| Age | Commit message (Collapse) | Author |
|---|
|
|
|
No longer has any users, so remove it
|
|
|
|
|
|
|
|
|
|
|
|
There was a bug for awhile in previous kernels where we weren't
computing dirent name lengths correctly and we weren't zeroing out
padding at the end of dirents (due to struct bch_dirent changing size by
adding __attribute__((aligned)), and not updating other code to use
offsetof).
This patch fixes dirents with junk at the end, by going off of the
dirent's hash.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
|
|
|
|
|
|
|
|
Convert generic_file_buffered_read() to get pages to read from in
batches, and then copy data to userspace from many pages at once - in
particular, we now don't touch any cachelines that might be contended
while we're in the loop to copy data to userspace.
This is is a performance improvement on workloads that do buffered reads
with large blocksizes, and a very large performance improvement if that
file is also being accessed concurrently by different threads.
On smaller reads (512 bytes), there's a very small performance
improvement (1%, within the margin of error).
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
|
|
This is prep work for changing generic_file_buffered_read() to use
find_get_pages_contig() to batch up all the pagecache lookups.
This patch should be functionally identical to the existing code and
changes as little as of the flow control as possible. More refactoring
could be done, this patch is intended to be relatively minimal.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Prep work for extents insert hook removal
|
|
|
|
|
|
|
|
|
|
|
|
This lifts the restriction that 0 size extents must not overlap with
other extents, which means we can now sort extents and non extents the
same way, and will let us simplify a bunch of other stuff as well.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
|
|
|
|
also cleanups of btree node offsets
|
|
shouldn't be any reason for an actual binary search here
|
|
working on getting rid of all the reasons bch2_insert_fixup_extent() can
fail/stop partway, which is needed for other refactorings.
One of the reasons we could have to bail out is if we're splitting a
compressed extent we might need to add to our disk reservation - but we
can check that before actually starting the insert.
|
|
Add an option, gc_reserve_bytes, to set the copygc reserve as a size
instead of a percent
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
|
|
|
|
Due to compression, the different replicas of a replicated extent don't
necessarily have to take up the same amount of space - so replicated
data sector counts shouldn't be stored divided by the number of
replicas.
|
|
|
|
|
|
|
|
|
|
gc lock must be held while invalidating buckets - fixes
"1f7a95698e bcachefs: Invalidate buckets when writing to alloc btree"
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
|
|
was returning wrong value
|
|
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
|
|
Also improve error reporting - only return an error from
bch2_journal_flush_seq() if we had an error writing that entry (i.e. not
if there was an error with a newer entry).
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
|
|
|
|
|
|
|
|
|
|
Prep work for persistent alloc information. Refactoring also lets us
make free_inc much smaller, which means a lot fewer buckets stranded on
freelists.
|
|
|
|
journal_buf_switch is called from the foreground when getting a journal
reservation and thus is somewhat latency sensitive;
bch2_bucket_seq_cleanup has to run infrequently but is a bit expensive
when it does run.
Call it from the journal write path instead, and punt the journal write
to worqueue context.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
|
|
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
|
