media: smipcie: panic on DVB feed restart
Hi,
Restarting VDR 2.8.2 with two DVBSky S952 V3 cards (1ade:3038, 4254:0552) caused this panic on Debian 6.12.107-1, x86-64:
#PF: supervisor read access in kernel mode
RIP: dvb_dmx_swfilter_packets+0x5d/0x90 [dvb_core] Call Trace: smi_dma_xfer+0x154/0x210 [smipcie] process_one_work+0x177/0x330 bh_worker+0x17b/0x1a0 Kernel panic - not syncing: Fatal exception in interrupt
The kernel was not tainted; smipcie was unmodified. VDR has local userspace changes. A redacted panic trace is attached. The diagnostic restart script is
available on request; there is no standalone mainline reproducer.
In drivers/media/pci/smipcie/smipcie-main.c, smi_start_feed() queues work with the previous _int_status. This may process a stale completion on restart; the exact IRQ sequence was not captured. smi_dma_xfer() also accepts lengths exceeding its 192,512-byte buffer. A zero length register is interpreted as 4 MiB and passed to the demux unchecked.
The change below clears the saved status, enables work before IRQ/DMA, and rejects oversized lengths. The same logic, built for the Debian kernel, passed ten service restarts and a reboot. A stubbed function test over all 22-bit lengths passed without oversized demux calls; it does not test IRQ concurrency.
The diff is against mainline fd179f8a05be3ccae366b9b96e176b51fbe54aab, where both paths remain. This mainline adaptation has not been built or hardware-tested.
(quoted diff) +0 -0
b/drivers/media/pci/smipcie/smipcie-main.c
drivers/media/pci/smipcie/smipcie-main.c +0 -0
@@ -310,8 +310,15 @@ "DMA CH0 engine complete length mismatched,
finish data=%d !\n",
finishedData);
}
- dvb_dmx_swfilter_packets(&port->demux,
- port->cpu_addr[0], (finishedData / 188));
+ /* Reject lengths exceeding the DMA buffer. */
+ if (finishedData > SMI_TS_DMA_BUF_SIZE)
+ dev_warn_ratelimited(&dev->pci_dev->dev,
+ "DMA CH0 invalid length %u,
dropping completion\n",
+ finishedData);
+ else
+ dvb_dmx_swfilter_packets(&port->demux,
+ port->cpu_addr[0],
+ finishedData / 188);
/*dvb_dmx_swfilter(&port->demux,
port->cpu_addr[0], finishedData);*/
}
@@ -333,8 +340,15 @@
"DMA CH1 engine complete length mismatched,
finish data=%d !\n",
finishedData);
}
- dvb_dmx_swfilter_packets(&port->demux,
- port->cpu_addr[1], (finishedData / 188));
+ /* Reject lengths exceeding the DMA buffer. */
+ if (finishedData > SMI_TS_DMA_BUF_SIZE)
+ dev_warn_ratelimited(&dev->pci_dev->dev,
+ "DMA CH1 invalid length %u,
dropping completion\n",
+ finishedData);
+ else
+ dvb_dmx_swfilter_packets(&port->demux,
+ port->cpu_addr[1],
+ finishedData / 188);
/*dvb_dmx_swfilter(&port->demux,
port->cpu_addr[1], finishedData);*/
}
@@ -821,9 +835,11 @@
if (port->users++ == 0) {
dmaManagement = smi_config_DMA(port); smi_port_clearInterrupt(port);
+ /* Clear stale status; let the IRQ queue work. */
+ port->_int_status = 0;
+ enable_work(&port->bh_work);
smi_port_enableInterrupt(port); smi_write(port->DMA_MANAGEMENT, dmaManagement);
- enable_and_queue_work(system_bh_wq, &port->bh_work);
}
return port->users;
}
Redacted netconsole excerpt. Host model/BIOS, UID/PID and the unrelated module inventory are omitted. Fault registers and trace are preserved; the original capture is retained locally.
[ 97.726900] VDR_DIAGNOSTIC_SERVICE_RESTART_BEGIN [ 101.991897] BUG: unable to handle page fault for address: ffffcbce408e4000 [ 101.991940] #PF: supervisor read access in kernel mode [ 101.991955] #PF: error_code(0x0000) - not-present page [ 101.991968] PGD 100000067 P4D 100000067 PUD 1001f5067 PMD 107ec0067 PTE 0 [ 101.991993] Oops: Oops: 0000 [#1] PREEMPT SMP PTI [ 101.992010] CPU: 3 UID: [omitted] PID: [omitted] Comm: vdr Not tainted 6.12.107+deb13-amd64 #1 Debian 6.12.107-1 [host model and BIOS omitted] [ 101.992046] RIP: 0010:dvb_dmx_swfilter_packets+0x5d/0x90 [dvb_core] [ 101.992089] Code: 49 8d 6d ff eb 24 66 66 2e 0f 1f 84 00 00 00 00 00 66 66 2e 0f 1f 84 00 00 00 00 00 90 48 81 c3 bc 00 00 00 48 83 ed 01 72 1d <80> 3b 47 75 ee 48 89 de 4c 89 e7 48 81 c3 bc 00 00 00 e8 7c f9 ff [ 101.992110] RSP: 0018:ffffcbce401a0ea8 EFLAGS: 00010002 [ 101.992125] RAX: ffff897c034505c8 RBX: ffffcbce408e4000 RCX: 0000000000000000 [ 101.992139] RDX: ffff897c034505c8 RSI: ffffcbce408e3f44 RDI: ffff897c03450418 [ 101.992152] RBP: 0000000000005325 R08: 0000000000000000 R09: 0000000000000000 [ 101.992165] R10: 0000000000000001 R11: 0000000000000000 R12: ffff897c03450418 [ 101.992178] R13: 0000000000005726 R14: 0000000000000246 R15: ffff897c034506c8 [ 101.992192] FS: 00007fc3fe66bf40(0000) GS:ffff897d17b80000(0000) knlGS:0000000000000000 [ 101.992207] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 101.992221] CR2: ffffcbce408e4000 CR3: 0000000106388003 CR4: 00000000001726f0 [ 101.992236] Call Trace: [ 101.992250] <IRQ> [ 101.992264] smi_dma_xfer+0x154/0x210 [smipcie] [ 101.992283] process_one_work+0x177/0x330 [ 101.992308] bh_worker+0x17b/0x1a0 [ 101.992327] tasklet_action+0x10/0x30 [ 101.992343] handle_softirqs+0xcf/0x280 [ 101.992358] ? __pfx_read_tsc+0x10/0x10 [ 101.992373] do_softirq.part.0+0x3b/0x60 [ 101.992387] </IRQ> [ 101.992396] <TASK> [ 101.992405] __local_bh_enable_ip+0x60/0x70 [ 101.992420] fpu_clone+0xf8/0x400 [ 101.992440] copy_thread+0x14d/0x2a0 [ 101.992457] copy_process+0x1c47/0x2740 [ 101.992476] ? __alloc_pages_noprof+0x16b/0x310 [ 101.992494] kernel_clone+0xbd/0x440 [ 101.992511] __do_sys_clone3+0xe4/0x130 [ 101.992532] do_syscall_64+0x87/0x1b0 [ 101.992549] ? handle_mm_fault+0x1bb/0x2c0 [ 101.992566] ? do_user_addr_fault+0x36c/0x620 [ 101.992582] ? arch_exit_to_user_mode_prepare.isra.0+0x16/0xa0 [ 101.992598] entry_SYSCALL_64_after_hwframe+0x76/0x7e [ 101.992617] RIP: 0033:0x7fc3fe22c529 [ 101.992654] Code: 90 b8 01 00 00 00 b9 01 00 00 00 eb ec 0f 1f 40 00 b8 ea ff ff ff 48 85 ff 74 28 48 85 d2 74 23 49 89 c8 b8 b3 01 00 00 0f 05 <48> 85 c0 7c 14 74 01 c3 31 ed 4c 89 c7 ff d2 48 89 c7 b8 3c 00 00 [ 101.992674] RSP: 002b:00007ffeca7f25e8 EFLAGS: 00000202 ORIG_RAX: 00000000000001b3 [ 101.992691] RAX: ffffffffffffffda RBX: 00007fc3fe1ae620 RCX: 00007fc3fe22c529 [ 101.992705] RDX: 00007fc3fe1ae620 RSI: 0000000000000058 RDI: 00007ffeca7f2640 [ 101.992717] RBP: 00007fc3d0ff96c0 R08: 00007fc3d0ff96c0 R09: 00007ffeca7f2737 [ 101.992730] R10: 0000000000000008 R11: 0000000000000202 R12: fffffffffffffe90 [ 101.992743] R13: 0000000000000000 R14: 00007ffeca7f2640 R15: 00007fc3d07f9000 [ 101.992760] </TASK> [module inventory omitted] [ 101.993088] CR2: ffffcbce408e4000 [ 101.993101] ---[ end trace 0000000000000000 ]--- [ 101.993113] RIP: 0010:dvb_dmx_swfilter_packets+0x5d/0x90 [dvb_core] [ 101.993146] Code: 49 8d 6d ff eb 24 66 66 2e 0f 1f 84 00 00 00 00 00 66 66 2e 0f 1f 84 00 00 00 00 00 90 48 81 c3 bc 00 00 00 48 83 ed 01 72 1d <80> 3b 47 75 ee 48 89 de 4c 89 e7 48 81 c3 bc 00 00 00 e8 7c f9 ff [ 101.993166] RSP: 0018:ffffcbce401a0ea8 EFLAGS: 00010002 [ 101.993180] RAX: ffff897c034505c8 RBX: ffffcbce408e4000 RCX: 0000000000000000 [ 101.993193] RDX: ffff897c034505c8 RSI: ffffcbce408e3f44 RDI: ffff897c03450418 [ 101.993206] RBP: 0000000000005325 R08: 0000000000000000 R09: 0000000000000000 [ 101.993219] R10: 0000000000000001 R11: 0000000000000000 R12: ffff897c03450418 [ 101.993232] R13: 0000000000005726 R14: 0000000000000246 R15: ffff897c034506c8 [ 101.993245] FS: 00007fc3fe66bf40(0000) GS:ffff897d17b80000(0000) knlGS:0000000000000000 [ 101.993260] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 101.993272] CR2: ffffcbce408e4000 CR3: 0000000106388003 CR4: 00000000001726f0 [ 101.993286] Kernel panic - not syncing: Fatal exception in interrupt [ 101.993374] Kernel Offset: 0x34400000 from 0xffffffff81000000 (relocation range: 0xffffffff80000000-0xffffffffbfffffff) [ 101.993395] Rebooting in 30 seconds..