Commit Graph

10 Commits

Author SHA1 Message Date
Rob Bradford
db5582f7d1 block_util: Preserve ordering of sync file backend completions
For the synchronous backends efficiently preserve the order for
completion requests through the use of VecDequeue. Preserving the order
is not required but is beneficial as it matches the existing
optimisation that looks to match completions and requests.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2023-01-10 17:30:25 +00:00
Rob Bradford
ce51755109 block_util: Avoid intermediate completion queue allocation
Rather than aggregate the completion list into an intermediate vector
instead adjust the API to provide one completion item at a time.

With DHAT this shows the number of heap allocations has decreased.

Before:

    dhat: Total:     623,852 bytes in 8,157 blocks

After:

    dhat: Total:     380,444 bytes in 3,469 blocks

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2023-01-10 17:30:25 +00:00
Rob Bradford
e7a51cde78 block_util: Pass a reference to the slice of iovecs
Rather than passing the vector of iovecs for the I/O to act on pass a
reference to the slice of values inside them. This removes the explicit
container type from the API allowing the use of e.g. SmallVec.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2023-01-10 17:30:25 +00:00
Bo Chen
a9ec0f33c0 misc: Fix clippy issues
Signed-off-by: Bo Chen <chen.bo@intel.com>
2022-11-02 09:41:43 +01:00
Wei Liu
af770c814b block_util: provide and use AsyncAdaptor trait
The observation is that the code in question was used to bridge
synchronized and asynchronized code.

We can group the functions for that purpose under an adaptor trait. To
limit the scope of locking, the users of the trait are required to
implement a method to return a MutexGuard for the underlying file.

This then allows us to use concrete types (QcowFile and Vhdx) in code,
which is easier to read than a bunch of traits.

No functional change intended.

Signed-off-by: Wei Liu <liuwe@microsoft.com>
2021-12-06 15:20:29 +01:00
Wei Liu
e1151482fc block_util: handle synchronized read/write/fsync idiomatically
Previously mutex (semaphore) and file were separated. The code needed to
create artificial scopes to use mutex to protect file.

Rewrite the code to be idiomatic. The file itself is turned into a trait
object and placed inside the mutex. This requires providing a new
ReadWriteSeekFile trait to unify all helper functions.

The rewrite further simplified vhdx_sync code. The original code
contained two mutex'es for no apparent reason.

No functional change intended.

Signed-off-by: Wei Liu <liuwe@microsoft.com>
2021-12-06 09:23:51 +00:00
Wei Liu
3e536f91eb block_util: drop disk_size
It is only used by qcow_sync code. Merge it to its caller.

No functional change.

Signed-off-by: Wei Liu <liuwe@microsoft.com>
2021-12-06 09:23:51 +00:00
Sebastien Boeuf
4918c1ca7f block_util, vmm: Propagate error on QcowDiskSync creation
Instead of panicking with an expect() function, the QcowDiskSync::new
function now propagates the error properly. This ensures the VMM will
not panic, which might be the source of weird errors if only one thread
exits while the VMM continues to run.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2021-08-11 16:44:28 -07:00
Sebastien Boeuf
060df69f07 block_util: Factorize common code for synchronous implementations
Since QCOW and RAW synchronous implementation are very close, it makes
sense to introduce some common functions that can be shared between
these two.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2021-01-22 16:10:34 +00:00
Sebastien Boeuf
12e20effd7 block_util: Port synchronous QCOW file to AsyncIo trait
Based on the synchronous QCOW file implementation present in the qcow
crate, we created a new qcow_sync module in block_util that ports this
synchronous implementation to the AsyncIo trait.

The point is to reuse virtio-blk asynchronous implementation for both
synchronous and asynchronous backends.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2021-01-22 16:10:34 +00:00