Pere Diaz Bou
e2276c2e9d
O_DIRECT disable on WAL
2024-11-13 13:45:42 +00:00
Pere Diaz Bou
cbfb45e55b
more trace logs
2024-11-13 11:31:22 +00:00
Pere Diaz Bou
78118bae81
fmt
2024-11-13 09:34:36 +00:00
Pere Diaz Bou
8d1d2d36cc
core/linux: aligned wal header read
2024-11-12 17:53:59 +00:00
Pere Diaz Bou
7d8fc80f4b
use len instead of size
2024-11-07 14:01:35 +01:00
Pere Diaz Bou
0b46648c0e
write wal header on init
2024-11-07 13:29:57 +01:00
Pere Diaz Bou
5207e49399
remove extra borrow mut
2024-11-05 15:29:54 +01:00
Pere Diaz Bou
0f4270b48f
rebase submit entry
2024-11-05 15:29:54 +01:00
Pere Diaz Bou
70a4ccd8bb
fix linux completion match
2024-11-05 15:29:54 +01:00
Pere Diaz Bou
fc65c5096d
cacheflush state machine
2024-11-05 15:29:54 +01:00
Pere Diaz Bou
f009eb35c6
suspendable checkpoint
2024-11-05 15:29:54 +01:00
Arpit Saxena
b7debabd81
Wrap IoUring to ensure pending_ops is always correctly updated
...
Adds a struct WrappedIOUring which contains a IoUring and a pending_ops
field. Entry submission and popping from the queue is done through
functions operating on it, which also maintains pending_ops count
NOTE: This is a bit weird/hacky since in get_completion we create a
CompletionQueue and just call its next(). If it were a normal iterator
it would always return the same first item. However it is a queue posing
as an iterator which makes this work.
2024-09-28 00:10:05 +05:30
Arpit Saxena
a7d735d368
Maintain pending io_uring ops count for early return on run_once
...
Add a `pending_ops` field to `InnerLinuxIO` struct which is incremented
for each operation submitted to the ring and decremented when they are
taken off the completion queue. With this, we can exit from run_once if
there are no pending operations. Otherwise, in that case, it would hang
indefinitely due to call of `ring.submit_and_wait(1)`
2024-09-27 23:32:19 +05:30
mazchew
671b61ba19
add time querying to io trait
2024-08-25 22:55:15 +08:00
mazchew
156005694a
added rng to io trait for simulation
2024-08-22 23:24:02 +08:00
Pere Diaz Bou
d088640855
core: cqe result
...
Signed-off-by: Pere Diaz Bou <pere-altea@hotmail.com >
2024-07-31 17:27:02 +02:00
Pere Diaz Bou
84bf0ea96a
core: remove a bunch of warnings
...
Signed-off-by: Pere Diaz Bou <pere-altea@hotmail.com >
2024-07-31 17:27:02 +02:00
Pere Diaz Bou
845a1ea175
core: cacheflush and fix *Completion casting
2024-07-31 17:27:02 +02:00
Pekka Enberg
351242561d
Kill anyhow usage
...
Switch anyhow to explicit `LimboError` type using thiserror crate, which
lets us make error handling more structured.
2024-07-25 17:15:08 +03:00
gandeevanr
c391b7b9a6
implemented the drop trait for all the limbo_core::File implementations
2024-07-17 11:34:47 -07:00
gandeevanr
9ef212edd3
Added support to disallow multiple processes from opening the same database file in linux and darwin
2024-07-17 11:34:47 -07:00
Joan Martinez
b4cd8e39b5
core: validate result from CQ
2024-07-16 14:08:49 +02:00
Pekka Enberg
207ec5ce92
core: Enable direct I/O optimistically
...
...but keep going if the filesystem does not support it. Fixes Limbo on
older Linux kernels that have io_uring but don't support direct I/O on
all filesystems, such as tmpfs or encryptfs.
2024-07-16 14:48:38 +03:00
Pekka Enberg
812a8b9ea2
core: Switch io_ring to Readv and Writev
...
The Readv and Writev opcodes are available on all Linux kernel versions
so let's switch to them.
2024-07-16 14:24:12 +03:00
Ishan Jain
c171b63ec2
use O_DIRECT for file read operations
2024-07-09 11:41:20 +05:30
Kunal Singh
00c26286ce
fix: lint warnings
2024-07-08 22:43:11 +05:30
Piotr Jastrzebski
1004215853
Remove submit from LinuxIO write
...
Signed-off-by: Piotr Jastrzebski <haaawk@gmail.com >
2024-07-07 09:24:44 +02:00
Rutvik Patel
fd9af2739e
wait for the read completion
...
The submit_and_wait command waits for `n` completion events.
We were providing an incorrect argument, which caused the issue.
2024-07-06 13:48:58 +05:30
Pekka Enberg
2684341bac
Initial pass on deterministic simulator
...
The simulator does not do much, but does discover failures on Linux with
io_uring, though, so it's a start.
2024-06-27 18:32:04 +03:00
Pere Diaz Bou
377073e158
core: I/O write path
2024-06-19 21:26:40 +03:00
Pekka Enberg
ed9f3e6d1e
Single-threaded architecture
...
Use Rc instead of Arc and replace the concurrent LRU with
single-threaded SIEVE.
Fixes #23
Fixes #29
2024-03-03 12:44:45 +02:00
Pekka Enberg
8c2926e4d1
core/linux: Make io_uring larger
2024-01-24 17:25:33 +02:00
Pekka Enberg
d2a925698d
core/linux: Don't block in run_once()
2024-01-24 17:25:19 +02:00
Pekka Enberg
0b9b6cfd29
Asynchronous I/O
2024-01-24 08:14:53 +02:00
Pekka Enberg
2bb56ade4b
Tracing improvements
2024-01-14 15:47:54 +02:00
Pekka Enberg
44dcd48071
Fix WebAssembly build
2024-01-14 15:29:27 +02:00
Pekka Enberg
08b40f2af7
core/linux: Move I/O completion to io_uring
...
Move the Completion::complete() call to the io_uring machinery. We still
essentially do synchronous I/O because of the call to submit_and_wait()
in pread(), but let's take one step at a time...
2024-01-13 09:11:35 +02:00
Pekka Enberg
5ea214083a
core: Switch to using Arc<Completion>
...
We need to be able to leak a reference for io_uring so switch to
Arc<Completion> in the code.
2024-01-13 09:03:21 +02:00
Pekka Enberg
509b322f8a
Fix build on Linux
2023-11-11 09:21:13 +02:00
Pekka Enberg
2267ee121b
Introduce Completion type
2023-11-09 21:11:48 +02:00
Pekka Enberg
de6f327a4a
Move raw buffer to I/O module
2023-09-30 14:19:55 +03:00
Pekka Enberg
1e1e096a48
Refactor I/O and storage layers
2023-09-16 09:34:17 +03:00
Pekka Enberg
b49c242266
Rename PageSource to File
2023-09-10 22:01:45 +03:00
Pekka Enberg
fe41f46bc0
I/O trait
2023-09-10 21:57:36 +03:00
Pekka Enberg
58eab38066
I/O layer cleanups
2023-09-10 21:02:23 +03:00
