Cause most software related desktop freezes? mc_heap_unmap
Having used 3 different versions of Catalyst including Catalyst 13.4 I keep periodically getting freezes of the desktop when either running OpenGL Apps or playing Video or both. When I say freeze I mean that the entire desktop including mouse pointer stops! However you can ssh into the box and do whatever including rebooting to restore operation.
It sounds like others in the community are also seeing this with no explanation. In my case every freeze I have seen (about 30 over the past year on 3 different computers) has been related to what appears to be a bug in the fglrx heap manager. Every time I get the identical information to what I post below in the /var/log/messages file.
I googled this and others have reporting seeing it too. I however do not see any recent references on this forum. I want to report it to AMD but it appears they only want reproducible problems and I have no consistent way to reproduce it YET the problem certainly occurs though usually at the worst times.
It seems like apps that use lots of of large textures trigger it more readily as does multiple OpenGL using apps like OpenGL Desktop effects, an OpenGL game, and OpenGL video running all at the same time. However the problem has occurred in simple scenarios also. Sometimes I will go for weeks without seeing it and other times it will bear its ugly head 3 times in a week all under similar usage.
My question to the community is are others seeing this? Does anyone have a scenario that frequently and consistently causes this? as this could be documented as the way to reproduce.
For anyone seeing this post who has periodic desktop freezes please check your /var/log/messages file or archived messages files for a similar trap report to what I have posted below.
I feel that most random non hardware related desktop freezes (at least with Radeon 7000 hardware) are related to this one bug and it would benefit all us users and AMD to get this Gremlin tracked down and fixed.
The hardware of all 3 of my systems is AMD Dual Opteron 43XX or FX8350 CPU's, 32GB of ECC memory, and either Radeon 7750 or 7770 Graphics boards. The systems are other than this issue absolutely stable.
The Catalyst I currently have installed is under OpenSuse 12.3 and from http://geeko.ioda.net/mirror/amd-fgl...4-3.x86_64.rpm download which is Catalyst 13.4 put into an easy to install RPM. I have however seen this with prior Catalysts and ones installed from the AMD downloads.
Here is the dump I get in /var/log/messages:
[ 418.614549] KBD BUG in ../../../../../../../../drivers/2d/lnx/fgl/drm/kernel/mc_heap_manager.c at line: 596!
[ 418.614568] int3: 0000 [#1] PREEMPT SMP
[ 418.614571] Modules linked in: bnep bluetooth fuse nfsv3 nfs_acl nfsv4 auth_rpcgss nfs fscache lockd sunrpc snd_hda_codec_hdmi mperf kvm_amd kvm crc32c_intel ghash_clmulni_intel snd_hda_codec_realtek aesni_intel ablk_helper cryptd lrw aes_x86_64 xts gf128mul microcode snd_hda_intel snd_hda_codec snd_hwdep snd_pcm k10temp usb_storage pcspkr fam15h_power amd64_edac_mod edac_core edac_mce_amd r8169 snd_seq sp5100_tco i2c_piix4 eeepc_wmi asus_wmi sparse_keymap rfkill video pci_hotplug sg sr_mod cdrom fglrx(PO) snd_timer snd_seq_device mxm_wmi snd amd_iommu_v2 soundcore snd_page_alloc wmi button raid1 edd autofs4 xhci_hcd processor thermal_sys scsi_dh_alua scsi_dh_emc scsi_dh_rdac scsi_dh_hp_sw scsi_dh ata_generic pata_atiixp
[ 418.614597] CPU 4
[ 418.614599] Pid: 2102, comm: rvt Tainted: P O 3.7.10-1.1-desktop #1 To be filled by O.E.M. To be filled by O.E.M./SABERTOOTH 990FX R2.0
[ 418.614601] RIP: 0010:[<ffffffffa011a35b>] [<ffffffffa011a35b>] __mc_heap_unmap_from_gart_space+0x22b/0x250 [fglrx]
[ 418.614656] RSP: 0018:ffff8807c5763ac8 EFLAGS: 00000296
[ 418.614657] RAX: 0000000000000011 RBX: ffff8806af1b04c0 RCX: 00000000000031a2
[ 418.614658] RDX: 0000000000000010 RSI: 0000000000000046 RDI: 0000000000000286
[ 418.614659] RBP: ffff8807e5548630 R08: 000000000000000a R09: 00000000000004d5
[ 418.614660] R10: 0000000000000000 R11: 00000000000004d4 R12: ffff8807e5548660
[ 418.614661] R13: 0000000000000000 R14: 0000000000000000 R15: ffff8808150ec000
[ 418.614662] FS: 00007f88f595b740(0000) GS:ffff88083ed00000(0000) knlGS:0000000000000000
[ 418.614663] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 418.614664] CR2: 00007f88f0d84a08 CR3: 00000007efd98000 CR4: 00000000000407e0
[ 418.614665] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[ 418.614666] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
[ 418.614667] Process rvt (pid: 2102, threadinfo ffff8807c5762000, task ffff8807cfc54040)
[ 418.614668] Stack:
[ 418.614669] ffff8806af1b04c0 0000000003800000 000000ff90100000 0000000001800000
[ 418.614672] ffff880700000003 0000000100000000 0000000000000000 ffffffffa0123e62
[ 418.614674] 00000466cfc54040 ffffffffa02be3f9 ffff8807fe85b1c0 00000000000132c0
[ 418.614676] Call Trace:
[ 418.614737] [<ffffffffa0114fb6>] mc_heap_unmap_from_gart_space+0x26/0x70 [fglrx]
[ 418.614787] [<ffffffffa010c83b>] MCIL_UnmapFromGartSpace+0x4b/0x90 [fglrx]
[ 418.614850] [<ffffffffa0184300>] _ZN2OS9gart_FreeEP7CMMPool+0xc0/0x100 [fglrx]
[ 418.614963] [<ffffffffa018d76f>] _ZN12CMMHeap_GART16dropPoolFromHeapEP7CMMPool+0x2f/0x70 [fglrx]
[ 418.615083] [<ffffffffa018d8c7>] _ZN12CMMHeap_GART10shrinkHeapEb+0xa7/0xb0 [fglrx]
[ 418.615200] [<ffffffffa017e6f6>] _ZN3MSF19free_surf_resourcesEP10CMMSurfaceP9CMMCli entP9CMMDriver+0x1a6/0x240 [fglrx]
[ 418.615304] [<ffffffffa017dc41>] _ZN3MSF9free_surfEP9CMMDriverP10CMMSurface+0x81/0x100 [fglrx]
[ 418.615407] [<ffffffffa01742c8>] CMMFreeSurface+0x1c8/0x280 [fglrx]
[ 418.615690] [<ffffffffa01946ae>] _Z8uCWDDEQCmjjPvjS_+0xd2e/0x1240 [fglrx]
[ 418.615801] [<ffffffffa012949f>] firegl_cmmqs_CWDDE_32+0x36f/0x480 [fglrx]
[ 418.615858] [<ffffffffa0127d3e>] firegl_cmmqs_CWDDE32+0x8e/0x140 [fglrx]
[ 418.615912] [<ffffffffa00fb34d>] firegl_ioctl+0x1ed/0x250 [fglrx]
[ 418.615949] [<ffffffffa00eb9b9>] ip_firegl_unlocked_ioctl+0x9/0x10 [fglrx]
[ 418.615954] [<ffffffff8118051f>] do_vfs_ioctl+0x8f/0x530
[ 418.615958] [<ffffffff81180a60>] sys_ioctl+0xa0/0xc0
[ 418.615962] [<ffffffff8159ebad>] system_call_fastpath+0x1a/0x1f
[ 418.615967] [<00007f88f4258f27>] 0x7f88f4258f26
[ 418.615967] Code: ff c8 7f 0c c7 43 18 00 00 00 00 e9 6d fe ff ff ba 54 02 00 00 48 c7 c6 e0 99 2c a0 48 c7 c7 9e d1 2b a0 31 c0 e8 96 79 fd ff cc <eb> d7 ba 51 02 00 00 48 c7 c6 e0 99 2c a0 48 c7 c7 9e d1 2b a0
[ 418.615983] RIP [<ffffffffa011a35b>] __mc_heap_unmap_from_gart_space+0x22b/0x250 [fglrx]
[ 418.616009] RSP <ffff8807c5763ac8>
[ 418.616024] ---[ end trace 129a58055eda177f ]---
Tags for this Thread