On 07/25/2012 02:44 PM, Sen Wang wrote: > 2012/7/25 Paolo Bonzini <pbonzini@xxxxxxxxxx>: >> Il 25/07/2012 10:29, Wang Sen ha scritto: >>> When using the commands below to write some data to a virtio-scsi LUN of the >>> QEMU guest(32-bit) with 1G physical memory(qemu -m 1024), the qemu will crash. >>> >>> # sudo mkfs.ext4 /dev/sdb (/dev/sdb is the virtio-scsi LUN.) >>> # sudo mount /dev/sdb /mnt >>> # dd if=/dev/zero of=/mnt/file bs=1M count=1024 >>> >>> In current implementation, sg_set_buf is called to add buffers to sg list which >>> is put into the virtqueue eventually. But there are some HighMem pages in >>> table->sgl can not get virtual address by sg_virt. So, sg_virt(sg_elem) may >>> return NULL value. This will cause QEMU exit when virtqueue_map_sg is called >>> in QEMU because an invalid GPA is passed by virtqueue. >> >> Heh, I was compiling (almost) the same patch as we speak. :) > > Uh, what a coincidence! :) > >> >> I've never seen QEMU crash; the VM would more likely just fail to boot >> with a panic. But it's the same bug anyway. > > I never met this before. How this situation happens? > >> >>> My solution is using sg_set_page instead of sg_set_buf. >>> >>> I have tested the patch on my workstation. QEMU would not crash any more. >>> >>> Signed-off-by: Wang Sen <senwang@xxxxxxxxxxxxxxxxxx> >>> --- >>> drivers/scsi/virtio_scsi.c | 3 ++- >>> 1 file changed, 2 insertions(+), 1 deletion(-) >>> >>> diff --git a/drivers/scsi/virtio_scsi.c b/drivers/scsi/virtio_scsi.c >>> index 1b38431..fc5c88a 100644 >>> --- a/drivers/scsi/virtio_scsi.c >>> +++ b/drivers/scsi/virtio_scsi.c >>> @@ -198,7 +198,8 @@ 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_set_page(&sg[idx++], sg_page(sg_elem), sg_elem->length, >>> + sg_elem->offset); >> >> This can simply be >> >> sg[idx++] = *sg_elem; >> > > Yes, I saw your another E-mail. I think you're right. Simply calling > sg_set_page can not handle > the flag bits correctly. So, I'll repost the patch soon. Thank you! > No this code is correct, though you will need to make sure to properly terminate the destination sg_list. But since old code was using sg_set_buf(), than it means it was properly terminated before, and there for this code is good as is please don't touch it. Thanks Boaz >> Can you repost it with this change, and also add stable@xxxxxxxxxxxxxxx >> to the Cc? Thanks very much! >> >> Paolo >> -- >> 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 > > > -- 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