On Tue, 4 Mar 2014, David Laight wrote: > > An alternative is to work around this particular problem by identifying > > the code that submits the 3-sector SG element, and changing it to use > > an even number of sectors. But obviously that doesn't solve the > > underlying issue. > > Actually most of the block layer code could be taught to split requests > into multiple URBs. Or even resubmit bounced requests split in two. > Although splitting requests won't help the network code. It won't help the mass-storage code either. If an SG element is shorter than the maxpacket length, splitting it up isn't going to make it longer. What _would_ help would be for the block layer to split up both this element and the next one, and then allocate a maxpacket-sized bounce buffer to hold the last part of this element plus the first part of the next. But _that_ won't help the network code. Alan Stern -- 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
