Skip to contentSkip to conversation

: Unmap fragment page once iterator is done

PATCH Applied Wedson Almeida Filho 3 participants 6 messages Jun 24, 2013

Subsystems (5)
Networking Network drivers SCSI Batman advanced +1 more subsystems

v2 of 3 Applied Latest is v3 →

lore
Share & tools
lore

Wedson Almeida Filho

author
#0 lore raw

PATCH v2 1/1 : Unmap fragment page once iterator is done

Callers of skb_seq_read() are currently forced to call skb_abort_seq_read() even when consuming all the data because the last call to skb_seq_read (the one that returns 0 to indicate the end) fails to unmap the last fragment page.

With this patch callers will be allowed to traverse the SKB data by calling skb_prepare_seq_read() once and repeatedly calling skb_seq_read() as originally intended (and documented in the original commit 677e90eda), that is, only call skb_abort_seq_read() if the sequential read is actually aborted.

Signed-off-by Wedson Almeida Filho <[email protected]>
---
 drivers/scsi/libiscsi_tcp.c |    1 -
 net/batman-adv/main.c       |    1 -
 net/core/skbuff.c           |    7 ++++++-
 3 files changed, 6 insertions(+), 3 deletions(-)
drivers/scsi/libiscsi_tcp.c +0 -1
index 552e8a2..448eae8 100644
@@ -906,7 +906,6 @@ int iscsi_tcp_recv_skb(struct iscsi_conn *conn, struct sk_buff *skb,
 			ISCSI_DBG_TCP(conn, "no more data avail. Consumed %d\n",
 				      consumed);
 			*status = ISCSI_TCP_SKB_DONE;
-			skb_abort_seq_read(&seq);
 			goto skb_done;
 		}
 		BUG_ON(segment->copied >= segment->size);
net/batman-adv/main.c +0 -1
index 51aafd6..08125f3 100644
@@ -473,7 +473,6 @@ __be32 batadv_skb_crc32(struct sk_buff *skb, u8 *payload_ptr)
 		crc = crc32c(crc, data, len);
 		consumed += len;
 	}
-	skb_abort_seq_read(&st);
 
 	return htonl(crc);
 }
net/core/skbuff.c +6 -1
index cfd777b..26ea1cf 100644
@@ -2554,8 +2554,13 @@ unsigned int skb_seq_read(unsigned int consumed, const u8 **data,
 	unsigned int block_limit, abs_offset = consumed + st->lower_offset;
 	skb_frag_t *frag;
 
-	if (unlikely(abs_offset >= st->upper_offset))
+	if (unlikely(abs_offset >= st->upper_offset)) {
+		if (st->frag_data) {
+			kunmap_atomic(st->frag_data);
+			st->frag_data = NULL;
+		}
 		return 0;
+	}
 
 next_skb:
 	block_limit = skb_headlen(st->cur_skb) + st->stepped_offset;
Signature · 1 line
1.7.9.5
maintainer Applied
#1 lore raw

From: Wedson Almeida Filho <[email protected]>
Date: Sun, 23 Jun 2013 23:33:48 -0700

Quoted 10 lines

Callers of skb_seq_read() are currently forced to call skb_abort_seq_read() even when consuming all the data because the last call to skb_seq_read (the one that returns 0 to indicate the end) fails to unmap the last fragment page.

With this patch callers will be allowed to traverse the SKB data by calling skb_prepare_seq_read() once and repeatedly calling skb_seq_read() as originally intended (and documented in the original commit 677e90eda), that is, only call skb_abort_seq_read() if the sequential read is actually aborted.

Signed-off-by Wedson Almeida Filho <[email protected]>

Applied.

Please format your subject lines as:

[PATCH vx] net: xxxx

Because anything inside of "[]" are deleted by the automated patch processing tools, and we now use lowercase for the subsystem prefixes.

Thanks.

#2 lore raw
In reply to Wedson Almeida Filho

Hello.

Quoted 7 lines · Wedson Almeida Filho

Callers of skb_seq_read() are currently forced to call skb_abort_seq_read() even when consuming all the data because the last call to skb_seq_read (the one that returns 0 to indicate the end) fails to unmap the last fragment page.

With this patch callers will be allowed to traverse the SKB data by calling skb_prepare_seq_read() once and repeatedly calling skb_seq_read() as originally intended (and documented in the original commit 677e90eda), that is, only call

Please also specify the summary line of that commit.

Quoted 3 lines

skb_abort_seq_read() if the sequential read is actually aborted.

Signed-off-by Wedson Almeida Filho <[email protected]>

WBR, Sergei

author
#3 lore raw

The summary line of the original commit is "[NET]: Zerocopy sequential reading of skb data".

Quoted 32 lines · Sergei Shtylyov

Hello.

Quoted 8 lines · Wedson Almeida Filho

Callers of skb_seq_read() are currently forced to call skb_abort_seq_read()
even when consuming all the data because the last call to skb_seq_read (the
one that returns 0 to indicate the end) fails to unmap the last fragment page.

Quoted 6 lines

With this patch callers will be allowed to traverse the SKB data by calling
skb_prepare_seq_read() once and repeatedly calling skb_seq_read() as originally
intended (and documented in the original commit 677e90eda), that is, only call

Please also specify the summary line of that commit.

Quoted 4 lines

skb_abort_seq_read() if the sequential read is actually aborted.

Signed-off-by Wedson Almeida Filho <[email protected]>

WBR, Sergei

#4 lore raw

Hello.

Quoted 2 lines · Wedson Almeida Filho

The summary line of the original commit is "[NET]: Zerocopy sequential reading of skb data".

That's interesting but please update your patch.

WBR, Sergei

maintainer
#5 lore raw
In reply to Wedson Almeida Filho

From: Wedson Almeida Filho <[email protected]>
Date: Mon, 24 Jun 2013 15:47:18 -0700

Quoted 2 lines

The summary line of the original commit is "[NET]: Zerocopy sequential reading of skb data".

He's telling you to put this in the commit message, and resubmit the patch.