On Tuesday 30 December 2008, Bruno Prémont wrote: > On Tue, 30 December 2008 Bruno Prémont wrote: > > Secondary note, with ATA subsystem it is possible to discover > > hot-plugged second SATA drive with issuing > > echo '- - -' > host0/scsi_host/host0/scan > > though the same action causes the access to first SATA drive to fail > > after removing second drive. (unfortunately BIOS does only allows > > configuring the chipset in IDE mode - so not possible to use SATA_VIA > > or AHCI). I guess this is caused by some incorrect detection of speed > > settings. > > > > What is the way to ask driver to rescan with IDE driver? > I found: > Documentation/ide/warm-plug-howto.txt > > > Doing just: > echo -n "1" > /sys/class/ide_port/idex/scan > causes the following BUG(): > <<<< during boot > Dec 30 18:19:50 venus [ 3.909946] Uniform Multi-Platform E-IDE driver > Dec 30 18:19:50 venus [ 3.913075] via82cxxx 0000:00:0f.0: VIA cx700 (rev 00) IDE UDMA133 > Dec 30 18:19:50 venus [ 3.916069] via82cxxx 0000:00:0f.0: IDE controller (0x1106:0x0581 rev 0x00) > Dec 30 18:19:50 venus [ 3.919138] via82cxxx 0000:00:0f.0: not 100% native mode: will probe irqs later > Dec 30 18:19:50 venus [ 3.924945] ide0: BM-DMA at 0xff00-0xff07 > Dec 30 18:19:50 venus [ 3.927990] ide1: BM-DMA at 0xff08-0xff0f > Dec 30 18:19:50 venus [ 3.930901] Probing IDE interface ide0... > Dec 30 18:19:50 venus [ 4.050047] Marking TSC unstable due to cpufreq changes > Dec 30 18:19:50 venus [ 4.370099] hda: FUJITSU MHY2250BH, ATA DISK drive > Dec 30 18:19:50 venus [ 4.390026] Clocksource tsc unstable (delta = -215834270 ns) > Dec 30 18:19:50 venus [ 4.730309] hda: host max PIO5 wanted PIO255(auto-tune) selected PIO4 > Dec 30 18:19:50 venus [ 4.730462] hda: UDMA/100 mode selected > Dec 30 18:19:50 venus [ 4.735541] Probing IDE interface ide1... > Dec 30 18:19:50 venus [ 5.420127] hdd: , ATA DISK drive > Dec 30 18:19:50 venus [ 5.480288] hdd: host max PIO5 wanted PIO255(auto-tune) selected PIO2 > Dec 30 18:19:50 venus [ 5.480330] hdd: UDMA/66 mode selected > Dec 30 18:19:50 venus [ 5.485099] ide0 at 0x1f0-0x1f7,0x3f6 on irq 14 > Dec 30 18:19:50 venus [ 5.565605] ide1 at 0x170-0x177,0x376 on irq 15 > Dec 30 18:19:50 venus [ 5.573072] ide-gd driver 1.18 > Dec 30 18:19:50 venus [ 5.577684] hda: max request size: 512KiB > Dec 30 18:19:50 venus [ 5.582339] hda: 488397168 sectors (250059 MB) w/8192KiB Cache, CHS=30401/255/63 > Dec 30 18:19:50 venus [ 5.592346] hda: cache flushes supported > Dec 30 18:19:50 venus [ 5.597500] hda: hda1 hda2 hda3 hda4 hda5 hda6 > Dec 30 18:19:50 venus [ 5.646208] hdd: max request size: 128KiB > Dec 30 18:19:50 venus [ 5.651361] hdd: 8043840 sectors (4118 MB) w/1KiB Cache, CHS=7980/16/63 > Dec 30 18:19:50 venus [ 5.656901] hdd: hdd1 > -- snip -- > <<<< some time later, triggering the scan > Dec 30 21:35:55 venus [ 7866.335247] Probing IDE interface ide0... > Dec 30 21:35:55 venus [ 7866.650105] hda: FUJITSU MHY2250BH, ATA DISK drive > Dec 30 21:35:56 venus [ 7867.370055] hda: host max PIO5 wanted PIO255(auto-tune) selected PIO4 > Dec 30 21:35:56 venus [ 7867.370221] hda: UDMA/100 mode selected > Dec 30 21:35:56 venus [ 7867.370375] BUG: unable to handle kernel NULL pointer dereference at 0000000c > Dec 30 21:35:56 venus [ 7867.370400] IP: [<c0248bf6>] elv_may_queue+0x6/0x20 > Dec 30 21:35:56 venus [ 7867.370440] *pde = 00000000 > Dec 30 21:35:56 venus [ 7867.370459] Oops: 0000 [#1] > Dec 30 21:35:56 venus [ 7867.375572] last sysfs file: /sys/devices/pci0000:00/0000:00:0f.0/ide0/ide_port/ide0/scan > Dec 30 21:35:56 venus [ 7867.380014] Modules linked in: snd_hda_intel snd_pcm snd_timer snd soundcore snd_page_alloc via_agp agpgart > Dec 30 21:35:56 venus [ 7867.380014] > Dec 30 21:35:56 venus [ 7867.380014] Pid: 1855, comm: bash Not tainted (2.6.28-ide #3) CX700+W697HG > Dec 30 21:35:56 venus [ 7867.380014] EIP: 0060:[<c0248bf6>] EFLAGS: 00010046 CPU: 0 > Dec 30 21:35:56 venus [ 7867.380014] EIP is at elv_may_queue+0x6/0x20 > Dec 30 21:35:56 venus [ 7867.380014] EAX: 00000000 EBX: 00000000 ECX: 00000000 EDX: 00000000 > Dec 30 21:35:56 venus [ 7867.380014] ESI: 00000000 EDI: 00000010 EBP: f6bb3e28 ESP: f6bb3e24 > Dec 30 21:35:56 venus [ 7867.380014] DS: 007b ES: 007b FS: 0000 GS: 0033 SS: 0068 > Dec 30 21:35:56 venus [ 7867.380014] Process bash (pid: 1855, ti=f6bb3000 task=f7110d80 task.ti=f6bb3000) > Dec 30 21:35:56 venus [ 7867.380014] Stack: > Dec 30 21:35:56 venus [ 7867.380014] 00000000 f6bb3e50 c024b40b 00000000 000000e1 f7005890 00000078 00000282 > Dec 30 21:35:56 venus [ 7867.380014] 00000000 f73d0020 00000000 f6bb3e80 c024b82c 00000010 00000000 f6bb3e7c > Dec 30 21:35:56 venus [ 7867.380014] 00000282 00000001 000000e1 000000e1 00000001 f73d0020 f6bb3ee4 f6bb3e8c > Dec 30 21:35:56 venus [ 7867.380014] Call Trace: > Dec 30 21:35:56 venus [ 7867.380014] [<c024b40b>] ? get_request+0x1b/0x240 > Dec 30 21:35:56 venus [ 7867.380014] [<c024b82c>] ? get_request_wait+0x1c/0xd0 > Dec 30 21:35:56 venus [ 7867.380014] [<c024bc48>] ? blk_get_request+0x18/0x30 > Dec 30 21:35:56 venus [ 7867.380014] [<c02e2d3a>] ? ide_raw_taskfile+0x2a/0x90 > Dec 30 21:35:56 venus [ 7867.380014] [<c02e2e21>] ? taskfile_lib_get_identify+0x61/0x70 > Dec 30 21:35:56 venus [ 7867.380014] [<c02e5ff0>] ? ide_acpi_port_init_devices+0x90/0xe0 > Dec 30 21:35:56 venus [ 7867.380014] [<c02e2379>] ? ide_port_scan+0x39/0x50 > Dec 30 21:35:56 venus [ 7867.380014] [<c02e23d1>] ? store_scan+0x41/0x60 > Dec 30 21:35:56 venus [ 7867.380014] [<c02e2390>] ? store_scan+0x0/0x60 > Dec 30 21:35:56 venus [ 7867.380014] [<c02cc03c>] ? dev_attr_store+0x2c/0x40 > Dec 30 21:35:56 venus [ 7867.380014] [<c01a468d>] ? sysfs_write_file+0x9d/0x100 > Dec 30 21:35:56 venus [ 7867.380014] [<c016b215>] ? vfs_write+0x95/0x120 > Dec 30 21:35:56 venus [ 7867.380014] [<c01a45f0>] ? sysfs_write_file+0x0/0x100 > Dec 30 21:35:56 venus [ 7867.380014] [<c016b34d>] ? sys_write+0x3d/0x70 > Dec 30 21:35:56 venus [ 7867.380014] [<c0103b01>] ? sysenter_do_call+0x12/0x25 > Dec 30 21:35:56 venus [ 7867.380014] Code: bf 00 00 00 00 55 8b 40 0c 89 e5 8b 00 8b 48 34 85 c9 74 04 89 d0 ff d1 5d c3 8d 74 26 00 8d bc 27 00 00 00 00 55 89 e5 53 89 c3 <8b> 40 0c 8b 00 8b 48 38 31 c0 85 c9 74 04 89 d8 ff d1 5b 5d c3 > Dec 30 21:35:56 venus [ 7867.380014] EIP: [<c0248bf6>] elv_may_queue+0x6/0x20 SS:ESP 0068:f6bb3e24 > Dec 30 21:35:56 venus [ 7867.380014] ---[ end trace b966cf347b1167e4 ]--- > Dec 30 21:36:15 venus [ 7886.330025] ide_timer_expiry: hwgroup->drive was NULL > Dec 30 21:36:29 venus [ 7899.770239] Buffer I/O error on device hda3, logical block 115511 > Dec 30 21:36:29 venus [ 7899.777078] lost page write due to I/O error on hda3 > Dec 30 21:36:29 venus [ 7899.783856] Buffer I/O error on device hda3, logical block 115512 > Dec 30 21:36:29 venus [ 7899.790570] lost page write due to I/O error on hda3 > > Same effect with: > cd /sys/class/ide_port/ide0/ > sync; sleep 1; echo -n 1 > delete_devices; sleep 1; echo -n 1 > scan Thanks for the report, the following patch fixes the OOPS for me: From: Bartlomiej Zolnierkiewicz <bzolnier@xxxxxxxxx> Subject: [PATCH] ide: fix ide_port_scan() to do ACPI setup after initializing request queues This makes ide_port_scan()'s behavior match ide_host_register()'s one and fixes OOPS in elv_may_queue() during port re-scan. Reported-by: Bruno Prémont <bonbons@xxxxxxxxxxxxxxxxx> Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@xxxxxxxxx> --- drivers/ide/ide-probe.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) Index: b/drivers/ide/ide-probe.c =================================================================== --- a/drivers/ide/ide-probe.c +++ b/drivers/ide/ide-probe.c @@ -1694,8 +1694,8 @@ void ide_port_scan(ide_hwif_t *hwif) hwif->present = 1; ide_port_tune_devices(hwif); - ide_acpi_port_init_devices(hwif); ide_port_setup_devices(hwif); + ide_acpi_port_init_devices(hwif); hwif_register_devices(hwif); ide_proc_port_register_devices(hwif); } -- To unsubscribe from this list: send the line "unsubscribe linux-ide" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html