Skip to contentSkip to conversation

small rings clean up

PATCH Applied Pavel Begunkov 3 participants 12 messages Mar 13, 2024 io_uring

lore
Share & tools
lore
#1 lore raw

PATCH 1/3 io_uring: simplify io_mem_alloc return values

In reply to Pavel Begunkov

io_mem_alloc() returns a pointer on success and a pointer-encoded error otherwise. However, it can only fail with -ENOMEM, just return NULL on failure. PTR_ERR is usually pretty error prone.

---
 io_uring/io_uring.c | 14 +++++---------
 io_uring/kbuf.c     |  4 ++--
 2 files changed, 7 insertions(+), 11 deletions(-)
io_uring/io_uring.c +5 -9
index e7d7a456b489..1d0eac0cc8aa 100644
@@ -2802,12 +2802,8 @@ static void io_rings_free(struct io_ring_ctx *ctx)
 void *io_mem_alloc(size_t size)
 {
 	gfp_t gfp = GFP_KERNEL_ACCOUNT | __GFP_ZERO | __GFP_NOWARN | __GFP_COMP;
-	void *ret;
 
-	ret = (void *) __get_free_pages(gfp, get_order(size));
-	if (ret)
-		return ret;
-	return ERR_PTR(-ENOMEM);
+	return (void *) __get_free_pages(gfp, get_order(size));
 }
 
 static unsigned long rings_size(struct io_ring_ctx *ctx, unsigned int sq_entries,
@@ -3762,8 +3758,8 @@ static __cold int io_allocate_scq_urings(struct io_ring_ctx *ctx,
 	else
 		rings = io_rings_map(ctx, p->cq_off.user_addr, size);
 
-	if (IS_ERR(rings))
-		return PTR_ERR(rings);
+	if (!rings)
+		return -ENOMEM;
 
 	ctx->rings = rings;
 	if (!(ctx->flags & IORING_SETUP_NO_SQARRAY))
@@ -3787,9 +3783,9 @@ static __cold int io_allocate_scq_urings(struct io_ring_ctx *ctx,
 	else
 		ptr = io_sqes_map(ctx, p->sq_off.user_addr, size);
 
-	if (IS_ERR(ptr)) {
+	if (!ptr) {
 		io_rings_free(ctx);
-		return PTR_ERR(ptr);
+		return -ENOMEM;
 	}
 
 	ctx->sq_sqes = ptr;
io_uring/kbuf.c +2 -2
index 9be42bff936b..0677eae92e79 100644
@@ -627,8 +627,8 @@ static int io_alloc_pbuf_ring(struct io_ring_ctx *ctx,
 	ibf = io_lookup_buf_free_entry(ctx, ring_size);
 	if (!ibf) {
 		ptr = io_mem_alloc(ring_size);
-		if (IS_ERR(ptr))
-			return PTR_ERR(ptr);
+		if (!ptr)
+			return -ENOMEM;
 
 		/* Allocate and store deferred free entry */
 		ibf = kmalloc(sizeof(*ibf), GFP_KERNEL_ACCOUNT);
Signature · 1 line
2.43.0
#9 lore raw

PATCH 2/3 io_uring: simplify io_pages_free

In reply to Pavel Begunkov

We never pass a null (top-level) pointer, remove the check.

---
 io_uring/io_uring.c | 6 +-----
 1 file changed, 1 insertion(+), 5 deletions(-)
io_uring/io_uring.c +1 -5
index 1d0eac0cc8aa..cfe1413511ca 100644
@@ -2696,13 +2696,9 @@ void io_mem_free(void *ptr)
 
 static void io_pages_free(struct page ***pages, int npages)
 {
-	struct page **page_array;
+	struct page **page_array = *pages;
 	int i;
 
-	if (!pages)
-		return;
-
-	page_array = *pages;
 	if (!page_array)
 		return;
 
Signature · 1 line
2.43.0
#10 lore raw

PATCH 3/3 io_uring/kbuf: rename is_mapped

In reply to Pavel Begunkov

In buffer lists we have ->is_mapped as well as ->is_mmap, it's pretty hard to stay sane double checking which one means what, and in the long run there is a high chance of an eventual bug. Rename ->is_mapped into ->is_buf_ring.

---
 io_uring/kbuf.c | 20 ++++++++++----------
 io_uring/kbuf.h |  2 +-
 2 files changed, 11 insertions(+), 11 deletions(-)
io_uring/kbuf.c +10 -10
index 0677eae92e79..4cf742749870 100644
@@ -199,7 +199,7 @@ void __user *io_buffer_select(struct io_kiocb *req, size_t *len,
 
 	bl = io_buffer_get_list(ctx, req->buf_index);
 	if (likely(bl)) {
-		if (bl->is_mapped)
+		if (bl->is_buf_ring)
 			ret = io_ring_buffer_select(req, len, bl, issue_flags);
 		else
 			ret = io_provided_buffer_select(req, len, bl);
@@ -253,7 +253,7 @@ static int __io_remove_buffers(struct io_ring_ctx *ctx,
 	if (!nbufs)
 		return 0;
 
-	if (bl->is_mapped) {
+	if (bl->is_buf_ring) {
 		i = bl->buf_ring->tail - bl->head;
 		if (bl->is_mmap) {
 			/*
@@ -274,7 +274,7 @@ static int __io_remove_buffers(struct io_ring_ctx *ctx,
 		}
 		/* make sure it's seen as empty */
 		INIT_LIST_HEAD(&bl->buf_list);
-		bl->is_mapped = 0;
+		bl->is_buf_ring = 0;
 		return i;
 	}
 
@@ -361,7 +361,7 @@ int io_remove_buffers(struct io_kiocb *req, unsigned int issue_flags)
 	if (bl) {
 		ret = -EINVAL;
 		/* can't use provide/remove buffers command on mapped buffers */
-		if (!bl->is_mapped)
+		if (!bl->is_buf_ring)
 			ret = __io_remove_buffers(ctx, bl, p->nbufs);
 	}
 	io_ring_submit_unlock(ctx, issue_flags);
@@ -519,7 +519,7 @@ int io_provide_buffers(struct io_kiocb *req, unsigned int issue_flags)
 		}
 	}
 	/* can't add buffers via this command for a mapped buffer ring */
-	if (bl->is_mapped) {
+	if (bl->is_buf_ring) {
 		ret = -EINVAL;
 		goto err;
 	}
@@ -575,7 +575,7 @@ static int io_pin_pbuf_ring(struct io_uring_buf_reg *reg,
 	bl->buf_pages = pages;
 	bl->buf_nr_pages = nr_pages;
 	bl->buf_ring = br;
-	bl->is_mapped = 1;
+	bl->is_buf_ring = 1;
 	bl->is_mmap = 0;
 	return 0;
 error_unpin:
@@ -642,7 +642,7 @@ static int io_alloc_pbuf_ring(struct io_ring_ctx *ctx,
 	}
 	ibf->inuse = 1;
 	bl->buf_ring = ibf->mem;
-	bl->is_mapped = 1;
+	bl->is_buf_ring = 1;
 	bl->is_mmap = 1;
 	return 0;
 }
@@ -688,7 +688,7 @@ int io_register_pbuf_ring(struct io_ring_ctx *ctx, void __user *arg)
 	bl = io_buffer_get_list(ctx, reg.bgid);
 	if (bl) {
 		/* if mapped buffer ring OR classic exists, don't allow */
-		if (bl->is_mapped || !list_empty(&bl->buf_list))
+		if (bl->is_buf_ring || !list_empty(&bl->buf_list))
 			return -EEXIST;
 	} else {
 		free_bl = bl = kzalloc(sizeof(*bl), GFP_KERNEL);
@@ -730,7 +730,7 @@ int io_unregister_pbuf_ring(struct io_ring_ctx *ctx, void __user *arg)
 	bl = io_buffer_get_list(ctx, reg.bgid);
 	if (!bl)
 		return -ENOENT;
-	if (!bl->is_mapped)
+	if (!bl->is_buf_ring)
 		return -EINVAL;
 
 	__io_remove_buffers(ctx, bl, -1U);
@@ -757,7 +757,7 @@ int io_register_pbuf_status(struct io_ring_ctx *ctx, void __user *arg)
 	bl = io_buffer_get_list(ctx, buf_status.buf_group);
 	if (!bl)
 		return -ENOENT;
-	if (!bl->is_mapped)
+	if (!bl->is_buf_ring)
 		return -EINVAL;
 
 	buf_status.head = bl->head;
io_uring/kbuf.h +1 -1
index 5218bfd79e87..1c7b654ee726 100644
@@ -26,7 +26,7 @@ struct io_buffer_list {
 	__u16 mask;
 
 	/* ring mapped provided buffers */
-	__u8 is_mapped;
+	__u8 is_buf_ring;
 	/* ring mapped provided buffers, but mmap'ed by application */
 	__u8 is_mmap;
 	/* bl is visible from an RCU point of view for lookup */
Signature · 1 line
2.43.0
#0 lore raw

cover small rings clean up

Some disconnected cleanup patches split out of a larger series as they make sense by themselves.

Pavel Begunkov (3):
io_uring: simplify io_mem_alloc return values
io_uring: simplify io_pages_free
io_uring/kbuf: rename is_mapped

 io_uring/io_uring.c | 20 ++++++--------------
 io_uring/kbuf.c     | 24 ++++++++++++------------
 io_uring/kbuf.h     |  2 +-
 3 files changed, 19 insertions(+), 27 deletions(-)
Signature · 1 line
2.43.0
maintainer Applied
#11 lore raw
Quoted 9 lines · Pavel Begunkov

Some disconnected cleanup patches split out of a larger series as they make sense by themselves.

Pavel Begunkov (3):
io_uring: simplify io_mem_alloc return values
io_uring: simplify io_pages_free
io_uring/kbuf: rename is_mapped

[...]

Applied, thanks!

[1/3] io_uring: simplify io_mem_alloc return values

      commit: ccd2254dfe739e37876e984c44caa6c7f318c728
[2/3] io_uring: simplify io_pages_free
      commit: 91e0c2c9640c959f1f50f6f80a6a7e93562a9272
[3/3] io_uring/kbuf: rename is_mapped
      commit: 585abfacf0c33b11ab66a69ba884a70c9459a63c

Best regards,

Signature · 1 line
Jens Axboe
maintainer
#2 lore raw

Re: [PATCH 1/3] io_uring: simplify io_mem_alloc return values

In reply to Pavel Begunkov
Quoted 3 lines · Pavel Begunkov

io_mem_alloc() returns a pointer on success and a pointer-encoded error otherwise. However, it can only fail with -ENOMEM, just return NULL on failure. PTR_ERR is usually pretty error prone.

I take that back, this is buggy - the io_rings_map() and friends return an error pointer. So better to keep it consistent. Dropped this one.

Signature · 1 line
Jens Axboe
#3 lore raw

Re: [PATCH 1/3] io_uring: simplify io_mem_alloc return values

Quoted 7 lines · Jens Axboe
Quoted 3 lines · Pavel Begunkov

io_mem_alloc() returns a pointer on success and a pointer-encoded error otherwise. However, it can only fail with -ENOMEM, just return NULL on failure. PTR_ERR is usually pretty error prone.

I take that back, this is buggy - the io_rings_map() and friends return an error pointer. So better to keep it consistent. Dropped this one.

Oh crap, you're right. Did something trigger it? Because tests are suspiciously silent, I'd assume it's not really tested, e.g. passing misaligned uptr

Signature · 1 line
Pavel Begunkov
maintainer
#4 lore raw

Re: [PATCH 1/3] io_uring: simplify io_mem_alloc return values

Quoted 11 lines · Pavel Begunkov
Quoted 7 lines · Jens Axboe
Quoted 3 lines · Pavel Begunkov

io_mem_alloc() returns a pointer on success and a pointer-encoded error otherwise. However, it can only fail with -ENOMEM, just return NULL on failure. PTR_ERR is usually pretty error prone.

I take that back, this is buggy - the io_rings_map() and friends return an error pointer. So better to keep it consistent. Dropped this one.

Oh crap, you're right. Did something trigger it? Because tests are suspiciously silent, I'd assume it's not really tested, e.g. passing misaligned uptr

I was doing a cleanup on top, and noticed it while doing that. None of the tests triggered it.

Signature · 1 line
Jens Axboe
#5 lore raw

Re: [PATCH 1/3] io_uring: simplify io_mem_alloc return values

In reply to Pavel Begunkov
Quoted 29 lines · Pavel Begunkov

io_mem_alloc() returns a pointer on success and a pointer-encoded error otherwise. However, it can only fail with -ENOMEM, just return NULL on failure. PTR_ERR is usually pretty error prone.

---
 io_uring/io_uring.c | 14 +++++---------
 io_uring/kbuf.c     |  4 ++--
 2 files changed, 7 insertions(+), 11 deletions(-)
io_uring/io_uring.c +1 -5
index e7d7a456b489..1d0eac0cc8aa 100644
@@ -2802,12 +2802,8 @@ static void io_rings_free(struct io_ring_ctx *ctx)
 void *io_mem_alloc(size_t size)
 {
 	gfp_t gfp = GFP_KERNEL_ACCOUNT | __GFP_ZERO | __GFP_NOWARN | __GFP_COMP;
-	void *ret;
 
-	ret = (void *) __get_free_pages(gfp, get_order(size));
-	if (ret)
-		return ret;
-	return ERR_PTR(-ENOMEM);
+	return (void *) __get_free_pages(gfp, get_order(size));
 }
 
 static unsigned long rings_size(struct io_ring_ctx *ctx, unsigned int sq_entries,

io_uring/io_uring.c static __cold int io_allocate_scq_urings(struct io_ring_ctx *ctx, · line 3763

Quoted 8 lines
io_uring/io_uring.c +2 -2
@@ -3762,8 +3758,8 @@ static __cold int io_allocate_scq_urings(struct io_ring_ctx *ctx,
 	else
 		rings = io_rings_map(ctx, p->cq_off.user_addr, size);
 
-	if (IS_ERR(rings))
-		return PTR_ERR(rings);
+	if (!rings)
+		return -ENOMEM;
 

Sorry, I started reviewing this, got excited about the error path quick fix, and didn't finish the review before it got it.

I think this change is broken for the ctx->flags & IORING_SETUP_NO_MMAP case, because io_rings_map returns ERR_PTR, and not NULL. In addition, io_rings_map might fail for multiple reasons, and we want to propagate the different error codes up here.

Signature · 1 line
Gabriel Krisman Bertazi
maintainer
#6 lore raw

Re: [PATCH 1/3] io_uring: simplify io_mem_alloc return values

Quoted 43 lines · Gabriel Krisman Bertazi
Quoted 29 lines · Pavel Begunkov

io_mem_alloc() returns a pointer on success and a pointer-encoded error otherwise. However, it can only fail with -ENOMEM, just return NULL on failure. PTR_ERR is usually pretty error prone.

---
 io_uring/io_uring.c | 14 +++++---------
 io_uring/kbuf.c     |  4 ++--
 2 files changed, 7 insertions(+), 11 deletions(-)
io_uring/io_uring.c +1 -5
index e7d7a456b489..1d0eac0cc8aa 100644
@@ -2802,12 +2802,8 @@ static void io_rings_free(struct io_ring_ctx *ctx)
 void *io_mem_alloc(size_t size)
 {
 	gfp_t gfp = GFP_KERNEL_ACCOUNT | __GFP_ZERO | __GFP_NOWARN | __GFP_COMP;
-	void *ret;
 
-	ret = (void *) __get_free_pages(gfp, get_order(size));
-	if (ret)
-		return ret;
-	return ERR_PTR(-ENOMEM);
+	return (void *) __get_free_pages(gfp, get_order(size));
 }
 
 static unsigned long rings_size(struct io_ring_ctx *ctx, unsigned int sq_entries,

io_uring/io_uring.c static __cold int io_allocate_scq_urings(struct io_ring_ctx *ctx, · line 3763

Quoted 8 lines
io_uring/io_uring.c +2 -2
@@ -3762,8 +3758,8 @@ static __cold int io_allocate_scq_urings(struct io_ring_ctx *ctx,
 	else
 		rings = io_rings_map(ctx, p->cq_off.user_addr, size);
 
-	if (IS_ERR(rings))
-		return PTR_ERR(rings);
+	if (!rings)
+		return -ENOMEM;
 

Sorry, I started reviewing this, got excited about the error path quick fix, and didn't finish the review before it got it.

Quoted 4 lines

I think this change is broken for the ctx->flags & IORING_SETUP_NO_MMAP case, because io_rings_map returns ERR_PTR, and not NULL. In addition, io_rings_map might fail for multiple reasons, and we want to propagate the different error codes up here.

Yeah, see my reply from some hours ago. I dropped it back then.

Signature · 1 line
Jens Axboe
#7 lore raw

Re: [PATCH 1/3] io_uring: simplify io_mem_alloc return values

Quoted 44 lines · Jens Axboe
Quoted 43 lines · Gabriel Krisman Bertazi
Quoted 29 lines · Pavel Begunkov

io_mem_alloc() returns a pointer on success and a pointer-encoded error otherwise. However, it can only fail with -ENOMEM, just return NULL on failure. PTR_ERR is usually pretty error prone.

---
 io_uring/io_uring.c | 14 +++++---------
 io_uring/kbuf.c     |  4 ++--
 2 files changed, 7 insertions(+), 11 deletions(-)
io_uring/io_uring.c +1 -5
index e7d7a456b489..1d0eac0cc8aa 100644
@@ -2802,12 +2802,8 @@ static void io_rings_free(struct io_ring_ctx *ctx)
 void *io_mem_alloc(size_t size)
 {
 	gfp_t gfp = GFP_KERNEL_ACCOUNT | __GFP_ZERO | __GFP_NOWARN | __GFP_COMP;
-	void *ret;
 
-	ret = (void *) __get_free_pages(gfp, get_order(size));
-	if (ret)
-		return ret;
-	return ERR_PTR(-ENOMEM);
+	return (void *) __get_free_pages(gfp, get_order(size));
 }
 
 static unsigned long rings_size(struct io_ring_ctx *ctx, unsigned int sq_entries,

io_uring/io_uring.c static __cold int io_allocate_scq_urings(struct io_ring_ctx *ctx, · line 3763

Quoted 8 lines
io_uring/io_uring.c +2 -2
@@ -3762,8 +3758,8 @@ static __cold int io_allocate_scq_urings(struct io_ring_ctx *ctx,
 	else
 		rings = io_rings_map(ctx, p->cq_off.user_addr, size);
 
-	if (IS_ERR(rings))
-		return PTR_ERR(rings);
+	if (!rings)
+		return -ENOMEM;
 

Sorry, I started reviewing this, got excited about the error path quick fix, and didn't finish the review before it got it.

Quoted 6 lines
Quoted 4 lines

I think this change is broken for the ctx->flags & IORING_SETUP_NO_MMAP case, because io_rings_map returns ERR_PTR, and not NULL. In addition, io_rings_map might fail for multiple reasons, and we want to propagate the different error codes up here.

Yeah, see my reply from some hours ago. I dropped it back then.

ah, thanks. I've configured lei to fetch the io_uring list every few hours. This ended up fetching part of the thread at first, and I only saw it dropped in the next fetch, after I sent the email. sorry for the noise.

Signature · 1 line
Gabriel Krisman Bertazi
maintainer
#8 lore raw

Re: [PATCH 1/3] io_uring: simplify io_mem_alloc return values

Quoted 51 lines · Gabriel Krisman Bertazi
Quoted 49 lines · Jens Axboe
Quoted 47 lines · Gabriel Krisman Bertazi
Quoted 29 lines · Pavel Begunkov

io_mem_alloc() returns a pointer on success and a pointer-encoded error otherwise. However, it can only fail with -ENOMEM, just return NULL on failure. PTR_ERR is usually pretty error prone.

---
 io_uring/io_uring.c | 14 +++++---------
 io_uring/kbuf.c     |  4 ++--
 2 files changed, 7 insertions(+), 11 deletions(-)
io_uring/io_uring.c +1 -5
index e7d7a456b489..1d0eac0cc8aa 100644
@@ -2802,12 +2802,8 @@ static void io_rings_free(struct io_ring_ctx *ctx)
 void *io_mem_alloc(size_t size)
 {
 	gfp_t gfp = GFP_KERNEL_ACCOUNT | __GFP_ZERO | __GFP_NOWARN | __GFP_COMP;
-	void *ret;
 
-	ret = (void *) __get_free_pages(gfp, get_order(size));
-	if (ret)
-		return ret;
-	return ERR_PTR(-ENOMEM);
+	return (void *) __get_free_pages(gfp, get_order(size));
 }
 
 static unsigned long rings_size(struct io_ring_ctx *ctx, unsigned int sq_entries,

io_uring/io_uring.c static __cold int io_allocate_scq_urings(struct io_ring_ctx *ctx, · line 3763

Quoted 8 lines
io_uring/io_uring.c +2 -2
@@ -3762,8 +3758,8 @@ static __cold int io_allocate_scq_urings(struct io_ring_ctx *ctx,
 	else
 		rings = io_rings_map(ctx, p->cq_off.user_addr, size);
 
-	if (IS_ERR(rings))
-		return PTR_ERR(rings);
+	if (!rings)
+		return -ENOMEM;
 

Sorry, I started reviewing this, got excited about the error path quick fix, and didn't finish the review before it got it.

I think this change is broken for the ctx->flags & IORING_SETUP_NO_MMAP case, because io_rings_map returns ERR_PTR, and not NULL. In addition, io_rings_map might fail for multiple reasons, and we want to propagate the different error codes up here.

Quoted 5 lines
Quoted 1 line

Yeah, see my reply from some hours ago. I dropped it back then.

ah, thanks. I've configured lei to fetch the io_uring list every few hours. This ended up fetching part of the thread at first, and I only saw it dropped in the next fetch, after I sent the email. sorry for the noise.

Oh it's fine, I'd rather have one extra review than none at all :-)

Signature · 1 line
Jens Axboe