There's at least one drive (MaxDigitalData OOS14000G) such that if it receives a large amount of I/O while entering an idle power state will first exit idle before responding, including causing SMART temperature requests to be delayed. This causes the drivetemp request to exceed its timeout of 1 second. Signed-off-by: Russell Harmon <russ@xxxxxx> --- Here's a test case reproducing the issue: $ time cat /sys/class/hwmon/hwmon9/temp1_input 28000 cat /sys/class/hwmon/hwmon9/temp1_input 0.00s user 0.00s system 7% cpu 0.023 total $ dd if=/dev/sdep of=/dev/null bs=1M iflag=direct & # Generate background load $ ./openSeaChest_PowerControl -d /dev/sdep --transitionPower idle_a $ time cat /sys/class/hwmon/hwmon9/temp1_input 0 cat /sys/class/hwmon/hwmon9/temp1_input 0.00s user 0.00s system 0% cpu 3.154 total $ dmesg -t sd 11:0:1:0: attempting task abort!scmd(0x00000000ef8da38c), outstanding for 2098 ms & timeout 1000 ms sd 11:0:1:0: [sdep] tag#4639 CDB: ATA command pass through(16) 85 08 0e 00 d5 00 01 00 e0 00 4f 00 c2 00 b0 00 scsi target11:0:1: handle(0x0009), sas_address(0x4433221105000000), phy(5) scsi target11:0:1: enclosure logical id(0x500062b202d7ea80), slot(6) scsi target11:0:1: enclosure level(0x0000), connector name( ) sd 11:0:1:0: task abort: SUCCESS scmd(0x00000000ef8da38c) sd 11:0:1:0: Power-on or device reset occurred $ time cat /sys/class/hwmon/hwmon9/temp1_input 28000 cat /sys/class/hwmon/hwmon9/temp1_input 0.00s user 0.00s system 48% cpu 0.005 total drivers/hwmon/drivetemp.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/hwmon/drivetemp.c b/drivers/hwmon/drivetemp.c index 2a4ec55ddb47..291d91f68646 100644 --- a/drivers/hwmon/drivetemp.c +++ b/drivers/hwmon/drivetemp.c @@ -194,7 +194,7 @@ static int drivetemp_scsi_command(struct drivetemp_data *st, scsi_cmd[14] = ata_command; err = scsi_execute_cmd(st->sdev, scsi_cmd, op, st->smartdata, - ATA_SECT_SIZE, HZ, 5, NULL); + ATA_SECT_SIZE, 10 * HZ, 5, NULL); if (err > 0) err = -EIO; return err; -- 2.34.1
