2011/10/11 Amit Sahrawat <amit.sahrawat83@xxxxxxxxx>: > On Tue, Oct 11, 2011 at 4:11 PM, Hannes Reinecke <hare@xxxxxxx> wrote: >> On 10/11/2011 10:26 AM, Amit Sahrawat wrote: >>> >>> SCSI: Retrieve Cache Mode Using SG_ATA_16 if normal routine fails >>> >>> It has been observed that a number of USB HDD's do not respond correctly >>> to SCSI mode sense command(retrieve caching pages) which results in their >>> Write Cache being discarded by queue requests i.e., WCE if left set to >>> '0'(disabled). So, in order to identify the devices correctly - give it >>> a last try using SG_ATA_16 after failure from normal routine. >>> >> I'd be _VERY_ cautious when trying something like this. >> >> ATA_16 is a rather recent command, and most older drives simply can't cope >> with this. Instead they just kick the error handler and go offline >> eventually. > I tried with a number of USB HDD(around 6 with different > manufacturer's, 1 USB SSD and pen drives) - they are responding > correctly with this. >> >> Shouldn't we rather use a sensible default, ie assume WCE is enabled? > If WCE is enabled by default - that will mark QUEUE ORDERING to > QUEUE_ORDERED_DRAIN_FLUSH(Preflush-Postflush)(ie., flushing will occur > for all the Mass Storage devices - which in turn will keep on calling > SYNC_CACHE for these devices(which do not have write cache)) - ie., an > extra overhead for the devices which do not have this routine(Pen > drives being widely used among these) > Is there any identification(unique information) for "older drives" ? > Please share your opinion more. I think that it is good approach. Have you tested journaling test on XFS ? For example, while tiny several files is written, Is there corruption fail when suddenly unplugging usb HDD ? And you should check write performance. If you change a little, It will be better~ let me check more. > >> >> Cheers, >> >> Hannes >> -- >> Dr. Hannes Reinecke zSeries & Storage >> hare@xxxxxxx +49 911 74053 688 >> SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg >> GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg) >> > Regards, > Amit Sahrawat > -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
