Hello,
On Fri, Apr 1, 2011 at 1:52 PM, Leon Woestenberg
<leon.woestenberg@xxxxxxxxx> wrote:
>
> I see the sata_sil24 / Sil 3132 fail on my system.
>
To add evidence, I tested against a Samsung 2.5" SATA drive instead of
an Intel SSD.
I also enabled libata debug logging.
The controller just seems to stop generating interrupts or similar,
and commands start to time-out.
Here is a kernel log dump of dd if=/dev/sda of=/dev/null bs=32k count=30000
Timeout occured at [ 168.899156]:
[ 138.665674] ata_sg_clean: unmapping 2 sg elements
[ 138.674312] ata_scsi_dump_cdb: CDB (2:0,0,0) 28 00 00 00 93 c0 00 01 00
[ 138.680923] ata_scsi_translate: ENTER
[ 138.684577] scsi_10_lba_len: ten-byte command
[ 138.688927] ata_sg_setup: ENTER, ata2
[ 138.692581] ata_sg_setup: 2 sg elements mapped
[ 138.697020] ata_scsi_translate: EXIT
[ 138.701368] ata_scsi_dump_cdb: CDB (2:0,0,0) 28 00 00 00 94 c0 00 01 00
[ 138.707980] ata_scsi_translate: ENTER
[ 138.711634] scsi_10_lba_len: ten-byte command
[ 138.715984] ata_sg_setup: ENTER, ata2
[ 138.719637] ata_sg_setup: 2 sg elements mapped
[ 138.724077] ata_scsi_translate: EXIT
[ 168.899156] ata_scsi_timed_out: ENTER
[ 168.902817] ata_scsi_timed_out: EXIT, ret=0
[ 168.906991] ata_scsi_timed_out: ENTER
[ 168.910644] ata_scsi_timed_out: EXIT, ret=0
[ 168.914835] ata_scsi_error: ENTER
[ 168.918153] __ata_port_freeze: ata2 port frozen
[ 168.922688] ata_eh_link_autopsy: ENTER
[ 168.926441] ata_eh_link_autopsy: EXIT
[ 168.930105] ata2.00: exception Emask 0x0 SAct 0x3 SErr 0x0 action 0x6 frozen
[ 168.937156] ata2.00: failed command: READ FPDMA QUEUED
[ 168.942306] ata2.00: cmd 60/00:00:c0:93:00/01:00:00:00:00/40 tag 0
ncq 131072 in
[ 168.942311] res 40/00:00:00:00:00/00:00:00:00:00/00 Emask
0x4 (timeout)
[ 168.957086] ata2.00: status: { DRDY }
[ 168.960746] ata2.00: failed command: READ FPDMA QUEUED
[ 168.965894] ata2.00: cmd 60/00:08:c0:94:00/01:00:00:00:00/40 tag 1
ncq 131072 in
[ 168.965899] res 40/00:00:00:00:00/00:00:00:00:00/00 Emask
0x4 (timeout)
[ 168.980673] ata2.00: status: { DRDY }
[ 168.984335] ata_eh_recover: ENTER
[ 168.987651] __ata_port_freeze: ata2 port frozen
[ 168.992180] ata2: hard resetting link
[ 171.195167] sil24_softreset: ENTER
[ 171.211173] ata_dev_classify: found ATA device by sig
[ 171.216229] sil24_softreset: EXIT, class=1
[ 171.220332] ata_eh_thaw_port: ata2 port thawed
[ 171.224771] ata_std_postreset: ENTER
[ 171.228351] ata2: SATA link up 3.0 Gbps (SStatus 123 SControl 0)
[ 171.234354] ata_std_postreset: EXIT
[ 171.237843] ata_eh_revalidate_and_attach: ENTER
[ 171.242375] ata2.00: ata_dev_read_id: ENTER
[ 171.246559] ata_sg_setup: ENTER, ata2
[ 171.250213] ata_sg_setup: 1 sg elements mapped
[ 176.335176] __ata_port_freeze: ata2 port frozen
[ 176.339715] ata_sg_clean: unmapping 1 sg elements
[ 176.344412] ata2.00: qc timeout (cmd 0xec)
[ 176.348517] ata2.00: failed to IDENTIFY (I/O error, err_mask=0x4)
[ 176.354611] ata2.00: revalidation failed (errno=-5)
[ 176.359487] ata_eh_revalidate_and_attach: EXIT rc=-5
[ 176.364454] __ata_port_freeze: ata2 port frozen
[ 176.368982] ata2: hard resetting link
[ 178.511168] sil24_softreset: ENTER
[ 178.527169] ata_dev_classify: found ATA device by sig
[ 178.532225] sil24_softreset: EXIT, class=1
[ 178.536327] ata_eh_thaw_port: ata2 port thawed
[ 178.540767] ata_std_postreset: ENTER
[ 178.544348] ata2: SATA link up 3.0 Gbps (SStatus 123 SControl 0)
[ 178.550350] ata_std_postreset: EXIT
[ 178.553840] ata_eh_revalidate_and_attach: ENTER
[ 178.558371] ata2.00: ata_dev_read_id: ENTER
[ 178.562554] ata_sg_setup: ENTER, ata2
[ 178.566208] ata_sg_setup: 1 sg elements mapped
[ 188.567185] __ata_port_freeze: ata2 port frozen
[ 188.571722] ata_sg_clean: unmapping 1 sg elements
[ 188.576419] ata2.00: qc timeout (cmd 0xec)
[ 188.580524] ata2.00: failed to IDENTIFY (I/O error, err_mask=0x4)
Do we have support for interrupt polling under libata?
Next I will investigate the interrupt request registers, maybe a
interrupt race condition in the driver?
Regards,
Leon.
--
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
