On 02/25/2011 11:02 PM, Scott Matthews wrote:
Is there a method to enable "write back" cache instead of "Assuming drive cache: write through"?
The USB enclosure probably doesn't report the drive's cache status properly. The kernel doesn't modify the drive's write cache state, that message only affects the kernel's behavior for things like deciding whether to issue cache flush requests. Since the device doesn't report the state properly, it probably wouldn't handle the commands to set the write cache state properly either (though you could try with sdparm).
However, the write cache is most likely enabled already.
I'm getting less than 40MBs (megabytes) write performance on two USB 3.0 Hard Drives and suspect it maybe the cache policy. This is on an Ubuntu 10.04 box. For reads: hdparm -Tt /dev/sdd1 WD my book 3TB /dev/sdd1: Timing cached reads: 13266 MB in 1.99 seconds = 6666.76 MB/sec Timing buffered disk reads: 376 MB in 3.00 seconds = 125.16 MB/sec Seagate goflex 3TB /dev/sdd1: Timing cached reads: 13328 MB in 1.99 seconds = 6697.37 MB/sec Timing buffered disk reads: 444 MB in 3.00 seconds = 147.85 MB/sec Sample dmesg when a single drive is plugged in: [ 267.676207] usb 5-1: new SuperSpeed USB device using xhci_hcd and address 2 [ 267.693777] xhci_hcd 0000:02:00.0: WARN: short transfer on control ep [ 267.694151] xhci_hcd 0000:02:00.0: WARN: short transfer on control ep [ 267.694526] xhci_hcd 0000:02:00.0: WARN: short transfer on control ep [ 267.695027] xhci_hcd 0000:02:00.0: WARN: short transfer on control ep [ 267.695231] usb 5-1: configuration #1 chosen from 1 choice [ 267.705478] Initializing USB Mass Storage driver... [ 267.705675] scsi6 : SCSI emulation for USB Mass Storage devices [ 267.705829] usb-storage: device found at 2 [ 267.705832] usb-storage: waiting for device to settle before scanning [ 267.705851] usbcore: registered new interface driver usb-storage [ 267.705855] USB Mass Storage support registered. [ 272.704427] usb-storage: device scan complete [ 272.704828] scsi 6:0:0:0: Direct-Access WD My Book 1130 1012 PQ: 0 ANSI: 6 [ 272.705159] scsi 6:0:0:1: Enclosure WD SES Device 1012 PQ: 0 ANSI: 6 [ 272.705643] sd 6:0:0:0: Attached scsi generic sg3 type 0 [ 272.705772] scsi 6:0:0:1: Attached scsi generic sg4 type 13 [ 285.369621] sd 6:0:0:0: [sdd] 732558336 4096-byte logical blocks: (3.00 TB/2.72 TiB) [ 285.369778] sd 6:0:0:0: [sdd] Write Protect is off [ 285.369780] sd 6:0:0:0: [sdd] Mode Sense: 47 00 10 08 [ 285.369783] sd 6:0:0:0: [sdd] Assuming drive cache: write through [ 285.370104] sd 6:0:0:0: [sdd] 732558336 4096-byte logical blocks: (3.00 TB/2.72 TiB) [ 285.370262] sd 6:0:0:0: [sdd] Assuming drive cache: write through [ 285.370266] sdd: sdd1 [ 285.379510] sd 6:0:0:0: [sdd] 732558336 4096-byte logical blocks: (3.00 TB/2.72 TiB) [ 285.379845] sd 6:0:0:0: [sdd] Assuming drive cache: write through [ 285.379851] sd 6:0:0:0: [sdd] Attached SCSI disk [ 285.415200] ses 6:0:0:1: Attached Enclosure device [ 285.439954] xhci_hcd 0000:02:00.0: WARN: Stalled endpoint -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@xxxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
-- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
