block: Make supported_write_flags a per-bds property

Pre-patch, .supported_write_flags lives at the driver level, which
means we are blindly declaring that all block devices using a
given driver will either equally support FUA, or that we need a
fallback at the block layer.  But there are drivers where FUA
support is a per-block decision: the NBD block driver is dependent
on the remote server advertising NBD_FLAG_SEND_FUA (and has
fallback code to duplicate the flush that the block layer would do
if NBD had not set .supported_write_flags); and the iscsi block
driver is dependent on the mode sense bits advertised by the
underlying device (and is currently silently ignoring FUA requests
if the underlying device does not support FUA).

The fix is to make supported flags as a per-BDS option, set during
.bdrv_open().  This patch moves the variable and fixes NBD and iscsi
to set it only conditionally; later patches will then further
simplify the NBD driver to quit duplicating work done at the block
layer, as well as tackle the fact that SCSI does not support FUA
semantics on WRITESAME(10/16) but only on WRITE(10/16).

Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
This commit is contained in:
Eric Blake 2016-05-03 16:39:06 -06:00 committed by Kevin Wolf
parent 2928abce6d
commit 4df863f336
6 changed files with 17 additions and 14 deletions

View File

@ -841,9 +841,10 @@ static int coroutine_fn bdrv_driver_pwritev(BlockDriverState *bs,
if (drv->bdrv_co_writev_flags) {
ret = drv->bdrv_co_writev_flags(bs, sector_num, nb_sectors, qiov,
flags);
flags & bs->supported_write_flags);
flags &= ~bs->supported_write_flags;
} else if (drv->bdrv_co_writev) {
assert(drv->supported_write_flags == 0);
assert(!bs->supported_write_flags);
ret = drv->bdrv_co_writev(bs, sector_num, nb_sectors, qiov);
} else {
BlockAIOCB *acb;
@ -862,9 +863,7 @@ static int coroutine_fn bdrv_driver_pwritev(BlockDriverState *bs,
}
emulate_flags:
if (ret == 0 && (flags & BDRV_REQ_FUA) &&
!(drv->supported_write_flags & BDRV_REQ_FUA))
{
if (ret == 0 && (flags & BDRV_REQ_FUA)) {
ret = bdrv_co_flush(bs);
}

View File

@ -456,8 +456,11 @@ iscsi_co_writev_flags(BlockDriverState *bs, int64_t sector_num, int nb_sectors,
struct IscsiTask iTask;
uint64_t lba;
uint32_t num_sectors;
bool fua;
bool fua = flags & BDRV_REQ_FUA;
if (fua) {
assert(iscsilun->dpofua);
}
if (!is_request_lun_aligned(sector_num, nb_sectors, iscsilun)) {
return -EINVAL;
}
@ -472,7 +475,6 @@ iscsi_co_writev_flags(BlockDriverState *bs, int64_t sector_num, int nb_sectors,
num_sectors = sector_qemu2lun(nb_sectors, iscsilun);
iscsi_co_init_iscsitask(iscsilun, &iTask);
retry:
fua = iscsilun->dpofua && (flags & BDRV_REQ_FUA);
if (iscsilun->use_16_for_rw) {
iTask.task = iscsi_write16_task(iscsilun->iscsi, iscsilun->lun, lba,
NULL, num_sectors * iscsilun->block_size,
@ -1548,6 +1550,9 @@ static int iscsi_open(BlockDriverState *bs, QDict *options, int flags,
task = NULL;
iscsi_modesense_sync(iscsilun);
if (iscsilun->dpofua) {
bs->supported_write_flags = BDRV_REQ_FUA;
}
/* Check the write protect flag of the LUN if we want to write */
if (iscsilun->type == TYPE_DISK && (flags & BDRV_O_RDWR) &&
@ -1841,7 +1846,6 @@ static BlockDriver bdrv_iscsi = {
.bdrv_co_write_zeroes = iscsi_co_write_zeroes,
.bdrv_co_readv = iscsi_co_readv,
.bdrv_co_writev_flags = iscsi_co_writev_flags,
.supported_write_flags = BDRV_REQ_FUA,
.bdrv_co_flush_to_disk = iscsi_co_flush,
#ifdef __linux__

View File

@ -414,6 +414,9 @@ int nbd_client_init(BlockDriverState *bs,
logout("Failed to negotiate with the NBD server\n");
return ret;
}
if (client->nbdflags & NBD_FLAG_SEND_FUA) {
bs->supported_write_flags = BDRV_REQ_FUA;
}
qemu_co_mutex_init(&client->send_mutex);
qemu_co_mutex_init(&client->free_sema);

View File

@ -471,7 +471,6 @@ static BlockDriver bdrv_nbd = {
.bdrv_file_open = nbd_open,
.bdrv_co_readv = nbd_co_readv,
.bdrv_co_writev_flags = nbd_co_writev_flags,
.supported_write_flags = BDRV_REQ_FUA,
.bdrv_close = nbd_close,
.bdrv_co_flush_to_os = nbd_co_flush,
.bdrv_co_discard = nbd_co_discard,
@ -490,7 +489,6 @@ static BlockDriver bdrv_nbd_tcp = {
.bdrv_file_open = nbd_open,
.bdrv_co_readv = nbd_co_readv,
.bdrv_co_writev_flags = nbd_co_writev_flags,
.supported_write_flags = BDRV_REQ_FUA,
.bdrv_close = nbd_close,
.bdrv_co_flush_to_os = nbd_co_flush,
.bdrv_co_discard = nbd_co_discard,
@ -509,7 +507,6 @@ static BlockDriver bdrv_nbd_unix = {
.bdrv_file_open = nbd_open,
.bdrv_co_readv = nbd_co_readv,
.bdrv_co_writev_flags = nbd_co_writev_flags,
.supported_write_flags = BDRV_REQ_FUA,
.bdrv_close = nbd_close,
.bdrv_co_flush_to_os = nbd_co_flush,
.bdrv_co_discard = nbd_co_discard,

View File

@ -204,6 +204,7 @@ static int raw_open(BlockDriverState *bs, QDict *options, int flags,
Error **errp)
{
bs->sg = bs->file->bs->sg;
bs->supported_write_flags = BDRV_REQ_FUA;
if (bs->probed && !bdrv_is_read_only(bs)) {
fprintf(stderr,
@ -250,7 +251,6 @@ BlockDriver bdrv_raw = {
.bdrv_create = &raw_create,
.bdrv_co_readv = &raw_co_readv,
.bdrv_co_writev_flags = &raw_co_writev_flags,
.supported_write_flags = BDRV_REQ_FUA,
.bdrv_co_write_zeroes = &raw_co_write_zeroes,
.bdrv_co_discard = &raw_co_discard,
.bdrv_co_get_block_status = &raw_co_get_block_status,

View File

@ -158,8 +158,6 @@ struct BlockDriver {
int coroutine_fn (*bdrv_co_pwritev)(BlockDriverState *bs,
uint64_t offset, uint64_t bytes, QEMUIOVector *qiov, int flags);
int supported_write_flags;
/*
* Efficiently zero a region of the disk image. Typically an image format
* would use a compact metadata representation to implement this. This
@ -445,6 +443,8 @@ struct BlockDriverState {
/* Alignment requirement for offset/length of I/O requests */
unsigned int request_alignment;
/* Flags honored during pwrite (so far: BDRV_REQ_FUA) */
unsigned int supported_write_flags;
/* the following member gives a name to every node on the bs graph. */
char node_name[32];