechanism for that.
@benh: possibly there is not (yet) an option for that. Don't know if there is a mechanism for that.
@ema: technically possible, but there is no bootparameter for that.
...
...
@@ -93,9 +93,9 @@ livepatch: is patching arbitrary code and needs know where it is located.
@ema: All the other distros have it on.
@enh: Main reason is for debugging and tracing, so enable it on all architectues where we have BTF enabled?
@
b
enh: Main reason is for debugging and tracing, so enable it on all architectu
r
es where we have BTF enabled?
Agreement: enable it on all architec
u
tes where BTF is enabled. @benh followsup on the MR, @ema will take care.
Agreement: enable it on all architect
ur
es where BTF is enabled. @benh followsup on the MR, @ema will take care.
*
[
\#1101207
](
https://bugs.debian.org/1101207
)
...
...
@@ -106,11 +106,11 @@ Action: close the bug
@ukleinek: maybe a hardware problem with RAM? Can be.
@ema: could be a mem
e
ory problem.
@ema: could be a memory problem.
-> Try a memory checker, to see if they have some bad ram.
@benh: we could disas
el
ble the code.
@benh: we could disas
sem
ble the code.
We could look deeper in the code, look into exact failure but first to test hardware
...
...
@@ -144,9 +144,9 @@ Action. (@carnil ask for the above informations)
-
more information about full kernel log
-
ask if it is a regresson? (maybe it is ....)
-
ask if it is a regress
i
on? (maybe it is ....)
-
do n
o
ewer kernel work?
-
do newer kernel
s
work?
-
error reported is quite uncommon.
...
...
@@ -156,7 +156,7 @@ Action: @ema take action to ask for more information.
*
[
\#1101546
](
https://bugs.debian.org/1101546
)
@waldi: firware possibl
l
y not loaded?
@waldi: fir
m
ware possibly not loaded?
@benh: suspect BIOS update did reset things harder and made wifi work again
...
...
@@ -164,9 +164,9 @@ Action: @ema take action to ask for more information.
Action: possible ask for kernel log for working system and nonworking state?
@waldi: the reporter does not have the correct firmware, fails to load every firmware it tries? Linux is not able to load the fir
w
ware.
@waldi: the reporter does not have the correct firmware, fails to load every firmware it tries? Linux is not able to load the fir
m
ware.
-> We do not have log for a non
o
wrking state.
-> We do not have log for a nonw
o
rking state.
Linux tries to detect the HW and fails with timeout.
...
...
@@ -180,19 +180,19 @@ User should use first reportbug
Ask for a proper report, if nothing comes back close the bugreport
(
ma
ybe close the bug
as fixed in bookworm-backports)
ma
rk
as fixed in bookworm-backports)
@ukleinek takes action.
*
[
\#1101940
](
https://bugs.debian.org/1101940
)
amdgpu is not only thing broken, we need more logs
amdgpu is not
the
only thing broken, we need more logs
should get more information / full.
upstream issue which look similar.
related issues are linked, no
t not
resolution as far we can see.
related issues are linked, no resolution as far we can see.
@benh: amdgpu dkms found bug about the adding of modprobe file, filled upstream
...
...
@@ -204,8 +204,6 @@ related issues are linked, not not resolution as far we can see.
Close the bug, there is nothing to do with with that
-> close the bug.
-> Action: close the bug (@carnil)
### New version 6.15
...
...
@@ -222,15 +220,15 @@ Close the bug, there is nothing to do with with that
*
modprobe.d bug report
one problem: implementation either requires debianutils from trixie or
`modprobe -c`
one problem: implementation either requires debianutils from trixie
(for multi-directory support in run-parts)
or
`modprobe -c`
considering about
`modprobe -c`
and filter out things out.
`run-parts`
does not properly skip directories
which
do
es
not exist -> @waldi has to fill a bug
`run-parts`
does not properly skip directories
that
do not exist -> @waldi has to fill a bug
*
@ukleinek takes arm64 specific MRs
-
@ukleinek: [arm64] Enable
`ARM_CCA_GUEST`
as module important for both flawours? -> so mostly useful
l
for cloud images
but ne
abled for both.
-
@ukleinek: [arm64] Enable
`ARM_CCA_GUEST`
as module important for both flawours? -> so mostly useful for cloud images
so should be en
abled for both.
*
Enable
`CONFIG_HZ_1000`
...
...
@@ -240,7 +238,7 @@ So just take into consideration just for `debian/latest` not trixie.
Keep it out of trixie
@benh: not saying we should defintively enable them as suggested. They should just be considered as well.
@benh: not saying we should defin
i
tively enable them as suggested. They should just be considered as well.
Not urgent, we can see how it moves on.
...
...
...
...