On Thu, 2008-05-01 at 18:59 +0300, Boaz Harrosh wrote: > On Thu, May 01 2008 at 18:47 +0300, James Bottomley <James.Bottomley@xxxxxxxxxxxxxxxxxxxxx> wrote: > > On Thu, 2008-05-01 at 18:36 +0300, Boaz Harrosh wrote: > >> On Thu, May 01 2008 at 18:22 +0300, James Bottomley <James.Bottomley@xxxxxxxxxxxxxxxxxxxxx> wrote: > >>> On Thu, 2008-05-01 at 17:56 +0300, Boaz Harrosh wrote: > >>>> The call to pci_map_page() on an ISA case will it not bounce the > >>>> buffer? > >>> Only if the system isn't a PCI_DMA_BUS_IS_PHYS one. > >>> > >>> That means basically either has an iommu/gart or has some silly swiotlb > >>> to emulate one. > >>> > >>> That means for standard x86 isa systems the answer is no: the buffer has > >>> to be allocated within the isa region because pci_map is simply a > >>> virt_to_phys. > >>> > >>> James > >>> > >> OK Thanks, so the second patch then. That will solve it. > > > > Um, your second patch is only changing isd200; it's hard to see how that > > can fix a gdth problem ... > > > > James > > > > > > Sorry I meant the second version of the patch that uses scsi_get_command() > > http://marc.info/?l=linux-scsi&m=120965483126936&w=2 No ... it still needs to use scsi_allocate_command() and pass the gfp flags for DMA if necessary. James -- 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
