Bug 58621

Summary:	radeon 7770 kernel crash when changing powwr profile
Product:	Drivers	Reporter:	rafael castillo (jrch2k10)
Component:	Video(DRI - non Intel)	Assignee:	drivers_video-dri
Status:	NEW ---
Severity:	blocking	CC:	alan, alexdeucher, szg00000
Priority:	P1
Hardware:	All
OS:	Linux
Kernel Version:	3.10-rc3	Subsystem:
Regression:	No	Bisected commit-id:
Attachments:	capture of the crash from my nexus dmidecode lscpi -vvvv dmesg output of the crash fix

Description rafael castillo 2013-05-22 04:52:57 UTC

My radeon 7770 from msi crashes the kernel at any try to change the freq of the gpu been medium or high profile(is always in low) and dynpm never kicks in whatever you do, it always stay at low.

Attaching image of the crash if you need anything else please include some instructions. Have years without debuging a kernel

Comment 1 rafael castillo 2013-05-22 04:55:20 UTC

Created attachment 102211 [details]
capture of the crash from my nexus

Crash capture

Comment 2 rafael castillo 2013-05-30 01:55:00 UTC

still reproducible in 3.10-rc3, attached my dmesg/dmidecode/lspci when the crash happened

Comment 3 rafael castillo 2013-05-30 01:55:37 UTC

Created attachment 102901 [details]
dmidecode

Comment 4 rafael castillo 2013-05-30 01:56:07 UTC

Created attachment 102911 [details]
lscpi -vvvv

Comment 5 rafael castillo 2013-05-30 01:56:53 UTC

Created attachment 102921 [details]
dmesg output of the crash

Comment 6 Alan 2013-11-13 20:19:06 UTC

[  257.496852] radeon 0000:01:00.0: failed to get a new IB (-35)
[  257.498006] BUG: unable to handle kernel NULL pointer dereference at           (null)
[  257.499132] IP: [<ffffffffa00c1e36>] si_vm_set_page+0x2c6/0x420 [radeon]
[  257.500250] PGD 4233d6067 PUD 42332e067 PMD 0 
[  257.501340] Oops: 0002 [#1] PREEMPT SMP 
[  257.502418] Modules linked in: ebtable_nat ebtables ipt_MASQUERADE iptable_nat nf_nat_ipv4 nf_nat xt_CHECKSUM bridge stp llc btusb bluetooth hid_a4tech mxm_wmi kvm_amd rtl8180 eeprom_93cx6 kvm crc32_pclmul crc32c_intel ghash_clmulni_intel mac80211 cdc_acm aesni_intel aes_x86_64 snd_emu10k1 lrw cfg80211 gf128mul glue_helper snd_util_mem ablk_helper snd_ac97_codec cryptd rfkill snd_rawmidi fam15h_power pcspkr edac_core ac97_bus sp5100_tco k10temp r8169 i2c_piix4 radeon ttm xhci_hcd drm_kms_helper wmi
[  257.508026] CPU: 4 PID: 2369 Comm: X Tainted: G        W    3.10.0-rc3 #1
[  257.509333] Hardware name: MSI MS-7693/970A-G46 (MS-7693), BIOS V1.11 10/31/2012
[  257.510533] task: ffff880426ff8000 ti: ffff880425896000 task.ti: ffff880425896000
[  257.511640] RIP: 0010:[<ffffffffa00c1e36>]  [<ffffffffa00c1e36>] si_vm_set_page+0x2c6/0x420 [radeon]
[  257.512749] RSP: 0018:ffff880425897aa8  EFLAGS: 00010202
[  257.513827] RAX: 0000000000000061 RBX: ffff880425897b48 RCX: 0000000000000000
[  257.514992] RDX: 0000000000000840 RSI: 000000000dfb8000 RDI: 0000000000000000
[  257.516074] RBP: ffff8804283ce000 R08: 000000000dfb8000 R09: 00000000000ffffe
[  257.517168] R10: 0000000000000000 R11: 0000000024200840 R12: 0000000000000420
[  257.518338] R13: 0000000000387b00 R14: 0000000000001000 R15: 0000000000000001
[  257.519536] FS:  00007eff0ee7e880(0000) GS:ffff88043ed00000(0000) knlGS:0000000000000000
[  257.520657] CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[  257.521691] CR2: 0000000000000000 CR3: 00000004211a3000 CR4: 00000000000407e0
[  257.522813] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[  257.523891] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
[  257.524966] Stack:
[  257.526086]  00000000ffffffdd ffff8804283ce000 000000000dfb8000 0000000000000614
[  257.527308]  ffff8804283ce000 0000000000180000 ffff880421032380 0000000000001f80
[  257.528422]  ffff8804283ce000 000000000e3d8000 0000000000387b00 ffffffffa0067d4d
[  257.529451] Call Trace:
[  257.530519]  [<ffffffffa0067d4d>] ? radeon_vm_bo_update_pte+0x39d/0x560 [radeon]
[  257.531604]  [<ffffffffa00c1915>] ? si_ib_parse+0x3c5/0x5f0 [radeon]
[  257.532649]  [<ffffffffa007a8cc>] ? radeon_cs_ioctl+0x80c/0x8d0 [radeon]
[  257.533765]  [<ffffffff8140c44d>] ? drm_ioctl+0x49d/0x5a0
[  257.534957]  [<ffffffff810c6045>] ? rcu_eqs_enter_common.isra.43+0x1e5/0x220
[  257.536061]  [<ffffffff812d8004>] ? avc_has_perm_flags+0x74/0x150
[  257.537141]  [<ffffffff8117696c>] ? do_vfs_ioctl+0x2ec/0x4d0
[  257.538192]  [<ffffffff812defce>] ? file_has_perm+0x8e/0xa0
[  257.539171]  [<ffffffff81176bd8>] ? SyS_ioctl+0x88/0xa0
[  257.540211]  [<ffffffff8179b0cf>] ? tracesys+0xdd/0xe2
[  257.541202] Code: fe ff ff 66 90 48 89 f1 41 89 f0 48 c1 e9 20 48 89 cf 8b 4b 08 4c 8b 53 18 44 8d 59 01 44 89 5b 08 41 89 d3 41 81 cb 00 00 20 24 <45> 89 1c 8a 8b 4b 08 4c 8b 53 18 44 8d 59 01 44 89 5b 08 45 89 
[  257.543604] RIP  [<ffffffffa00c1e36>] si_vm_set_page+0x2c6/0x420 [radeon]
[  257.544841]  RSP <ffff880425897aa8>

Comment 7 Alex Deucher 2013-11-13 20:30:37 UTC

Created attachment 114571 [details]
fix

This patch should avoid the null pointer dereference.

Comment 8 rafael castillo 2013-11-13 23:34:43 UTC

well dpm fixed it for me entirely, if you wanna close the bug once alan tested it, is fine

thanks for your time