@@ -1953,6 +1953,19 @@ uint32_t blk_get_request_alignment(BlockBackend *blk)
return bs ? bs->bl.request_alignment : BDRV_SECTOR_SIZE;
}
+/* Returns the maximum hardware transfer length, in bytes; guaranteed nonzero */
+uint64_t blk_get_max_hw_transfer(BlockBackend *blk)
+{
+ BlockDriverState *bs = blk_bs(blk);
+ uint64_t max = INT_MAX;
+
+ if (bs) {
+ max = MIN_NON_ZERO(max, bs->bl.max_hw_transfer);
+ max = MIN_NON_ZERO(max, bs->bl.max_transfer);
+ }
+ return ROUND_DOWN(max, blk_get_request_alignment(blk));
+}
+
/* Returns the maximum transfer length, in bytes; guaranteed nonzero */
uint32_t blk_get_max_transfer(BlockBackend *blk)
{
@@ -1232,7 +1232,7 @@ static void raw_refresh_limits(BlockDriverState *bs, Error **errp)
int ret = sg_get_max_transfer_length(s->fd);
if (ret > 0 && ret <= BDRV_REQUEST_MAX_BYTES) {
- bs->bl.max_transfer = pow2floor(ret);
+ bs->bl.max_hw_transfer = pow2floor(ret);
}
ret = sg_get_max_segments(s->fd);
@@ -127,6 +127,8 @@ static void bdrv_merge_limits(BlockLimits *dst, const BlockLimits *src)
{
dst->opt_transfer = MAX(dst->opt_transfer, src->opt_transfer);
dst->max_transfer = MIN_NON_ZERO(dst->max_transfer, src->max_transfer);
+ dst->max_hw_transfer = MIN_NON_ZERO(dst->max_hw_transfer,
+ src->max_hw_transfer);
dst->opt_mem_alignment = MAX(dst->opt_mem_alignment,
src->opt_mem_alignment);
dst->min_mem_alignment = MAX(dst->min_mem_alignment,
@@ -179,7 +179,7 @@ static int scsi_handle_inquiry_reply(SCSIGenericReq *r, SCSIDevice *s, int len)
(r->req.cmd.buf[1] & 0x01)) {
page = r->req.cmd.buf[2];
if (page == 0xb0) {
- uint32_t max_transfer = blk_get_max_transfer(s->conf.blk);
+ uint64_t max_transfer = blk_get_max_hw_transfer(s->conf.blk);
uint32_t max_iov = blk_get_max_iov(s->conf.blk);
assert(max_transfer);
@@ -695,6 +695,13 @@ typedef struct BlockLimits {
* clamped down. */
uint32_t max_transfer;
+ /* Maximal hardware transfer length in bytes. Applies whenever
+ * transfers to the device bypass the kernel I/O scheduler, for
+ * example with SG_IO. If larger than max_transfer or if zero,
+ * blk_get_max_hw_transfer will fall back to max_transfer.
+ */
+ uint64_t max_hw_transfer;
+
/* memory alignment, in bytes so that no bounce buffer is needed */
size_t min_mem_alignment;
@@ -208,6 +208,7 @@ void blk_eject(BlockBackend *blk, bool eject_flag);
int blk_get_flags(BlockBackend *blk);
uint32_t blk_get_request_alignment(BlockBackend *blk);
uint32_t blk_get_max_transfer(BlockBackend *blk);
+uint64_t blk_get_max_hw_transfer(BlockBackend *blk);
int blk_get_max_iov(BlockBackend *blk);
void blk_set_guest_block_size(BlockBackend *blk, int align);
void *blk_try_blockalign(BlockBackend *blk, size_t size);
For block host devices, I/O can happen through either the kernel file descriptor I/O system calls (preadv/pwritev, io_submit, io_uring) or the SCSI passthrough ioctl SG_IO. In the latter case, the size of each transfer can be limited by the HBA, while for file descriptor I/O the kernel is able to split and merge I/O in smaller pieces as needed. Applying the HBA limits to file descriptor I/O results in more system calls and suboptimal performance, so this patch splits the max_transfer limit in two: max_transfer remains valid and is used in general, while max_hw_transfer is limited to the maximum hardware size. max_hw_transfer can then be included by the scsi-generic driver in the block limits page, to ensure that the stricter hardware limit is used. Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> --- block/block-backend.c | 13 +++++++++++++ block/file-posix.c | 2 +- block/io.c | 2 ++ hw/scsi/scsi-generic.c | 2 +- include/block/block_int.h | 7 +++++++ include/sysemu/block-backend.h | 1 + 6 files changed, 25 insertions(+), 2 deletions(-)