Created attachment 481169 [details] traceback (photo of text console) Description of problem: kernel oops some time after XFCE screensaver starts Version-Release number of selected component (if applicable): xorg-x11-drv-ati-6.14.0-2.20110204gita27b5dbd9.fc15.i686 How reproducible: unknown Steps to Reproduce: 1. Install F15 alpha XFCE Desktop from Live CD 2. yum update 3. login and wait Actual results: kernel oops implicating ttm_bo_ref; will attach photo of console showing traceback Expected results: no bug/oops Additional info:01:00.0 VGA compatible controller [0300]: ATI Technologies Inc Radeon RV200 QW [Radeon 7500] [1002:5157]
Created attachment 481170 [details] Xorg.0.log after reboot
Created attachment 481171 [details] output from 'dmesg' after reboot
Created attachment 481172 [details] /var/log/messages after reboot
Created attachment 481267 [details] serial console log This is reproducible. XFCE Desktop > Applications > Preferences > Screensaver; Random screensaver (leave default selections checked), Blank after 2 minutes, Cycle after 1 minute; Advanced tab > Power saver off (want screensaver to loop forever, trying a random screensaver once per minute); File > Blank now. Wait; in this particular case, 7 minutes.
Recursive oops, 2 deep, the first one is: [ 346.124012] kernel BUG at drivers/gpu/drm/ttm/ttm_bo.c:272! [ 346.124012] invalid opcode: 0000 [#1] SMP [ 346.124012] last sysfs file: /sys/devices/pci0000:00/0000:00:01.0/0000:01:00.0/drm/card0/card0-VGA-1/status [ 346.124012] Modules linked in: sco bnep l2cap bluetooth rfkill asb100 hwmon_vid sunrpc 8021q garp stp llc p4_clockmod ip6t_REJECT nf_conntrack_ipv6 nf_defrag_ipv6 ip6table_filter ip6_tables snd_cmipci gameport snd_seq snd_pcm snd_page_alloc snd_opl3_lib snd_timer snd_hwdep snd_mpu401_uart snd_rawmidi microcode snd_seq_device iTCO_wdt e100 snd soundcore mii i2c_i801 iTCO_vendor_support uinput ipv6 uas usb_storage radeon ttm drm_kms_helper drm i2c_algo_bit i2c_core [last unloaded: mperf] [ 346.124012] [ 346.124012] Pid: 288, comm: kworker/0:3 Not tainted 2.6.38-0.rc6.git4.1.fc15.i686 #1 System Manufacturer System Name/P4B266 [ 346.124012] EIP: 0060:[<f78de2e7>] EFLAGS: 00010202 CPU: 0 [ 346.124012] EIP is at ttm_bo_ref_bug+0x8/0xa [ttm] [ 346.124012] EAX: f23b5254 EBX: 00000001 ECX: f78de2df EDX: 00000054 [ 346.124012] ESI: f78de2df EDI: f23b5254 EBP: f44afeec ESP: f44afeec [ 346.124012] DS: 007b ES: 007b FS: 00d8 GS: 00e0 SS: 0068 [ 346.124012] Process kworker/0:3 (pid: 288, ti=f44ae000 task=f66ebed0 task.ti=f44ae000) [ 346.124012] Stack: [ 346.124012] f44aff00 c05cb8b8 f78de2df f23b522c 00000000 f44aff0c f78dec94 f5167380 [ 346.124012] f44aff2c f78df0a3 00000000 00000000 0000b8fb f23b5200 f66c2440 f5403240 [ 346.124012] f44aff48 f79d7e45 00000000 00000000 00000000 f2323b00 f66c2440 f44aff5c [ 346.124012] Call Trace: [ 346.124012] [<c05cb8b8>] kref_sub+0x3c/0x46 [ 346.124012] [<f78de2df>] ? ttm_bo_ref_bug+0x0/0xa [ttm] [ 346.124012] [<f78dec94>] ttm_bo_list_ref_sub+0x22/0x25 [ttm] [ 346.124012] [<f78df0a3>] ttm_bo_reserve+0x63/0x6d [ttm]
This is reproducible on a recent card, too: 01:00.0 VGA compatible controller [0300]: ATI Technologies Inc Redwood PRO [Radeon HD 5500 Series] [1002:68da] [ 19.239] (--) PCI:*(0:1:0:0) 1002:68da:1682:3083 rev 0, Mem @ 0xd0000000/268435456, 0xfdfc0000/131072, I/O @ 0x0000bc00/256, BIOS @ 0x????????/131072 [ 19.274] (--) RADEON(0): Chipset: "ATI Radeon HD 5500 Series" (ChipID = 0x68da) The key is to boot with only 1GB of memory: add " mem=1023m" to the kernel boot command line [note all lower-case.] The kernel was 2.6.38-0.rc6.git4.1.fc15.i686 [and not -PAE], the CPU is processor : 0 vendor_id : AuthenticAMD cpu family : 15 model : 107 model name : AMD Athlon(tm) 64 X2 Dual Core Processor 4800+ stepping : 2 cpu MHz : 2499.649 cache size : 512 KB It took about 12 minutes to crash.
Created attachment 490973 [details] console photo of BUG with traceback and info Another card, another crash, this time on [RV280]: [ 52.794] (--) RADEON(0): Chipset: "ATI Radeon 9250 5960 (AGP)" (ChipID = 0x5960) with F15 beta RC2 LiveCD: Linux version 2.6.38.2-9.fc15.i686 (mockbuild.fedoraproject.org) (gcc version 4.6.0 20110329 (Red Hat 4.6.0-1) (GCC) ) #1 SMP Wed Mar 30 16:54:01 UTC 2011 and this Xserver: X.Org X Server 1.10.0 Release Date: 2011-2-25 [ 52.282] Build Date: 30 March 2011 08:08:44PM [ 52.282] Build ID: xorg-x11-server 1.10.0-7.fc15 [ 52.282] Current version of pixman: 0.20.2 and this visible symptom: BUG: Unable to handle kernel paging request at fffffffc
Added [Redwood][RV280] tags adjacent to [RV200] in Summary, reflecting Comment #6 and Comment #7. Such a wide range of cards suggests that the bug is in [ttm] independent of which card, particularly because one common feature is low kernel RAM (1023m of Comment #6.)
I've built a scratch kernel at it will end up at: http://kojipkgs.fedoraproject.org/scratch/airlied/task_3098801/ it should be finished in a few hours, though it may disappear in a few days.
Good news: kernel-2.6.38.7-30.bz680651.fc15.i686 survived 110 minutes [normal termination and no OOPS] running xscreensaver under XFCE on a 1GB machine with RV280 (PCI 1002:5960:174b:0130). Also: xorg-x11-drv-ati-6.14.1-1.20110504gita6d2dba6.fc15.i686 mesa-dri-drivers-7.11-0.11.20110525.0.fc15.i686 xorg-x11-server-Xorg-1.10.1-14.fc15.i686 The previous non-patched kernel-2.6.38.6-27.fc15.i686 was getting an OOPS [Unable to handle kernel paging request at fffffffc] after about half an hour.
Fixed in 2.6.38.8-31
*** Bug 716495 has been marked as a duplicate of this bug. ***