On Tue, May 26, 2009 at 12:31:20AM +0200, Andries E. Brouwer wrote: > On Tue, May 26, 2009 at 12:47:55AM +0300, Michael S. Tsirkin wrote: > > On Mon, May 25, 2009 at 05:08:12PM -0400, Alan Stern wrote: > >> On Mon, 25 May 2009, Michael S. Tsirkin wrote: > >> > >>>> So apparently this is a bug in the device; it doesn't respond correctly > >>>> to the first READ command. But since it does respond correctly to > >>>> later commands, everything works okay thereafter. You ought to be able > >>>> to recover from the error by running > >>>> > >>>> blockdev --rereadpt /dev/sdb > >>>> > >>>> manually. > >>> > >>> Yes, this helps. > >>> Would it make sense for kernel to retry automatically? > >>> Why doesn't it? > >> > >> I don't know the details in this case. Most likely the error code > >> (Logical Block Address Out of Range) is interpreted as a fatal > >> non-retryable error. For other sorts of errors, the kernel does retry. > >> > >>>> As far as I can tell, this has nothing to do with any user programs in > >>>> the distribution. It appears to be entirely the device's fault. > >>>> > >>>> Alan Stern > >>> > >>> BTW, any idea how come I later get errors apparently from amiga fs? > >> > >> Not a clue. Unless it was some odd side effect of the partition code > >> trying to interpret an uninitialized buffer. > >> > >> Alan Stern > > Comments: > > The partition reading code tries one by one all partition types that > you have configured. If amigafs is one of these, also that will be tried. > > I can imagine that your device is a slow starter and needs some time > to initialize itself, so that the first, or the first few, commands fail, > and after that all is well. > > I hate retrying code. Such code is the reason that boot time increases > all the time. Usually a read failure is missing media or so, and retrying > is totally meaningless. Only at the lowest level should a retry be done, > and preferably not blindly "let us try a few times" but only when there > is a positive reason to expect that a second time might fare better. Yea. But I don't know where that is. I noted that if I stick a loop around check_partition I just get the error reported each time. > (Amiga is not usually the first or second to be tried, so if it fails > then either the first few reads fail, or the first read failure is cached.) Maybe it shouldn't printk stuff on IO failure when probing for partitions then. > Andries > > > > So, the following works for me as a work-around. But it's probably not > > the appropriate way to solve the problem. Or is it? Can someone who > > understands partitions and filesystems tell? > > > > > > > > block: retry on I/O error when reading partition table > > > > Retry once on an I/O error when reading the partition table: > > there's no much to loose, and this helps with some disk on > > key devices I have. > > > > Signed-off-by: Michael S. Tsirkin <mst@xxxxxxxxxx> -- MST -- 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
