Bug 116724 (shmat)

Summary: X segfaults with kernel-2.6.3-1.100 with SHM_IPC errors
Product: [Fedora] Fedora Reporter: Kaj J. Niemi <kajtzu>
Component: kernelAssignee: Dave Jones <davej>
Status: CLOSED RAWHIDE QA Contact:
Severity: medium Docs Contact:
Priority: medium    
Version: rawhideCC: hardawayd, mharris, pfrields, walkera
Target Milestone: ---   
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2004-02-25 14:48:45 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Bug Depends On:    
Bug Blocks: 114961    
Attachments:
Description Flags
XFree86.0.log, doesn't seem to much interesting related to the segfaults none

Description Kaj J. Niemi 2004-02-24 18:43:09 UTC
Description of problem:
After rebooting with kernel-2.6.3-1.100 both rhgb started and run
level 5 X servers segfault. Everything works fine with kernel-2.6.3-1.97.

I'll attach XFree86.0.log, below is the output from messages:

Feb 24 20:30:35 d100 gdm[2268]: gdm_slave_xioerror_handler: Fatal X
error - Restarting :0
Feb 24 20:30:37 d100 gdm[2272]: gdm_slave_xioerror_handler: Fatal X
error - Restarting :0
Feb 24 20:30:43 d100 gdm[2276]: gdm_slave_xioerror_handler: Fatal X
error - Restarting :0
Feb 24 20:30:48 d100 gdm[2280]: gdm_slave_xioerror_handler: Fatal X
error - Restarting :0
Feb 24 20:30:48 d100 gdm[2213]: deal_with_x_crashes: Running the
XKeepsCrashing script


Version-Release number of selected component (if applicable):
XFree86-4.3.0-59
kernel-2.6.3-1.100

How reproducible:
Always.

Steps to Reproduce:
1. reboot with 2.6.3-1.100
2. wait for rhgb or run level 5 X to start
3.

Comment 1 Kaj J. Niemi 2004-02-24 18:44:59 UTC
Created attachment 98003 [details]
XFree86.0.log, doesn't seem to much interesting related to the segfaults

Comment 2 Mike A. Harris 2004-02-24 20:05:15 UTC
This seems rather significant to me actually:

(EE) RADEON(0): shmat() call retruned errno 1013

So does:

(WW) RADEON(0): Video BIOS not detected in PCI space!
(WW) RADEON(0): Attempting to read Video BIOS from legacy ISA space!

Comment 3 Kaj J. Niemi 2004-02-24 20:15:24 UTC
Whoops, I skimmed it for The Big Error.. that EE was so suddle ;-) My bad.

Related, Warren Togami reported the same with 2.6.3-1.99 in his blog.

Maybe related, 2.6.3-1.100 lost the USB controllers as well. I noticed
it when kudzu asked if I really wanted to remove them.

Comment 4 Gene Czarcinski 2004-02-24 21:25:18 UTC
Confirm ... have the same problem on Radeon 9200 (R200) with pciid
1106:3059

Comment 5 John Ellson 2004-02-24 21:33:49 UTC
(I realize that the NVIDEA drivers taint the kernel, but this might be
related.)

The NVIDIA-Linux-x86-1.0-5336-pkg1 drivers compiled OK with 1.97,
but fail to compile with 1.100.   The error message was:
"Shared memory test failed (Invalid argument): please check that your
kernel has CONFIG_SYSVIPC enabled."
   

Comment 6 Steve Bergman 2004-02-25 00:15:17 UTC
I have a problem that looks to be related.

mplayer dies with the error:

X11 error: BadAccess (attempt to access private resource denied) 

From an strace:
2578  shmget(IPC_PRIVATE, 230400, IPC_CREAT|0777) = 1081356
2578  shmat(1081356, 0, 0)              = -1 EINVAL (Invalid argument)

My rawhide is coompletely up to date, except for openssl-devel which
has a dependency issue.

Comment 7 Sammy 2004-02-25 01:51:21 UTC
I am using XFree86 -59 with 2.6.3 arjanv kernel (compiled myself) 
with no problems (Radeon 7200). There were some new radeonfb upgrades on top 
of 2.6.3. I would think that this is more of a kernel problem. 

Comment 8 Bill Nottingham 2004-02-25 04:32:07 UTC
*** Bug 116756 has been marked as a duplicate of this bug. ***

Comment 9 Matthew Eastman 2004-02-25 07:06:21 UTC
I have the exact problem since I've upgraded kernels and I have a
Radeon M7-C; perhaps there is some kind of conflict between this
particular kernel and the Radeon drivers.

Comment 10 Warren Togami 2004-02-25 09:17:46 UTC
This seems to be fixed for me on 2.6.3-1.106.  Try it for yourself
when it hits rawhide.

Comment 11 Mike A. Harris 2004-02-25 11:05:28 UTC
*** Bug 116725 has been marked as a duplicate of this bug. ***

Comment 12 Mike A. Harris 2004-02-25 11:12:10 UTC
*** Bug 116756 has been marked as a duplicate of this bug. ***

Comment 13 Kaj J. Niemi 2004-02-25 12:47:23 UTC
Works with 2.6.3-1.106.

Comment 14 John Ellson 2004-02-25 14:41:57 UTC
NVIDIA 5336 drivers on 1.106 seem OK.

Comment 15 Dave Jones 2004-02-25 14:48:45 UTC
should be fixed now.


Comment 16 Don Hardaway 2004-02-25 15:37:43 UTC
How about drivers for the netgear WG511 wireless card that uses prism
duette or gt chip set.  What is it's status since I can only get my
old 802.11 card to work with core 2

Comment 17 Mike A. Harris 2004-02-27 07:42:15 UTC
Dave) Yep, I've gotten numerous confirmations that the new kernel
fixes the shmat bug.  Thanks.