- 12 Apr, 2021 37 commits
-
-
Pavel Begunkov authored
Now __io_uring_cancel() and __io_uring_files_cancel() are very similar and mostly differ by how we count requests, merge them and allow tctx_inflight() to handle counting. Signed-off-by:
Pavel Begunkov <asml.silence@gmail.com> Link: https://lore.kernel.org/r/1a5986a97df4dc1378f3fe0ca1eb483dbcf42112.1618101759.git.asml.silence@gmail.comSigned-off-by:
Jens Axboe <axboe@kernel.dk>
-
Pavel Begunkov authored
Instead of keeping requests in a inflight_list, just track them with a per tctx atomic counter. Apart from it being much easier and more consistent with task cancel, it frees ->inflight_entry from being shared between iopoll and cancel-track, so less headache for us. Signed-off-by:
-
Pavel Begunkov authored
Move tracked inflight number check up the stack into __io_uring_files_cancel() so it's similar to task cancel. Will be used for further cleaning. Signed-off-by:
-
Pavel Begunkov authored
hash_del() works well with non-hashed nodes, there's no need to check if it is hashed first. Signed-off-by:
-
Pavel Begunkov authored
Remove error parameter from io_poll_complete(), 0 is always passed, and do a bit of cleaning on top. Signed-off-by:
-
Pavel Begunkov authored
io_poll_complete() always fills an event (even an overflowed one), so we always should do io_cqring_ev_posted() afterwards. And that's what is currently happening, because second EPOLLONESHOT check is always true, it can't return !done for oneshots. Remove those branching, it's much easier to read. Signed-off-by:
-
Peter Zijlstra authored
Do not include private headers and do not frob in internals. On top of that, while the previous code restores the affinity, it doesn't ensure the task actually moves there if it was running, leading to the fun situation that it can be observed running outside of its allowed mask for potentially significant time. Use the proper API instead. Signed-off-by:
Peter Zijlstra (Intel) <peterz@infradead.org> Link: https://lore.kernel.org/r/YG7QkiUzlEbW85TU@hirez.programming.kicks-ass.netSigned-off-by:
-
Jens Axboe authored
We currently allow racy updates to multishot requests, but we can end up double adding the poll request if both completion and update does it. Ensure that we skip re-add on the update side if someone else is completing it. Fixes: b69de288 ("io_uring: allow events and user_data update of running poll requests") Reported-by:
Joakim Hassila <joj@mac.com> Signed-off-by:
-
Hao Xu authored
Leverage XOR to simplify the code in __io_worker_busy. Signed-off-by:
Hao Xu <haoxu@linux.alibaba.com> Link: https://lore.kernel.org/r/1617678525-3129-1-git-send-email-haoxu@linux.alibaba.comSigned-off-by:
-
Pavel Begunkov authored
The splice/tee comment in io_prep_async_work() isn't relevant since the section was moved, delete it. Signed-off-by:
-
Pavel Begunkov authored
Add struct io_fixed_file representing a single registered file, first to hide ugly struct file **, which may be misleading, and secondly to retype it to unsigned long as conversions to it and back to file * for handling and masking FFS_* flags are getting nasty. Signed-off-by:
-
Pavel Begunkov authored
Introduce a heler io_free_file_tables() doing all the cleaning, there are several places where it's hand coded. Also move all allocations into io_sqe_alloc_file_tables() and rename it, so all of it is in one place. Signed-off-by:
-
Pavel Begunkov authored
There is no reason why we would want to fully quiesce ring on IORING_REGISTER_FILES, if it's already registered we fail. Signed-off-by:
-
Pavel Begunkov authored
Set FFS_* flags (e.g. FFS_ASYNC_READ) not only in initial registration but also on registered files update. Not a bug, but may miss getting profit out of the feature. Signed-off-by:
-
Pavel Begunkov authored
Set MSG_NOSIGNAL and REQ_F_NOWAIT in send/recv prep routines and don't duplicate it in all four send/recv handlers. Signed-off-by:
-
Pavel Begunkov authored
Don't put linked timeout req in io_async_find_and_cancel() but do it in io_link_timeout_fn(), so we have only one point for that and won't have to do it differently as it's now (put vs put_deferred). Btw, improve a bit io_async_find_and_cancel()'s locking. Signed-off-by:
-
Pavel Begunkov authored
Overflowed CQEs doesn't lock requests anymore, so we don't care so much about cancelling them, so kill cq_overflow_flushed and simplify the code. Signed-off-by:
-
Pavel Begunkov authored
Add timeout and poll ->comletion_lock annotations for Sparse, makes life easier while looking at the functions. Signed-off-by:
-
Pavel Begunkov authored
Kill unused forward declarations for io_ring_file_put() and io_queue_next(). Also btw rename the first one. Signed-off-by:
-
Pavel Begunkov authored
It's a bit more convenient for us to store a registered buffer end address instead of length, see struct io_mapped_ubuf, as it allow to not recompute it every time. Signed-off-by:
-
Pavel Begunkov authored
Replace a hand-coded overflow check with a specialised function. Even though compilers are smart enough to generate identical binary (i.e. check carry bit), but it's more foolproof and conveys the intention better. Signed-off-by:
-
Pavel Begunkov authored
Remove extra tctx==NULL checks that are already done by io_async_cancel_one(). Signed-off-by:
-
Pavel Begunkov authored
No users of io_uring_ctx::hash_wait left, kill it. Signed-off-by:
-
Pavel Begunkov authored
Instead of io_put_req() to drop not a final ref, use req_ref_put(), which is slimmer and will also check the invariant. Signed-off-by:
-
Pavel Begunkov authored
io_ring_exit_work() already does uring_lock lock/unlock, no need to repeat it for lock waiting trick in io_ring_ctx_free(). Move the waiting with comments and spinlocking into io_ring_exit_work. Signed-off-by:
-
Pavel Begunkov authored
rsrc_data refs should always be valid for potential submitters, io_rsrc_ref_quiesce() restores it before unlocking, so percpu_ref_is_dying() check in io_sqe_files_unregister() does nothing and misleading. Concurrent quiesce is prevented with struct io_rsrc_data::quiesce. Signed-off-by:
-
Pavel Begunkov authored
Reuse io_rsrc_node_destroy() in __io_rsrc_put_work(). Also move it to a more appropriate place -- to the other node routines, and remove forward declaration. Signed-off-by:
-
Pavel Begunkov authored
If we're going to ever support multiple types of resources we need shared rsrc nodes to not bloat requests, that is implemented in this patch. It also gives a nicer API and saves one pointer dereference in io_req_set_rsrc_node(). We may say that all requests bound to a resource belong to one and only one rsrc node, and considering that nodes are removed and recycled strictly in-order, this separates requests into generations, where generation are changed on each node switch (i.e. io_rsrc_node_switch()). The API is simple, io_rsrc_node_switch() switches to a new generation if needed, and also optionally kills a passed in io_rsrc_data. Each call to io_rsrc_node_switch() have to be preceded with io_rsrc_node_switch_start(). The start function is idempotent and should not necessarily be followed by switch. One difference is that once a node was set it will always retain a valid rsrc node, even on unregister. It may be a nuisance at the moment, but makes much sense for multiple types of resources. Another thing changed is that nodes are bound to/associated with a io_rsrc_data later just before killing (i.e. switching). Signed-off-by:
-
Pavel Begunkov authored
Pass rsrc_node into io_queue_rsrc_removal() explicitly. Just a simple preparation patch, makes following changes nicer. Signed-off-by:
-
Pavel Begunkov authored
io_rsrc_node's callback operates only on a single io_rsrc_data and only with its resources, so rsrc_put() callback is actually a property of io_rsrc_data. Move it there, it makes code much nicecr. Signed-off-by:
-
Pavel Begunkov authored
io_rsrc_node_get() and io_rsrc_node_set() are always used together, merge them into one so most users don't even see io_rsrc_node and don't need to care about it. It helped to catch io_sqe_files_register() inferring rsrc data argument for get and set differently, not a problem but a good sign. Signed-off-by:
-
Pavel Begunkov authored
Keep it consistent with update and use io_rsrc_node_prealloc() + io_rsrc_node_get() in io_sqe_files_register() as well, that will be used in future patches, not as error prone and allows to deduplicate rsrc_node init. Signed-off-by:
-
Pavel Begunkov authored
Replace queue_delayed_work() with mod_delayed_work() in io_rsrc_node_ref_zero() as the later one can schedule a new work, and cleanup it further for better readability. Signed-off-by:
-
Pavel Begunkov authored
Keep resource related structs' and functions' naming consistent, in particular use "io_rsrc" prefix for everything. Signed-off-by:
-
Jens Axboe authored
With using task_work_cancel(), we're potentially canceling task_work that isn't related to this specific io_wq. Use the newly added task_work_cancel_match() to ensure that we only remove and cancel work items that are specific to this io_wq. Fixes: 685fe7fe ("io-wq: eliminate the need for a manager thread") Signed-off-by:
-
Jens Axboe authored
The only exported helper we have right now is task_work_cancel(), which cancels any task_work from a given task where func matches the queued work item. This is a bit too coarse for some use cases. Add a task_work_cancel_match() that allows to more specifically target individual work items outside of purely the callback function used. task_work_cancel() can be trivially implemented on top of that, hence do so. Reviewed-by:
Oleg Nesterov <oleg@redhat.com> Signed-off-by:
-
Jens Axboe authored
Joakim reports that in some conditions he sees a multishot poll request being canceled, and that it coincides with getting -EALREADY on modification. As part of the poll update procedure, there's a small window where the request is marked as canceled, and if this coincides with the event actually triggering, then we can get a spurious -ECANCELED and termination of the multishot request. Don't mark the poll request as being canceled for update. We also don't care if we race on removal unless it's a one-shot request, we can safely updated for either case. Fixes: b69de288 ("io_uring: allow events and user_data update of running poll requests") Reported-by:
-
- 11 Apr, 2021 3 commits
-
-
Pavel Begunkov authored
We are safe with overflows in io_sqe_buffer_register() because it will just yield alloc failure, but it's nicer to check explicitly. Signed-off-by:
-
Jens Axboe authored
Now that we have any worker being attached to the original task as threads, accounting of CPU time is directly attributed to the original task as well. This means that we no longer have to restrict SQPOLL to needing elevated privileges, as it's really no different from just having the task spawn a busy looping thread in userspace. Reported-by:
Stefano Garzarella <sgarzare@redhat.com> Signed-off-by:
-
Jens Axboe authored
io-wq relies on a manager thread to create/fork new workers, as needed. But there's really no strong need for it anymore. We have the following cases that fork a new worker: 1) Work queue. This is done from the task itself always, and it's trivial to create a worker off that path, if needed. 2) All workers have gone to sleep, and we have more work. This is called off the sched out path. For this case, use a task_work items to queue a fork-worker operation. 3) Hashed work completion. Don't think we need to do anything off this case. If need be, it could just use approach 2 as well. Part of this change is incrementing the running worker count before the fork, to avoid cases where we observe we need a worker and then queue creation of one. Then new work comes in, we fork a new one. That last queue operation should have waited for the previous worker to come up, it's quite possible we don't even need it. Hence move the worker running from before we fork it off to more efficiently handle that case. Signed-off-by:
-
