sctp: add GSO support

sctp GSO requires it and sctp can be compiled as a module, so export this function.

Signed-off-by Marcelo Ricardo Leitner <[email protected]>

---
 net/core/skbuff.c | 1 +
 1 file changed, 1 insertion(+)

net/core/skbuff.c +0 -0

index b2df375ec9c2173a8132b8efa1c3062f0510284b..704b69682085dec77f3d0f990aaf0024afd705b9 100644

Quoted 13 lines

sctp GSO requires it and sctp can be compiled as a module, so export this function.

Signed-off-by Marcelo Ricardo Leitner <[email protected]>

---
  net/core/skbuff.c | 1 +
  1 file changed, 1 insertion(+)

net/core/skbuff.c +0 -0

index b2df375ec9c2173a8132b8efa1c3062f0510284b..704b69682085dec77f3d0f990aaf0024afd705b9 100644

net/core/skbuff.c done: · line 3318

Quoted 7 lines

net/core/skbuff.c +1 -0

@@ -3312,6 +3312,7 @@ done:
  	NAPI_GRO_CB(skb)->same_flow = 1;
  	return 0;
  }
+EXPORT_SYMBOL_GPL(skb_gro_receive);
 
  void __init skb_init(void)
  {

Normally, all the offloading support belongs in vmlinux, so this export is not needed.

For instance, we support GRO IPV6 even if IPv6 is not enabled on the host.

Ie all these net/ipv6 files are included in vmlinux if CONFIG_INET is enabled

ip6_offload.o tcpv6_offload.o udp_offload.o exthdrs_offload.o

Sent: 27 January 2016 17:07
This patchset is merely a RFC for the moment. There are some controversial points that I'd like to discuss before actually proposing the patches.

From: Marcelo Ricardo Leitner

Quoted 4 lines

Sent: 27 January 2016 17:07
This patchset is merely a RFC for the moment. There are some controversial points that I'd like to discuss before actually proposing the patches.

You also need to look at how a 'user' can actually get SCTP to merge data chunks in the first place.

With Nagle disabled (and it probably has to be since the data flow is unlikely to be 'command-response' or 'unidirectional bulk') it is currently almost impossible to get more than one chunk into an ethernet frame.

Support for MSG_MORE would help.

Given the current implementation you can get almost the required behaviour by turning nagle off and on repeatedly.

I did wonder whether the queued data could actually be picked up be a Heartbeat chunk that is probing a different remote address (which would be bad news).

Sent: 28 January 2016 15:53

Quoted 14 lines · David Laight

From: Marcelo Ricardo Leitner

Quoted 4 lines

Sent: 27 January 2016 17:07
This patchset is merely a RFC for the moment. There are some controversial points that I'd like to discuss before actually proposing the patches.

You also need to look at how a 'user' can actually get SCTP to merge data chunks in the first place.

With Nagle disabled (and it probably has to be since the data flow is unlikely to be 'command-response' or 'unidirectional bulk') it is currently almost impossible to get more than one chunk into an ethernet frame.

Quoted 4 lines

Support for MSG_MORE would help.

Given the current implementation you can get almost the required behaviour by turning nagle off and on repeatedly.

That's pretty much expected, I think. Without Nagle, if bandwidth and cwnd allow, segment will be sent. GSO by itself shouldn't cause a buffering to protect from that.

If something causes a bottleneck, tx may get queue up. Like if I do a stress test in my system, generally receiver side is slower than sender, so I end up having tx buffers pretty easily. It mimics bandwidth restrictions.

There is also the case of sending large data chunks, where sctp_sendmsg() will segment it into smaller chunks already.

But yes, agreed, MSG_MORE is at least a welcomed compliment here, specially for applications generating a train of chunks. Will put that in my ToDo here, thanks.

Quoted 3 lines

I did wonder whether the queued data could actually be picked up be a Heartbeat chunk that is probing a different remote address (which would be bad news).

I don't follow. You mean if a heartbeat may get stuck in queue or if sending of a heartbeat can end up carrying additional data by accident?

From: Marcelo Ricardo Leitner

Quoted 4 lines

Sent: 27 January 2016 17:07
This patchset is merely a RFC for the moment. There are some controversial points that I'd like to discuss before actually proposing the patches.

You also need to look at how a 'user' can actually get SCTP to merge data chunks in the first place.

With Nagle disabled (and it probably has to be since the data flow is unlikely to be 'command-response' or 'unidirectional bulk') it is currently almost impossible to get more than one chunk into an ethernet frame.

Support for MSG_MORE would help.

Given the current implementation you can get almost the required behaviour by turning nagle off and on repeatedly.

I did wonder whether the queued data could actually be picked up be a Heartbeat chunk that is probing a different remote address (which would be bad news).

David

--
To unsubscribe from this list: send the line "unsubscribe linux-sctp" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

From: 'Marcelo Ricardo Leitner'

Quoted 21 lines

Sent: 28 January 2016 15:53

Quoted 14 lines · David Laight

From: Marcelo Ricardo Leitner

Quoted 4 lines

Sent: 27 January 2016 17:07
This patchset is merely a RFC for the moment. There are some controversial points that I'd like to discuss before actually proposing the patches.

You also need to look at how a 'user' can actually get SCTP to merge data chunks in the first place.

With Nagle disabled (and it probably has to be since the data flow is unlikely to be 'command-response' or 'unidirectional bulk') it is currently almost impossible to get more than one chunk into an ethernet frame.

Quoted 4 lines

Support for MSG_MORE would help.

Given the current implementation you can get almost the required behaviour by turning nagle off and on repeatedly.

Quoted 8 lines

That's pretty much expected, I think. Without Nagle, if bandwidth and cwnd allow, segment will be sent. GSO by itself shouldn't cause a buffering to protect from that.

If something causes a bottleneck, tx may get queue up. Like if I do a stress test in my system, generally receiver side is slower than sender, so I end up having tx buffers pretty easily. It mimics bandwidth restrictions.

Imagine a using M2UA to connect local machines (one running mtp3, the other mtp2). Configure two linksets of 16 signalling links and perform a double-reflect loopback test.
The SCTP connection won't ever saturate, so every msu ends up in its own ethernet packet.
It is easy to generate 1000's of ethernet frames/sec on a single connection.

(We do this with something not entirely quite like M2UA over TCP, even then it is very hard to get multiple message into a single ethernet frame.)

Quoted 2 lines

There is also the case of sending large data chunks, where sctp_sendmsg() will segment it into smaller chunks already.

I presume they are merged before being passed to the receiving socket?

Quoted 3 lines

But yes, agreed, MSG_MORE is at least a welcomed compliment here, specially for applications generating a train of chunks. Will put that in my ToDo here, thanks.

I've posted a patch in the past for MSG_MORE, didn't quite work.

Quoted 6 lines

Quoted 3 lines

I did wonder whether the queued data could actually be picked up be a Heartbeat chunk that is probing a different remote address (which would be bad news).

I don't follow. You mean if a heartbeat may get stuck in queue or if sending of a heartbeat can end up carrying additional data by accident?

My suspicion was that the heartbeat would carry the queued data.

Quoted 11 lines · David Laight

From: Marcelo Ricardo Leitner

Quoted 4 lines

Sent: 27 January 2016 17:07
This patchset is merely a RFC for the moment. There are some controversial points that I'd like to discuss before actually proposing the patches.

You also need to look at how a 'user' can actually get SCTP to merge data chunks in the first place.

Quoted 6 lines

With Nagle disabled (and it probably has to be since the data flow is unlikely to be 'command-response' or 'unidirectional bulk') it is currently almost impossible to get more than one chunk into an ethernet frame.

Support for MSG_MORE would help.

What about adding support for the explicit EOR mode as specified in
https://tools.ietf.org/html/rfc6458#section-8.1.26

Best regards
Michael

Quoted 15 lines

Given the current implementation you can get almost the required behaviour by turning nagle off and on repeatedly.

I did wonder whether the queued data could actually be picked up be a Heartbeat chunk that is probing a different remote address (which would be bad news).

David

--
To unsubscribe from this list: send the line "unsubscribe linux-sctp" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

Quoted 16 lines · Michael Tuexen

Quoted 16 lines · David Laight

From: Marcelo Ricardo Leitner

Quoted 4 lines

Sent: 27 January 2016 17:07
This patchset is merely a RFC for the moment. There are some controversial points that I'd like to discuss before actually proposing the patches.

You also need to look at how a 'user' can actually get SCTP to merge data chunks in the first place.

With Nagle disabled (and it probably has to be since the data flow is unlikely to be 'command-response' or 'unidirectional bulk') it is currently almost impossible to get more than one chunk into an ethernet frame.

Quoted 3 lines

Quoted 1 line

Support for MSG_MORE would help.

What about adding support for the explicit EOR mode as specified in
https://tools.ietf.org/html/rfc6458#section-8.1.26

Seizing the moment to clarify my understanding on that. :) Such multiple calls to send system calls will result in a single data chunk. Is that so? That's what I get from that text and also from this

snippet:
"Sending a message using sendmsg() is atomic unless explicit end of record (EOR) marking is enabled on the socket specified by sd (see Section 8.1.26)."

Best regards,
Marcelo

Quoted 18 lines

Best regards
Michael

Quoted 15 lines

Given the current implementation you can get almost the required behaviour by turning nagle off and on repeatedly.

I did wonder whether the queued data could actually be picked up be a Heartbeat chunk that is probing a different remote address (which would be bad news).

David

--
To unsubscribe from this list: send the line "unsubscribe linux-sctp" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

Quoted 24 lines · Marcelo Ricardo Leitner

Quoted 19 lines · Michael Tuexen

Quoted 11 lines · David Laight

From: Marcelo Ricardo Leitner

Quoted 4 lines

Sent: 27 January 2016 17:07
This patchset is merely a RFC for the moment. There are some controversial points that I'd like to discuss before actually proposing the patches.

You also need to look at how a 'user' can actually get SCTP to merge data chunks in the first place.

Quoted 6 lines

With Nagle disabled (and it probably has to be since the data flow is unlikely to be 'command-response' or 'unidirectional bulk') it is currently almost impossible to get more than one chunk into an ethernet frame.

Support for MSG_MORE would help.

What about adding support for the explicit EOR mode as specified in
https://tools.ietf.org/html/rfc6458#section-8.1.26

Quoted 3 lines

Seizing the moment to clarify my understanding on that. :) Such multiple calls to send system calls will result in a single data chunk. Is that so? That's what I get from that text and also from this

No. It results in a single user message. This means you can send a user message larger than the send buffer size. How the user message is fragmented in DATA chunks is transparent to the upper layer.

Does this make things clearer?

Best regards
Michael

Quoted 27 lines

snippet:
"Sending a message using sendmsg() is atomic unless explicit end of record (EOR) marking is enabled on the socket specified by sd (see Section 8.1.26)."

Best regards,
Marcelo

Quoted 18 lines

Best regards
Michael

Quoted 15 lines

Given the current implementation you can get almost the required behaviour by turning nagle off and on repeatedly.

I did wonder whether the queued data could actually be picked up be a Heartbeat chunk that is probing a different remote address (which would be bad news).

David

--
To unsubscribe from this list: send the line "unsubscribe linux-sctp" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

Quoted 24 lines · Michael Tuexen

Quoted 24 lines · Marcelo Ricardo Leitner

Quoted 19 lines · Michael Tuexen

Quoted 11 lines · David Laight

From: Marcelo Ricardo Leitner

Quoted 4 lines

Sent: 27 January 2016 17:07
This patchset is merely a RFC for the moment. There are some controversial points that I'd like to discuss before actually proposing the patches.

You also need to look at how a 'user' can actually get SCTP to merge data chunks in the first place.

Quoted 6 lines

With Nagle disabled (and it probably has to be since the data flow is unlikely to be 'command-response' or 'unidirectional bulk') it is currently almost impossible to get more than one chunk into an ethernet frame.

Support for MSG_MORE would help.

What about adding support for the explicit EOR mode as specified in
https://tools.ietf.org/html/rfc6458#section-8.1.26

Quoted 8 lines

Quoted 3 lines

Seizing the moment to clarify my understanding on that. :) Such multiple calls to send system calls will result in a single data chunk. Is that so? That's what I get from that text and also from this

No. It results in a single user message. This means you can send a user message larger than the send buffer size. How the user message is fragmented in DATA chunks is transparent to the upper layer.

Does this make things clearer?

I think so, yes. So it allows delaying setting the Ending fragment bit until the application set SCTP_EOR. All the rest before this stays as
before: first send() will generate a chunk with Beginning bit set and may generate some other middle-fragments (no B nor E bit set) if necessary, second to N-1 call to send will generate only middle fragments, while the last send, with SCTP_EOF, will then set the Ending fragment in the last one. Right?

Thanks,
Marcelo

Quoted 31 lines

Best regards
Michael

Quoted 27 lines

snippet:
"Sending a message using sendmsg() is atomic unless explicit end of record (EOR) marking is enabled on the socket specified by sd (see Section 8.1.26)."

Best regards,
Marcelo

Quoted 18 lines

Best regards
Michael

Quoted 15 lines

Given the current implementation you can get almost the required behaviour by turning nagle off and on repeatedly.

I did wonder whether the queued data could actually be picked up be a Heartbeat chunk that is probing a different remote address (which would be bad news).

David

--
To unsubscribe from this list: send the line "unsubscribe linux-sctp" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

Quoted 36 lines · Marcelo Ricardo Leitner

Quoted 24 lines · Michael Tuexen

Quoted 24 lines · Marcelo Ricardo Leitner

Quoted 19 lines · Michael Tuexen

Quoted 11 lines · David Laight

From: Marcelo Ricardo Leitner

Quoted 4 lines

Sent: 27 January 2016 17:07
This patchset is merely a RFC for the moment. There are some controversial points that I'd like to discuss before actually proposing the patches.

You also need to look at how a 'user' can actually get SCTP to merge data chunks in the first place.

Quoted 6 lines

With Nagle disabled (and it probably has to be since the data flow is unlikely to be 'command-response' or 'unidirectional bulk') it is currently almost impossible to get more than one chunk into an ethernet frame.

Support for MSG_MORE would help.

What about adding support for the explicit EOR mode as specified in
https://tools.ietf.org/html/rfc6458#section-8.1.26

Quoted 8 lines

Quoted 3 lines

Seizing the moment to clarify my understanding on that. :) Such multiple calls to send system calls will result in a single data chunk. Is that so? That's what I get from that text and also from this

No. It results in a single user message. This means you can send a user message larger than the send buffer size. How the user message is fragmented in DATA chunks is transparent to the upper layer.

Does this make things clearer?

Quoted 7 lines

I think so, yes. So it allows delaying setting the Ending fragment bit until the application set SCTP_EOR. All the rest before this stays as
before: first send() will generate a chunk with Beginning bit set and may generate some other middle-fragments (no B nor E bit set) if necessary, second to N-1 call to send will generate only middle fragments, while the last send, with SCTP_EOF, will then set the Ending fragment in the last one. Right?

Yes. But there are no restrictions on the user data provided in send() calls and DATA chunks. So you can
send(100000 byte, no SCTP_EOR)
resulting in one DATA chunk with the B bit, several with no B and no E bit. send(100000 byte, no SCTP_EOR)
resulting in several chunks with no B and no E bit.
send(100000 byte, SCTP_EOR)
resulting in several chunks with no B and no E bit and one (the last) chunk with the E bit.

On the other hand you can do
send(1 byte, no SCTP_EOR)
resulting in a single DATA chunk with the E bit set.
send(1 byte, no SCTP_EOR)
send(1 byte, no SCTP_EOR)
send(1 byte, no SCTP_EOR)
send(1 byte, no SCTP_EOR)
send(1 byte, no SCTP_EOR)
All resulting in a single DATA chunk with 5 bytes user data and no B or E bit. (For example if Nagle is enabled and only after the last send call the SACK arrives). send(1 byte, SCTP_EOR)
results in a single DATA chunk with the E bist set.

Quoted 53 lines · Michael Tuexen

Quoted 36 lines · Marcelo Ricardo Leitner

Quoted 24 lines · Michael Tuexen

Quoted 24 lines · Marcelo Ricardo Leitner

Quoted 19 lines · Michael Tuexen

Quoted 11 lines · David Laight

From: Marcelo Ricardo Leitner

Quoted 4 lines

Sent: 27 January 2016 17:07
This patchset is merely a RFC for the moment. There are some controversial points that I'd like to discuss before actually proposing the patches.

You also need to look at how a 'user' can actually get SCTP to merge data chunks in the first place.

Quoted 6 lines

With Nagle disabled (and it probably has to be since the data flow is unlikely to be 'command-response' or 'unidirectional bulk') it is currently almost impossible to get more than one chunk into an ethernet frame.

Support for MSG_MORE would help.

What about adding support for the explicit EOR mode as specified in
https://tools.ietf.org/html/rfc6458#section-8.1.26

Quoted 8 lines

Quoted 3 lines

Seizing the moment to clarify my understanding on that. :) Such multiple calls to send system calls will result in a single data chunk. Is that so? That's what I get from that text and also from this

No. It results in a single user message. This means you can send a user message larger than the send buffer size. How the user message is fragmented in DATA chunks is transparent to the upper layer.

Does this make things clearer?

Quoted 7 lines

I think so, yes. So it allows delaying setting the Ending fragment bit until the application set SCTP_EOR. All the rest before this stays as
before: first send() will generate a chunk with Beginning bit set and may generate some other middle-fragments (no B nor E bit set) if necessary, second to N-1 call to send will generate only middle fragments, while the last send, with SCTP_EOF, will then set the Ending fragment in the last one. Right?

Yes. But there are no restrictions on the user data provided in send() calls and DATA chunks. So you can
send(100000 byte, no SCTP_EOR)
resulting in one DATA chunk with the B bit, several with no B and no E bit. send(100000 byte, no SCTP_EOR)
resulting in several chunks with no B and no E bit.
send(100000 byte, SCTP_EOR)
resulting in several chunks with no B and no E bit and one (the last) chunk with the E bit.

Quoted 12 lines

On the other hand you can do
send(1 byte, no SCTP_EOR)
resulting in a single DATA chunk with the E bit set.
send(1 byte, no SCTP_EOR)
send(1 byte, no SCTP_EOR)
send(1 byte, no SCTP_EOR)
send(1 byte, no SCTP_EOR)
send(1 byte, no SCTP_EOR)
All resulting in a single DATA chunk with 5 bytes user data and no B or E bit. (For example if Nagle is enabled and only after the last send call the SACK arrives). send(1 byte, SCTP_EOR)
results in a single DATA chunk with the E bist set.

Cool, thanks Michael. It will be quite fun to mix this with MSG_MORE logic, I think :)

Best regards,
Marcelo

Sent: 28 January 2016 20:56

Quoted 3 lines · David Laight

From: 'Marcelo Ricardo Leitner'

Quoted 2 lines

Sent: 28 January 2016 15:53
On Thu, Jan 28, 2016 at 01:51:02PM +0000, David Laight wrote:

Quoted 26 lines

Quoted 10 lines

Quoted 9 lines

With Nagle disabled (and it probably has to be since the data flow is unlikely to be 'command-response' or 'unidirectional bulk') it is currently almost impossible to get more than one chunk into an ethernet frame.

Support for MSG_MORE would help.

Given the current implementation you can get almost the required behaviour by turning nagle off and on repeatedly.

Quoted 8 lines

That's pretty much expected, I think. Without Nagle, if bandwidth and cwnd allow, segment will be sent. GSO by itself shouldn't cause a buffering to protect from that.

If something causes a bottleneck, tx may get queue up. Like if I do a stress test in my system, generally receiver side is slower than sender, so I end up having tx buffers pretty easily. It mimics bandwidth restrictions.

Imagine a using M2UA to connect local machines (one running mtp3, the other mtp2). Configure two linksets of 16 signalling links and perform a double-reflect loopback test.
The SCTP connection won't ever saturate, so every msu ends up in its own ethernet packet.
It is easy to generate 1000's of ethernet frames/sec on a single connection.

Quoted 4 lines

(We do this with something not entirely quite like M2UA over TCP, even then it is very hard to get multiple message into a single ethernet frame.)

Agreed, GSO won't help much in there without a corking feature like MSG_MORE or that on/off switch on Nagle you mentioned.

Thing about this (and also GRO) is on identifying how much can be spent on waiting for the next chunk/packet without causing issues to the application. Nagle is there and helps quite a lot, but timing-sensitive applications will turn it off.

Sent: 28 January 2016 20:56

Quoted 8 lines

Quoted 6 lines

Quoted 3 lines

I did wonder whether the queued data could actually be picked up be a Heartbeat chunk that is probing a different remote address (which would be bad news).

I don't follow. You mean if a heartbeat may get stuck in queue or if sending of a heartbeat can end up carrying additional data by accident?

My suspicion was that the heartbeat would carry the queued data.

I'm afraid I'm still not following, sorry. You mean that this GSO patch would cause the heartbeat to carry queued data? If yes, no, because for SCTP side of it it mangles the packet size and make it look bigger instead of handling multiple packets. It will then break this large sctp_packet into several sk_buff and glue them together as if they were GROed, allowing skb_segment to just split them back. The reason the sctp_packet is generated, being it due to user data or control chunks like heartbeats, is not modified.

From: 'Marcelo Ricardo Leitner' [mailto:[email protected]]
Sent: 28 January 2016 20:56

Quoted 5 lines

Quoted 3 lines

But yes, agreed, MSG_MORE is at least a welcomed compliment here, specially for applications generating a train of chunks. Will put that in my ToDo here, thanks.

I've posted a patch in the past for MSG_MORE, didn't quite work.

Ahh cool. Can you share the archive link please? Maybe I can take it from there then.

From: 'Marcelo Ricardo Leitner'

Quoted 5 lines

Sent: 28 January 2016 20:56

Quoted 3 lines · David Laight

From: 'Marcelo Ricardo Leitner'

Quoted 2 lines

Sent: 28 January 2016 15:53
On Thu, Jan 28, 2016 at 01:51:02PM +0000, David Laight wrote:

...

Quoted 31 lines

Quoted 26 lines

Quoted 10 lines

Quoted 9 lines

With Nagle disabled (and it probably has to be since the data flow is unlikely to be 'command-response' or 'unidirectional bulk') it is currently almost impossible to get more than one chunk into an ethernet frame.

Support for MSG_MORE would help.

Given the current implementation you can get almost the required behaviour by turning nagle off and on repeatedly.

Quoted 8 lines

That's pretty much expected, I think. Without Nagle, if bandwidth and cwnd allow, segment will be sent. GSO by itself shouldn't cause a buffering to protect from that.

If something causes a bottleneck, tx may get queue up. Like if I do a stress test in my system, generally receiver side is slower than sender, so I end up having tx buffers pretty easily. It mimics bandwidth restrictions.

Imagine a using M2UA to connect local machines (one running mtp3, the other mtp2). Configure two linksets of 16 signalling links and perform a double-reflect loopback test.
The SCTP connection won't ever saturate, so every msu ends up in its own ethernet packet.
It is easy to generate 1000's of ethernet frames/sec on a single connection.

Quoted 4 lines

(We do this with something not entirely quite like M2UA over TCP, even then it is very hard to get multiple message into a single ethernet frame.)

Quoted 7 lines

Agreed, GSO won't help much in there without a corking feature like MSG_MORE or that on/off switch on Nagle you mentioned.

Thing about this (and also GRO) is on identifying how much can be spent on waiting for the next chunk/packet without causing issues to the application. Nagle is there and helps quite a lot, but timing-sensitive applications will turn it off.

Nagle is only any good for unidirectional data (think file transfer) and command-response (think telnet or rlogin), for anything else it generates 200ms+ delays.
The SIGTRAN protocols (M3UA etc) are all specified to use SCTP and have no real relationship between the send and receive data. SCTP is being used to replace 64kb/s links so a few milliseconds of delay (1ms is 8 byte times) probably don't matter.
The Nagle timeout is, however far too long - so has to be disabled.

So never mind GSO to improve 'bulk' output, the big SCTP performance issue (as I see it) is the inability for a lot of workloads to ever get more than one small data chunk into an ethernet frame.

Which other (application) protocols are using SCTP? It doesn't seem an appropriate protocol for bulk data.

This patch enables SCTP to do GSO.

SCTP has this pecualiarty that its packets cannot be just segmented to (P)MTU. Its chunks must be contained in IP segments, padding respected. So we can't just generate a big skb, set gso_size to the fragmentation point and deliver it to IP layer.

Instead, this patch proposes that SCTP build a skb as it would be if it was received using GRO. That is, there will be a cover skb with the headers (incluing SCTP one) and children ones containing the actual SCTP chunks, already segmented in a way that respects SCTP RFCs and MTU.

This way SCTP can benefit from GSO and instead of passing several packets through the stack, it can pass a single large packet if there are enough data queued and cwnd allows.

Main points that need help:
- Usage of skb_gro_receive()
It fits nicely in there and properly handles offsets/lens, though the name means another thing. If you agree with this usage, we can rename it to something like skb_coalesce

- Checksum handling
Why only packets with checksum offloaded can be GSOed? Most of the NICs doesn't support SCTP CRC offloading and this will nearly defeat this feature. If checksum is being computed in sw, it doesn't really matter if it's earlier or later, right?
This patch hacks skb_needs_check() to allow using GSO with sw-computed checksums.
Also the meaning of UNNECESSARY and NONE are quite foggy to me yet and its usage may be wrong.

- gso_size = 1
There is skb_is_gso() all over the stack and it basically checks for non-zero skb_shinfo(skb)->gso_size. Setting it to 1 is the hacky way I found to keep skb_is_gso() working while being able to signal to skb_segment() that it shouldn't use gso_size but instead the fragment sizes themselves. skb_segment() will mainly just unpack the skb then.

- socket / gso max values
usage of sk_setup_caps() still needs a review

Signed-off-by Marcelo Ricardo Leitner <[email protected]>

---
 include/linux/netdev_features.h |   7 +-
 include/linux/netdevice.h       |   1 +
 net/core/dev.c                  |   6 +-
 net/core/skbuff.c               |  12 +-
 net/ipv4/af_inet.c              |   1 +
 net/sctp/offload.c              |  53 +++++++
 net/sctp/output.c               | 338 +++++++++++++++++++++++++---------------
 net/sctp/socket.c               |   2 +
 8 files changed, 292 insertions(+), 128 deletions(-)

include/linux/netdev_features.h +5 -2

index d9654f0eecb3519383441afa6b131ff9a5898485..f678998841f1800e0f2fe416a79935197d4ed305 100644
@@ -48,8 +48,9 @@ enum {
        NETIF_F_GSO_UDP_TUNNEL_BIT,     /* ... UDP TUNNEL with TSO */
        NETIF_F_GSO_UDP_TUNNEL_CSUM_BIT,/* ... UDP TUNNEL with TSO & CSUM */
        NETIF_F_GSO_TUNNEL_REMCSUM_BIT, /* ... TUNNEL with TSO & REMCSUM */
+       NETIF_F_GSO_SCTP_BIT,           /* ... SCTP fragmentation */
        /**/NETIF_F_GSO_LAST =          /* last bit, see GSO_MASK */
-               NETIF_F_GSO_TUNNEL_REMCSUM_BIT,
+               NETIF_F_GSO_SCTP_BIT,
 
        NETIF_F_FCOE_CRC_BIT,           /* FCoE CRC32 */
        NETIF_F_SCTP_CRC_BIT,           /* SCTP checksum offload */

@@ -119,6 +120,7 @@ enum {
 #define NETIF_F_GSO_UDP_TUNNEL __NETIF_F(GSO_UDP_TUNNEL)
 #define NETIF_F_GSO_UDP_TUNNEL_CSUM __NETIF_F(GSO_UDP_TUNNEL_CSUM)
 #define NETIF_F_GSO_TUNNEL_REMCSUM __NETIF_F(GSO_TUNNEL_REMCSUM)
+#define NETIF_F_GSO_SCTP       __NETIF_F(GSO_SCTP)
 #define NETIF_F_HW_VLAN_STAG_FILTER __NETIF_F(HW_VLAN_STAG_FILTER)
 #define NETIF_F_HW_VLAN_STAG_RX        __NETIF_F(HW_VLAN_STAG_RX)
 #define NETIF_F_HW_VLAN_STAG_TX        __NETIF_F(HW_VLAN_STAG_TX)

@@ -144,7 +146,8 @@ enum {
 
 /* List of features with software fallbacks. */
 #define NETIF_F_GSO_SOFTWARE   (NETIF_F_TSO | NETIF_F_TSO_ECN | \
-                                NETIF_F_TSO6 | NETIF_F_UFO)
+                                NETIF_F_TSO6 | NETIF_F_UFO | \
+                                NETIF_F_GSO_SCTP)
 
 /* List of IP checksum features. Note that NETIF_F_ HW_CSUM should not be
  * set in features when NETIF_F_IP_CSUM or NETIF_F_IPV6_CSUM are set--

include/linux/netdevice.h +1 -0

index 289c2314d76668b8357728382bb33d6828617458..ce14fab858bf96dd0f85aca237350c8d8317756e 100644
@@ -3928,6 +3928,7 @@ static inline bool net_gso_ok(netdev_features_t features, int gso_type)
        BUILD_BUG_ON(SKB_GSO_UDP_TUNNEL != (NETIF_F_GSO_UDP_TUNNEL >> NETIF_F_GSO_SHIFT));
        BUILD_BUG_ON(SKB_GSO_UDP_TUNNEL_CSUM != (NETIF_F_GSO_UDP_TUNNEL_CSUM >> NETIF_F_GSO_SHIFT));
        BUILD_BUG_ON(SKB_GSO_TUNNEL_REMCSUM != (NETIF_F_GSO_TUNNEL_REMCSUM >> NETIF_F_GSO_SHIFT));
+       BUILD_BUG_ON(SKB_GSO_SCTP    != (NETIF_F_GSO_SCTP >> NETIF_F_GSO_SHIFT));
 
        return (features & feature) = feature;
 }

net/core/dev.c +0 -0

index 8cba3d852f251c503b193823b71b27aaef3fb3ae..9583284086967c0746de5f553535e25e125714a5 100644
@@ -2680,7 +2680,11 @@ EXPORT_SYMBOL(skb_mac_gso_segment);
 static inline bool skb_needs_check(struct sk_buff *skb, bool tx_path)
 {

net/core/skbuff.c +1 -2

index 704b69682085dec77f3d0f990aaf0024afd705b9..96f223f8d769d2765fd64348830c76cb222906c8 100644
@@ -3017,8 +3017,16 @@ struct sk_buff *skb_segment(struct sk_buff *head_skb,
                int size;
 
                len = head_skb->len - offset;
-               if (len > mss)
-                       len = mss;
+               if (len > mss) {

net/core/skbuff.c +0 -0

                hsize = skb_headlen(head_skb) - offset;
                if (hsize < 0)

net/ipv4/af_inet.c +1 -0

index 5c5db6636704daa0c49fc13e84b2c5b282a44ed3..ec1c779bb664d1399d74f2bd7016e30b648ce47d 100644
@@ -1220,6 +1220,7 @@ static struct sk_buff *inet_gso_segment(struct sk_buff *skb,
                       SKB_GSO_UDP_TUNNEL |
                       SKB_GSO_UDP_TUNNEL_CSUM |
                       SKB_GSO_TUNNEL_REMCSUM |
+                      SKB_GSO_SCTP |
                       0)))
                goto out;
 

net/sctp/offload.c +0 -0

index 7080a6318da7110c1688dd0c5bb240356dbd0cd3..3b96035fa180a4e7195f7b6e7a8be7b97c8f8b26 100644

net/sctp/offload.c +20 -0

+static struct sk_buff *sctp_gso_segment(struct sk_buff *skb,
+                                       netdev_features_t features)
+{
+       struct sk_buff *segs = ERR_PTR(-EINVAL);
+       struct sctphdr *sh;
+
+       sh = sctp_hdr(skb);
+       if (!pskb_may_pull(skb, sizeof(*sh)))
+               goto out;
+
+       __skb_pull(skb, sizeof(*sh));
+
+       if (skb_gso_ok(skb, features | NETIF_F_GSO_ROBUST)) {
+               /* Packet is from an untrusted source, reset gso_segs. */
+               int type = skb_shinfo(skb)->gso_type;
+
+               if (unlikely(type &
+                            ~(SKB_GSO_SCTP | SKB_GSO_DODGY |
+                              0) ||
+                            !(type & (SKB_GSO_SCTP))))

net/sctp/offload.c +5 -0

+out:
+       return segs;
+}
+
 static const struct net_offload sctp_offload = {
        .callbacks = {
+               .gso_segment = sctp_gso_segment,
        },
 };
 

net/sctp/output.c +215 -123

index 9d610eddd19ef2320fc34ae9d91e7426ae5f50f9..5e619b1b7b47737447bce746b2420bac3427fde4 100644
@@ -381,12 +381,14 @@ int sctp_packet_transmit(struct sctp_packet *packet)
        struct sctp_transport *tp = packet->transport;
        struct sctp_association *asoc = tp->asoc;
        struct sctphdr *sh;
-       struct sk_buff *nskb;
+       struct sk_buff *nskb = NULL, *head = NULL;
        struct sctp_chunk *chunk, *tmp;
-       struct sock *sk;
+       struct sock *sk = asoc->base.sk;
        int err = 0;
        int padding;            /* How much padding do we need?  */
+       int pkt_size;
        __u8 has_data = 0;
+       int gso = 0;
        struct dst_entry *dst;
        unsigned char *auth = NULL;     /* pointer to auth in skb data */
 

@@ -396,37 +398,44 @@ int sctp_packet_transmit(struct sctp_packet *packet)
        if (list_empty(&packet->chunk_list))
                return err;
 
-       /* Set up convenience variables... */
+       /* TODO: double check this */
        chunk = list_entry(packet->chunk_list.next, struct sctp_chunk, list);
        sk = chunk->skb->sk;
+       dst_hold(tp->dst);
+       sk_setup_caps(sk, tp->dst);
+
+       if (packet->size > tp->pathmtu) {
+               WARN_ON(packet->ipfragok);
+               if (sk_can_gso(sk)) {
+                       gso = 1;
+                       pkt_size = packet->overhead;
+               } else {
+                       /* Something nasty happened */
+                       /* FIXME */
+                       printk("Damn, we can't GSO and packet is too big %d for pmtu %d.\n",
+                              packet->size, tp->pathmtu);
+                       goto nomem;
+               }
+       } else {
+               pkt_size = packet->size;
+       }
 
-       /* Allocate the new skb.  */
-       nskb = alloc_skb(packet->size + MAX_HEADER, GFP_ATOMIC);
-       if (!nskb)
+       /* Allocate the head skb, or main one if not in GSO */
+       head = alloc_skb(pkt_size + MAX_HEADER, GFP_ATOMIC);
+       if (!head)
                goto nomem;
+       if (gso) {
+               NAPI_GRO_CB(head)->last = head;
+       } else {
+               nskb = head;
+       }
 
        /* Make sure the outbound skb has enough header room reserved. */
-       skb_reserve(nskb, packet->overhead + MAX_HEADER);
-
-       /* Set the owning socket so that we know where to get the
-        * destination IP address.
-        */
-       sctp_packet_set_owner_w(nskb, sk);
-
-       if (!sctp_transport_dst_check(tp)) {
-               sctp_transport_route(tp, NULL, sctp_sk(sk));
-               if (asoc && (asoc->param_flags & SPP_PMTUD_ENABLE)) {
-                       sctp_assoc_sync_pmtu(sk, asoc);
-               }
-       }
-       dst = dst_clone(tp->dst);
-       if (!dst)
-               goto no_route;
-       skb_dst_set(nskb, dst);
+       skb_reserve(head, packet->overhead + MAX_HEADER);
 
        /* Build the SCTP header.  */
-       sh = (struct sctphdr *)skb_push(nskb, sizeof(struct sctphdr));
-       skb_reset_transport_header(nskb);
+       sh = (struct sctphdr *)skb_push(head, sizeof(struct sctphdr));
+       skb_reset_transport_header(head);
        sh->source = htons(packet->source_port);
        sh->dest   = htons(packet->destination_port);
 

@@ -441,90 +450,164 @@ int sctp_packet_transmit(struct sctp_packet *packet)
        sh->vtag     = htonl(packet->vtag);
        sh->checksum = 0;
 
-       /**
-        * 6.10 Bundling
-        *
-        *    An endpoint bundles chunks by simply including multiple
-        *    chunks in one outbound SCTP packet.  ...
+       /* Set the owning socket so that we know where to get the
+        * destination IP address.
         */
+       sctp_packet_set_owner_w(head, sk);
 
-       /**
-        * 3.2  Chunk Field Descriptions
-        *
-        * The total length of a chunk (including Type, Length and
-        * Value fields) MUST be a multiple of 4 bytes.  If the length
-        * of the chunk is not a multiple of 4 bytes, the sender MUST
-        * pad the chunk with all zero bytes and this padding is not
-        * included in the chunk length field.  The sender should
-        * never pad with more than 3 bytes.
-        *
-        * [This whole comment explains WORD_ROUND() below.]
-        */
+       if (!sctp_transport_dst_check(tp)) {
+               sctp_transport_route(tp, NULL, sctp_sk(sk));
+               if (asoc && (asoc->param_flags & SPP_PMTUD_ENABLE)) {
+                       sctp_assoc_sync_pmtu(sk, asoc);
+               }
+       }
+       dst = dst_clone(tp->dst);
+       if (!dst)
+               goto no_route;
+       skb_dst_set(head, dst);
 
        pr_debug("***sctp_transmit_packet***\n");
 
-       list_for_each_entry_safe(chunk, tmp, &packet->chunk_list, list) {
-               list_del_init(&chunk->list);
-               if (sctp_chunk_is_data(chunk)) {
-                       /* 6.3.1 C4) When data is in flight and when allowed
-                        * by rule C5, a new RTT measurement MUST be made each
-                        * round trip.  Furthermore, new RTT measurements
-                        * SHOULD be made no more than once per round-trip
-                        * for a given destination transport address.
-                        */
-
-                       if (!chunk->resent && !tp->rto_pending) {
-                               chunk->rtt_in_progress = 1;
-                               tp->rto_pending = 1;
+       do {
+               /* Set up convenience variables... */
+               chunk = list_entry(packet->chunk_list.next, struct sctp_chunk, list);
+               WARN_ON(sk != chunk->skb->sk); /* XXX */
+
+               /* Calculate packet size, so it fits in PMTU. Leave
+                * other chunks for the next packets. */
+               if (gso) {
+                       pkt_size = packet->overhead;
+                       list_for_each_entry(chunk, &packet->chunk_list, list) {
+                               int padded = WORD_ROUND(chunk->skb->len);
+                               if (pkt_size + padded > tp->pathmtu)
+                                       break;
+                               pkt_size += padded;
                        }
 
-                       has_data = 1;
+                       /* Allocate the new skb.  */
+                       nskb = alloc_skb(pkt_size + MAX_HEADER, GFP_ATOMIC);
+
+                       /* Make sure the outbound skb has enough header room reserved. */
+                       if (nskb)
+                               skb_reserve(nskb, packet->overhead + MAX_HEADER);
                }
+               if (!nskb)
+                       goto nomem;
+
+               /**
+                * 3.2  Chunk Field Descriptions
+                *
+                * The total length of a chunk (including Type, Length and
+                * Value fields) MUST be a multiple of 4 bytes.  If the length
+                * of the chunk is not a multiple of 4 bytes, the sender MUST
+                * pad the chunk with all zero bytes and this padding is not
+                * included in the chunk length field.  The sender should
+                * never pad with more than 3 bytes.
+                *
+                * [This whole comment explains WORD_ROUND() below.]
+                */
+
+               pkt_size -= packet->overhead;
+               list_for_each_entry_safe(chunk, tmp, &packet->chunk_list, list) {
+                       list_del_init(&chunk->list);
+                       if (sctp_chunk_is_data(chunk)) {
+                               /* 6.3.1 C4) When data is in flight and when allowed
+                                * by rule C5, a new RTT measurement MUST be made each
+                                * round trip.  Furthermore, new RTT measurements
+                                * SHOULD be made no more than once per round-trip
+                                * for a given destination transport address.
+                                */
+
+                               if (!chunk->resent && !tp->rto_pending) {
+                                       chunk->rtt_in_progress = 1;
+                                       tp->rto_pending = 1;
+                               }
+
+                               has_data = 1;
+                       }
+
+                       padding = WORD_ROUND(chunk->skb->len) - chunk->skb->len;
+                       if (padding)
+                               memset(skb_put(chunk->skb, padding), 0, padding);
 
-               padding = WORD_ROUND(chunk->skb->len) - chunk->skb->len;
-               if (padding)
-                       memset(skb_put(chunk->skb, padding), 0, padding);
+                       /* if this is the auth chunk that we are adding,
+                        * store pointer where it will be added and put
+                        * the auth into the packet.
+                        */
+                       if (chunk = packet->auth) {
+                               auth = skb_tail_pointer(nskb);
+                       }
+
+                       memcpy(skb_put(nskb, chunk->skb->len),
+                                      chunk->skb->data, chunk->skb->len);
+
+                       pr_debug("*** Chunk:%p[%s] %s 0x%x, length:%d, chunk->skb->len:%d, "
+                                "rtt_in_progress:%d\n", chunk,
+                                sctp_cname(SCTP_ST_CHUNK(chunk->chunk_hdr->type)),
+                                chunk->has_tsn ? "TSN" : "No TSN",
+                                chunk->has_tsn ? ntohl(chunk->subh.data_hdr->tsn) : 0,
+                                ntohs(chunk->chunk_hdr->length), chunk->skb->len,
+                                chunk->rtt_in_progress);
+
+                       /*
+                        * If this is a control chunk, this is our last
+                        * reference. Free data chunks after they've been
+                        * acknowledged or have failed.
+                        * Re-queue auth chunks if needed.
+                        */
+                       pkt_size -= WORD_ROUND(chunk->skb->len);
+
+                       if (chunk = packet->auth && !list_empty(&packet->chunk_list))
+                               list_add(&chunk->list, &packet->chunk_list);
+                       else if (!sctp_chunk_is_data(chunk))
+                               sctp_chunk_free(chunk);
 
-               /* if this is the auth chunk that we are adding,
-                * store pointer where it will be added and put
-                * the auth into the packet.
+                       if (!pkt_size)
+                               break;
+               }
+
+               /* SCTP-AUTH, Section 6.2
+                *    The sender MUST calculate the MAC as described in RFC2104 [2]
+                *    using the hash function H as described by the MAC Identifier and
+                *    the shared association key K based on the endpoint pair shared key
+                *    described by the shared key identifier.  The 'data' used for the
+                *    computation of the AUTH-chunk is given by the AUTH chunk with its
+                *    HMAC field set to zero (as shown in Figure 6) followed by all
+                *    chunks that are placed after the AUTH chunk in the SCTP packet.
                 */
-               if (chunk = packet->auth)
-                       auth = skb_tail_pointer(nskb);
-
-               memcpy(skb_put(nskb, chunk->skb->len),
-                              chunk->skb->data, chunk->skb->len);
-
-               pr_debug("*** Chunk:%p[%s] %s 0x%x, length:%d, chunk->skb->len:%d, "
-                        "rtt_in_progress:%d\n", chunk,
-                        sctp_cname(SCTP_ST_CHUNK(chunk->chunk_hdr->type)),
-                        chunk->has_tsn ? "TSN" : "No TSN",
-                        chunk->has_tsn ? ntohl(chunk->subh.data_hdr->tsn) : 0,
-                        ntohs(chunk->chunk_hdr->length), chunk->skb->len,
-                        chunk->rtt_in_progress);
-
-               /*
-                * If this is a control chunk, this is our last
-                * reference. Free data chunks after they've been
-                * acknowledged or have failed.
+               if (auth)
+                       sctp_auth_calculate_hmac(asoc, nskb,
+                                               (struct sctp_auth_chunk *)auth,
+                                               GFP_ATOMIC);
+
+               /* Set up the IP options.  */
+               /* BUG: not implemented
+                * For v4 this all lives somewhere in sk->sk_opt...
                 */
-               if (!sctp_chunk_is_data(chunk))
-                       sctp_chunk_free(chunk);
-       }
 
-       /* SCTP-AUTH, Section 6.2
-        *    The sender MUST calculate the MAC as described in RFC2104 [2]
-        *    using the hash function H as described by the MAC Identifier and
-        *    the shared association key K based on the endpoint pair shared key
-        *    described by the shared key identifier.  The 'data' used for the
-        *    computation of the AUTH-chunk is given by the AUTH chunk with its
-        *    HMAC field set to zero (as shown in Figure 6) followed by all
-        *    chunks that are placed after the AUTH chunk in the SCTP packet.
-        */
-       if (auth)
-               sctp_auth_calculate_hmac(asoc, nskb,
-                                       (struct sctp_auth_chunk *)auth,
-                                       GFP_ATOMIC);
+               /* Dump that on IP!  */
+               if (asoc) {
+                       asoc->stats.opackets++;
+                       if (asoc->peer.last_sent_to != tp)
+                               /* Considering the multiple CPU scenario, this is a
+                                * "correcter" place for last_sent_to.  --xguo
+                                */
+                               asoc->peer.last_sent_to = tp;
+               }
+
+
+               if (!gso ||
+                   skb_shinfo(head)->gso_segs >= sk->sk_gso_max_segs)
+//                 head->len + asoc->pathmtu >= sk->sk_gso_max_size)
+                       break;
+
+               if (skb_gro_receive(&head, nskb))
+                       goto nomem;
+               skb_shinfo(head)->gso_segs++;
+               /* FIXME: below is a lie */
+               skb_shinfo(head)->gso_size = 1;
+               nskb = NULL;
+       } while (!list_empty(&packet->chunk_list));
 
        /* 2) Calculate the Adler-32 checksum of the whole packet,
         *    including the SCTP common header and all the

@@ -532,16 +615,21 @@ int sctp_packet_transmit(struct sctp_packet *packet)
         *
         * Note: Adler-32 is no longer applicable, as has been replaced
         * by CRC32-C as described in <draft-ietf-tsvwg-sctpcsum-02.txt>.
+        *
+        * If it's a GSO packet, it's postponed to sctp_skb_segment.
         */
-       if (!sctp_checksum_disable) {
+       if (!sctp_checksum_disable || gso) {
                if (!(dst->dev->features & NETIF_F_SCTP_CRC) ||
                    (dst_xfrm(dst) != NULL) || packet->ipfragok) {
-                       sh->checksum = sctp_compute_cksum(nskb, 0);
+                       if (!gso)
+                               sh->checksum = sctp_compute_cksum(head, 0);
+                       else
+                               head->ip_summed = CHECKSUM_UNNECESSARY;
                } else {
                        /* no need to seed pseudo checksum for SCTP */
-                       nskb->ip_summed = CHECKSUM_PARTIAL;
-                       nskb->csum_start = skb_transport_header(nskb) - nskb->head;
-                       nskb->csum_offset = offsetof(struct sctphdr, checksum);
+                       head->ip_summed = CHECKSUM_PARTIAL;
+                       head->csum_start = skb_transport_header(head) - head->head;
+                       head->csum_offset = offsetof(struct sctphdr, checksum);
                }
        }
 

@@ -557,22 +645,7 @@ int sctp_packet_transmit(struct sctp_packet *packet)
         * Note: The works for IPv6 layer checks this bit too later
         * in transmission.  See IP6_ECN_flow_xmit().
         */
-       tp->af_specific->ecn_capable(nskb->sk);
-
-       /* Set up the IP options.  */
-       /* BUG: not implemented
-        * For v4 this all lives somewhere in sk->sk_opt...
-        */
-
-       /* Dump that on IP!  */
-       if (asoc) {
-               asoc->stats.opackets++;
-               if (asoc->peer.last_sent_to != tp)
-                       /* Considering the multiple CPU scenario, this is a
-                        * "correcter" place for last_sent_to.  --xguo
-                        */
-                       asoc->peer.last_sent_to = tp;
-       }
+       tp->af_specific->ecn_capable(head->sk);
 
        if (has_data) {
                struct timer_list *timer;

@@ -589,16 +662,23 @@ int sctp_packet_transmit(struct sctp_packet *packet)
                }
        }
 
-       pr_debug("***sctp_transmit_packet*** skb->len:%d\n", nskb->len);
+       pr_debug("***sctp_transmit_packet*** skb->len:%d\n", head->len);
 
-       nskb->ignore_df = packet->ipfragok;
-       tp->af_specific->sctp_xmit(nskb, tp);
+       head->ignore_df = packet->ipfragok;
+       printk("%s %d %d %d\n", __func__, head->len,
+              packet->transport->pathmtu,
+              packet->transport->pathmtu - packet->overhead);
+       if (gso)
+               skb_shinfo(head)->gso_type = SKB_GSO_SCTP;
+       tp->af_specific->sctp_xmit(head, tp);
 
 out:
        sctp_packet_reset(packet);
+       sk_dst_reset(sk); /* FIXME: double check */
        return err;
 no_route:
        kfree_skb(nskb);
+       kfree_skb(head);
 
        if (asoc)
                IP_INC_STATS(sock_net(asoc->base.sk), IPSTATS_MIB_OUTNOROUTES);

@@ -635,7 +715,7 @@ nomem:
 static sctp_xmit_t sctp_packet_can_append_data(struct sctp_packet *packet,
                                           struct sctp_chunk *chunk)
 {
-       size_t datasize, rwnd, inflight, flight_size;
+       size_t datasize, rwnd, inflight, flight_size, maxsize;
        struct sctp_transport *transport = packet->transport;
        struct sctp_association *asoc = transport->asoc;
        struct sctp_outq *q = &asoc->outqueue;

@@ -705,7 +785,15 @@ static sctp_xmit_t sctp_packet_can_append_data(struct sctp_packet *packet,
        /* Check whether this chunk and all the rest of pending data will fit
         * or delay in hopes of bundling a full sized packet.
         */
-       if (chunk->skb->len + q->out_qlen >= transport->pathmtu - packet->overhead)
+       if (packet->ipfragok) {
+               /* Means chunk won't fit and needs fragmentation at
+                * transport level, so we can't do GSO.
+                */
+               maxsize = transport->pathmtu;
+       } else {
+               maxsize = transport->dst->dev->gso_max_size;
+       }
+       if (chunk->skb->len + q->out_qlen >= maxsize - packet->overhead)
                /* Enough data queued to fill a packet */
                return SCTP_XMIT_OK;
 

@@ -764,6 +852,8 @@ static sctp_xmit_t sctp_packet_will_fit(struct sctp_packet *packet,
 
        /* Decide if we need to fragment or resubmit later. */
        if (too_big) {
+               struct net_device *dev = packet->transport->dst->dev;
+
                /* It's OK to fragmet at IP level if any one of the following
                 * is true:
                 *      1. The packet is empty (meaning this chunk is greater

@@ -779,9 +869,11 @@ static sctp_xmit_t sctp_packet_will_fit(struct sctp_packet *packet,
                         * actually hit this condition
                         */
                        packet->ipfragok = 1;
-               } else {
+               } else if (psize + chunk_len > dev->gso_max_size - packet->overhead) {
+                       /* Hit GSO limit, gotta flush */
                        retval = SCTP_XMIT_PMTU_FULL;
                }
+               /* Otherwise it will fit in the GSO packet */
        }
 
        return retval;

net/sctp/socket.c +2 -1

index 5ca2ebfe0be83882fcb841de6fa8029b6455ef85..064e5d375e612f2ec745f384d35f0e4c6b96212c 100644
@@ -4001,6 +4001,8 @@ static int sctp_init_sock(struct sock *sk)
                return -ESOCKTNOSUPPORT;
        }
 
+       sk->sk_gso_type = SKB_GSO_SCTP;
+
        /* Initialize default send parameters. These parameters can be
         * modified with the SCTP_DEFAULT_SEND_PARAM socket option.
         */
--

2.5.0

Quoted 32 lines · Marcelo Ricardo Leitner

This patch enables SCTP to do GSO.

SCTP has this pecualiarty that its packets cannot be just segmented to (P)MTU. Its chunks must be contained in IP segments, padding respected. So we can't just generate a big skb, set gso_size to the fragmentation point and deliver it to IP layer.

Instead, this patch proposes that SCTP build a skb as it would be if it was received using GRO. That is, there will be a cover skb with the headers (incluing SCTP one) and children ones containing the actual SCTP chunks, already segmented in a way that respects SCTP RFCs and MTU.

This way SCTP can benefit from GSO and instead of passing several packets through the stack, it can pass a single large packet if there are enough data queued and cwnd allows.

Main points that need help:
- Usage of skb_gro_receive()
It fits nicely in there and properly handles offsets/lens, though the name means another thing. If you agree with this usage, we can rename it to something like skb_coalesce

- Checksum handling
Why only packets with checksum offloaded can be GSOed? Most of the NICs doesn't support SCTP CRC offloading and this will nearly defeat this feature. If checksum is being computed in sw, it doesn't really matter if it's earlier or later, right?
This patch hacks skb_needs_check() to allow using GSO with sw-computed checksums.
Also the meaning of UNNECESSARY and NONE are quite foggy to me yet and its usage may be wrong.

Quoted 6 lines

- gso_size = 1
There is skb_is_gso() all over the stack and it basically checks for non-zero skb_shinfo(skb)->gso_size. Setting it to 1 is the hacky way I found to keep skb_is_gso() working while being able to signal to skb_segment() that it shouldn't use gso_size but instead the fragment sizes themselves. skb_segment() will mainly just unpack the skb then.

Instead of 1 why not use 0xFFFF? It is a value that can never be used for a legitimate segment size since IP total length is a 16 bit value and includes the IP header in the size.

Quoted 72 lines

- socket / gso max values
usage of sk_setup_caps() still needs a review

Signed-off-by Marcelo Ricardo Leitner <[email protected]>

---
 include/linux/netdev_features.h |   7 +-
 include/linux/netdevice.h       |   1 +
 net/core/dev.c                  |   6 +-
 net/core/skbuff.c               |  12 +-
 net/ipv4/af_inet.c              |   1 +
 net/sctp/offload.c              |  53 +++++++
 net/sctp/output.c               | 338 +++++++++++++++++++++++++---------------
 net/sctp/socket.c               |   2 +
 8 files changed, 292 insertions(+), 128 deletions(-)

include/linux/netdev_features.h +5 -2

index d9654f0eecb3519383441afa6b131ff9a5898485..f678998841f1800e0f2fe416a79935197d4ed305 100644
@@ -48,8 +48,9 @@ enum {
        NETIF_F_GSO_UDP_TUNNEL_BIT,     /* ... UDP TUNNEL with TSO */
        NETIF_F_GSO_UDP_TUNNEL_CSUM_BIT,/* ... UDP TUNNEL with TSO & CSUM */
        NETIF_F_GSO_TUNNEL_REMCSUM_BIT, /* ... TUNNEL with TSO & REMCSUM */
+       NETIF_F_GSO_SCTP_BIT,           /* ... SCTP fragmentation */
        /**/NETIF_F_GSO_LAST =          /* last bit, see GSO_MASK */
-               NETIF_F_GSO_TUNNEL_REMCSUM_BIT,
+               NETIF_F_GSO_SCTP_BIT,
 
        NETIF_F_FCOE_CRC_BIT,           /* FCoE CRC32 */
        NETIF_F_SCTP_CRC_BIT,           /* SCTP checksum offload */

@@ -119,6 +120,7 @@ enum {
 #define NETIF_F_GSO_UDP_TUNNEL __NETIF_F(GSO_UDP_TUNNEL)
 #define NETIF_F_GSO_UDP_TUNNEL_CSUM __NETIF_F(GSO_UDP_TUNNEL_CSUM)
 #define NETIF_F_GSO_TUNNEL_REMCSUM __NETIF_F(GSO_TUNNEL_REMCSUM)
+#define NETIF_F_GSO_SCTP       __NETIF_F(GSO_SCTP)
 #define NETIF_F_HW_VLAN_STAG_FILTER __NETIF_F(HW_VLAN_STAG_FILTER)
 #define NETIF_F_HW_VLAN_STAG_RX        __NETIF_F(HW_VLAN_STAG_RX)
 #define NETIF_F_HW_VLAN_STAG_TX        __NETIF_F(HW_VLAN_STAG_TX)

@@ -144,7 +146,8 @@ enum {
 
 /* List of features with software fallbacks. */
 #define NETIF_F_GSO_SOFTWARE   (NETIF_F_TSO | NETIF_F_TSO_ECN | \
-                                NETIF_F_TSO6 | NETIF_F_UFO)
+                                NETIF_F_TSO6 | NETIF_F_UFO | \
+                                NETIF_F_GSO_SCTP)
 
 /* List of IP checksum features. Note that NETIF_F_ HW_CSUM should not be
  * set in features when NETIF_F_IP_CSUM or NETIF_F_IPV6_CSUM are set--

include/linux/netdevice.h +1 -0

index 289c2314d76668b8357728382bb33d6828617458..ce14fab858bf96dd0f85aca237350c8d8317756e 100644
@@ -3928,6 +3928,7 @@ static inline bool net_gso_ok(netdev_features_t features, int gso_type)
        BUILD_BUG_ON(SKB_GSO_UDP_TUNNEL != (NETIF_F_GSO_UDP_TUNNEL >> NETIF_F_GSO_SHIFT));
        BUILD_BUG_ON(SKB_GSO_UDP_TUNNEL_CSUM != (NETIF_F_GSO_UDP_TUNNEL_CSUM >> NETIF_F_GSO_SHIFT));
        BUILD_BUG_ON(SKB_GSO_TUNNEL_REMCSUM != (NETIF_F_GSO_TUNNEL_REMCSUM >> NETIF_F_GSO_SHIFT));
+       BUILD_BUG_ON(SKB_GSO_SCTP    != (NETIF_F_GSO_SCTP >> NETIF_F_GSO_SHIFT));
 
        return (features & feature) = feature;
 }

net/core/dev.c +3 -1

index 8cba3d852f251c503b193823b71b27aaef3fb3ae..9583284086967c0746de5f553535e25e125714a5 100644
@@ -2680,7 +2680,11 @@ EXPORT_SYMBOL(skb_mac_gso_segment);
 static inline bool skb_needs_check(struct sk_buff *skb, bool tx_path)
 {
        if (tx_path)
-               return skb->ip_summed != CHECKSUM_PARTIAL;
+               /* FIXME: Why only packets with checksum offloading are
+                * supported for GSO?
+                */

net/core/dev.c EXPORT_SYMBOL(skb_mac_gso_segment); · line 2690

Quoted 5 lines

net/core/dev.c +2 -0

+               return skb->ip_summed != CHECKSUM_PARTIAL &&
+                      skb->ip_summed != CHECKSUM_UNNECESSARY;
        else
                return skb->ip_summed = CHECKSUM_NONE;
 }

Tom Herbert just got rid of the use of CHECKSUM_UNNECESSARY in the transmit path a little while ago. Please don't reintroduce it.

Quoted 16 lines

net/core/skbuff.c +6 -2

index 704b69682085dec77f3d0f990aaf0024afd705b9..96f223f8d769d2765fd64348830c76cb222906c8 100644
@@ -3017,8 +3017,16 @@ struct sk_buff *skb_segment(struct sk_buff *head_skb,
                int size;
 
                len = head_skb->len - offset;
-               if (len > mss)
-                       len = mss;
+               if (len > mss) {
+                       /* FIXME: A define is surely welcomed, but maybe
+                        * shinfo->txflags is better for this flag, but
+                        * we need to expand it then
+                        */
+                       if (mss = 1)

net/core/skbuff.c struct sk_buff *skb_segment(struct sk_buff *head_skb, · line 3030

Quoted 5 lines

net/core/skbuff.c +4 -0

+                               len = list_skb->len;
+                       else
+                               len = mss;
+               }
 

Using 0xFFFF here as a flag with the MSS value would likely be much more readable.

Quoted 22 lines

net/core/skbuff.c +0 -0

                hsize = skb_headlen(head_skb) - offset;
                if (hsize < 0)

net/ipv4/af_inet.c +1 -0

index 5c5db6636704daa0c49fc13e84b2c5b282a44ed3..ec1c779bb664d1399d74f2bd7016e30b648ce47d 100644
@@ -1220,6 +1220,7 @@ static struct sk_buff *inet_gso_segment(struct sk_buff *skb,
                       SKB_GSO_UDP_TUNNEL |
                       SKB_GSO_UDP_TUNNEL_CSUM |
                       SKB_GSO_TUNNEL_REMCSUM |
+                      SKB_GSO_SCTP |
                       0)))
                goto out;
 

net/sctp/offload.c +0 -0

index 7080a6318da7110c1688dd0c5bb240356dbd0cd3..3b96035fa180a4e7195f7b6e7a8be7b97c8f8b26 100644
@@ -36,8 +36,61 @@
 #include <net/sctp/checksum.h>
 #include <net/protocol.h>
 

net/sctp/offload.c line 44

Quoted 6 lines

net/sctp/offload.c +6 -0

+static __le32 sctp_gso_make_checksum(struct sk_buff *skb)
+{
+       skb->ip_summed = CHECKSUM_NONE;
+       return sctp_compute_cksum(skb, skb_transport_offset(skb));
+}
+

I really despise the naming of this bit here. SCTP does not use a checksum. It uses a CRC. Please don't call this a checksum as it
will just make the code really confusing. I think the name should be
something like gso_make_crc32c.

I think we need to address the CRC issues before we can really get into segmentation. Specifically we need to be able to offload SCTP and FCoE in software since they both use the CHECKSUM_PARTIAL value and then we can start cleaning up more of this mess and move onto segmentation.

Quoted 20 lines

net/sctp/offload.c +20 -0

+static struct sk_buff *sctp_gso_segment(struct sk_buff *skb,
+                                       netdev_features_t features)
+{
+       struct sk_buff *segs = ERR_PTR(-EINVAL);
+       struct sctphdr *sh;
+
+       sh = sctp_hdr(skb);
+       if (!pskb_may_pull(skb, sizeof(*sh)))
+               goto out;
+
+       __skb_pull(skb, sizeof(*sh));
+
+       if (skb_gso_ok(skb, features | NETIF_F_GSO_ROBUST)) {
+               /* Packet is from an untrusted source, reset gso_segs. */
+               int type = skb_shinfo(skb)->gso_type;
+
+               if (unlikely(type &
+                            ~(SKB_GSO_SCTP | SKB_GSO_DODGY |
+                              0) ||
+                            !(type & (SKB_GSO_SCTP))))

net/sctp/offload.c line 74

Quoted 10 lines

net/sctp/offload.c +10 -0

+                       goto out;
+
+               /* This should not happen as no NIC has SCTP GSO
+                * offloading, it's always via software and thus we
+                * won't send a large packet down the stack.
+                */
+               WARN_ONCE(1, "SCTP segmentation offloading to NICs is not supported.");
+               goto out;
+       }
+

So what you are going to end up needing here is some way to tell the hardware that you are doing the checksum no matter what. There is no value in you computing a 1's compliment checksum for the payload if you aren't going to use it. What you can probably do is just clear the standard checksum flags and then OR in NETIF_F_HW_CSUM if NETIF_F_SCTP_CRC is set and that should get skb_segment to skip offloading the checksum.

One other bit that will make this more complicated is if we ever get around to supporting SCTP in tunnels. Then we will need to sort out how things like remote checksum offload should impact SCTP, and how to deal with needing to compute both a CRC and 1's compliment checksum. What we would probably need to do is check for encap_hdr_csum and if it is set and we are doing SCTP then we would need to clear the NETIF_F_HW_CSUM, NETIF_F_IP_CSUM, and NETIF_F_IPV6_CSUM flags.

Quoted 7 lines

net/sctp/offload.c +7 -0

+       segs = skb_segment(skb, features);
+       if (IS_ERR(segs))
+               goto out;
+
+       /* All that is left is update SCTP CRC if necessary */
+       for (skb = segs; skb; skb = skb->next) {
+               if (skb->ip_summed != CHECKSUM_PARTIAL) {

net/sctp/offload.c line 86

Quoted 5 lines

net/sctp/offload.c +5 -0

+                       sh = sctp_hdr(skb);
+                       sh->checksum = sctp_gso_make_checksum(skb);
+               }
+       }
+

Okay, so it looks like you are doing the right thing here and leaving this as CHECKSUM_PARTIAL.

net/sctp/offload.c line 95

Quoted 9 lines

net/sctp/offload.c +5 -0

+out:
+       return segs;
+}
+
 static const struct net_offload sctp_offload = {
        .callbacks = {
+               .gso_segment = sctp_gso_segment,
        },
 };

Quoted 41 lines · Alexander Duyck

Quoted 32 lines · Marcelo Ricardo Leitner

This patch enables SCTP to do GSO.

SCTP has this pecualiarty that its packets cannot be just segmented to (P)MTU. Its chunks must be contained in IP segments, padding respected. So we can't just generate a big skb, set gso_size to the fragmentation point and deliver it to IP layer.

Instead, this patch proposes that SCTP build a skb as it would be if it was received using GRO. That is, there will be a cover skb with the headers (incluing SCTP one) and children ones containing the actual SCTP chunks, already segmented in a way that respects SCTP RFCs and MTU.

This way SCTP can benefit from GSO and instead of passing several packets through the stack, it can pass a single large packet if there are enough data queued and cwnd allows.

Main points that need help:
- Usage of skb_gro_receive()
It fits nicely in there and properly handles offsets/lens, though the name means another thing. If you agree with this usage, we can rename it to something like skb_coalesce

- Checksum handling
Why only packets with checksum offloaded can be GSOed? Most of the NICs doesn't support SCTP CRC offloading and this will nearly defeat this feature. If checksum is being computed in sw, it doesn't really matter if it's earlier or later, right?
This patch hacks skb_needs_check() to allow using GSO with sw-computed checksums.
Also the meaning of UNNECESSARY and NONE are quite foggy to me yet and its usage may be wrong.

Quoted 6 lines

- gso_size = 1
There is skb_is_gso() all over the stack and it basically checks for non-zero skb_shinfo(skb)->gso_size. Setting it to 1 is the hacky way I found to keep skb_is_gso() working while being able to signal to skb_segment() that it shouldn't use gso_size but instead the fragment sizes themselves. skb_segment() will mainly just unpack the skb then.

Quoted 3 lines

Instead of 1 why not use 0xFFFF? It is a value that can never be used for a legitimate segment size since IP total length is a 16 bit value and includes the IP header in the size.

Just felt that 1 was unpractical. But perhaps with no hard restriction like the one for 0xFFFF you said. I can replace it, 0xFFFF is better.

Quoted 72 lines

Quoted 72 lines

- socket / gso max values
usage of sk_setup_caps() still needs a review

Signed-off-by Marcelo Ricardo Leitner <[email protected]>

---
 include/linux/netdev_features.h |   7 +-
 include/linux/netdevice.h       |   1 +
 net/core/dev.c                  |   6 +-
 net/core/skbuff.c               |  12 +-
 net/ipv4/af_inet.c              |   1 +
 net/sctp/offload.c              |  53 +++++++
 net/sctp/output.c               | 338 +++++++++++++++++++++++++---------------
 net/sctp/socket.c               |   2 +
 8 files changed, 292 insertions(+), 128 deletions(-)

include/linux/netdev_features.h +5 -2

index d9654f0eecb3519383441afa6b131ff9a5898485..f678998841f1800e0f2fe416a79935197d4ed305 100644
@@ -48,8 +48,9 @@ enum {
        NETIF_F_GSO_UDP_TUNNEL_BIT,     /* ... UDP TUNNEL with TSO */
        NETIF_F_GSO_UDP_TUNNEL_CSUM_BIT,/* ... UDP TUNNEL with TSO & CSUM */
        NETIF_F_GSO_TUNNEL_REMCSUM_BIT, /* ... TUNNEL with TSO & REMCSUM */
+       NETIF_F_GSO_SCTP_BIT,           /* ... SCTP fragmentation */
        /**/NETIF_F_GSO_LAST =          /* last bit, see GSO_MASK */
-               NETIF_F_GSO_TUNNEL_REMCSUM_BIT,
+               NETIF_F_GSO_SCTP_BIT,
 
        NETIF_F_FCOE_CRC_BIT,           /* FCoE CRC32 */
        NETIF_F_SCTP_CRC_BIT,           /* SCTP checksum offload */

@@ -119,6 +120,7 @@ enum {
 #define NETIF_F_GSO_UDP_TUNNEL __NETIF_F(GSO_UDP_TUNNEL)
 #define NETIF_F_GSO_UDP_TUNNEL_CSUM __NETIF_F(GSO_UDP_TUNNEL_CSUM)
 #define NETIF_F_GSO_TUNNEL_REMCSUM __NETIF_F(GSO_TUNNEL_REMCSUM)
+#define NETIF_F_GSO_SCTP       __NETIF_F(GSO_SCTP)
 #define NETIF_F_HW_VLAN_STAG_FILTER __NETIF_F(HW_VLAN_STAG_FILTER)
 #define NETIF_F_HW_VLAN_STAG_RX        __NETIF_F(HW_VLAN_STAG_RX)
 #define NETIF_F_HW_VLAN_STAG_TX        __NETIF_F(HW_VLAN_STAG_TX)

@@ -144,7 +146,8 @@ enum {
 
 /* List of features with software fallbacks. */
 #define NETIF_F_GSO_SOFTWARE   (NETIF_F_TSO | NETIF_F_TSO_ECN | \
-                                NETIF_F_TSO6 | NETIF_F_UFO)
+                                NETIF_F_TSO6 | NETIF_F_UFO | \
+                                NETIF_F_GSO_SCTP)
 
 /* List of IP checksum features. Note that NETIF_F_ HW_CSUM should not be
  * set in features when NETIF_F_IP_CSUM or NETIF_F_IPV6_CSUM are set--

include/linux/netdevice.h +1 -0

index 289c2314d76668b8357728382bb33d6828617458..ce14fab858bf96dd0f85aca237350c8d8317756e 100644
@@ -3928,6 +3928,7 @@ static inline bool net_gso_ok(netdev_features_t features, int gso_type)
        BUILD_BUG_ON(SKB_GSO_UDP_TUNNEL != (NETIF_F_GSO_UDP_TUNNEL >> NETIF_F_GSO_SHIFT));
        BUILD_BUG_ON(SKB_GSO_UDP_TUNNEL_CSUM != (NETIF_F_GSO_UDP_TUNNEL_CSUM >> NETIF_F_GSO_SHIFT));
        BUILD_BUG_ON(SKB_GSO_TUNNEL_REMCSUM != (NETIF_F_GSO_TUNNEL_REMCSUM >> NETIF_F_GSO_SHIFT));
+       BUILD_BUG_ON(SKB_GSO_SCTP    != (NETIF_F_GSO_SCTP >> NETIF_F_GSO_SHIFT));
 
        return (features & feature) = feature;
 }

net/core/dev.c +3 -1

index 8cba3d852f251c503b193823b71b27aaef3fb3ae..9583284086967c0746de5f553535e25e125714a5 100644
@@ -2680,7 +2680,11 @@ EXPORT_SYMBOL(skb_mac_gso_segment);
 static inline bool skb_needs_check(struct sk_buff *skb, bool tx_path)
 {
        if (tx_path)
-               return skb->ip_summed != CHECKSUM_PARTIAL;
+               /* FIXME: Why only packets with checksum offloading are
+                * supported for GSO?
+                */

Quoted 8 lines

net/core/dev.c EXPORT_SYMBOL(skb_mac_gso_segment); · line 2690

Quoted 5 lines

net/core/dev.c +2 -0

+               return skb->ip_summed != CHECKSUM_PARTIAL &&
+                      skb->ip_summed != CHECKSUM_UNNECESSARY;
        else
                return skb->ip_summed = CHECKSUM_NONE;
 }

Tom Herbert just got rid of the use of CHECKSUM_UNNECESSARY in the transmit path a little while ago. Please don't reintroduce it.

Can you give me some pointers on that? I cannot find such change. skb_needs_check() seems to be like that since beginning.

Quoted 16 lines

Quoted 16 lines

net/core/skbuff.c +6 -2

index 704b69682085dec77f3d0f990aaf0024afd705b9..96f223f8d769d2765fd64348830c76cb222906c8 100644
@@ -3017,8 +3017,16 @@ struct sk_buff *skb_segment(struct sk_buff *head_skb,
                int size;
 
                len = head_skb->len - offset;
-               if (len > mss)
-                       len = mss;
+               if (len > mss) {
+                       /* FIXME: A define is surely welcomed, but maybe
+                        * shinfo->txflags is better for this flag, but
+                        * we need to expand it then
+                        */
+                       if (mss = 1)

Quoted 8 lines

net/core/skbuff.c struct sk_buff *skb_segment(struct sk_buff *head_skb, · line 3030

Quoted 5 lines

net/core/skbuff.c +4 -0

+                               len = list_skb->len;
+                       else
+                               len = mss;
+               }
 

Using 0xFFFF here as a flag with the MSS value would likely be much more readable.

Either way it will be replaced by a define/name instead.

Quoted 22 lines

Quoted 22 lines

net/core/skbuff.c +0 -0

                hsize = skb_headlen(head_skb) - offset;
                if (hsize < 0)

net/ipv4/af_inet.c +1 -0

index 5c5db6636704daa0c49fc13e84b2c5b282a44ed3..ec1c779bb664d1399d74f2bd7016e30b648ce47d 100644
@@ -1220,6 +1220,7 @@ static struct sk_buff *inet_gso_segment(struct sk_buff *skb,
                       SKB_GSO_UDP_TUNNEL |
                       SKB_GSO_UDP_TUNNEL_CSUM |
                       SKB_GSO_TUNNEL_REMCSUM |
+                      SKB_GSO_SCTP |
                       0)))
                goto out;
 

net/sctp/offload.c +0 -0

index 7080a6318da7110c1688dd0c5bb240356dbd0cd3..3b96035fa180a4e7195f7b6e7a8be7b97c8f8b26 100644
@@ -36,8 +36,61 @@
 #include <net/sctp/checksum.h>
 #include <net/protocol.h>
 

Quoted 11 lines

net/sctp/offload.c line 44

Quoted 6 lines

net/sctp/offload.c +6 -0

+static __le32 sctp_gso_make_checksum(struct sk_buff *skb)
+{
+       skb->ip_summed = CHECKSUM_NONE;
+       return sctp_compute_cksum(skb, skb_transport_offset(skb));
+}
+

I really despise the naming of this bit here. SCTP does not use a checksum. It uses a CRC. Please don't call this a checksum as it
will just make the code really confusing. I think the name should be
something like gso_make_crc32c.

Agreed. SCTP code still references it as 'cksum'. I'll change that in another patch.

Quoted 5 lines

I think we need to address the CRC issues before we can really get into segmentation. Specifically we need to be able to offload SCTP and FCoE in software since they both use the CHECKSUM_PARTIAL value and then we can start cleaning up more of this mess and move onto segmentation.

Hm? The mess on CRC issues here is caused by this patch alone. It's good as it is today. And a good part of this mess is caused by trying to GSO without offloading CRC too.

Or you mean that SCTP and FCoE should stop using CHECKSUM_* at all?

Quoted 31 lines

Quoted 20 lines

net/sctp/offload.c +20 -0

+static struct sk_buff *sctp_gso_segment(struct sk_buff *skb,
+                                       netdev_features_t features)
+{
+       struct sk_buff *segs = ERR_PTR(-EINVAL);
+       struct sctphdr *sh;
+
+       sh = sctp_hdr(skb);
+       if (!pskb_may_pull(skb, sizeof(*sh)))
+               goto out;
+
+       __skb_pull(skb, sizeof(*sh));
+
+       if (skb_gso_ok(skb, features | NETIF_F_GSO_ROBUST)) {
+               /* Packet is from an untrusted source, reset gso_segs. */
+               int type = skb_shinfo(skb)->gso_type;
+
+               if (unlikely(type &
+                            ~(SKB_GSO_SCTP | SKB_GSO_DODGY |
+                              0) ||
+                            !(type & (SKB_GSO_SCTP))))

net/sctp/offload.c line 74

Quoted 10 lines

net/sctp/offload.c +10 -0

+                       goto out;
+
+               /* This should not happen as no NIC has SCTP GSO
+                * offloading, it's always via software and thus we
+                * won't send a large packet down the stack.
+                */
+               WARN_ONCE(1, "SCTP segmentation offloading to NICs is not supported.");
+               goto out;
+       }
+

Quoted 7 lines

So what you are going to end up needing here is some way to tell the hardware that you are doing the checksum no matter what. There is no value in you computing a 1's compliment checksum for the payload if you aren't going to use it. What you can probably do is just clear the standard checksum flags and then OR in NETIF_F_HW_CSUM if NETIF_F_SCTP_CRC is set and that should get skb_segment to skip offloading the checksum.

Interesting, ok

Quoted 7 lines

One other bit that will make this more complicated is if we ever get around to supporting SCTP in tunnels. Then we will need to sort out how things like remote checksum offload should impact SCTP, and how to deal with needing to compute both a CRC and 1's compliment checksum. What we would probably need to do is check for encap_hdr_csum and if it is set and we are doing SCTP then we would need to clear the NETIF_F_HW_CSUM, NETIF_F_IP_CSUM, and NETIF_F_IPV6_CSUM flags.

Yup. And that includes on storing pointers to where to store each of it.

Quoted 7 lines

Quoted 7 lines

net/sctp/offload.c +7 -0

+       segs = skb_segment(skb, features);
+       if (IS_ERR(segs))
+               goto out;
+
+       /* All that is left is update SCTP CRC if necessary */
+       for (skb = segs; skb; skb = skb->next) {
+               if (skb->ip_summed != CHECKSUM_PARTIAL) {

Quoted 8 lines

net/sctp/offload.c line 86

Quoted 5 lines

net/sctp/offload.c +5 -0

+                       sh = sctp_hdr(skb);
+                       sh->checksum = sctp_gso_make_checksum(skb);
+               }
+       }
+

Okay, so it looks like you are doing the right thing here and leaving this as CHECKSUM_PARTIAL.

Actually no then. sctp_gso_make_checksum() replaces it:

net/sctp/offload.c +4 -0

+static __le32 sctp_gso_make_checksum(struct sk_buff *skb)
+{
+       skb->ip_summed = CHECKSUM_NONE;
+       return sctp_compute_cksum(skb, skb_transport_offset(skb));

Why again would have to leave it as CHECKSUM_PARTIAL? IP header?

Quoted 2 lines · Marcelo Ricardo Leitner

Quoted 1 line · Alexander Duyck

On Wed, Jan 27, 2016 at 9:06 AM, Marcelo Ricardo Leitner

Quoted 21 lines

Quoted 12 lines

Quoted 12 lines

net/core/dev.c +3 -1

index 8cba3d852f251c503b193823b71b27aaef3fb3ae..9583284086967c0746de5f553535e25e125714a5 100644
@@ -2680,7 +2680,11 @@ EXPORT_SYMBOL(skb_mac_gso_segment);
  static inline bool skb_needs_check(struct sk_buff *skb, bool tx_path)
  {
         if (tx_path)
-               return skb->ip_summed != CHECKSUM_PARTIAL;
+               /* FIXME: Why only packets with checksum offloading are
+                * supported for GSO?
+                */

Quoted 8 lines

net/core/dev.c EXPORT_SYMBOL(skb_mac_gso_segment); · line 2690

Quoted 5 lines

net/core/dev.c +2 -0

+               return skb->ip_summed != CHECKSUM_PARTIAL &&
+                      skb->ip_summed != CHECKSUM_UNNECESSARY;
         else
                 return skb->ip_summed = CHECKSUM_NONE;
  }

Tom Herbert just got rid of the use of CHECKSUM_UNNECESSARY in the transmit path a little while ago. Please don't reintroduce it.

Quoted 2 lines

Can you give me some pointers on that? I cannot find such change. skb_needs_check() seems to be like that since beginning.

Maybe you need to update your kernel. All this stuff was changed in December and has been this way for a little while now.

Commits:
7a6ae71b24905 "net: Elaborate on checksum offload interface description" 253aab0597d9e "fcoe: Use CHECKSUM_PARTIAL to indicate CRC offload" 53692b1de419c "sctp: Rename NETIF_F_SCTP_CSUM to NETIF_F_SCTP_CRC"

The main reason I even noticed it is because of some of the work I did on the Intel NIC offloads.

Quoted 33 lines

Quoted 8 lines

Quoted 8 lines

net/sctp/offload.c +0 -0

index 7080a6318da7110c1688dd0c5bb240356dbd0cd3..3b96035fa180a4e7195f7b6e7a8be7b97c8f8b26 100644
@@ -36,8 +36,61 @@
  #include <net/sctp/checksum.h>
  #include <net/protocol.h>
 

Quoted 11 lines

net/sctp/offload.c line 44

Quoted 6 lines

net/sctp/offload.c +6 -0

+static __le32 sctp_gso_make_checksum(struct sk_buff *skb)
+{
+       skb->ip_summed = CHECKSUM_NONE;
+       return sctp_compute_cksum(skb, skb_transport_offset(skb));
+}
+

I really despise the naming of this bit here. SCTP does not use a checksum. It uses a CRC. Please don't call this a checksum as it
will just make the code really confusing. I think the name should be
something like gso_make_crc32c.

Agreed. SCTP code still references it as 'cksum'. I'll change that in another patch.

Quoted 5 lines

I think we need to address the CRC issues before we can really get into segmentation. Specifically we need to be able to offload SCTP and FCoE in software since they both use the CHECKSUM_PARTIAL value and then we can start cleaning up more of this mess and move onto segmentation.

Hm? The mess on CRC issues here is caused by this patch alone. It's good as it is today. And a good part of this mess is caused by trying to GSO without offloading CRC too.

Quoted 1 line

Or you mean that SCTP and FCoE should stop using CHECKSUM_* at all?

Well after Tom's change both SCTP and FCoE use CHECKSUM_PARTIAL. CHECKSUM_PARTIAL is what is used to indicate to the hardware that a checksum offload has been requested so that is what is looked for at the driver level.

My concern with all this is that we should probably be looking at coming up with a means of offloading this in software when skb_checksum_help is called. Right now validate_xmit_skb doesn't have any understanding of what to do with SCTP or FCoE and will try to just compute a checksum for them.

Quoted 49 lines

Quoted 31 lines

Quoted 20 lines

net/sctp/offload.c +20 -0

+static struct sk_buff *sctp_gso_segment(struct sk_buff *skb,
+                                       netdev_features_t features)
+{
+       struct sk_buff *segs = ERR_PTR(-EINVAL);
+       struct sctphdr *sh;
+
+       sh = sctp_hdr(skb);
+       if (!pskb_may_pull(skb, sizeof(*sh)))
+               goto out;
+
+       __skb_pull(skb, sizeof(*sh));
+
+       if (skb_gso_ok(skb, features | NETIF_F_GSO_ROBUST)) {
+               /* Packet is from an untrusted source, reset gso_segs. */
+               int type = skb_shinfo(skb)->gso_type;
+
+               if (unlikely(type &
+                            ~(SKB_GSO_SCTP | SKB_GSO_DODGY |
+                              0) ||
+                            !(type & (SKB_GSO_SCTP))))

net/sctp/offload.c line 74

Quoted 10 lines

net/sctp/offload.c +10 -0

+                       goto out;
+
+               /* This should not happen as no NIC has SCTP GSO
+                * offloading, it's always via software and thus we
+                * won't send a large packet down the stack.
+                */
+               WARN_ONCE(1, "SCTP segmentation offloading to NICs is not supported.");
+               goto out;
+       }
+

Quoted 7 lines

So what you are going to end up needing here is some way to tell the hardware that you are doing the checksum no matter what. There is no value in you computing a 1's compliment checksum for the payload if you aren't going to use it. What you can probably do is just clear the standard checksum flags and then OR in NETIF_F_HW_CSUM if NETIF_F_SCTP_CRC is set and that should get skb_segment to skip offloading the checksum.

Interesting, ok

Quoted 7 lines

One other bit that will make this more complicated is if we ever get around to supporting SCTP in tunnels. Then we will need to sort out how things like remote checksum offload should impact SCTP, and how to deal with needing to compute both a CRC and 1's compliment checksum. What we would probably need to do is check for encap_hdr_csum and if it is set and we are doing SCTP then we would need to clear the NETIF_F_HW_CSUM, NETIF_F_IP_CSUM, and NETIF_F_IPV6_CSUM flags.

Quoted 1 line

Yup. And that includes on storing pointers to where to store each of it.

Actually the pointers bit is easy. The csum_start and csum_offset values should be set up after you have segmented the skb and should be updated after the skb has been segmented. If nothing else you can probably take a look at the TCP code tcp_gso_segment and tcp4_gso_segment for inspiration. Basically you need to make sure that you set the ip_summed, csum_start, and csum_offset values for your first frame before you start segmenting it into multiple frames.

Quoted 22 lines

Quoted 7 lines

Quoted 7 lines

net/sctp/offload.c +7 -0

+       segs = skb_segment(skb, features);
+       if (IS_ERR(segs))
+               goto out;
+
+       /* All that is left is update SCTP CRC if necessary */
+       for (skb = segs; skb; skb = skb->next) {
+               if (skb->ip_summed != CHECKSUM_PARTIAL) {

Quoted 8 lines

net/sctp/offload.c line 86

Quoted 5 lines

net/sctp/offload.c +5 -0

+                       sh = sctp_hdr(skb);
+                       sh->checksum = sctp_gso_make_checksum(skb);
+               }
+       }
+

Okay, so it looks like you are doing the right thing here and leaving this as CHECKSUM_PARTIAL.

Actually no then. sctp_gso_make_checksum() replaces it:

net/sctp/offload.c +4 -0

+static __le32 sctp_gso_make_checksum(struct sk_buff *skb)
+{
+       skb->ip_summed = CHECKSUM_NONE;
+       return sctp_compute_cksum(skb, skb_transport_offset(skb));

Quoted 1 line

Why again would have to leave it as CHECKSUM_PARTIAL? IP header?

My earlier comment is actually incorrect. This section is pretty much broken since CHECKSUM_PARTIAL only reflects a 1's compliment checksum in the case of skb_segment so whatever the value it is worthless. CHECKSUM_PARTIAL is used to indicate if a given frame needs to be offloaded. It is meant to let the device know that it still needs to compute a checksum or CRC beginning at csum_start and then storing the new value at csum_offset. However for skb_segment it is actually referring to a 1's compliment checksum and if it returns CHECKSUM_NONE it means it is stored in skb->csum which would really wreck things for you since that was your skb->csum_start and skb->csum_offset values. I have a patch to change this so that we update a checksum in the SKB_GSO_CB, but I wasn't planning on submitting that until net-next opens.

In the case of SCTP you probably don't even need to bother checking the value since it is meaningless as skb_segment doesn't know how to do an SCTP checksum anyway. To that end for now what you could do is just set NETIF_F_HW_CSUM. This way skb_segment won't go and try to compute a 1's compliment checksum on the payload since there is no actual need for it.

One other bit you will need to do is to check the value of SCTP_CRC outside of skb_segment. You might look at how __skb_udp_tunnel_segment does this to populate its own offload_csum boolean value, though you would want to use features, not skb->dev->features as that is a bit of a workaround since features is stripped by hw_enc_features in some paths if I recall correctly.

Once the frames are segmented and if you don't support the offload you could then call gso_make_crc32c() or whatever you want to name it to perform the CRC calculation and populate the field. One question by

the way. Don't you need to initialize the checksum value to 0 before you compute it? I think you might have missed that step when you were setting this up.

Em 30-01-2016 02:07, Alexander Duyck escreveu:

Quoted 7 lines

Quoted 4 lines · Marcelo Ricardo Leitner

Quoted 2 lines · Alexander Duyck

On Wed, Jan 27, 2016 at 9:06 AM, Marcelo Ricardo Leitner

...

Quoted 37 lines

Quoted 25 lines

Quoted 15 lines

Quoted 15 lines

net/core/dev.c +0 -0

index 8cba3d852f251c503b193823b71b27aaef3fb3ae..9583284086967c0746de5f553535e25e125714a5 100644

net/core/dev.c +4 -1

@@ -2680,7 +2680,11 @@ EXPORT_SYMBOL(skb_mac_gso_segment);
  static inline bool skb_needs_check(struct sk_buff *skb, bool tx_path)
  {
         if (tx_path)
-               return skb->ip_summed != CHECKSUM_PARTIAL;
+               /* FIXME: Why only packets with checksum offloading are
+                * supported for GSO?
+                */
+               return skb->ip_summed != CHECKSUM_PARTIAL &&

Quoted 8 lines

net/core/dev.c EXPORT_SYMBOL(skb_mac_gso_segment); · line 2690

Quoted 4 lines

net/core/dev.c +1 -0

+                      skb->ip_summed != CHECKSUM_UNNECESSARY;
         else
                 return skb->ip_summed = CHECKSUM_NONE;
  }

Tom Herbert just got rid of the use of CHECKSUM_UNNECESSARY in the transmit path a little while ago. Please don't reintroduce it.

Quoted 2 lines

Can you give me some pointers on that? I cannot find such change. skb_needs_check() seems to be like that since beginning.

Maybe you need to update your kernel. All this stuff was changed in December and has been this way for a little while now.

Commits:
7a6ae71b24905 "net: Elaborate on checksum offload interface description" 253aab0597d9e "fcoe: Use CHECKSUM_PARTIAL to indicate CRC offload" 53692b1de419c "sctp: Rename NETIF_F_SCTP_CSUM to NETIF_F_SCTP_CRC"

Quoted 2 lines

The main reason I even noticed it is because of some of the work I did on the Intel NIC offloads.

Ok I have those here, but my need here is different. I want to do GSO with packets that won't have CRC offloaded, so I shouldn't use CHECKSUM_PARTIAL but something else.

...

Quoted 38 lines

Quoted 38 lines

Quoted 10 lines

Quoted 10 lines

net/sctp/offload.c +0 -0

index 7080a6318da7110c1688dd0c5bb240356dbd0cd3..3b96035fa180a4e7195f7b6e7a8be7b97c8f8b26 100644

net/sctp/offload.c +0 -0

@@ -36,8 +36,61 @@
  #include <net/sctp/checksum.h>
  #include <net/protocol.h>
 

Quoted 12 lines

net/sctp/offload.c line 44

Quoted 6 lines

net/sctp/offload.c +6 -0

+static __le32 sctp_gso_make_checksum(struct sk_buff *skb)
+{
+       skb->ip_summed = CHECKSUM_NONE;
+       return sctp_compute_cksum(skb, skb_transport_offset(skb));
+}
+

I really despise the naming of this bit here. SCTP does not use a checksum. It uses a CRC. Please don't call this a checksum as it
will just make the code really confusing. I think the name should be
something like gso_make_crc32c.

Agreed. SCTP code still references it as 'cksum'. I'll change that in another patch.

Quoted 5 lines

I think we need to address the CRC issues before we can really get into segmentation. Specifically we need to be able to offload SCTP and FCoE in software since they both use the CHECKSUM_PARTIAL value and then we can start cleaning up more of this mess and move onto segmentation.

Hm? The mess on CRC issues here is caused by this patch alone. It's good as it is today. And a good part of this mess is caused by trying to GSO without offloading CRC too.

Quoted 7 lines

Quoted 1 line

Or you mean that SCTP and FCoE should stop using CHECKSUM_* at all?

Well after Tom's change both SCTP and FCoE use CHECKSUM_PARTIAL. CHECKSUM_PARTIAL is what is used to indicate to the hardware that a checksum offload has been requested so that is what is looked for at the driver level.

SCTP was actually already using CHECKSUM_PARTIAL. That patch was just a rename in an attempt to make this crc difference more evident. Yet I'll continue the rename within sctp code.

Quoted 5 lines

My concern with all this is that we should probably be looking at coming up with a means of offloading this in software when skb_checksum_help is called. Right now validate_xmit_skb doesn't have any understanding of what to do with SCTP or FCoE and will try to just compute a checksum for them.

My worry is placed a bit earlier than that, I think. Currently I just cannot do GSO with packets that doesn't have checksum/crc offloaded too because validate_xmit_skb() will complain.

As NICs hardly have sctp crc offloading capabilities, I'm thinking it makes sense to do GSO even without crc offloaded. After all, it doesn't matter much in which stage we are computing the crc as we are computing it anyway.

Quoted 57 lines

Quoted 54 lines

Quoted 34 lines

Quoted 22 lines

net/sctp/offload.c +21 -0

+static struct sk_buff *sctp_gso_segment(struct sk_buff *skb,
+                                       netdev_features_t features)
+{
+       struct sk_buff *segs = ERR_PTR(-EINVAL);
+       struct sctphdr *sh;
+
+       sh = sctp_hdr(skb);
+       if (!pskb_may_pull(skb, sizeof(*sh)))
+               goto out;
+
+       __skb_pull(skb, sizeof(*sh));
+
+       if (skb_gso_ok(skb, features | NETIF_F_GSO_ROBUST)) {
+               /* Packet is from an untrusted source, reset gso_segs.
 */
+               int type = skb_shinfo(skb)->gso_type;
+
+               if (unlikely(type &
+                            ~(SKB_GSO_SCTP | SKB_GSO_DODGY |
+                              0) ||
+                            !(type & (SKB_GSO_SCTP))))
+                       goto out;

net/sctp/offload.c line 74

Quoted 10 lines

net/sctp/offload.c +9 -0

+
+               /* This should not happen as no NIC has SCTP GSO
+                * offloading, it's always via software and thus we
+                * won't send a large packet down the stack.
+                */
+               WARN_ONCE(1, "SCTP segmentation offloading to NICs is
 not supported.");
+               goto out;
+       }
+

Quoted 7 lines

So what you are going to end up needing here is some way to tell the hardware that you are doing the checksum no matter what. There is no value in you computing a 1's compliment checksum for the payload if you aren't going to use it. What you can probably do is just clear the standard checksum flags and then OR in NETIF_F_HW_CSUM if NETIF_F_SCTP_CRC is set and that should get skb_segment to skip offloading the checksum.

Interesting, ok

Quoted 7 lines

One other bit that will make this more complicated is if we ever get around to supporting SCTP in tunnels. Then we will need to sort out how things like remote checksum offload should impact SCTP, and how to deal with needing to compute both a CRC and 1's compliment checksum. What we would probably need to do is check for encap_hdr_csum and if it is set and we are doing SCTP then we would need to clear the NETIF_F_HW_CSUM, NETIF_F_IP_CSUM, and NETIF_F_IPV6_CSUM flags.

Quoted 1 line

Yup. And that includes on storing pointers to where to store each of it.

Quoted 7 lines

Actually the pointers bit is easy. The csum_start and csum_offset values should be set up after you have segmented the skb and should be updated after the skb has been segmented. If nothing else you can probably take a look at the TCP code tcp_gso_segment and tcp4_gso_segment for inspiration. Basically you need to make sure that you set the ip_summed, csum_start, and csum_offset values for your first frame before you start segmenting it into multiple frames.

Ah yes, ok, that's for now, when not doing crc offloading with some chksum offloading (tunnel) too.

Quoted 24 lines

Quoted 24 lines

Quoted 8 lines

Quoted 8 lines

net/sctp/offload.c +8 -0

+       segs = skb_segment(skb, features);
+       if (IS_ERR(segs))
+               goto out;
+
+       /* All that is left is update SCTP CRC if necessary */
+       for (skb = segs; skb; skb = skb->next) {
+               if (skb->ip_summed != CHECKSUM_PARTIAL) {
+                       sh = sctp_hdr(skb);

Quoted 8 lines

net/sctp/offload.c line 86

Quoted 4 lines

net/sctp/offload.c +4 -0

+                       sh->checksum = sctp_gso_make_checksum(skb);
+               }
+       }
+

Okay, so it looks like you are doing the right thing here and leaving this as CHECKSUM_PARTIAL.

Actually no then. sctp_gso_make_checksum() replaces it:

net/sctp/offload.c +4 -0

+static __le32 sctp_gso_make_checksum(struct sk_buff *skb)
+{
+       skb->ip_summed = CHECKSUM_NONE;
+       return sctp_compute_cksum(skb, skb_transport_offset(skb));

Quoted 16 lines

Quoted 1 line

Why again would have to leave it as CHECKSUM_PARTIAL? IP header?

My earlier comment is actually incorrect. This section is pretty much broken since CHECKSUM_PARTIAL only reflects a 1's compliment checksum in the case of skb_segment so whatever the value it is worthless. CHECKSUM_PARTIAL is used to indicate if a given frame needs to be offloaded. It is meant to let the device know that it still needs to compute a checksum or CRC beginning at csum_start and then storing the new value at csum_offset. However for skb_segment it is actually referring to a 1's compliment checksum and if it returns CHECKSUM_NONE it means it is stored in skb->csum which would really wreck things for you since that was your skb->csum_start and skb->csum_offset values. I have a patch to change this so that we update a checksum in the SKB_GSO_CB, but I wasn't planning on submitting that until net-next opens.

sctp currently ignores skb->csum. It doesn't mess with the crc but computing it is at least not optimal, yes.

Quoted 6 lines

In the case of SCTP you probably don't even need to bother checking the value since it is meaningless as skb_segment doesn't know how to do an SCTP checksum anyway. To that end for now what you could do is just set NETIF_F_HW_CSUM. This way skb_segment won't go and try to compute a 1's compliment checksum on the payload since there is no actual need for it.

Nice, ok.

Quoted 9 lines

One other bit you will need to do is to check the value of SCTP_CRC outside of skb_segment. You might look at how __skb_udp_tunnel_segment does this to populate its own offload_csum boolean value, though you would want to use features, not skb->dev->features as that is a bit of a workaround since features is stripped by hw_enc_features in some paths if I recall correctly.

Quoted 3 lines

Once the frames are segmented and if you don't support the offload you could then call gso_make_crc32c() or whatever you want to name it to perform the CRC calculation and populate the field. One question by

Hmmm.. does it mean that we can use CHECKSUM_PARTIAL then even if CRC offloading is not possible then? Because the packet will not be offloaded in the end, yes, but this solves my questions above. Then while doing GSO, it re-evaluates if it can offload crc or not?

Quoted 3 lines

the way. Don't you need to initialize the checksum value to 0 before you compute it? I think you might have missed that step when you were setting this up.

It's fine :) sctp_compute_cksum will replace it with zeroes, calculate, and put back the old value, which then we overwrite with the new one at sctp_gso_segment.

Quoted 7 lines · Marcelo Ricardo Leitner

Em 30-01-2016 02:07, Alexander Duyck escreveu:

Quoted 6 lines

Quoted 3 lines · Marcelo Ricardo Leitner

Quoted 1 line · Alexander Duyck

Quoted 7 lines

Quoted 1 line

Quoted 1 line

Quoted 1 line

On Wed, Jan 27, 2016 at 9:06 AM, Marcelo Ricardo Leitner

...

Quoted 3 lines

Quoted 3 lines

Quoted 3 lines

Quoted 3 lines

net/core/dev.c +0 -0

index

Quoted 5 lines

Quoted 5 lines

Quoted 5 lines

Quoted 5 lines

net/core/dev.c EXPORT_SYMBOL(skb_mac_gso_segment); · line 2680

Quoted 5 lines

100644

net/core/dev.c +0 -0

@@ -2680,7 +2680,11 @@ EXPORT_SYMBOL(skb_mac_gso_segment);
   static inline bool skb_needs_check(struct sk_buff *skb, bool 

Quoted 4 lines

Quoted 4 lines

Quoted 4 lines

Quoted 4 lines

net/core/dev.c EXPORT_SYMBOL(skb_mac_gso_segment);

Quoted 4 lines

net/core/dev.c +1 -1

   {
          if (tx_path)
-               return skb->ip_summed != CHECKSUM_PARTIAL;
+               /* FIXME: Why only packets with checksum 

Quoted 17 lines

Quoted 17 lines

Quoted 15 lines

Quoted 11 lines

net/core/dev.c EXPORT_SYMBOL(skb_mac_gso_segment); · line 2690

Quoted 7 lines

net/core/dev.c +4 -0

+                * supported for GSO?
+                */
+               return skb->ip_summed != CHECKSUM_PARTIAL &&
+                      skb->ip_summed != CHECKSUM_UNNECESSARY;
          else
                  return skb->ip_summed = CHECKSUM_NONE;
   }

Tom Herbert just got rid of the use of CHECKSUM_UNNECESSARY in the transmit path a little while ago. Please don't reintroduce it.

Can you give me some pointers on that? I cannot find such change. skb_needs_check() seems to be like that since beginning.

Quoted 5 lines

Quoted 5 lines

Maybe you need to update your kernel. All this stuff was changed in December and has been this way for a little while now.

Commits:
7a6ae71b24905 "net: Elaborate on checksum offload interface

Quoted 8 lines

Quoted 5 lines

253aab0597d9e "fcoe: Use CHECKSUM_PARTIAL to indicate CRC offload" 53692b1de419c "sctp: Rename NETIF_F_SCTP_CSUM to NETIF_F_SCTP_CRC"

The main reason I even noticed it is because of some of the work I did on the Intel NIC offloads.

Ok I have those here, but my need here is different. I want to do

Quoted 1 line

packets that won't have CRC offloaded, so I shouldn't use

Quoted 1 line

but something else.

CHECKSUM_NONE if you don't want to have any of the CRC or checksums offloaded. However as I mentioned before you will want to fake it then since skb_segment assumes it is doing a 1's compliment checksum so you will want to pass NET_F_HW_CSUM as a feature flag and then set CHECKSUM_NONE after the frame has been segmented.

Quoted 5 lines

...

Quoted 3 lines

Quoted 3 lines

Quoted 3 lines

Quoted 3 lines

net/sctp/offload.c +0 -0

index

Quoted 7 lines

Quoted 7 lines

Quoted 7 lines

Quoted 7 lines

Quoted 7 lines

100644

net/sctp/offload.c +0 -0

@@ -36,8 +36,61 @@
   #include <net/sctp/checksum.h>
   #include <net/protocol.h>
 

Quoted 11 lines

Quoted 11 lines

Quoted 11 lines

Quoted 11 lines

net/sctp/offload.c line 44

Quoted 6 lines

net/sctp/offload.c +6 -0

+static __le32 sctp_gso_make_checksum(struct sk_buff *skb)
+{
+       skb->ip_summed = CHECKSUM_NONE;
+       return sctp_compute_cksum(skb, skb_transport_offset(skb));
+}
+

I really despise the naming of this bit here. SCTP does not use a checksum. It uses a CRC. Please don't call this a checksum as it
will just make the code really confusing. I think the name

Quoted 6 lines

Quoted 6 lines

Quoted 6 lines

Quoted 1 line

something like gso_make_crc32c.

Agreed. SCTP code still references it as 'cksum'. I'll change that in another patch.

Quoted 8 lines

Quoted 8 lines

Quoted 8 lines

Quoted 5 lines

I think we need to address the CRC issues before we can really get into segmentation. Specifically we need to be able to offload SCTP and FCoE in software since they both use the CHECKSUM_PARTIAL value and then we can start cleaning up more of this mess and move onto segmentation.

Hm? The mess on CRC issues here is caused by this patch alone.

Quoted 1 line

Quoted 1 line

Quoted 1 line

as it is today. And a good part of this mess is caused by trying

Quoted 11 lines

Quoted 9 lines

Quoted 3 lines

without offloading CRC too.

Or you mean that SCTP and FCoE should stop using CHECKSUM_* at all?

Well after Tom's change both SCTP and FCoE use CHECKSUM_PARTIAL. CHECKSUM_PARTIAL is what is used to indicate to the hardware that a checksum offload has been requested so that is what is looked for at the driver level.

Quoted 3 lines

SCTP was actually already using CHECKSUM_PARTIAL. That patch was just a rename in an attempt to make this crc difference more evident. Yet I'll continue the rename within sctp code.

Yeah it was FCoE that was doing something different.

Quoted 8 lines

Quoted 5 lines

My concern with all this is that we should probably be looking at coming up with a means of offloading this in software when skb_checksum_help is called. Right now validate_xmit_skb doesn't have any understanding of what to do with SCTP or FCoE and will try to just compute a checksum for them.

My worry is placed a bit earlier than that, I think. Currently I

Quoted 2 lines

do GSO with packets that doesn't have checksum/crc offloaded too because validate_xmit_skb() will complain.

That is probably because you are passing CHECKSUM_PARTIAL instead of CHECKSUM_NONE.

Quoted 1 line

As NICs hardly have sctp crc offloading capabilities, I'm thinking

Quoted 2 lines

sense to do GSO even without crc offloaded. After all, it doesn't matter much in which stage we are computing the crc as we are computing it

Agreed. You will need to support CHECKSUM_PARTIAL being passed to a device that doesn't support SCTP first. That way you can start looking at just always setting CHECKSUM_PARTIAL in the transport layer which is really needed if you want to do SCO (SCTP Segmentation Offload) in the first place. Once you have that you could then start looking at doing the SCO since from that point on you should already be in good shape to address those type of issues. You should probably use the csum_offset value in the skb in order to flag if this is possibly SCTP. As far as I know for now there shouldn't be any other protocols that are using the same offset, and if needed you can actually parse the headers to verify if the frame is actually SCTP.

Quoted 6 lines

Quoted 6 lines

Quoted 6 lines

Quoted 6 lines

Quoted 6 lines

net/sctp/offload.c +6 -0

+static struct sk_buff *sctp_gso_segment(struct sk_buff *skb,
+                                       netdev_features_t features)
+{
+       struct sk_buff *segs = ERR_PTR(-EINVAL);
+       struct sctphdr *sh;
+

Quoted 8 lines

Quoted 8 lines

Quoted 8 lines

Quoted 8 lines

net/sctp/offload.c

Quoted 8 lines

net/sctp/offload.c +8 -0

+       sh = sctp_hdr(skb);
+       if (!pskb_may_pull(skb, sizeof(*sh)))
+               goto out;
+
+       __skb_pull(skb, sizeof(*sh));
+
+       if (skb_gso_ok(skb, features | NETIF_F_GSO_ROBUST)) {
+               /* Packet is from an untrusted source, reset 

Quoted 6 lines

Quoted 6 lines

Quoted 6 lines

Quoted 6 lines

Quoted 6 lines

net/sctp/offload.c +5 -0

 */
+               int type = skb_shinfo(skb)->gso_type;
+
+               if (unlikely(type &
+                            ~(SKB_GSO_SCTP | SKB_GSO_DODGY |
+                              0) ||

Quoted 8 lines

Quoted 8 lines

Quoted 8 lines

Quoted 8 lines

net/sctp/offload.c

Quoted 8 lines

net/sctp/offload.c +8 -0

+                            !(type & (SKB_GSO_SCTP))))
+                       goto out;
+
+               /* This should not happen as no NIC has SCTP GSO
+                * offloading, it's always via software and thus we
+                * won't send a large packet down the stack.
+                */
+               WARN_ONCE(1, "SCTP segmentation offloading to 

Quoted 15 lines

Quoted 15 lines

Quoted 15 lines

Quoted 13 lines

net/sctp/offload.c line 74

Quoted 4 lines

net/sctp/offload.c +3 -0

 not supported.");
+               goto out;
+       }
+

So what you are going to end up needing here is some way to tell the hardware that you are doing the checksum no matter what. There is no value in you computing a 1's compliment checksum for the payload if you aren't going to use it. What you can probably do is just clear the standard checksum flags and then OR in NETIF_F_HW_CSUM if NETIF_F_SCTP_CRC is set and that should get skb_segment to skip offloading the checksum.

Quoted 5 lines

Quoted 5 lines

Quoted 5 lines

Interesting, ok

Quoted 3 lines

One other bit that will make this more complicated is if we ever get around to supporting SCTP in tunnels. Then we will need to sort out how things like remote checksum offload should impact SCTP, and

Quoted 7 lines

Quoted 7 lines

Quoted 7 lines

Quoted 4 lines

deal with needing to compute both a CRC and 1's compliment checksum. What we would probably need to do is check for encap_hdr_csum and if it is set and we are doing SCTP then we would need to clear the NETIF_F_HW_CSUM, NETIF_F_IP_CSUM, and NETIF_F_IPV6_CSUM flags.

Yup. And that includes on storing pointers to where to store each

Quoted 11 lines

Quoted 9 lines

Actually the pointers bit is easy. The csum_start and csum_offset values should be set up after you have segmented the skb and should be updated after the skb has been segmented. If nothing else you can probably take a look at the TCP code tcp_gso_segment and tcp4_gso_segment for inspiration. Basically you need to make sure that you set the ip_summed, csum_start, and csum_offset values for your first frame before you start segmenting it into multiple frames.

Quoted 1 line

Ah yes, ok, that's for now, when not doing crc offloading with some

Quoted 1 line

offloading (tunnel) too.

Actually that would be regardless of tunnel offloading. We don't store the outer checksum offsets. If we need outer checksum we restore them after the fact since the inner checksum offsets are needed as part of the inner header TCP checksum computation.

Quoted 42 lines

Quoted 40 lines

Quoted 24 lines

Quoted 8 lines

Quoted 8 lines

net/sctp/offload.c +8 -0

+       segs = skb_segment(skb, features);
+       if (IS_ERR(segs))
+               goto out;
+
+       /* All that is left is update SCTP CRC if necessary */
+       for (skb = segs; skb; skb = skb->next) {
+               if (skb->ip_summed != CHECKSUM_PARTIAL) {
+                       sh = sctp_hdr(skb);

Quoted 8 lines

net/sctp/offload.c line 86

Quoted 4 lines

net/sctp/offload.c +4 -0

+                       sh->checksum = sctp_gso_make_checksum(skb);
+               }
+       }
+

Okay, so it looks like you are doing the right thing here and leaving this as CHECKSUM_PARTIAL.

Actually no then. sctp_gso_make_checksum() replaces it:

net/sctp/offload.c +4 -0

+static __le32 sctp_gso_make_checksum(struct sk_buff *skb)
+{
+       skb->ip_summed = CHECKSUM_NONE;
+       return sctp_compute_cksum(skb, skb_transport_offset(skb));

Quoted 1 line

Why again would have to leave it as CHECKSUM_PARTIAL? IP header?

My earlier comment is actually incorrect. This section is pretty much broken since CHECKSUM_PARTIAL only reflects a 1's compliment checksum in the case of skb_segment so whatever the value it is worthless. CHECKSUM_PARTIAL is used to indicate if a given frame needs to be offloaded. It is meant to let the device know that it still needs to compute a checksum or CRC beginning at csum_start and then storing the new value at csum_offset. However for skb_segment it is actually referring to a 1's compliment checksum and if it returns CHECKSUM_NONE it means it is stored in skb->csum which would really wreck things for you since that was your skb->csum_start and skb->csum_offset values. I have a patch to change this so that we update a checksum in the SKB_GSO_CB, but I wasn't planning on submitting that until net-next opens.

Quoted 1 line

sctp currently ignores skb->csum. It doesn't mess with the crc but

Quoted 1 line

it is at least not optimal, yes.

Actually sctp sets csum_start and csum_offset if it sets CHECKSUM_PARTIAL. So it does mess with skb->csum since it is contained in a union with those two fields.

Quoted 19 lines

Quoted 6 lines

In the case of SCTP you probably don't even need to bother checking the value since it is meaningless as skb_segment doesn't know how to do an SCTP checksum anyway. To that end for now what you could do is just set NETIF_F_HW_CSUM. This way skb_segment won't go and try to compute a 1's compliment checksum on the payload since there is no actual need for it.

Nice, ok.

Quoted 9 lines

One other bit you will need to do is to check the value of SCTP_CRC outside of skb_segment. You might look at how __skb_udp_tunnel_segment does this to populate its own offload_csum boolean value, though you would want to use features, not skb->dev->features as that is a bit of a workaround since features is stripped by hw_enc_features in some paths if I recall correctly.

Quoted 7 lines

Quoted 3 lines

Once the frames are segmented and if you don't support the offload you could then call gso_make_crc32c() or whatever you want to name it to perform the CRC calculation and populate the field. One question by

Hmmm.. does it mean that we can use CHECKSUM_PARTIAL then even if CRC offloading is not possible then? Because the packet will not be

Quoted 1 line

the end, yes, but this solves my questions above. Then while doing

Quoted 1 line

re-evaluates if it can offload crc or not?

If you compute the CRC you set CHECKSUM_NONE, if you want the device to do it on transmit you should set CHECKSUM_PARTIAL.

Quoted 6 lines

Quoted 3 lines

the way. Don't you need to initialize the checksum value to 0 before you compute it? I think you might have missed that step when you were setting this up.

It's fine :) sctp_compute_cksum will replace it with zeroes,

Quoted 2 lines

put back the old value, which then we overwrite with the new one at sctp_gso_segment.

Right but there are scenarios where this will be offloaded isn't there? You would probably be better off setting the CRC to 0 before you start segmentation and then that way you can either just set csum_offset, csum_start and ip_summed if the lower device supports SCTP CRC offload, otherwise you can just compute it without the need to write the 0 into the header.