Bug 13988

Summary: X server crashes after starting program with DRI_PRIME=1
Product: Mageia Reporter: Nikita Krupenko <krnekit>
Component: RPM PackagesAssignee: Mageia Bug Squad <bugsquad>
Status: RESOLVED FIXED QA Contact:
Severity: normal    
Priority: Normal CC: thierry.vignaud
Version: CauldronKeywords: NEEDINFO
Target Milestone: ---   
Hardware: x86_64   
OS: Linux   
Whiteboard:
Source RPM: x11-server-1.16.0-3.mga5.src.rpm CVE:
Status comment:
Attachments: Log of crashed X server
Output of xrandr
Output of lspci

Description Nikita Krupenko 2014-08-25 23:18:09 CEST
Description of problem:
I have laptop with AMD A8-5557M APU (Radeon HD 8550G) and discrete videocard Radeon HD 8750M. X server crashing when I try to start program on my discrete card using radeon driver.
I deleted xorg.conf to made both videocards be available (separate bug 13986).
fglrx driver also installed but not used due a too recent X.
Some information about using PRIME I've found in this wiki: https://wiki.archlinux.org/index.php/PRIME

Version-Release number of selected component (if applicable):
x11-server-1.16.0-3.mga5
kernel-desktop-3.15.6-1.mga5
mesa-10.3.0-0.rc1.2.mga5.tainted
x11-driver-video-ati-1:7.4.0-1.mga5

Steps to Reproduce:
1. Get list of videocards using this command:
$ xrandr --listproviders
2. Use xrandr to tell discrete card to output to integrated card:
$ xrandr --setprovideroffloadsink 0x41 0x78
3. Then I launch some program on discrete card:
$ DRI_PRIME=1 glxinfo
4. See X server crash


Reproducible: 

Steps to Reproduce:
Comment 1 Nikita Krupenko 2014-08-25 23:18:48 CEST
Created attachment 5370 [details]
Log of crashed X server
Comment 2 Nikita Krupenko 2014-08-25 23:19:40 CEST
Created attachment 5371 [details]
Output of xrandr
Comment 3 Nikita Krupenko 2014-08-25 23:20:32 CEST
Created attachment 5372 [details]
Output of lspci
Nikita Krupenko 2014-08-25 23:23:00 CEST

Attachment 5370 mime type: application/octet-stream => text/plain

Comment 4 Thierry Vignaud 2014-08-26 10:04:37 CEST
Could you get a GDB trace?
(preferably from a remote ssh connection or scripted -- as you cannot run gdb on xserver from a X terminal as gdb would suspend the X server and thus you would never be able to debug it)

Keywords: (none) => NEEDINFO
CC: (none) => thierry.vignaud

Comment 5 Nikita Krupenko 2014-08-27 19:47:17 CEST
Though I installed debuginfo foll all required libraries and even for kernel, I see thin in bt:

0x000000000055aad4 in ?? ()
(gdb) bt
#0  0x000000000055aad4 in ?? ()
#1  0x0000000000f75200 in ?? ()
#2  0x000000000055ab0d in ?? ()
#3  0x00007fff606192b8 in ?? ()
#4  0x000000000055cacd in ?? ()
#5  0x00007fff606192c0 in ?? ()
#6  0x00007fff606192c8 in ?? ()
#7  0x0000000000ec0e20 in ?? ()
#8  0x0000000000f75200 in ?? ()
#9  0x00000000000002db in ?? ()
#10 0x00000000021f3f80 in ?? ()
#11 0x0000000000f75200 in ?? ()
#12 0x000000000055d78a in ?? ()
#13 0x0000000000f75200 in ?? ()
#14 0x0000000000f0c980 in ?? ()
#15 0x0000000000f75200 in ?? ()
#16 0x00000000004e5c61 in ?? ()
#17 0x00000000001b0001 in ?? ()
#18 0x0000000000000000 in ?? ()
(gdb) c
Continuing.

Program received signal SIGABRT, Aborted.
0x00007fc926e4b917 in __GI_raise (sig=sig@entry=6) at ../nptl/sysdeps/unix/sysv/linux/raise.c:56
56        return INLINE_SYSCALL (tgkill, 3, pid, selftid, sig);
(gdb) bt
#0  0x00007fc926e4b917 in __GI_raise (sig=sig@entry=6) at ../nptl/sysdeps/unix/sysv/linux/raise.c:56
#1  0x00007fc926e4d088 in __GI_abort () at abort.c:89
#2  0x00000000005910de in ?? ()
#3  0x0000000000000001 in ?? ()
#4  0x00000000004763bc in ?? ()
#5  0x00000000005c03e8 in ?? ()
#6  0x00000000005968a7 in ?? ()
#7  0x00007fff606192b0 in ?? ()
#8  0x00000000005976c8 in ?? ()
#9  0x0000003000000018 in ?? ()
#10 0x00007fff60618ca0 in ?? ()
#11 0x00007fff60618bd0 in ?? ()
#12 0x0000003000000008 in ?? ()
#13 0x00007fff60618ca0 in ?? ()
#14 0x00007fff60618bd0 in ?? ()
#15 0x0000000000000000 in ?? ()
(gdb)
Comment 6 Nikita Krupenko 2014-08-27 19:48:37 CEST
I used gdbserver to make this trace.
Comment 7 Samuel Verschelde 2015-05-19 14:06:37 CEST
Is this bug still present in Mageia 5RC?
Comment 8 Nikita Krupenko 2015-05-19 15:50:56 CEST
No, it does not crashes X server anymore.
Comment 9 Samuel Verschelde 2015-05-19 15:53:06 CEST
Closing then

Status: NEW => RESOLVED
Resolution: (none) => FIXED