All of lore.kernel.org
 help / color / mirror / Atom feed
From: Stefan Roesch <shr@fb.com>
To: <io-uring@vger.kernel.org>, <kernel-team@fb.com>
Cc: <shr@fb.com>, <joshi.k@samsung.com>
Subject: [PATCH v4 4/6] liburing: index large CQE's correctly
Date: Mon, 25 Apr 2022 11:51:26 -0700	[thread overview]
Message-ID: <20220425185128.2537966-5-shr@fb.com> (raw)
In-Reply-To: <20220425185128.2537966-1-shr@fb.com>

Large CQE's need to take into account that each CQE has double the size.
When the CQE array is indexed, the offset into the array needs to be
changed accordingly.

Signed-off-by: Stefan Roesch <shr@fb.com>
---
 src/include/liburing.h | 18 ++++++++++++++++--
 src/queue.c            |  6 +++++-
 2 files changed, 21 insertions(+), 3 deletions(-)

diff --git a/src/include/liburing.h b/src/include/liburing.h
index c01c231..317963c 100644
--- a/src/include/liburing.h
+++ b/src/include/liburing.h
@@ -188,6 +188,16 @@ int __io_uring_get_cqe(struct io_uring *ring,
 
 #define LIBURING_UDATA_TIMEOUT	((__u64) -1)
 
+/*
+ * Calculates the step size for CQE iteration.
+ * 	For standard CQE's its 1, for big CQE's its two.
+ */
+#define io_uring_cqe_shift(ring)					\
+	(!!((ring)->flags & IORING_SETUP_CQE32))
+
+#define io_uring_cqe_index(ring,ptr,mask)				\
+	(((ptr) & (mask)) << io_uring_cqe_shift(ring))
+
 #define io_uring_for_each_cqe(ring, head, cqe)				\
 	/*								\
 	 * io_uring_smp_load_acquire() enforces the order of tail	\
@@ -195,7 +205,7 @@ int __io_uring_get_cqe(struct io_uring *ring,
 	 */								\
 	for (head = *(ring)->cq.khead;					\
 	     (cqe = (head != io_uring_smp_load_acquire((ring)->cq.ktail) ? \
-		&(ring)->cq.cqes[head & (*(ring)->cq.kring_mask)] : NULL)); \
+		&(ring)->cq.cqes[io_uring_cqe_index(ring, head, *(ring)->cq.kring_mask)] : NULL)); \
 	     head++)							\
 
 /*
@@ -844,6 +854,10 @@ static inline int __io_uring_peek_cqe(struct io_uring *ring,
 	int err = 0;
 	unsigned available;
 	unsigned mask = *ring->cq.kring_mask;
+	int shift = 0;
+
+	if (ring->flags & IORING_SETUP_CQE32)
+		shift = 1;
 
 	do {
 		unsigned tail = io_uring_smp_load_acquire(ring->cq.ktail);
@@ -854,7 +868,7 @@ static inline int __io_uring_peek_cqe(struct io_uring *ring,
 		if (!available)
 			break;
 
-		cqe = &ring->cq.cqes[head & mask];
+		cqe = &ring->cq.cqes[(head & mask) << shift];
 		if (!(ring->features & IORING_FEAT_EXT_ARG) &&
 				cqe->user_data == LIBURING_UDATA_TIMEOUT) {
 			if (cqe->res < 0)
diff --git a/src/queue.c b/src/queue.c
index 2f85756..4ad41fc 100644
--- a/src/queue.c
+++ b/src/queue.c
@@ -132,6 +132,10 @@ unsigned io_uring_peek_batch_cqe(struct io_uring *ring,
 {
 	unsigned ready;
 	bool overflow_checked = false;
+	int shift = 0;
+
+	if (ring->flags & IORING_SETUP_CQE32)
+		shift = 1;
 
 again:
 	ready = io_uring_cq_ready(ring);
@@ -144,7 +148,7 @@ again:
 		count = count > ready ? ready : count;
 		last = head + count;
 		for (;head != last; head++, i++)
-			cqes[i] = &ring->cq.cqes[head & mask];
+			cqes[i] = &ring->cq.cqes[(head & mask) << shift];
 
 		return count;
 	}
-- 
2.30.2


  parent reply	other threads:[~2022-04-25 18:51 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-04-25 18:51 [PATCH v4 0/6]liburing: add support for large CQE sizes Stefan Roesch
2022-04-25 18:51 ` [PATCH v4 1/6] liburing: Update io_uring.h with large CQE kernel changes Stefan Roesch
2022-04-25 18:51 ` [PATCH v4 2/6] liburing: increase mmap size for large CQE's Stefan Roesch
2022-04-25 18:51 ` [PATCH v4 3/6] liburing: return correct ring " Stefan Roesch
2022-04-25 18:51 ` Stefan Roesch [this message]
2022-04-25 18:51 ` [PATCH v4 5/6] liburing: add large CQE tests to nop test Stefan Roesch
2022-04-25 18:51 ` [PATCH v4 6/6] liburing: Test all configurations with NOP test Stefan Roesch

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20220425185128.2537966-5-shr@fb.com \
    --to=shr@fb.com \
    --cc=io-uring@vger.kernel.org \
    --cc=joshi.k@samsung.com \
    --cc=kernel-team@fb.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.