Github messages for voidlinux
 help / color / mirror / Atom feed
* [ISSUE] stripped .so files in mesa-*-dri-dbg
@ 2019-10-05  2:17 voidlinux-github
  2019-10-07 23:20 ` Stripped " voidlinux-github
                   ` (9 more replies)
  0 siblings, 10 replies; 11+ messages in thread
From: voidlinux-github @ 2019-10-05  2:17 UTC (permalink / raw)
  To: ml

[-- Attachment #1: Type: text/plain, Size: 3423 bytes --]

New issue by blah1 on void-packages repository

https://github.com/void-linux/void-packages/issues/15104

Description:
<!-- Don't request update of package. We have a script for that. https://alpha.de.repo.voidlinux.org/void-updates/void-updates.txt . However, a quality pull request may help. -->
### System

* xuname:  
  `Void 5.2.18_1 x86_64-musl GenuineIntel uptodate rrrnFFFFFF`
* package:  
  `mesa-intel-dri-dbg-19.1.7_3`
possibly also:
  `mesa-ati-dri-dbg-19.1.7_3`
  `mesa-nouveau-dri-dbg-19.1.7_3`
  `mesa-vmwgfx-dri-dbg-19.1.7_3`

### Expected behavior
Using gdb to debug a crash in /usr/lib/xorg/modules/drivers/i915_dri.so shows the relevant functions where the crash occured.
/usr/lib/debug/usr/lib/xorg/modules/i915_dri.so is more than a megabyte in size and not stripped

### Actual behavior
gdb shows "#0  0x00007ffff6d86cd9 in ?? () from /usr/lib/xorg/modules/drivers/i915_dri.so" and similar lines for relevant functions
`file /usr/lib/debug/usr/lib/xorg/modules/drivers/i915_dri.so` says that the file is stripped
`ls -l /usr/lib/debug/usr/lib/xorg/modules/drivers/i915_dri.so` says that the file is only 2768 bytes

### Steps to reproduce the behavior
Run `file /usr/lib/debug/usr/lib/xorg/modules/drivers/i915_dri.so` and see if the file is stripped of debugging symbols
A list of stripped files in mesa-\*-dri-dbg packages can be acquired with `for j in ati nouveau intel vmwgfx ; do xbps-query -f mesa-"$j"-dri-dbg ; done |while read -r i ; do ls -l "$i" ; file "$i" ; done |grep -B1 ', stripped' |grep -v '^--$' |sed '2~2d'`
These are the mesa-*-dri-dbg files that I found that are stripped:
  `/usr/lib/debug/usr/lib/dri/nouveau_drv_video.so`
  `/usr/lib/debug/usr/lib/libXvMCnouveau.so.1.0.0`
  `/usr/lib/debug/usr/lib/vdpau/libvdpau_nouveau.so.1.0.0`
  `/usr/lib/debug/usr/lib/xorg/modules/drivers/nouveau_dri.so`
  `/usr/lib/debug/usr/lib/xorg/modules/drivers/nouveau_vieux_dri.so`
  `/usr/lib/debug/usr/lib/dri/r600_drv_video.so`
  `/usr/lib/debug/usr/lib/vdpau/libvdpau_r300.so.1.0.0`
  `/usr/lib/debug/usr/lib/vdpau/libvdpau_r600.so.1.0.0`
  `/usr/lib/debug/usr/lib/xorg/modules/drivers/r200_dri.so`
  `/usr/lib/debug/usr/lib/xorg/modules/drivers/r300_dri.so`
  `/usr/lib/debug/usr/lib/xorg/modules/drivers/r600_dri.so`
  `/usr/lib/debug/usr/lib/xorg/modules/drivers/i915_dri.so`
  `/usr/lib/debug/usr/lib/xorg/modules/drivers/i965_dri.so`
  `/usr/lib/debug/usr/lib/xorg/modules/drivers/vmwgfx_dri.so`
Take this with a grain of salt because I'm a novice at debugging and for all I know some of those are supposed to be stripped.

### Workaround(?)
In testing mesa-ati-dri-dbg for completeness, I found that using gdb to debug the same crash as before gave the message `warning: the debug information found in "/usr/lib/xorg/modules/drivers/radeon_dri.so" does not match "/usr/lib/xorg/modules/drivers/i915_dri.so" (CRC mismatch).`
gdb then gives backtrace lines such as `#0  0x00007ffff6d86cd9 in _mesa_reallocate_registers (prog=prog@entry=0x55555590b230) at ..src/mesa/program/prog_optimize.c:1103`
I don't know if these are the right function names and line numbers exactly, but it seems like installing mesa-ati-dri-dbg is a workaround?

I think my bug still stands, you shouldn't need to install mesa-ati-dri-dbg and mesa-ati-dri in order to debug mesa-intel-dri. It also seems suspicious that a number of debug object files are stripped.

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: Stripped .so files in mesa-*-dri-dbg
  2019-10-05  2:17 [ISSUE] stripped .so files in mesa-*-dri-dbg voidlinux-github
@ 2019-10-07 23:20 ` voidlinux-github
  2019-10-07 23:22 ` voidlinux-github
                   ` (8 subsequent siblings)
  9 siblings, 0 replies; 11+ messages in thread
From: voidlinux-github @ 2019-10-07 23:20 UTC (permalink / raw)
  To: ml

[-- Attachment #1: Type: text/plain, Size: 710 bytes --]

New comment by st3r4g on void-packages repository

https://github.com/void-linux/void-packages/issues/15104#issuecomment-539245536

Comment:
I think this is due to the fact that the `xorg/modules/drivers/*.so` are created as hard links by Mesa. Output of `ls masterdir/usr/lib/xorg/modules/drivers/ -i`:
> 394315 i915_dri.so
394315 i965_dri.so
394316 kms_swrast_dri.so
394316 nouveau_dri.so
394315 nouveau_vieux_dri.so
394315 r200_dri.so
394316 r300_dri.so
394316 r600_dri.so
394315 radeon_dri.so
394316 radeonsi_dri.so
394316 swrast_dri.so
394316 virtio_gpu_dri.so
394316 vmwgfx_dri.so

They probably result from https://gitlab.freedesktop.org/mesa/mesa/blob/master/bin/install_megadrivers.py.

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: Stripped .so files in mesa-*-dri-dbg
  2019-10-05  2:17 [ISSUE] stripped .so files in mesa-*-dri-dbg voidlinux-github
  2019-10-07 23:20 ` Stripped " voidlinux-github
@ 2019-10-07 23:22 ` voidlinux-github
  2019-10-07 23:25 ` voidlinux-github
                   ` (7 subsequent siblings)
  9 siblings, 0 replies; 11+ messages in thread
From: voidlinux-github @ 2019-10-07 23:22 UTC (permalink / raw)
  To: ml

[-- Attachment #1: Type: text/plain, Size: 867 bytes --]

New comment by st3r4g on void-packages repository

https://github.com/void-linux/void-packages/issues/15104#issuecomment-539245536

Comment:
I think this is due to the fact that the `xorg/modules/drivers/*.so` are created as hard links by Mesa. Output of `ls masterdir/usr/lib/xorg/modules/drivers/ -i`:
> 394315 i915_dri.so
394315 i965_dri.so
394316 kms_swrast_dri.so
394316 nouveau_dri.so
394315 nouveau_vieux_dri.so
394315 r200_dri.so
394316 r300_dri.so
394316 r600_dri.so
394315 radeon_dri.so
394316 radeonsi_dri.so
394316 swrast_dri.so
394316 virtio_gpu_dri.so
394316 vmwgfx_dri.so

They probably result from https://gitlab.freedesktop.org/mesa/mesa/blob/master/bin/install_megadrivers.py. This causes `xbps-src` to output lines like `objcopy: /destdir//mesa-ati-dri-19.2.0/usr/lib/xorg/modules/drivers/stDAOeBg: debuglink section already exists`

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: Stripped .so files in mesa-*-dri-dbg
  2019-10-05  2:17 [ISSUE] stripped .so files in mesa-*-dri-dbg voidlinux-github
  2019-10-07 23:20 ` Stripped " voidlinux-github
  2019-10-07 23:22 ` voidlinux-github
@ 2019-10-07 23:25 ` voidlinux-github
  2019-10-07 23:27 ` voidlinux-github
                   ` (6 subsequent siblings)
  9 siblings, 0 replies; 11+ messages in thread
From: voidlinux-github @ 2019-10-07 23:25 UTC (permalink / raw)
  To: ml

[-- Attachment #1: Type: text/plain, Size: 941 bytes --]

New comment by st3r4g on void-packages repository

https://github.com/void-linux/void-packages/issues/15104#issuecomment-539245536

Comment:
I think this is due to the fact that the `xorg/modules/drivers/*.so` are created as hard links by Mesa. Output of `ls masterdir/usr/lib/xorg/modules/drivers/ -i`:
> 394315 i915_dri.so
394315 i965_dri.so
394316 kms_swrast_dri.so
394316 nouveau_dri.so
394315 nouveau_vieux_dri.so
394315 r200_dri.so
394316 r300_dri.so
394316 r600_dri.so
394315 radeon_dri.so
394316 radeonsi_dri.so
394316 swrast_dri.so
394316 virtio_gpu_dri.so
394316 vmwgfx_dri.so

They probably result from https://gitlab.freedesktop.org/mesa/mesa/blob/master/bin/install_megadrivers.py. Then `xbps-src` will output lines like `objcopy: /destdir//mesa-ati-dri-19.2.0/usr/lib/xorg/modules/drivers/stDAOeBg: debuglink section already exists` because of a previous `attach_debug` which refers to the same underlying file.

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: Stripped .so files in mesa-*-dri-dbg
  2019-10-05  2:17 [ISSUE] stripped .so files in mesa-*-dri-dbg voidlinux-github
                   ` (2 preceding siblings ...)
  2019-10-07 23:25 ` voidlinux-github
@ 2019-10-07 23:27 ` voidlinux-github
  2019-10-07 23:32 ` voidlinux-github
                   ` (5 subsequent siblings)
  9 siblings, 0 replies; 11+ messages in thread
From: voidlinux-github @ 2019-10-07 23:27 UTC (permalink / raw)
  To: ml

[-- Attachment #1: Type: text/plain, Size: 868 bytes --]

New comment by st3r4g on void-packages repository

https://github.com/void-linux/void-packages/issues/15104#issuecomment-539245536

Comment:
I think this is due to the fact that the `xorg/modules/drivers/*.so` are created as hard links by Mesa. Output of `ls masterdir/usr/lib/xorg/modules/drivers/ -i1`:
> 394315 i915_dri.so
394315 i965_dri.so
394316 kms_swrast_dri.so
394316 nouveau_dri.so
394315 nouveau_vieux_dri.so
394315 r200_dri.so
394316 r300_dri.so
394316 r600_dri.so
394315 radeon_dri.so
394316 radeonsi_dri.so
394316 swrast_dri.so
394316 virtio_gpu_dri.so
394316 vmwgfx_dri.so

They probably result from https://gitlab.freedesktop.org/mesa/mesa/blob/master/bin/install_megadrivers.py. This causes `xbps-src` to output lines like `objcopy: /destdir//mesa-ati-dri-19.2.0/usr/lib/xorg/modules/drivers/stDAOeBg: debuglink section already exists`

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: Stripped .so files in mesa-*-dri-dbg
  2019-10-05  2:17 [ISSUE] stripped .so files in mesa-*-dri-dbg voidlinux-github
                   ` (3 preceding siblings ...)
  2019-10-07 23:27 ` voidlinux-github
@ 2019-10-07 23:32 ` voidlinux-github
  2019-10-07 23:33 ` voidlinux-github
                   ` (4 subsequent siblings)
  9 siblings, 0 replies; 11+ messages in thread
From: voidlinux-github @ 2019-10-07 23:32 UTC (permalink / raw)
  To: ml

[-- Attachment #1: Type: text/plain, Size: 949 bytes --]

New comment by st3r4g on void-packages repository

https://github.com/void-linux/void-packages/issues/15104#issuecomment-539245536

Comment:
I think this is due to the fact that the `xorg/modules/drivers/*.so` are created as hard links by Mesa. Output of `ls masterdir/usr/lib/xorg/modules/drivers/ -i1`:
> 394315 i915_dri.so
394315 i965_dri.so
394316 kms_swrast_dri.so
394316 nouveau_dri.so
394315 nouveau_vieux_dri.so
394315 r200_dri.so
394316 r300_dri.so
394316 r600_dri.so
394315 radeon_dri.so
394316 radeonsi_dri.so
394316 swrast_dri.so
394316 virtio_gpu_dri.so
394316 vmwgfx_dri.so

They result from https://gitlab.freedesktop.org/mesa/mesa/blob/master/src/mesa/drivers/dri/meson.build and https://gitlab.freedesktop.org/mesa/mesa/blob/master/bin/install_megadrivers.py. This causes `xbps-src` to output lines like `objcopy: /destdir//mesa-ati-dri-19.2.0/usr/lib/xorg/modules/drivers/stDAOeBg: debuglink section already exists`

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: Stripped .so files in mesa-*-dri-dbg
  2019-10-05  2:17 [ISSUE] stripped .so files in mesa-*-dri-dbg voidlinux-github
                   ` (4 preceding siblings ...)
  2019-10-07 23:32 ` voidlinux-github
@ 2019-10-07 23:33 ` voidlinux-github
  2019-10-07 23:36 ` voidlinux-github
                   ` (3 subsequent siblings)
  9 siblings, 0 replies; 11+ messages in thread
From: voidlinux-github @ 2019-10-07 23:33 UTC (permalink / raw)
  To: ml

[-- Attachment #1: Type: text/plain, Size: 1026 bytes --]

New comment by st3r4g on void-packages repository

https://github.com/void-linux/void-packages/issues/15104#issuecomment-539245536

Comment:
I think this is due to the fact that the `xorg/modules/drivers/*.so` are created as hard links by Mesa. Output of `ls masterdir/usr/lib/xorg/modules/drivers/ -i1`:
> 394315 i915_dri.so
394315 i965_dri.so
394316 kms_swrast_dri.so
394316 nouveau_dri.so
394315 nouveau_vieux_dri.so
394315 r200_dri.so
394316 r300_dri.so
394316 r600_dri.so
394315 radeon_dri.so
394316 radeonsi_dri.so
394316 swrast_dri.so
394316 virtio_gpu_dri.so
394316 vmwgfx_dri.so

They result from https://gitlab.freedesktop.org/mesa/mesa/blob/master/src/mesa/drivers/dri/meson.build and https://gitlab.freedesktop.org/mesa/mesa/blob/master/bin/install_megadrivers.py. This causes `xbps-src` to output lines like `objcopy: /destdir//mesa-ati-dri-19.2.0/usr/lib/xorg/modules/drivers/stDAOeBg: debuglink section already exists` because of a previous attach_debug which refers to the same underlying file.

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: Stripped .so files in mesa-*-dri-dbg
  2019-10-05  2:17 [ISSUE] stripped .so files in mesa-*-dri-dbg voidlinux-github
                   ` (5 preceding siblings ...)
  2019-10-07 23:33 ` voidlinux-github
@ 2019-10-07 23:36 ` voidlinux-github
  2019-10-08  8:19 ` voidlinux-github
                   ` (2 subsequent siblings)
  9 siblings, 0 replies; 11+ messages in thread
From: voidlinux-github @ 2019-10-07 23:36 UTC (permalink / raw)
  To: ml

[-- Attachment #1: Type: text/plain, Size: 1028 bytes --]

New comment by st3r4g on void-packages repository

https://github.com/void-linux/void-packages/issues/15104#issuecomment-539245536

Comment:
I think this is due to the fact that the `xorg/modules/drivers/*.so` are created as hard links by Mesa. Output of `ls masterdir/usr/lib/xorg/modules/drivers/ -i1`:
> 394315 i915_dri.so
394315 i965_dri.so
394316 kms_swrast_dri.so
394316 nouveau_dri.so
394315 nouveau_vieux_dri.so
394315 r200_dri.so
394316 r300_dri.so
394316 r600_dri.so
394315 radeon_dri.so
394316 radeonsi_dri.so
394316 swrast_dri.so
394316 virtio_gpu_dri.so
394316 vmwgfx_dri.so

They result from https://gitlab.freedesktop.org/mesa/mesa/blob/master/src/mesa/drivers/dri/meson.build and https://gitlab.freedesktop.org/mesa/mesa/blob/master/bin/install_megadrivers.py. This causes `xbps-src` to output lines like `objcopy: /destdir//mesa-ati-dri-19.2.0/usr/lib/xorg/modules/drivers/stDAOeBg: debuglink section already exists` because of a previous `attach_debug` which refers to the same underlying file.

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: Stripped .so files in mesa-*-dri-dbg
  2019-10-05  2:17 [ISSUE] stripped .so files in mesa-*-dri-dbg voidlinux-github
                   ` (6 preceding siblings ...)
  2019-10-07 23:36 ` voidlinux-github
@ 2019-10-08  8:19 ` voidlinux-github
  2019-12-27 22:19 ` [ISSUE] [CLOSED] " voidlinux-github
  2019-12-27 22:19 ` voidlinux-github
  9 siblings, 0 replies; 11+ messages in thread
From: voidlinux-github @ 2019-10-08  8:19 UTC (permalink / raw)
  To: ml

[-- Attachment #1: Type: text/plain, Size: 161 bytes --]

New comment by st3r4g on void-packages repository

https://github.com/void-linux/void-packages/issues/15104#issuecomment-539405236

Comment:
Duplicate of #12760

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: Stripped .so files in mesa-*-dri-dbg
  2019-10-05  2:17 [ISSUE] stripped .so files in mesa-*-dri-dbg voidlinux-github
                   ` (8 preceding siblings ...)
  2019-12-27 22:19 ` [ISSUE] [CLOSED] " voidlinux-github
@ 2019-12-27 22:19 ` voidlinux-github
  9 siblings, 0 replies; 11+ messages in thread
From: voidlinux-github @ 2019-12-27 22:19 UTC (permalink / raw)
  To: ml

[-- Attachment #1: Type: text/plain, Size: 164 bytes --]

New comment by xtraeme on void-packages repository

https://github.com/void-linux/void-packages/issues/15104#issuecomment-569355896

Comment:
This is now in master

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [ISSUE] [CLOSED] Stripped .so files in mesa-*-dri-dbg
  2019-10-05  2:17 [ISSUE] stripped .so files in mesa-*-dri-dbg voidlinux-github
                   ` (7 preceding siblings ...)
  2019-10-08  8:19 ` voidlinux-github
@ 2019-12-27 22:19 ` voidlinux-github
  2019-12-27 22:19 ` voidlinux-github
  9 siblings, 0 replies; 11+ messages in thread
From: voidlinux-github @ 2019-12-27 22:19 UTC (permalink / raw)
  To: ml

[-- Attachment #1: Type: text/plain, Size: 3426 bytes --]

Closed issue by blah1 on void-packages repository

https://github.com/void-linux/void-packages/issues/15104

Description:
<!-- Don't request update of package. We have a script for that. https://alpha.de.repo.voidlinux.org/void-updates/void-updates.txt . However, a quality pull request may help. -->
### System

* xuname:  
  `Void 5.2.18_1 x86_64-musl GenuineIntel uptodate rrrnFFFFFF`
* package:  
  `mesa-intel-dri-dbg-19.1.7_3`
possibly also:
  `mesa-ati-dri-dbg-19.1.7_3`
  `mesa-nouveau-dri-dbg-19.1.7_3`
  `mesa-vmwgfx-dri-dbg-19.1.7_3`

### Expected behavior
Using gdb to debug a crash in /usr/lib/xorg/modules/drivers/i915_dri.so shows the relevant functions where the crash occured.
/usr/lib/debug/usr/lib/xorg/modules/i915_dri.so is more than a megabyte in size and not stripped

### Actual behavior
gdb shows "#0  0x00007ffff6d86cd9 in ?? () from /usr/lib/xorg/modules/drivers/i915_dri.so" and similar lines for relevant functions
`file /usr/lib/debug/usr/lib/xorg/modules/drivers/i915_dri.so` says that the file is stripped
`ls -l /usr/lib/debug/usr/lib/xorg/modules/drivers/i915_dri.so` says that the file is only 2768 bytes

### Steps to reproduce the behavior
Run `file /usr/lib/debug/usr/lib/xorg/modules/drivers/i915_dri.so` and see if the file is stripped of debugging symbols
A list of stripped files in mesa-\*-dri-dbg packages can be acquired with `for j in ati nouveau intel vmwgfx ; do xbps-query -f mesa-"$j"-dri-dbg ; done |while read -r i ; do ls -l "$i" ; file "$i" ; done |grep -B1 ', stripped' |grep -v '^--$' |sed '2~2d'`
These are the mesa-*-dri-dbg files that I found that are stripped:
  `/usr/lib/debug/usr/lib/dri/nouveau_drv_video.so`
  `/usr/lib/debug/usr/lib/libXvMCnouveau.so.1.0.0`
  `/usr/lib/debug/usr/lib/vdpau/libvdpau_nouveau.so.1.0.0`
  `/usr/lib/debug/usr/lib/xorg/modules/drivers/nouveau_dri.so`
  `/usr/lib/debug/usr/lib/xorg/modules/drivers/nouveau_vieux_dri.so`
  `/usr/lib/debug/usr/lib/dri/r600_drv_video.so`
  `/usr/lib/debug/usr/lib/vdpau/libvdpau_r300.so.1.0.0`
  `/usr/lib/debug/usr/lib/vdpau/libvdpau_r600.so.1.0.0`
  `/usr/lib/debug/usr/lib/xorg/modules/drivers/r200_dri.so`
  `/usr/lib/debug/usr/lib/xorg/modules/drivers/r300_dri.so`
  `/usr/lib/debug/usr/lib/xorg/modules/drivers/r600_dri.so`
  `/usr/lib/debug/usr/lib/xorg/modules/drivers/i915_dri.so`
  `/usr/lib/debug/usr/lib/xorg/modules/drivers/i965_dri.so`
  `/usr/lib/debug/usr/lib/xorg/modules/drivers/vmwgfx_dri.so`
Take this with a grain of salt because I'm a novice at debugging and for all I know some of those are supposed to be stripped.

### Workaround(?)
In testing mesa-ati-dri-dbg for completeness, I found that using gdb to debug the same crash as before gave the message `warning: the debug information found in "/usr/lib/xorg/modules/drivers/radeon_dri.so" does not match "/usr/lib/xorg/modules/drivers/i915_dri.so" (CRC mismatch).`
gdb then gives backtrace lines such as `#0  0x00007ffff6d86cd9 in _mesa_reallocate_registers (prog=prog@entry=0x55555590b230) at ..src/mesa/program/prog_optimize.c:1103`
I don't know if these are the right function names and line numbers exactly, but it seems like installing mesa-ati-dri-dbg is a workaround?

I think my bug still stands, you shouldn't need to install mesa-ati-dri-dbg and mesa-ati-dri in order to debug mesa-intel-dri. It also seems suspicious that a number of debug object files are stripped.

^ permalink raw reply	[flat|nested] 11+ messages in thread

end of thread, other threads:[~2019-12-27 22:19 UTC | newest]

Thread overview: 11+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-10-05  2:17 [ISSUE] stripped .so files in mesa-*-dri-dbg voidlinux-github
2019-10-07 23:20 ` Stripped " voidlinux-github
2019-10-07 23:22 ` voidlinux-github
2019-10-07 23:25 ` voidlinux-github
2019-10-07 23:27 ` voidlinux-github
2019-10-07 23:32 ` voidlinux-github
2019-10-07 23:33 ` voidlinux-github
2019-10-07 23:36 ` voidlinux-github
2019-10-08  8:19 ` voidlinux-github
2019-12-27 22:19 ` [ISSUE] [CLOSED] " voidlinux-github
2019-12-27 22:19 ` voidlinux-github

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).