On Mon, 19 Jan 2015, Kenneth R. Crudup wrote: > Sorry for taking so long to get back to you guys about this (the 3.19-rc series has been > problematic for me in a couple of areas, so I'd let it go for a while): > > On Mon, 5 Jan 2015, Alan Stern wrote: > > > The patch I posted sets a general limit of 32 MB for USB drives that > > don't have a quirk flag for a smaller limit. > > Kenneth, have you tried that patch? Does it fix your problem? > > No, it fails in the same way (applied to 3.19-rc5): > > ---- > Jan 19 01:38:16 tosh-p75a kernel: [ 268.606258] sd 5:0:0:0: [sdc] FAILED Result: hostbyte=DID_ERROR driverbyte=DRIVER_OK > Jan 19 01:38:16 tosh-p75a kernel: [ 268.606262] sd 5:0:0:0: [sdc] CDB: > Jan 19 01:38:16 tosh-p75a kernel: [ 268.606263] Write(10): 2a 00 0f e8 e0 40 00 3c 00 00 > Jan 19 01:38:16 tosh-p75a kernel: [ 268.606268] blk_update_request: I/O error, dev sdc, sector 2135360000 > Jan 19 01:38:16 tosh-p75a kernel: [ 268.839597] sd 5:0:0:0: [sdc] FAILED Result: hostbyte=DID_ERROR driverbyte=DRIVER_OK > Jan 19 01:38:16 tosh-p75a kernel: [ 268.839603] sd 5:0:0:0: [sdc] CDB: > Jan 19 01:38:16 tosh-p75a kernel: [ 268.839605] Write(10): 2a 00 0f e9 68 40 00 3c 00 00 > Jan 19 01:38:16 tosh-p75a kernel: [ 268.839615] blk_update_request: I/O error, dev sdc, sector 2135638528 > <snip> > Jan 19 01:38:47 tosh-p75a kernel: [ 300.103494] usb 2-4.2: reset SuperSpeed USB device number 3 using xhci_hcd > Jan 19 01:38:47 tosh-p75a kernel: [ 300.115145] xhci_hcd 0000:00:14.0: xHCI xhci_drop_endpoint called with disabled ep ffff880448c9bf00 > Jan 19 01:38:47 tosh-p75a kernel: [ 300.115150] xhci_hcd 0000:00:14.0: xHCI xhci_drop_endpoint called with disabled ep ffff880448c9bf48 > Jan 19 01:38:48 tosh-p75a kernel: [ 300.443729] INFO: task kworker/u16:6:197 blocked for more than 30 seconds. > ---- > > FWIW, my quick fix gets me going again: > > ---- > diff --git a/block/blk-settings.c b/block/blk-settings.c > index 6ed2cbe..3a10bf6 100644 > --- a/block/blk-settings.c > +++ b/block/blk-settings.c > @@ -257,7 +257,8 @@ void blk_limits_max_hw_sectors(struct queue_limits *limits, unsigned int max_hw_ > __func__, max_hw_sectors); > } > > - limits->max_sectors = limits->max_hw_sectors = max_hw_sectors; > + limits->max_sectors = limits->max_hw_sectors = > + min_t(unsigned int, max_hw_sectors, 65535); > } > EXPORT_SYMBOL(blk_limits_max_hw_sectors); Are you certain the patch had been applied to the kernel you were testing? The patch adds a line that calls blk_queue_max_hw_sectors() with max_hw_sectors set to 65535, so it should have the same effect as your quick fix. Maybe you can add a printk statement to scsiglue.c:slave_configure() in drivers/usb/storage to make sure that the patch's call really does take place as intended. Alan Stern -- To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
