Skip to contentSkip to conversation

5.2-rc0.8: emacs segfaults?! x220, with 32-bit userland

DISCUSSION Pavel Machek 2 participants 13 messages May 19, 2019 – Jun 1, 2019

lore
Share & tools
lore
#0 lore raw

Hi!

emacs segfaults... when I attempt to exit it. And that did not use to be the case. Nothing suspect in the dmesg. Rest of the machine seems to work ok.

I'm using "l1tf=full" on kernel command line.

During one crash, I got some kind of debug output (from glibc?), but it scrolled away when I asked for dmesg :-(.

Ideas welcome...
Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#1 lore raw
Quoted 1 line · Pavel Machek

Hi!

Hi,

Quoted 3 lines

emacs segfaults... when I attempt to exit it. And that did not use to be the case. Nothing suspect in the dmesg. Rest of the machine seems to work ok.

…

Quoted 1 line

Ideas welcome...

I assume that this happens with -rc1, too. Could you please check if emacs still segfaults with commit

d9c9ce34ed5c8 ("x86/fpu: Fault-in user stack if copy_fpstate_to_sigframe() fails")

and commit

89833fab15d60 ("x86/fpu: Fix __user annotations")

?

Sebastian

#2 lore raw

Hi!

Quoted 9 lines
Quoted 3 lines

emacs segfaults... when I attempt to exit it. And that did not use to be the case. Nothing suspect in the dmesg. Rest of the machine seems to work ok.

…

Quoted 1 line

Ideas welcome...

I assume that this happens with -rc1, too. Could you please check if emacs still segfaults with commit

Quoted 7 lines

d9c9ce34ed5c8 ("x86/fpu: Fault-in user stack if copy_fpstate_to_sigframe() fails")

and commit

89833fab15d60 ("x86/fpu: Fix __user annotations")

?

I don't have reproducible test case :-(. I had two or three failures so far.

Best regards,

Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#3 lore raw
Quoted 1 line · Pavel Machek

Hi!

Hi,

Quoted 2 lines

I don't have reproducible test case :-(. I had two or three failures so far.

oki. Could you please send me the output of
dmesg | grep fpu

Sebastian

#4 lore raw

Hi!

I did not notice any new crashes.

Quoted 5 lines
Quoted 2 lines

I don't have reproducible test case :-(. I had two or three failures so far.

oki. Could you please send me the output of
dmesg | grep fpu

Here you go:

Using method `randr'.

pavel@duo:~$ uname -a
Linux duo 5.2.0-rc1+ #32 SMP Mon May 20 12:51:17 CEST 2019 x86_64
GNU/Linux
pavel@duo:~$ dmesg | grep fpu
[    0.000000] x86/fpu: Supporting XSAVE feature 0x001: 'x87 floating
point registers'
[    0.000000] x86/fpu: Supporting XSAVE feature 0x002: 'SSE
registers'
[    0.000000] x86/fpu: Supporting XSAVE feature 0x004: 'AVX
registers'
[    0.000000] x86/fpu: xstate_offset[2]:  576, xstate_sizes[2]:  256
[    0.000000] x86/fpu: Enabled xstate features 0x7, context size is
832 bytes, using 'standard' format.
pavel@duo:~$
model name	: Intel(R) Core(TM) i5-2520M CPU @ 2.50GHz

Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#5 lore raw

5.1.0-next-20190520 -- emacs segfaults on 32-bit machine Re: 5.2-rc0.8: emacs segfaults?! x220, with 32-bit userland

Hi!

Quoted 1 line

I did not notice any new crashes.

New crash now; different machine, way -next kernel... and I even have a backtrace.

Pavel

pavel@amd:~$ e cip/HOURS
Fatal error 11: Segmentation fault
Backtrace:
emacs[0x8138719]
emacs[0x8120446]
emacs[0x813758e]
emacs[0x81375fb]
/lib/i386-linux-gnu/i686/cmov/libpthread.so.0(+0xecb8)[0xb636ccb8]
/lib/i386-linux-gnu/i686/cmov/libc.so.6(+0xebbf3)[0xb6283bf3]
/lib/i386-linux-gnu/i686/cmov/libc.so.6(pselect+0xd3)[0xb627c363]
emacs[0x81ffdd9]
emacs[0x81c7ba4]
emacs[0x8127815]
emacs[0x812ada3]
emacs[0x812bdad]
emacs[0x812d838]
emacs[0x818b76c]
emacs[0x8120890]
emacs[0x818b66b]
emacs[0x8124b84]
emacs[0x8124e3f]
emacs[0x8059cb0]
/lib/i386-linux-gnu/i686/cmov/libc.so.6(__libc_start_main+0xf3)[0xb61b1a63]
emacs[0x805a76f]
^CSegmentation fault (core dumped)
pavel@amd:~$ uname -a
Linux amd 5.1.0-next-20190520 #285 SMP Mon May 20 13:20:20 CEST 2019
i686 GNU/Linux
pavel@amd:~$ dmesg | grep fpu
[    0.000000] x86/fpu: x87 FPU will use FXSAVE
"Genuine Intel(R) CPU           T2400  @ 1.83GHz"
Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#6 lore raw

Re: 5.1.0-next-20190520 -- emacs segfaults on 32-bit machine Re: 5.2-rc0.8: emacs segfaults?! x220, with 32-bit userland

Hi!

Quoted 4 lines
Quoted 1 line

I did not notice any new crashes.

New crash now; different machine, way -next kernel... and I even have a backtrace.

And one more crash, on x220:

pavel@duo:/data/l/linux-next-32$ uname -a
Linux duo 5.2.0-rc1+ #32 SMP Mon May 20 12:51:17 CEST 2019 x86_64
GNU/Linux
pavel@duo:~$ fg
e cip/HOURS
Fatal error 11: Segmentation fault
Backtrace:
emacs[0x8138719]
emacs[0x8120446]
emacs[0x813758e]
emacs[0x81375fb]
/lib/i386-linux-gnu/i686/cmov/libpthread.so.0(+0xecb8)[0xf634fcb8]
/lib/i386-linux-gnu/i686/cmov/libc.so.6(+0xebbf3)[0xf6266bf3]
/lib/i386-linux-gnu/i686/cmov/libc.so.6(pselect+0xd3)[0xf625f363]
emacs[0x81ffdd9]
emacs[0x81c7ba4]
emacs[0x8127815]
emacs[0x812ada3]
emacs[0x812bdad]
emacs[0x812d838]
emacs[0x818b76c]
emacs[0x8120890]
emacs[0x818b66b]
emacs[0x8124b84]
emacs[0x8124e3f]
emacs[0x8059cb0]
/lib/i386-linux-gnu/i686/cmov/libc.so.6(__libc_start_main+0xf3)[0xf6194a63]
emacs[0x805a76f]
^P/data/pavel/bin/e: line 2: 12713 Segmentation fault      (core
dumped) emacs -nw $*
pavel@duo:~$ fg

Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#7 lore raw

Re: 5.1.0-next-20190520 -- emacs segfaults on 32-bit machine Re: 5.2-rc0.8: emacs segfaults?! x220, with 32-bit userland

In reply to Pavel Machek
Quoted 1 line · Pavel Machek

Hi!

Hi,

Quoted 4 lines
Quoted 1 line

I did not notice any new crashes.

New crash now; different machine, way -next kernel... and I even have a backtrace.

could you please send me (offlist) your .config? Also, what kind of userland do you run? Something like Debian stable?

Quoted 1 line

Pavel

Sebastian

#8 lore raw

Re: 5.1.0-next-20190520 -- emacs segfaults on 32-bit machine Re: 5.2-rc0.8: emacs segfaults?! x220, with 32-bit userland

Quoted 11 lines · Sebastian Andrzej Siewior
Quoted 1 line · Pavel Machek

Hi!

Hi,

Quoted 4 lines
Quoted 1 line

I did not notice any new crashes.

New crash now; different machine, way -next kernel... and I even have a backtrace.

could you please send me (offlist) your .config? Also, what kind of userland do you run? Something like Debian stable?

Yep, debian stable.

cat /etc/debian_version
8.11

And now it happened again, and yes, emacs again:

emacs: ../../../../src/cairo-arc.c:189: _cairo_arc_in_direction:
Assertion `angle_max >= angle_min' failed.
Fatal error 6: Aborted

Backtrace:

emacs[0x8138719]
emacs[0x8120446]
emacs[0x813758e]
emacs[0x81375fb] /lib/i386-linux-gnu/i686/cmov/libpthread.so.0(+0xecb8)[0xf636ccb8] /lib/ld-linux.so.2(+0xc42)[0xf7f49c42] /lib/i386-linux-gnu/i686/cmov/libc.so.6(gsignal+0x47)[0xf61c6367] /lib/i386-linux-gnu/i686/cmov/libc.so.6(abort+0x143)[0xf61c7a23] /lib/i386-linux-gnu/i686/cmov/libc.so.6(+0x276c7)[0xf61bf6c7] /lib/i386-linux-gnu/i686/cmov/libc.so.6(+0x27777)[0xf61bf777] /usr/lib/i386-linux-gnu/libcairo.so.2(+0xee02)[0xf735ce02] /usr/lib/i386-linux-gnu/libcairo.so.2(+0x2803b)[0xf737603b] /usr/lib/i386-linux-gnu/libcairo.so.2(cairo_arc+0x5c)[0xf736d94c] /usr/lib/i386-linux-gnu/libgtk-3.so.0(+0x2558f5)[0xf78468f5] /usr/lib/i386-linux-gnu/libgtk-3.so.0(+0x255e61)[0xf7846e61] /usr/lib/i386-linux-gnu/libgtk-3.so.0(+0xb21e3)[0xf76a31e3] /usr/lib/i386-linux-gnu/libgtk-3.so.0(+0xb25ca)[0xf76a35ca] /usr/lib/i386-linux-gnu/libgtk-3.so.0(+0xb2f4f)[0xf76a3f4f] /usr/lib/i386-linux-gnu/libgtk-3.so.0(gtk_render_frame+0x11c)[0xf787051c] /usr/lib/i386-linux-gnu/libgtk-3.so.0(+0xf1e97)[0xf76e2e97] /usr/lib/i386-linux-gnu/libgtk-3.so.0(+0x1e44cd)[0xf77d54cd] /usr/lib/i386-linux-gnu/libgtk-3.so.0(+0x31bf3e)[0xf790cf3e] /usr/lib/i386-linux-gnu/libgobject-2.0.so.0(+0xc2e2)[0xf711d2e2] /usr/lib/i386-linux-gnu/libgobject-2.0.so.0(+0xd9c9)[0xf711e9c9] /usr/lib/i386-linux-gnu/libgobject-2.0.so.0(g_signal_emit_valist+0x3e3)[0xf7138463] /usr/lib/i386-linux-gnu/libgobject-2.0.so.0(g_signal_emit+0x25)[0xf71391e5] /usr/lib/i386-linux-gnu/libgtk-3.so.0(+0x32b037)[0xf791c037] /usr/lib/i386-linux-gnu/libgtk-3.so.0(+0x32cc89)[0xf791dc89] /usr/lib/i386-linux-gnu/libgtk-3.so.0(gtk_container_propagate_draw+0x24c)[0xf771f73c] /usr/lib/i386-linux-gnu/libgtk-3.so.0(+0x12e80e)[0xf771f80e] /usr/lib/i386-linux-gnu/libgtk-3.so.0(+0x16eb78)[0xf775fb78] /usr/lib/i386-linux-gnu/libgtk-3.so.0(+0x1e44cd)[0xf77d54cd] /usr/lib/i386-linux-gnu/libgtk-3.so.0(+0x31bf3e)[0xf790cf3e] /usr/lib/i386-linux-gnu/libgobject-2.0.so.0(+0xc2e2)[0xf711d2e2] /usr/lib/i386-linux-gnu/libgobject-2.0.so.0(+0xd9c9)[0xf711e9c9] /usr/lib/i386-linux-gnu/libgobject-2.0.so.0(g_signal_emit_valist+0x3e3)[0xf7138463] /usr/lib/i386-linux-gnu/libgobject-2.0.so.0(g_signal_emit+0x25)[0xf71391e5] /usr/lib/i386-linux-gnu/libgtk-3.so.0(+0x32b037)[0xf791c037] /usr/lib/i386-linux-gnu/libgtk-3.so.0(+0x32c86d)[0xf791d86d] /usr/lib/i386-linux-gnu/libgtk-3.so.0(+0x32cb09)[0xf791db09] /usr/lib/i386-linux-gnu/libgtk-3.so.0(gtk_container_propagate_draw+0x24c)[0xf771f73c]
...
Aborted (core dumped)

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#12 lore raw

Re: 5.1.0-next-20190520 -- emacs segfaults on 32-bit machine Re: 5.2-rc0.8: emacs segfaults?! x220, with 32-bit userland

In reply to Sebastian Andrzej Siewior
Quoted 11 lines · Sebastian Andrzej Siewior
Quoted 1 line · Pavel Machek

Hi!

Hi,

Quoted 4 lines
Quoted 1 line

I did not notice any new crashes.

New crash now; different machine, way -next kernel... and I even have a backtrace.

could you please send me (offlist) your .config? Also, what kind of userland do you run? Something like Debian stable?

You have configs in your inbox.
Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#9 lore raw

Re: 5.1.0-next-20190520 -- emacs segfaults on 32-bit machine Re: 5.2-rc0.8: emacs segfaults?! x220, with 32-bit userland

In reply to Pavel Machek
Quoted 4 lines · Pavel Machek
Quoted 2 lines

could you please send me (offlist) your .config? Also, what kind of userland do you run? Something like Debian stable?

Yep, debian stable.

Since we had a little bit of development recently, could you please check if

	http://lkml.kernel.org/r/[email protected]

makes any difference?

Sebastian

#10 lore raw

Re: 5.1.0-next-20190520 -- emacs segfaults on 32-bit machine Re: 5.2-rc0.8: emacs segfaults?! x220, with 32-bit userland

Quoted 11 lines · Sebastian Andrzej Siewior
Quoted 4 lines · Pavel Machek
Quoted 2 lines

could you please send me (offlist) your .config? Also, what kind of userland do you run? Something like Debian stable?

Yep, debian stable.

Since we had a little bit of development recently, could you please check if

	http://lkml.kernel.org/r/[email protected]

makes any difference?

Ok, I updated to -rc2 and applied that patch. I'll let you know if I see another emacs crash.
Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#11 lore raw

Re: 5.1.0-next-20190520 -- emacs segfaults on 32-bit machine Re: 5.2-rc0.8: emacs segfaults?! x220, with 32-bit userland

In reply to Sebastian Andrzej Siewior
Quoted 11 lines · Sebastian Andrzej Siewior
Quoted 4 lines · Pavel Machek
Quoted 2 lines

could you please send me (offlist) your .config? Also, what kind of userland do you run? Something like Debian stable?

Yep, debian stable.

Since we had a little bit of development recently, could you please check if

	http://lkml.kernel.org/r/[email protected]

makes any difference?

X version of emacs keeps crashing, but I guess that's something unrelated. Looks like this one is solved.
Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html