From: Wang Sen <senwang@xxxxxxxxxxxxxxxxxx> On a 32-bit guest with virtio-scsi devices and more than 1G physical memory, QEMU may crash or Linux will fail to boot. This bug happens when building the sg_list that is eventually put in the virtqueue. Each buffer from the original sg_list is added with sg_set_buf, but this will not work for HighMem pages in table->sgl. In that case, the original sg_list elements do not have a valid virtual address, but sg_set_buf will use sg_virt. For now, virtio_ring does not care about the form of the scatterlist and simply processes the first out_num + in_num consecutive elements of the sg[] array. However, it is better to create a well-formed scatterlist including the termination marker. http://lkml.indiana.edu/hypermail/linux/kernel/1207.3/00675.html discusses using value assignment vs. sg_set_page to copy the scatterlist. With sg_set_page, the driver would need to drop the marker manually in case it was left there by a previous request, and then use sg_mark_end to add the marker to the last entry. Value assignment instead will copy the last entry of the source sg_list to the destination list. The end marker that were set by blk_rq_map_sg() is copied too when the last entry of the source sg_list is copied to the the last entry in destination list. Cc: Stable kernel <stable@xxxxxxxxxxxxxxx> # 3.4: 4fe74b1: [SCSI] virtio-scsi: SCSI driver Signed-off-by: Wang Sen <senwang@xxxxxxxxxxxxxxxxxx> Signed-off-by: Paolo Bonzini <pbonzini@xxxxxxxxxx> --- drivers/scsi/virtio_scsi.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/drivers/scsi/virtio_scsi.c b/drivers/scsi/virtio_scsi.c index c7030fb..3e79a2f 100644 --- a/drivers/scsi/virtio_scsi.c +++ b/drivers/scsi/virtio_scsi.c @@ -331,7 +331,7 @@ static void virtscsi_map_sgl(struct scatterlist *sg, unsigned int *p_idx, int i; for_each_sg(table->sgl, sg_elem, table->nents, i) - sg_set_buf(&sg[idx++], sg_virt(sg_elem), sg_elem->length); + sg[idx++] = *sg_elem; *p_idx = idx; } -- 1.7.1 -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
