On 05/18/2015 01:38 PM, Haggai Eran wrote: > On 18 May 2015 at 18:31, Larry Finger wrote: >> On 05/17/2015 02:22 PM, Haggai Eran wrote: >>> >>> I added some debugging prints, trying to see more details about the >>> packet that fails the r8712_validate_recv_frame. I noticed I'm getting >>> many packets where recv_decache returns _FAIL. However, the last two >>> packets before the crash fail for different reasons. The first has the >>> ver field set to 3 (instead of zero). The second (the one that get's >>> freed and cause the crash apparently) has an unknown type (12). If I'm >>> not mistaken, 12 = WIFI_CTRL_TYPE | WIFI_DATA_TYPE. Is that possible? >>> >>> It could be that the packet headers are garbled though. >> >> >> I think the headers are garbled. Did you log the address of the skb at >> precvframe->u.hdr.pkt in r8712_free_recvframe() or orig_prframe->u.hdr.pct >> in recv_func(). > > I added prints of the skb pointer in every call to recv_func. Here are > the results: > > ... > [ 674.111771] recv_func: pcontext = 96335820, prframe->u.hdr.pkt = 9729fb40 > [ 674.118782] recv_func: pcontext = 963359b8, prframe->u.hdr.pkt = 9729f6c0 > [ 674.125777] recv_func: pcontext = 96335930, prframe->u.hdr.pkt = 9729f780 > [ 674.132769] recv_func: pcontext = 963358a8, prframe->u.hdr.pkt = 973d56c0 > [ 674.139753] recv_func: pcontext = 96335d70, prframe->u.hdr.pkt = 973d5000 > [ 674.146922] recv_func: pcontext = 963361b0, prframe->u.hdr.pkt = 973d5000 > [ 674.153961] recv_func: pcontext = 963360a0, prframe->u.hdr.pkt = 973d5000 > [ 674.161023] recv_func: pcontext = 96336128, prframe->u.hdr.pkt = 973d5000 > [ 674.168186] recv_func: pcontext = 96336018, prframe->u.hdr.pkt = 973d5000 > [ 674.175231] recv_func: pcontext = 96335f90, prframe->u.hdr.pkt = 973d5000 > [ 674.182141] r8712_validate_recv_frame: ver = 1 > [ 674.186814] recv_func: pcontext = 96335f08, prframe->u.hdr.pkt = 973d5000 > [ 674.193811] r8712_validate_recv_frame: ver = 1 > [ 674.198530] recv_func: pcontext = 963363d0, prframe->u.hdr.pkt = 973d5000 > [ 674.205434] r8712_validate_recv_frame: ver = 3 > [ 674.210018] Unable to handle kernel NULL pointer dereference at > virtual address 00000001 > [ 674.218209] pgd = 80004000 > [ 674.221025] [00000001] *pgd=00000000 > [ 674.224752] Internal error: Oops: 5 [#1] ARM > [ 674.229028] Modules linked in: rfcomm cfg80211 r8712u(C) btusb > bluetooth bcm2708_rng > [ 674.236857] CPU: 0 PID: 530 Comm: kworker/0:1 Tainted: G WC > 4.0.3 #1 > [ 674.244247] Hardware name: BCM2708 > [ 674.247663] task: 962cdee0 ti: 960fc000 task.ti: 960fc000 > [ 674.253082] PC is at put_page+0xc/0x68 > [ 674.256853] LR is at skb_release_data+0x6c/0xcc > [ 674.261388] pc : [<800933e0>] lr : [<80433fe4>] psr: 20000113 > [ 674.261388] sp : 960fdc18 ip : 960fdc28 fp : 960fdc24 > [ 674.272856] r10: 84d6cc00 r9 : 0000fff8 r8 : 00002f17 > [ 674.278079] r7 : 973d5000 r6 : 84dc7620 r5 : 84dc7620 r4 : 00000000 > [ 674.284602] r3 : 00000037 r2 : 00000000 r1 : 00000001 r0 : 00000001 > [ 674.291127] Flags: nzCv IRQs on FIQs on Mode SVC_32 ISA ARM > Segment kernel > [ 674.298432] Control: 00c5387d Table: 168a4008 DAC: 00000015 > [ 674.304179] Process kworker/0:1 (pid: 530, stack limit = 0x960fc188) > [ 674.310532] Stack: (0x960fdc18 to 0x960fe000) > [ 674.314893] dc00: > 960fdc44 960fdc28 > [ 674.323076] dc20: 80433fe4 800933e0 973d5000 96309010 96308520 > 96309010 960fdc5c 960fdc48 > [ 674.331257] dc40: 8043406c 80433f84 00000001 973d5000 960fdc74 > 960fdc60 8043413c 80434050 > [ 674.339437] dc60: 40000113 963363d0 960fdc84 960fdc78 80441c08 > 80434120 960fdcac 960fdc88 > [ 674.347618] dc80: 7f10ca70 80441bd0 00000000 96308520 963363d0 > 00000000 96309010 00002f17 > [ 674.355798] dca0: 960fdce4 960fdcb0 7f10d3f8 7f10ca50 960fdcd4 > 960fdcc0 80439aac 96308520 > [ 674.363978] dcc0: 963363d0 9630a520 00002f80 00002f17 0000fff8 > 84d6cc00 960fdd04 960fdce8 > [ 674.372157] dce0: 7f10eb84 7f10d36c 000000d2 963363d0 84dc7626 > 00000018 960fdd54 960fdd08 > [ 674.380338] dd00: 7f10c65c 7f10eb5c 96308ff0 963090d4 9729f840 > 9630b520 96309010 973d5000 > [ 674.388518] dd20: ffff2f00 00000002 808f4590 96309094 808f458c > 8093f820 00000000 96a32900 > [ 674.396698] dd40: 8093f840 40000000 960fdd7c 960fdd58 8001fbbc > 7f10c4b8 0000833e 00000000 > [ 674.404879] dd60: 00000000 00000102 960fc000 8093f840 960fddcc > 960fdd80 8001ffc0 8001fb48 > [ 674.413058] dd80: 8054e348 80052428 00000001 00000001 04208060 > 0001b61a 00000009 960fc000 > [ 674.421237] dda0: 00000000 00000000 80920c94 00000000 00000000 > 00000000 8003555c 00000000 > [ 674.429416] ddc0: 960fdde4 960fddd0 80020474 8001fea4 00000000 > 00000000 960fde0c 960fdde8 > [ 674.437598] dde0: 80057298 800203bc 960fde20 8054e760 60000013 > f200b200 960fde54 972ba1e0 > [ 674.445777] de00: 960fde1c 960fde10 800081e4 80057224 960fde7c > 960fde20 800127f8 800081cc > [ 674.453957] de20: 8054e75c 00000001 962cdee0 00000000 808f6668 > 969021e0 97051140 00000000 > [ 674.462140] de40: 972ba1e0 8003555c 00000000 960fde7c 960fde58 > 960fde68 8004b37c 8054e760 > [ 674.470319] de60: 60000013 ffffffff 00000000 808f6668 960fdeac > 960fde80 8003e738 8054e73c > [ 674.478499] de80: 00000001 00000000 8003e6cc 960fde98 97239140 > 962cdee0 808f6668 972ba1e0 > [ 674.486679] dea0: 960fded4 960fdeb0 80549748 8003e6d8 960fded8 > 960fc000 808f5834 808f5834 > [ 674.494860] dec0: 808f5864 00000008 960fdeec 960fded8 80549ad8 > 80549558 962cdee0 971e55a0 > [ 674.503039] dee0: 960fdf24 960fdef0 80035590 80549aa0 972c4940 > 971e55a0 800354cc 00000000 > [ 674.511220] df00: 972c4940 971e55a0 800354cc 00000000 00000000 > 00000000 960fdfac 960fdf28 > [ 674.519401] df20: 8003a080 800354d8 00000000 00000000 960fdf4c > 971e55a0 00000000 00000001 > [ 674.527581] df40: dead4ead ffffffff ffffffff 8093fd70 00000000 > 00000000 80646d2c 960fdf5c > [ 674.535759] df60: 960fdf5c 00000000 00000001 dead4ead ffffffff > ffffffff 8093fd70 00000000 > [ 674.543939] df80: 00000000 80646d2c 960fdf88 960fdf88 972c4940 > 80039fa0 00000000 00000000 > [ 674.552118] dfa0: 00000000 960fdfb0 8000e8f0 80039fac 00000000 > 00000000 00000000 00000000 > [ 674.560296] dfc0: 00000000 00000000 00000000 00000000 00000000 > 00000000 00000000 00000000 > [ 674.568474] dfe0: 00000000 00000000 00000000 00000000 00000013 > 00000000 00000000 00000000 > [ 674.576643] Backtrace: > [ 674.579129] [<800933d4>] (put_page) from [<80433fe4>] > (skb_release_data+0x6c/0xcc) > [ 674.586711] [<80433f78>] (skb_release_data) from [<8043406c>] > (skb_release_all+0x28/0x2c) > [ 674.594881] r7:96309010 r6:96308520 r5:96309010 r4:973d5000 > [ 674.600599] [<80434044>] (skb_release_all) from [<8043413c>] > (consume_skb+0x28/0x5c) > [ 674.608338] r4:973d5000 r3:00000001 > [ 674.611961] [<80434114>] (consume_skb) from [<80441c08>] > (__dev_kfree_skb_any+0x44/0x48) > [ 674.620045] r4:963363d0 r3:40000113 > [ 674.623891] [<80441bc4>] (__dev_kfree_skb_any) from [<7f10ca70>] > (r8712_free_recvframe+0x2c/0x94 [r8712u]) > [ 674.633827] [<7f10ca44>] (r8712_free_recvframe [r8712u]) from > [<7f10d3f8>] (recv_func+0x98/0x6f0 [r8712u]) > [ 674.643477] r8:00002f17 r7:96309010 r6:00000000 r5:963363d0 > r4:96308520 r3:00000000 > [ 674.651563] [<7f10d360>] (recv_func [r8712u]) from [<7f10eb84>] > (r8712_recv_entry+0x34/0x78 [r8712u]) > [ 674.660780] r10:84d6cc00 r9:0000fff8 r8:00002f17 r7:00002f80 > r6:9630a520 r5:963363d0 > [ 674.668666] r4:96308520 > [ 674.671495] [<7f10eb50>] (r8712_recv_entry [r8712u]) from > [<7f10c65c>] (recv_tasklet+0x1b0/0x324 [r8712u]) > [ 674.681145] r6:00000018 r5:84dc7626 r4:963363d0 r3:000000d2 > [ 674.687002] [<7f10c4ac>] (recv_tasklet [r8712u]) from [<8001fbbc>] > (tasklet_hi_action+0x80/0xdc) > [ 674.695785] r10:40000000 r9:8093f840 r8:96a32900 r7:00000000 > r6:8093f820 r5:808f458c > [ 674.703670] r4:96309094 > [ 674.706228] [<8001fb3c>] (tasklet_hi_action) from [<8001ffc0>] > (__do_softirq+0x128/0x290) > [ 674.714399] r8:8093f840 r7:960fc000 r6:00000102 r5:00000000 > r4:00000000 r3:0000833e > [ 674.722205] [<8001fe98>] (__do_softirq) from [<80020474>] > (irq_exit+0xc4/0x118) > [ 674.729509] r10:00000000 r9:8003555c r8:00000000 r7:00000000 > r6:00000000 r5:80920c94 > [ 674.737392] r4:00000000 > [ 674.739968] [<800203b0>] (irq_exit) from [<80057298>] > (__handle_domain_irq+0x80/0xe0) > [ 674.747793] r4:00000000 r3:00000000 > [ 674.751406] [<80057218>] (__handle_domain_irq) from [<800081e4>] > (asm_do_IRQ+0x24/0x28) > [ 674.759404] r8:972ba1e0 r7:960fde54 r6:f200b200 r5:60000013 > r4:8054e760 r3:960fde20 > [ 674.767228] [<800081c0>] (asm_do_IRQ) from [<800127f8>] (__irq_svc+0x38/0xb0) > [ 674.774361] Exception stack(0x960fde20 to 0x960fde68) > [ 674.779421] de20: 8054e75c 00000001 962cdee0 00000000 808f6668 > 969021e0 97051140 00000000 > [ 674.787600] de40: 972ba1e0 8003555c 00000000 960fde7c 960fde58 > 960fde68 8004b37c 8054e760 > [ 674.795771] de60: 60000013 ffffffff > [ 674.799294] [<8054e730>] (_raw_spin_unlock_irq) from [<8003e738>] > (finish_task_switch+0x6c/0x108) > [ 674.808157] r4:808f6668 r3:00000000 > [ 674.811772] [<8003e6cc>] (finish_task_switch) from [<80549748>] > (__schedule+0x1fc/0x548) > [ 674.819856] r7:972ba1e0 r6:808f6668 r5:962cdee0 r4:97239140 > [ 674.825572] [<8054954c>] (__schedule) from [<80549ad8>] (schedule+0x44/0x9c) > [ 674.832615] r8:00000008 r7:808f5864 r6:808f5834 r5:808f5834 > r4:960fc000 r3:960fded8 > [ 674.840438] [<80549a94>] (schedule) from [<80035590>] > (worker_thread+0xc4/0x4d0) > [ 674.847829] r4:971e55a0 r3:962cdee0 > [ 674.851442] [<800354cc>] (worker_thread) from [<8003a080>] > (kthread+0xe0/0x100) > [ 674.858747] r10:00000000 r9:00000000 r8:00000000 r7:800354cc > r6:971e55a0 r5:972c4940 > [ 674.866630] r4:00000000 > [ 674.869192] [<80039fa0>] (kthread) from [<8000e8f0>] > (ret_from_fork+0x14/0x24) > [ 674.876409] r7:00000000 r6:00000000 r5:80039fa0 r4:972c4940 > [ 674.882121] Code: 8009272c e1a0c00d e92dd800 e24cb004 (e5902000) > [ 674.888596] ---[ end trace 8b18691702087335 ]--- > [ 674.893371] Kernel panic - not syncing: Fatal exception in interrupt > [ 674.899744] ---[ end Kernel panic - not syncing: Fatal exception in interrupt > > The offsets are a little different, I guess because of the added > prints, and debugging features I enabled in the kernel. One thing I > notice is that the skb at 0x973d5000 gets reused a couple of times > before the crash. Also, this time the pointer being dereferenced is > NULL (0x1). OK, I will have to search further upstream to see how a faulty skb was provided. I have been testing r8712u on my x86_64 system with no difficulty. I checked the driver with Smatch and found a couple of array problems. These likely won't be the problem, but try the attached patches anyway. Larry