On Mon, 2021-03-08 at 15:59 +0530, Ritesh Raj Sarraf wrote:
> Hi,
>
> Just a follow-up question on this fix.
>
> Is it something that is a candidate for linux-stable ?
I guess that makes sense. Once it's in mainline you can also request
that yourself :)
johannes
On Sun, 2021-03-07 at 21:22 +0900, Hajime Tazaki wrote:
> Sorry that this email is going to be long. In summary, what Johannes
> said is right: what objcopy does is not sufficient, and with ld it
> transforms as we expected.
>
> More goes to below.
[snip]
Interesting, thanks for looking into
From: Johannes Berg
Ritesh reported a bug [1] against UML, noting that it crashed on
startup. The backtrace shows the following (heavily redacted):
(gdb) bt
...
#26 0x60015b5d in sem_init () at ipc/sem.c:268
#27 0x7f89906d92f7 in ?? () from /lib/x86_64-linux-gnu/libcom_err.so.2
On Thu, 2021-03-04 at 14:38 +0900, Hajime Tazaki wrote:
>
> objcopy (from binutils) can localize symbols (i.e., objcopy -L
> sem_init $orig_file $new_file).
This doesn't seem to be sufficient.
> It also does renaming symbols. But
> not sure this is the ideal solution.
Even that doesn't seem
> Ritesh, can you give the following a spin - it renames sem_init as
> um_sem_init for UML only?
FWIW, this fixes the issue in my reproducer, so should work here too:
diff --git a/ipc/util.h b/ipc/util.h
index 5766c61aed0e..cfed40ba983c 100644
--- a/ipc/util.h
+++ b/ipc/util.h
@@ -14,6 +14,7
On Fri, 2021-03-05 at 19:03 +, Anton Ivanov wrote:
>
> I thought of that, but surrendered to the "dark side" of the quick and ugly
> fix.
:)
> We can do that for the ipc/sem.c - it brings in uaccess.h which
> ultimately pulls uaccess from our asm tree. So if we do it there, it
> will end
On Wed, 2021-03-03 at 23:40 +0100, Johannes Berg wrote:
> Now libcom_err.so.2 is trying to call sem_init(), and that gets ... tada
> ... Linux's sem_init() instead of libpthread's.
>
> And then the crash.
FWIW, I can trivially reproduce this by simply force-loading
libcom_err.so:
On 5 March 2021 18:39:42 CET, Anton Ivanov
wrote:
>
>
>On 04/03/2021 07:47, Johannes Berg wrote:
>> On Thu, 2021-03-04 at 14:38 +0900, Hajime Tazaki wrote:
>>
>>>> Now, I don't know how to fix it (short of changing your nsswitch
>>>> configurat
On Fri, 2021-03-05 at 09:59 +, Anton Ivanov wrote:
>
> This is proving very "interesting" to try to chase down, because the
> "picking the wrong library" does not happen every time.
>
> F.E. yesterday my 5.10 builds were picking glibc memcpy and friends.
> Today with the same config and
On Thu, 2021-03-04 at 14:38 +0900, Hajime Tazaki wrote:
> > Now, I don't know how to fix it (short of changing your nsswitch
> > configuration) - maybe we could somehow rename sem_init()? Or maybe we
> > can somehow give the kernel binary a lower symbol resolution than the
> > libc/libpthread.
>
On Thu, 2021-03-04 at 07:28 +, Anton Ivanov wrote:
>
> > Now, I don't know how to fix it (short of changing your nsswitch
> > configuration) - maybe we could somehow rename sem_init()? Or maybe we
> > can somehow give the kernel binary a lower symbol resolution than the
> > libc/libpthread.
>
I think the problem is here:
> #24 0x6080f234 in ipc_init_ids (ids=0x60c60de8 )
> at ipc/util.c:119
> #25 0x60813c6d in sem_init_ns (ns=0x60d895bb ) at
> ipc/sem.c:254
> #26 0x60015b5d in sem_init () at ipc/sem.c:268
> #27 0x7f89906d92f7 in ?? () from
On Mon, 2019-09-30 at 14:19 +0100, Anton Ivanov wrote:
> All vector drivers now allow a BPF program to be loaded and
> associated with the RX socket in the host kernel.
>
> 1. The program can be loaded as an extra kernel command line
> option to any of the drivers.
>
> 2. The program can also be
Version 1.4 of my plugin didn't have support for 2.0, so it shouldn't
even have installed, but now I did add support for 2.0 in the git
version.
johannes
--
To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact
On Thu, 2009-07-23 at 00:12 +0930, Ron wrote:
Hi Johannes,
It was reported that dovecot-antispam currently has trouble on ia64
due to a missing prototype for getenv. You can find the original
report here: http://bugs.debian.org/537765
And the trivial patch I've applied to the debian
On Sat, 2006-05-13 at 15:13 -0700, Arias Hung wrote:
Attached as requested.
Thanks. Could you also add -s 1000 so I can see what messages are sent
and received to/from hald-runner?
And it's fine to just mail it to me privately too, probably no sense in
spamming the bug system.
Thanks,
Hey,
Since I wrote the code that seems to be causing this... I guess I should
look into this.
Can I get an strace that has the -f option on its command line?
Thanks,
johannes
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
On Fri, 2006-05-12 at 15:51 +0200, Martin Pitt wrote:
Thank you! BTW, do you happen to have any pointer about documentation
of that ioctl?
linux sources :)
drivers/input/evdev.c
lines 308 to 369 contain the code copying the bitmap to userspace
I've looked into it and I'm pretty sure my patch
On Fri, 2006-05-12 at 17:51 +0200, Martin Pitt wrote:
Please note that the crashes were reported on i386 and amd64, both of
which are little endian.
Yeah, I saw. Is hal a 64-bit program when running on amd64?
johannes
signature.asc
Description: This is a digitally signed message part
I also just got this when upgrading and it turned out that
my /var/lib/dpkg/alternatives/pc file was somehow broken. I purged
fp-compiler and deleted the file, and reinstalling now worked (I have no
other pc installed at all!)
This was the contents of the file:
auto
/usr/bin/pc
fpc.1.gz
20 matches
Mail list logo
