From 69f47505ee66afaa513305de0c1895a224e52c45 Mon Sep 17 00:00:00 2001 From: Vladimir Sementsov-Ogievskiy Date: Mon, 8 Apr 2019 19:26:17 +0300 Subject: block: avoid recursive block_status call if possible drv_co_block_status digs bs->file for additional, more accurate search for hole inside region, reported as DATA by bs since 5daa74a6ebc. This accuracy is not free: assume we have qcow2 disk. Actually, qcow2 knows, where are holes and where is data. But every block_status request calls lseek additionally. Assume a big disk, full of data, in any iterative copying block job (or img convert) we'll call lseek(HOLE) on every iteration, and each of these lseeks will have to iterate through all metadata up to the end of file. It's obviously ineffective behavior. And for many scenarios we don't need this lseek at all. However, lseek is needed when we have metadata-preallocated image. So, let's detect metadata-preallocation case and don't dig qcow2's protocol file in other cases. The idea is to compare allocation size in POV of filesystem with allocations size in POV of Qcow2 (by refcounts). If allocation in fs is significantly lower, consider it as metadata-preallocation case. 102 iotest changed, as our detector can't detect shrinked file as metadata-preallocation, which don't seem to be wrong, as with metadata preallocation we always have valid file length. Two other iotests have a slight change in their QMP output sequence: Active 'block-commit' returns earlier because the job coroutine yields earlier on a blocking operation. This operation is loading the refcount blocks in qcow2_detect_metadata_preallocation(). Suggested-by: Denis V. Lunev Signed-off-by: Vladimir Sementsov-Ogievskiy Signed-off-by: Kevin Wolf --- include/block/block.h | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) (limited to 'include') diff --git a/include/block/block.h b/include/block/block.h index 9b083e2..531cf59 100644 --- a/include/block/block.h +++ b/include/block/block.h @@ -156,10 +156,15 @@ typedef struct HDGeometry { * BDRV_BLOCK_EOF: the returned pnum covers through end of file for this * layer, set by block layer * - * Internal flag: + * Internal flags: * BDRV_BLOCK_RAW: for use by passthrough drivers, such as raw, to request * that the block layer recompute the answer from the returned * BDS; must be accompanied by just BDRV_BLOCK_OFFSET_VALID. + * BDRV_BLOCK_RECURSE: request that the block layer will recursively search for + * zeroes in file child of current block node inside + * returned region. Only valid together with both + * BDRV_BLOCK_DATA and BDRV_BLOCK_OFFSET_VALID. Should not + * appear with BDRV_BLOCK_ZERO. * * If BDRV_BLOCK_OFFSET_VALID is set, the map parameter represents the * host offset within the returned BDS that is allocated for the @@ -184,6 +189,7 @@ typedef struct HDGeometry { #define BDRV_BLOCK_RAW 0x08 #define BDRV_BLOCK_ALLOCATED 0x10 #define BDRV_BLOCK_EOF 0x20 +#define BDRV_BLOCK_RECURSE 0x40 #define BDRV_BLOCK_OFFSET_MASK BDRV_SECTOR_MASK typedef QSIMPLEQ_HEAD(BlockReopenQueue, BlockReopenQueueEntry) BlockReopenQueue; -- cgit v1.1 From 2b02fd81ded2f8fefbdacee431fc9fb006f92bc6 Mon Sep 17 00:00:00 2001 From: Julia Suvorova Date: Sun, 2 Jun 2019 23:17:09 +0300 Subject: block/linux-aio: Drop unused BlockAIOCB submission method Callback-based laio_submit() and laio_cancel() were left after rewriting Linux AIO backend to coroutines in hope that they would be used in other code that could bypass coroutines. They can be safely removed because they have not been used since that time. Signed-off-by: Julia Suvorova Reviewed-by: Stefan Hajnoczi Signed-off-by: Kevin Wolf --- include/block/raw-aio.h | 3 --- 1 file changed, 3 deletions(-) (limited to 'include') diff --git a/include/block/raw-aio.h b/include/block/raw-aio.h index ba223dd..0cb7cc7 100644 --- a/include/block/raw-aio.h +++ b/include/block/raw-aio.h @@ -50,9 +50,6 @@ LinuxAioState *laio_init(Error **errp); void laio_cleanup(LinuxAioState *s); int coroutine_fn laio_co_submit(BlockDriverState *bs, LinuxAioState *s, int fd, uint64_t offset, QEMUIOVector *qiov, int type); -BlockAIOCB *laio_submit(BlockDriverState *bs, LinuxAioState *s, int fd, - int64_t sector_num, QEMUIOVector *qiov, int nb_sectors, - BlockCompletionFunc *cb, void *opaque, int type); void laio_detach_aio_context(LinuxAioState *s, AioContext *old_context); void laio_attach_aio_context(LinuxAioState *s, AioContext *new_context); void laio_io_plug(BlockDriverState *bs, LinuxAioState *s); -- cgit v1.1 From 3036a626e9efad8ee5dc178b149184cf87bd06b0 Mon Sep 17 00:00:00 2001 From: Kenneth Heitke Date: Mon, 20 May 2019 11:40:30 -0600 Subject: nvme: add Get/Set Feature Timestamp support Signed-off-by: Kenneth Heitke Reviewed-by: Klaus Birkelund Jensen Signed-off-by: Kevin Wolf --- include/block/nvme.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'include') diff --git a/include/block/nvme.h b/include/block/nvme.h index 849a6f3..3ec8efc 100644 --- a/include/block/nvme.h +++ b/include/block/nvme.h @@ -581,6 +581,7 @@ enum NvmeIdCtrlOncs { NVME_ONCS_WRITE_ZEROS = 1 << 3, NVME_ONCS_FEATURES = 1 << 4, NVME_ONCS_RESRVATIONS = 1 << 5, + NVME_ONCS_TIMESTAMP = 1 << 6, }; #define NVME_CTRL_SQES_MIN(sqes) ((sqes) & 0xf) @@ -622,6 +623,7 @@ enum NvmeFeatureIds { NVME_INTERRUPT_VECTOR_CONF = 0x9, NVME_WRITE_ATOMICITY = 0xa, NVME_ASYNCHRONOUS_EVENT_CONF = 0xb, + NVME_TIMESTAMP = 0xe, NVME_SOFTWARE_PROGRESS_MARKER = 0x80 }; -- cgit v1.1 From 97896a4887a0a29c3314c5f0e9a82e269a6401fc Mon Sep 17 00:00:00 2001 From: Kevin Wolf Date: Thu, 2 May 2019 11:10:59 +0200 Subject: block: Add Error to blk_set_aio_context() Add an Error parameter to blk_set_aio_context() and use bdrv_child_try_set_aio_context() internally to check whether all involved nodes can actually support the AioContext switch. Signed-off-by: Kevin Wolf --- include/sysemu/block-backend.h | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'include') diff --git a/include/sysemu/block-backend.h b/include/sysemu/block-backend.h index 938de34..228fb3f 100644 --- a/include/sysemu/block-backend.h +++ b/include/sysemu/block-backend.h @@ -208,7 +208,8 @@ void blk_op_unblock(BlockBackend *blk, BlockOpType op, Error *reason); void blk_op_block_all(BlockBackend *blk, Error *reason); void blk_op_unblock_all(BlockBackend *blk, Error *reason); AioContext *blk_get_aio_context(BlockBackend *blk); -void blk_set_aio_context(BlockBackend *blk, AioContext *new_context); +int blk_set_aio_context(BlockBackend *blk, AioContext *new_context, + Error **errp); void blk_add_aio_context_notifier(BlockBackend *blk, void (*attached_aio_context)(AioContext *new_context, void *opaque), void (*detach_aio_context)(void *opaque), void *opaque); -- cgit v1.1 From d861ab3acf8dcf817e0c2335979b258847b69564 Mon Sep 17 00:00:00 2001 From: Kevin Wolf Date: Thu, 25 Apr 2019 14:25:10 +0200 Subject: block: Add BlockBackend.ctx This adds a new parameter to blk_new() which requires its callers to declare from which AioContext this BlockBackend is going to be used (or the locks of which AioContext need to be taken anyway). The given context is only stored and kept up to date when changing AioContexts. Actually applying the stored AioContext to the root node is saved for another commit. Signed-off-by: Kevin Wolf --- include/sysemu/block-backend.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include') diff --git a/include/sysemu/block-backend.h b/include/sysemu/block-backend.h index 228fb3f..733c495 100644 --- a/include/sysemu/block-backend.h +++ b/include/sysemu/block-backend.h @@ -76,7 +76,7 @@ typedef struct BlockBackendPublic { ThrottleGroupMember throttle_group_member; } BlockBackendPublic; -BlockBackend *blk_new(uint64_t perm, uint64_t shared_perm); +BlockBackend *blk_new(AioContext *ctx, uint64_t perm, uint64_t shared_perm); BlockBackend *blk_new_open(const char *filename, const char *reference, QDict *options, int flags, Error **errp); int blk_get_refcnt(BlockBackend *blk); -- cgit v1.1 From 307a5f60ebe0506a5b90f323fdcabb4333405484 Mon Sep 17 00:00:00 2001 From: Kevin Wolf Date: Mon, 29 Apr 2019 17:40:14 +0200 Subject: block: Add qdev_prop_drive_iothread property type Some qdev block devices have support for iothreads and take care of the AioContext they are running in, but most devices don't know about any of this. For the latter category, the qdev drive property must make sure that their BlockBackend is in the main AioContext. Unfortunately, while the current code just does the same thing for devices that do support iothreads, this is not correct and it would show as soon as we actually try to keep a consistent AioContext assignment across all nodes and users of a block graph subtree: If a node is already in a non-default AioContext because of one of its users, attaching a new device should still be possible if that device can work in the same AioContext. Switching the node back to the main context first and only then into the device AioContext causes failure (because the existing user wouldn't allow the switch to the main context). So devices that support iothreads need a different kind of drive property that leaves the node in its current AioContext, but by using this type, the device promises to check later that it can work with this context. This patch adds the qdev infrastructure that allows devices to signal that they handle iothreads and qdev should leave the AioContext alone. Signed-off-by: Kevin Wolf --- include/hw/block/block.h | 7 +++++-- include/hw/qdev-properties.h | 3 +++ 2 files changed, 8 insertions(+), 2 deletions(-) (limited to 'include') diff --git a/include/hw/block/block.h b/include/hw/block/block.h index d06f25a..6075390 100644 --- a/include/hw/block/block.h +++ b/include/hw/block/block.h @@ -45,8 +45,7 @@ static inline unsigned int get_physical_block_exp(BlockConf *conf) return exp; } -#define DEFINE_BLOCK_PROPERTIES(_state, _conf) \ - DEFINE_PROP_DRIVE("drive", _state, _conf.blk), \ +#define DEFINE_BLOCK_PROPERTIES_BASE(_state, _conf) \ DEFINE_PROP_BLOCKSIZE("logical_block_size", _state, \ _conf.logical_block_size), \ DEFINE_PROP_BLOCKSIZE("physical_block_size", _state, \ @@ -59,6 +58,10 @@ static inline unsigned int get_physical_block_exp(BlockConf *conf) ON_OFF_AUTO_AUTO), \ DEFINE_PROP_BOOL("share-rw", _state, _conf.share_rw, false) +#define DEFINE_BLOCK_PROPERTIES(_state, _conf) \ + DEFINE_PROP_DRIVE("drive", _state, _conf.blk), \ + DEFINE_BLOCK_PROPERTIES_BASE(_state, _conf) + #define DEFINE_BLOCK_CHS_PROPERTIES(_state, _conf) \ DEFINE_PROP_UINT32("cyls", _state, _conf.cyls, 0), \ DEFINE_PROP_UINT32("heads", _state, _conf.heads, 0), \ diff --git a/include/hw/qdev-properties.h b/include/hw/qdev-properties.h index b6758c8..1eae5ab 100644 --- a/include/hw/qdev-properties.h +++ b/include/hw/qdev-properties.h @@ -28,6 +28,7 @@ extern const PropertyInfo qdev_prop_blockdev_on_error; extern const PropertyInfo qdev_prop_bios_chs_trans; extern const PropertyInfo qdev_prop_fdc_drive_type; extern const PropertyInfo qdev_prop_drive; +extern const PropertyInfo qdev_prop_drive_iothread; extern const PropertyInfo qdev_prop_netdev; extern const PropertyInfo qdev_prop_pci_devfn; extern const PropertyInfo qdev_prop_blocksize; @@ -198,6 +199,8 @@ extern const PropertyInfo qdev_prop_pcie_link_width; DEFINE_PROP(_n, _s, _f, qdev_prop_netdev, NICPeers) #define DEFINE_PROP_DRIVE(_n, _s, _f) \ DEFINE_PROP(_n, _s, _f, qdev_prop_drive, BlockBackend *) +#define DEFINE_PROP_DRIVE_IOTHREAD(_n, _s, _f) \ + DEFINE_PROP(_n, _s, _f, qdev_prop_drive_iothread, BlockBackend *) #define DEFINE_PROP_MACADDR(_n, _s, _f) \ DEFINE_PROP(_n, _s, _f, qdev_prop_macaddr, MACAddr) #define DEFINE_PROP_ON_OFF_AUTO(_n, _s, _f, _d) \ -- cgit v1.1 From 4f71fb436a6fbec7a7b2360cdba741deeb24be67 Mon Sep 17 00:00:00 2001 From: Kevin Wolf Date: Fri, 26 Apr 2019 19:29:47 +0200 Subject: scsi-disk: Use qdev_prop_drive_iothread This makes use of qdev_prop_drive_iothread for scsi-disk so that the disk can be attached to a node that is already in the target AioContext. We need to check that the HBA actually supports iothreads, otherwise scsi-disk must make sure that the node is already in the main AioContext. This changes the error message for conflicting iothread settings. Previously, virtio-scsi produced the error message, now it comes from blk_set_aio_context(). Update a test case accordingly. Signed-off-by: Kevin Wolf --- include/hw/scsi/scsi.h | 1 + 1 file changed, 1 insertion(+) (limited to 'include') diff --git a/include/hw/scsi/scsi.h b/include/hw/scsi/scsi.h index acef25f..426566a 100644 --- a/include/hw/scsi/scsi.h +++ b/include/hw/scsi/scsi.h @@ -88,6 +88,7 @@ struct SCSIDevice int scsi_version; int default_scsi_version; bool needs_vpd_bl_emulation; + bool hba_supports_iothread; }; extern const VMStateDescription vmstate_scsi_device; -- cgit v1.1 From 132ada80c4a6fea7b67e8bb0a5fd299994d927c6 Mon Sep 17 00:00:00 2001 From: Kevin Wolf Date: Wed, 24 Apr 2019 17:41:46 +0200 Subject: block: Adjust AioContexts when attaching nodes So far, we only made sure that updating the AioContext of a node affected the whole subtree. However, if a node is newly attached to a new parent, we also need to make sure that both the subtree of the node and the parent are in the same AioContext. This tries to move the new child node to the parent AioContext and returns an error if this isn't possible. BlockBackends now actually apply their AioContext to their root node. Signed-off-by: Kevin Wolf --- include/block/block_int.h | 1 + 1 file changed, 1 insertion(+) (limited to 'include') diff --git a/include/block/block_int.h b/include/block/block_int.h index 1eebc7c..06df2bd 100644 --- a/include/block/block_int.h +++ b/include/block/block_int.h @@ -1163,6 +1163,7 @@ void hmp_drive_add_node(Monitor *mon, const char *optstr); BdrvChild *bdrv_root_attach_child(BlockDriverState *child_bs, const char *child_name, const BdrvChildRole *child_role, + AioContext *ctx, uint64_t perm, uint64_t shared_perm, void *opaque, Error **errp); void bdrv_root_unref_child(BdrvChild *child); -- cgit v1.1 From 42a65f02f9b380bd8074882d5844d4ea033389cc Mon Sep 17 00:00:00 2001 From: Kevin Wolf Date: Tue, 7 May 2019 18:31:38 +0200 Subject: block: Remove bdrv_set_aio_context() All callers of bdrv_set_aio_context() are eliminated now, they have moved to bdrv_try_set_aio_context() and related safe functions. Remove bdrv_set_aio_context(). With this, we can now know that the .set_aio_ctx callback must be present in bdrv_set_aio_context_ignore() because bdrv_can_set_aio_context() would have returned false previously, so instead of checking the condition, we can assert it. Signed-off-by: Kevin Wolf --- include/block/block.h | 9 --------- 1 file changed, 9 deletions(-) (limited to 'include') diff --git a/include/block/block.h b/include/block/block.h index 531cf59..13ea050 100644 --- a/include/block/block.h +++ b/include/block/block.h @@ -583,15 +583,6 @@ AioContext *bdrv_get_aio_context(BlockDriverState *bs); */ void bdrv_coroutine_enter(BlockDriverState *bs, Coroutine *co); -/** - * bdrv_set_aio_context: - * - * Changes the #AioContext used for fd handlers, timers, and BHs by this - * BlockDriverState and all its children. - * - * This function must be called with iothread lock held. - */ -void bdrv_set_aio_context(BlockDriverState *bs, AioContext *new_context); void bdrv_set_aio_context_ignore(BlockDriverState *bs, AioContext *new_context, GSList **ignore); int bdrv_try_set_aio_context(BlockDriverState *bs, AioContext *ctx, -- cgit v1.1 From d93e57268892d555d7c71ec30b25276b0d8132b6 Mon Sep 17 00:00:00 2001 From: Vladimir Sementsov-Ogievskiy Date: Tue, 23 Apr 2019 15:57:05 +0300 Subject: block/io: bdrv_pdiscard: support int64_t bytes parameter This fixes at least one overflow in qcow2_process_discards, which passes 64bit region length to bdrv_pdiscard where bytes (or sectors in the past) parameter is int since its introduction in 0b919fae. Signed-off-by: Vladimir Sementsov-Ogievskiy Signed-off-by: Kevin Wolf --- include/block/block.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'include') diff --git a/include/block/block.h b/include/block/block.h index 13ea050..f9415ed 100644 --- a/include/block/block.h +++ b/include/block/block.h @@ -434,8 +434,8 @@ void bdrv_drain_all(void); AIO_WAIT_WHILE(bdrv_get_aio_context(bs_), \ cond); }) -int bdrv_pdiscard(BdrvChild *child, int64_t offset, int bytes); -int bdrv_co_pdiscard(BdrvChild *child, int64_t offset, int bytes); +int bdrv_pdiscard(BdrvChild *child, int64_t offset, int64_t bytes); +int bdrv_co_pdiscard(BdrvChild *child, int64_t offset, int64_t bytes); int bdrv_has_zero_init_1(BlockDriverState *bs); int bdrv_has_zero_init(BlockDriverState *bs); bool bdrv_unallocated_blocks_are_zero(BlockDriverState *bs); -- cgit v1.1