summaryrefslogtreecommitdiff
path: root/io_uring
diff options
context:
space:
mode:
Diffstat (limited to 'io_uring')
-rw-r--r--io_uring/fdinfo.c48
-rw-r--r--io_uring/io_uring.c1
-rw-r--r--io_uring/memmap.c2
-rw-r--r--io_uring/net.c14
-rw-r--r--io_uring/uring_cmd.c5
-rw-r--r--io_uring/uring_cmd.h6
6 files changed, 42 insertions, 34 deletions
diff --git a/io_uring/fdinfo.c b/io_uring/fdinfo.c
index 9414ca6d101c..e0d6a59a89fa 100644
--- a/io_uring/fdinfo.c
+++ b/io_uring/fdinfo.c
@@ -86,13 +86,8 @@ static inline void napi_show_fdinfo(struct io_ring_ctx *ctx,
}
#endif
-/*
- * Caller holds a reference to the file already, we don't need to do
- * anything else to get an extra reference.
- */
-__cold void io_uring_show_fdinfo(struct seq_file *m, struct file *file)
+static void __io_uring_show_fdinfo(struct io_ring_ctx *ctx, struct seq_file *m)
{
- struct io_ring_ctx *ctx = file->private_data;
struct io_overflow_cqe *ocqe;
struct io_rings *r = ctx->rings;
struct rusage sq_usage;
@@ -106,7 +101,6 @@ __cold void io_uring_show_fdinfo(struct seq_file *m, struct file *file)
unsigned int sq_entries, cq_entries;
int sq_pid = -1, sq_cpu = -1;
u64 sq_total_time = 0, sq_work_time = 0;
- bool has_lock;
unsigned int i;
if (ctx->flags & IORING_SETUP_CQE32)
@@ -176,15 +170,7 @@ __cold void io_uring_show_fdinfo(struct seq_file *m, struct file *file)
seq_printf(m, "\n");
}
- /*
- * Avoid ABBA deadlock between the seq lock and the io_uring mutex,
- * since fdinfo case grabs it in the opposite direction of normal use
- * cases. If we fail to get the lock, we just don't iterate any
- * structures that could be going away outside the io_uring mutex.
- */
- has_lock = mutex_trylock(&ctx->uring_lock);
-
- if (has_lock && (ctx->flags & IORING_SETUP_SQPOLL)) {
+ if (ctx->flags & IORING_SETUP_SQPOLL) {
struct io_sq_data *sq = ctx->sq_data;
/*
@@ -206,7 +192,7 @@ __cold void io_uring_show_fdinfo(struct seq_file *m, struct file *file)
seq_printf(m, "SqTotalTime:\t%llu\n", sq_total_time);
seq_printf(m, "SqWorkTime:\t%llu\n", sq_work_time);
seq_printf(m, "UserFiles:\t%u\n", ctx->file_table.data.nr);
- for (i = 0; has_lock && i < ctx->file_table.data.nr; i++) {
+ for (i = 0; i < ctx->file_table.data.nr; i++) {
struct file *f = NULL;
if (ctx->file_table.data.nodes[i])
@@ -218,7 +204,7 @@ __cold void io_uring_show_fdinfo(struct seq_file *m, struct file *file)
}
}
seq_printf(m, "UserBufs:\t%u\n", ctx->buf_table.nr);
- for (i = 0; has_lock && i < ctx->buf_table.nr; i++) {
+ for (i = 0; i < ctx->buf_table.nr; i++) {
struct io_mapped_ubuf *buf = NULL;
if (ctx->buf_table.nodes[i])
@@ -228,7 +214,7 @@ __cold void io_uring_show_fdinfo(struct seq_file *m, struct file *file)
else
seq_printf(m, "%5u: <none>\n", i);
}
- if (has_lock && !xa_empty(&ctx->personalities)) {
+ if (!xa_empty(&ctx->personalities)) {
unsigned long index;
const struct cred *cred;
@@ -238,7 +224,7 @@ __cold void io_uring_show_fdinfo(struct seq_file *m, struct file *file)
}
seq_puts(m, "PollList:\n");
- for (i = 0; has_lock && i < (1U << ctx->cancel_table.hash_bits); i++) {
+ for (i = 0; i < (1U << ctx->cancel_table.hash_bits); i++) {
struct io_hash_bucket *hb = &ctx->cancel_table.hbs[i];
struct io_kiocb *req;
@@ -247,9 +233,6 @@ __cold void io_uring_show_fdinfo(struct seq_file *m, struct file *file)
task_work_pending(req->tctx->task));
}
- if (has_lock)
- mutex_unlock(&ctx->uring_lock);
-
seq_puts(m, "CqOverflowList:\n");
spin_lock(&ctx->completion_lock);
list_for_each_entry(ocqe, &ctx->cq_overflow_list, list) {
@@ -262,4 +245,23 @@ __cold void io_uring_show_fdinfo(struct seq_file *m, struct file *file)
spin_unlock(&ctx->completion_lock);
napi_show_fdinfo(ctx, m);
}
+
+/*
+ * Caller holds a reference to the file already, we don't need to do
+ * anything else to get an extra reference.
+ */
+__cold void io_uring_show_fdinfo(struct seq_file *m, struct file *file)
+{
+ struct io_ring_ctx *ctx = file->private_data;
+
+ /*
+ * Avoid ABBA deadlock between the seq lock and the io_uring mutex,
+ * since fdinfo case grabs it in the opposite direction of normal use
+ * cases.
+ */
+ if (mutex_trylock(&ctx->uring_lock)) {
+ __io_uring_show_fdinfo(ctx, m);
+ mutex_unlock(&ctx->uring_lock);
+ }
+}
#endif
diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c
index 541e65a1eebf..edda31a15c6e 100644
--- a/io_uring/io_uring.c
+++ b/io_uring/io_uring.c
@@ -636,6 +636,7 @@ static void __io_cqring_overflow_flush(struct io_ring_ctx *ctx, bool dying)
* to care for a non-real case.
*/
if (need_resched()) {
+ ctx->cqe_sentinel = ctx->cqe_cached;
io_cq_unlock_post(ctx);
mutex_unlock(&ctx->uring_lock);
cond_resched();
diff --git a/io_uring/memmap.c b/io_uring/memmap.c
index 76fcc79656b0..07f8a5cbd37e 100644
--- a/io_uring/memmap.c
+++ b/io_uring/memmap.c
@@ -116,7 +116,7 @@ static int io_region_init_ptr(struct io_mapped_region *mr)
void *ptr;
if (io_check_coalesce_buffer(mr->pages, mr->nr_pages, &ifd)) {
- if (ifd.nr_folios == 1) {
+ if (ifd.nr_folios == 1 && !PageHighMem(mr->pages[0])) {
mr->ptr = page_address(mr->pages[0]);
return 0;
}
diff --git a/io_uring/net.c b/io_uring/net.c
index 24040bc3916a..27f37fa2ef79 100644
--- a/io_uring/net.c
+++ b/io_uring/net.c
@@ -827,18 +827,24 @@ static inline bool io_recv_finish(struct io_kiocb *req, int *ret,
cflags |= IORING_CQE_F_SOCK_NONEMPTY;
if (sr->flags & IORING_RECVSEND_BUNDLE) {
- cflags |= io_put_kbufs(req, *ret, io_bundle_nbufs(kmsg, *ret),
+ size_t this_ret = *ret - sr->done_io;
+
+ cflags |= io_put_kbufs(req, *ret, io_bundle_nbufs(kmsg, this_ret),
issue_flags);
if (sr->retry)
cflags = req->cqe.flags | (cflags & CQE_F_MASK);
/* bundle with no more immediate buffers, we're done */
if (req->flags & REQ_F_BL_EMPTY)
goto finish;
- /* if more is available, retry and append to this one */
- if (!sr->retry && kmsg->msg.msg_inq > 0 && *ret > 0) {
+ /*
+ * If more is available AND it was a full transfer, retry and
+ * append to this one
+ */
+ if (!sr->retry && kmsg->msg.msg_inq > 0 && this_ret > 0 &&
+ !iov_iter_count(&kmsg->msg.msg_iter)) {
req->cqe.flags = cflags & ~CQE_F_MASK;
sr->len = kmsg->msg.msg_inq;
- sr->done_io += *ret;
+ sr->done_io += this_ret;
sr->retry = true;
return false;
}
diff --git a/io_uring/uring_cmd.c b/io_uring/uring_cmd.c
index a9ea7d29cdd9..430ed620ddfe 100644
--- a/io_uring/uring_cmd.c
+++ b/io_uring/uring_cmd.c
@@ -254,6 +254,11 @@ int io_uring_cmd(struct io_kiocb *req, unsigned int issue_flags)
return -EOPNOTSUPP;
issue_flags |= IO_URING_F_IOPOLL;
req->iopoll_completed = 0;
+ if (ctx->flags & IORING_SETUP_HYBRID_IOPOLL) {
+ /* make sure every req only blocks once */
+ req->flags &= ~REQ_F_IOPOLL_STATE;
+ req->iopoll_start = ktime_get_ns();
+ }
}
ret = file->f_op->uring_cmd(ioucmd, issue_flags);
diff --git a/io_uring/uring_cmd.h b/io_uring/uring_cmd.h
index b04686b6b5d2..e6a5142c890e 100644
--- a/io_uring/uring_cmd.h
+++ b/io_uring/uring_cmd.h
@@ -17,9 +17,3 @@ bool io_uring_try_cancel_uring_cmd(struct io_ring_ctx *ctx,
struct io_uring_task *tctx, bool cancel_all);
void io_cmd_cache_free(const void *entry);
-
-int io_uring_cmd_import_fixed_vec(struct io_uring_cmd *ioucmd,
- const struct iovec __user *uvec,
- size_t uvec_segs,
- int ddir, struct iov_iter *iter,
- unsigned issue_flags);
generated by cgit (git 2.34.1) at 2025-05-23 12:17:08 +0000