[PATCH mptcp-next] Squash to "selftests: mptcp: add splice io mode"

[PATCH mptcp-next] Squash to "selftests: mptcp: add splice io mode"

[Matthieu Baerts (NGI0)]

NIPA review assistant rightly spotted errors with splice() are simply...
ignored! Technically, mptcp_connect.sh will check the integrity of the
transfer, but still, that's clearly not a good practice!

The different possible errors are now caught, printed, and reported to
the caller.

Link: https://netdev-ai.bots.linux.dev/ai-review.html?id=93466fc9-574b-45fc-996e-13a19a1490a9
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
---
 tools/testing/selftests/net/mptcp/mptcp_connect.c | 34 +++++++++++++++++------
 1 file changed, 25 insertions(+), 9 deletions(-)

diff --git a/tools/testing/selftests/net/mptcp/mptcp_connect.c b/tools/testing/selftests/net/mptcp/mptcp_connect.c
index ff1a298d3469..34d8efe429a4 100644
--- a/tools/testing/selftests/net/mptcp/mptcp_connect.c
+++ b/tools/testing/selftests/net/mptcp/mptcp_connect.c
@@ -937,24 +937,40 @@ static int do_splice(const int infd, const int outfd, const size_t len,
 		     struct wstate *winfo)
 {
 	int pipefd[2];
-	ssize_t bytes;
+	ssize_t in_bytes, out_bytes;
 	int err;
 
 	err = pipe(pipefd);
-	if (err)
-		return err;
+	if (err) {
+		perror("pipe");
+		return 2;
+	}
 
-	while ((bytes = splice(infd, NULL, pipefd[1], NULL,
-			       len - winfo->total_len,
-			       SPLICE_F_MOVE | SPLICE_F_MORE)) > 0) {
-		splice(pipefd[0], NULL, outfd, NULL, bytes,
-		       SPLICE_F_MOVE | SPLICE_F_MORE);
+again:
+	in_bytes = splice(infd, NULL, pipefd[1], NULL, len - winfo->total_len,
+			  SPLICE_F_MOVE | SPLICE_F_MORE);
+	if (in_bytes < 0) {
+		perror("splice in");
+		err = 3;
+	} else if (in_bytes > 0) {
+		out_bytes = splice(pipefd[0], NULL, outfd, NULL, in_bytes,
+				   SPLICE_F_MOVE | SPLICE_F_MORE);
+		if (out_bytes < 0) {
+			perror("splice out");
+			err = 4;
+		} else if (in_bytes != out_bytes) {
+			fprintf(stderr, "Unexpected transfer: %zu vs %zu\n",
+				in_bytes, out_bytes);
+			err = 5;
+		} else {
+			goto again;
+		}
 	}
 
 	close(pipefd[0]);
 	close(pipefd[1]);
 
-	return 0;
+	return err;
 }
 
 static int copyfd_io_splice(int infd, int peerfd, int outfd, unsigned int size,

---
base-commit: 7d56f2ba4d372e9660633de02a01236209c9df8f
change-id: 20260130-mptcp-sft-no-ignore-splice-err-a35a97c217db

Best regards,
-- 
Matthieu Baerts (NGI0) <matttbe@kernel.org>

Re: [PATCH mptcp-next] Squash to "selftests: mptcp: add splice io mode"

[MPTCP CI]

Hi Matthieu,

Thank you for your modifications, that's great!

Our CI did some validations and here is its report:

- KVM Validation: normal (except selftest_mptcp_join): Success! ✅
- KVM Validation: normal (only selftest_mptcp_join): Success! ✅
- KVM Validation: debug (except selftest_mptcp_join): Unstable: 1 failed test(s): packetdrill_sockopts 🔴
- KVM Validation: debug (only selftest_mptcp_join): Success! ✅
- KVM Validation: btf-normal (only bpftest_all): Success! ✅
- KVM Validation: btf-debug (only bpftest_all): Success! ✅
- Task: https://github.com/multipath-tcp/mptcp_net-next/actions/runs/21512533499

Initiator: Patchew Applier
Commits: https://github.com/multipath-tcp/mptcp_net-next/commits/6384e7649754
Patchwork: https://patchwork.kernel.org/project/mptcp/list/?series=1048919


If there are some issues, you can reproduce them using the same environment as
the one used by the CI thanks to a docker image, e.g.:

    $ cd [kernel source code]
    $ docker run -v "${PWD}:${PWD}:rw" -w "${PWD}" --privileged --rm -it \
        --pull always mptcp/mptcp-upstream-virtme-docker:latest \
        auto-normal

For more details:

    https://github.com/multipath-tcp/mptcp-upstream-virtme-docker


Please note that despite all the efforts that have been already done to have a
stable tests suite when executed on a public CI like here, it is possible some
reported issues are not due to your modifications. Still, do not hesitate to
help us improve that ;-)

Cheers,
MPTCP GH Action bot
Bot operated by Matthieu Baerts (NGI0 Core)

Re: [PATCH mptcp-next] Squash to "selftests: mptcp: add splice io mode"

[Geliang Tang]

Hi Matt,

Thanks for this fix.

On Fri, 2026-01-30 at 11:17 +0100, Matthieu Baerts (NGI0) wrote:
> NIPA review assistant rightly spotted errors with splice() are
> simply...
> ignored! Technically, mptcp_connect.sh will check the integrity of
> the
> transfer, but still, that's clearly not a good practice!
> 
> The different possible errors are now caught, printed, and reported
> to
> the caller.
> 
> Link:
> https://netdev-ai.bots.linux.dev/ai-review.html?id=93466fc9-574b-45fc-996e-13a19a1490a9
> Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>

Looks good to me!

Reviewed-by: Geliang Tang <geliang@kernel.org>

> ---
>  tools/testing/selftests/net/mptcp/mptcp_connect.c | 34
> +++++++++++++++++------
>  1 file changed, 25 insertions(+), 9 deletions(-)
> 
> diff --git a/tools/testing/selftests/net/mptcp/mptcp_connect.c
> b/tools/testing/selftests/net/mptcp/mptcp_connect.c
> index ff1a298d3469..34d8efe429a4 100644
> --- a/tools/testing/selftests/net/mptcp/mptcp_connect.c
> +++ b/tools/testing/selftests/net/mptcp/mptcp_connect.c
> @@ -937,24 +937,40 @@ static int do_splice(const int infd, const int
> outfd, const size_t len,
>  		     struct wstate *winfo)
>  {
>  	int pipefd[2];
> -	ssize_t bytes;
> +	ssize_t in_bytes, out_bytes;

This line is longer than the previous one; their positions could be
swapped.

-Geliang

>  	int err;
>  
>  	err = pipe(pipefd);
> -	if (err)
> -		return err;
> +	if (err) {
> +		perror("pipe");
> +		return 2;
> +	}
>  
> -	while ((bytes = splice(infd, NULL, pipefd[1], NULL,
> -			       len - winfo->total_len,
> -			       SPLICE_F_MOVE | SPLICE_F_MORE)) > 0)
> {
> -		splice(pipefd[0], NULL, outfd, NULL, bytes,
> -		       SPLICE_F_MOVE | SPLICE_F_MORE);
> +again:
> +	in_bytes = splice(infd, NULL, pipefd[1], NULL, len - winfo-
> >total_len,
> +			  SPLICE_F_MOVE | SPLICE_F_MORE);
> +	if (in_bytes < 0) {
> +		perror("splice in");
> +		err = 3;
> +	} else if (in_bytes > 0) {
> +		out_bytes = splice(pipefd[0], NULL, outfd, NULL,
> in_bytes,
> +				   SPLICE_F_MOVE | SPLICE_F_MORE);
> +		if (out_bytes < 0) {
> +			perror("splice out");
> +			err = 4;
> +		} else if (in_bytes != out_bytes) {
> +			fprintf(stderr, "Unexpected transfer: %zu vs
> %zu\n",
> +				in_bytes, out_bytes);
> +			err = 5;
> +		} else {
> +			goto again;
> +		}
>  	}
>  
>  	close(pipefd[0]);
>  	close(pipefd[1]);
>  
> -	return 0;
> +	return err;
>  }
>  
>  static int copyfd_io_splice(int infd, int peerfd, int outfd,
> unsigned int size,
> 
> ---
> base-commit: 7d56f2ba4d372e9660633de02a01236209c9df8f
> change-id: 20260130-mptcp-sft-no-ignore-splice-err-a35a97c217db
> 
> Best regards,

Re: [PATCH mptcp-next] Squash to "selftests: mptcp: add splice io mode"

[Matthieu Baerts]

Hi Geliang,

On 30/01/2026 14:46, Geliang Tang wrote:
> Hi Matt,
> 
> Thanks for this fix.
> 
> On Fri, 2026-01-30 at 11:17 +0100, Matthieu Baerts (NGI0) wrote:
>> NIPA review assistant rightly spotted errors with splice() are
>> simply...
>> ignored! Technically, mptcp_connect.sh will check the integrity of
>> the
>> transfer, but still, that's clearly not a good practice!
>>
>> The different possible errors are now caught, printed, and reported
>> to
>> the caller.
>>
>> Link:
>> https://netdev-ai.bots.linux.dev/ai-review.html?id=93466fc9-574b-45fc-996e-13a19a1490a9
>> Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
> 
> Looks good to me!
> 
> Reviewed-by: Geliang Tang <geliang@kernel.org>
> 
>> ---
>>  tools/testing/selftests/net/mptcp/mptcp_connect.c | 34
>> +++++++++++++++++------
>>  1 file changed, 25 insertions(+), 9 deletions(-)
>>
>> diff --git a/tools/testing/selftests/net/mptcp/mptcp_connect.c
>> b/tools/testing/selftests/net/mptcp/mptcp_connect.c
>> index ff1a298d3469..34d8efe429a4 100644
>> --- a/tools/testing/selftests/net/mptcp/mptcp_connect.c
>> +++ b/tools/testing/selftests/net/mptcp/mptcp_connect.c
>> @@ -937,24 +937,40 @@ static int do_splice(const int infd, const int
>> outfd, const size_t len,
>>  		     struct wstate *winfo)
>>  {
>>  	int pipefd[2];
>> -	ssize_t bytes;
>> +	ssize_t in_bytes, out_bytes;
> 
> This line is longer than the previous one; their positions could be
> swapped.

Good point, I missed that. I just fixed that when applying the patch.

New patches for t/upstream:
- f09f3f41e540: Squash to "selftests: mptcp: add splice io mode"
- Results: 50e7f60d4f2a..6f5b41319f11 (export)

Tests are now in progress:

- export:
https://github.com/multipath-tcp/mptcp_net-next/commit/e34e5e3567f9b0fc679d6e6778c749684db62186/checks

Cheers,
Matt
-- 
Sponsored by the NGI0 Core fund.

[PATCH mptcp-next] Squash to "selftests: mptcp: add splice io mode"

[Geliang Tang]

From: Geliang Tang <tanggeliang@kylinos.cn>

Check the return value of splice().

Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 tools/testing/selftests/net/mptcp/mptcp_connect.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/tools/testing/selftests/net/mptcp/mptcp_connect.c b/tools/testing/selftests/net/mptcp/mptcp_connect.c
index ff1a298d3469..aaa2248ac76e 100644
--- a/tools/testing/selftests/net/mptcp/mptcp_connect.c
+++ b/tools/testing/selftests/net/mptcp/mptcp_connect.c
@@ -947,14 +947,16 @@ static int do_splice(const int infd, const int outfd, const size_t len,
 	while ((bytes = splice(infd, NULL, pipefd[1], NULL,
 			       len - winfo->total_len,
 			       SPLICE_F_MOVE | SPLICE_F_MORE)) > 0) {
-		splice(pipefd[0], NULL, outfd, NULL, bytes,
-		       SPLICE_F_MOVE | SPLICE_F_MORE);
+		bytes = splice(pipefd[0], NULL, outfd, NULL, bytes,
+			       SPLICE_F_MOVE | SPLICE_F_MORE);
+		if (bytes <= 0)
+			break;
 	}
 
 	close(pipefd[0]);
 	close(pipefd[1]);
 
-	return 0;
+	return bytes < 0 ? bytes : 0;
 }
 
 static int copyfd_io_splice(int infd, int peerfd, int outfd, unsigned int size,
-- 
2.51.0

Re: [PATCH mptcp-next] Squash to "selftests: mptcp: add splice io mode"

[Matthieu Baerts]

Hi Geliang,

On 30/01/2026 11:43, Geliang Tang wrote:
> From: Geliang Tang <tanggeliang@kylinos.cn>
> 
> Check the return value of splice().

Thank you for looking at that.

> Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
> ---
>  tools/testing/selftests/net/mptcp/mptcp_connect.c | 8 +++++---
>  1 file changed, 5 insertions(+), 3 deletions(-)
> 
> diff --git a/tools/testing/selftests/net/mptcp/mptcp_connect.c b/tools/testing/selftests/net/mptcp/mptcp_connect.c
> index ff1a298d3469..aaa2248ac76e 100644
> --- a/tools/testing/selftests/net/mptcp/mptcp_connect.c
> +++ b/tools/testing/selftests/net/mptcp/mptcp_connect.c
> @@ -947,14 +947,16 @@ static int do_splice(const int infd, const int outfd, const size_t len,
>  	while ((bytes = splice(infd, NULL, pipefd[1], NULL,
>  			       len - winfo->total_len,
>  			       SPLICE_F_MOVE | SPLICE_F_MORE)) > 0) {
> -		splice(pipefd[0], NULL, outfd, NULL, bytes,
> -		       SPLICE_F_MOVE | SPLICE_F_MORE);
> +		bytes = splice(pipefd[0], NULL, outfd, NULL, bytes,
> +			       SPLICE_F_MOVE | SPLICE_F_MORE);
> +		if (bytes <= 0)
> +			break;

Please check [1]: that's not enough: it is important to print the error
to be able to find out what went wrong in case of error. Also, the two
splice() should return the same value, and if "bytes" is 0 here, that's
not normal.

[1]
https://lore.kernel.org/mptcp/20260130-mptcp-sft-no-ignore-splice-err-v1-1-af0a458b6ec8@kernel.org

>  	}
>  
>  	close(pipefd[0]);
>  	close(pipefd[1]);
>  
> -	return 0;
> +	return bytes < 0 ? bytes : 0;

The returned value is supposed to be positive (that's the exit returned
value). Also, I don't think you are supposed to have bytes > 0 here, no?

>  }
>  
>  static int copyfd_io_splice(int infd, int peerfd, int outfd, unsigned int size,

Cheers,
Matt
-- 
Sponsored by the NGI0 Core fund.

Re: [PATCH mptcp-next] Squash to "selftests: mptcp: add splice io mode"

[Geliang Tang]

Hi Matt,

I just noticed that you also submitted a squash-to patch to fix this
issue. Your patch is better, so I've changed the status of this one to
Rejected.

Thanks,
-Geliang

On Fri, 2026-01-30 at 11:50 +0100, Matthieu Baerts wrote:
> Hi Geliang,
> 
> On 30/01/2026 11:43, Geliang Tang wrote:
> > From: Geliang Tang <tanggeliang@kylinos.cn>
> > 
> > Check the return value of splice().
> 
> Thank you for looking at that.
> 
> > Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
> > ---
> >  tools/testing/selftests/net/mptcp/mptcp_connect.c | 8 +++++---
> >  1 file changed, 5 insertions(+), 3 deletions(-)
> > 
> > diff --git a/tools/testing/selftests/net/mptcp/mptcp_connect.c
> > b/tools/testing/selftests/net/mptcp/mptcp_connect.c
> > index ff1a298d3469..aaa2248ac76e 100644
> > --- a/tools/testing/selftests/net/mptcp/mptcp_connect.c
> > +++ b/tools/testing/selftests/net/mptcp/mptcp_connect.c
> > @@ -947,14 +947,16 @@ static int do_splice(const int infd, const
> > int outfd, const size_t len,
> >  	while ((bytes = splice(infd, NULL, pipefd[1], NULL,
> >  			       len - winfo->total_len,
> >  			       SPLICE_F_MOVE | SPLICE_F_MORE)) >
> > 0) {
> > -		splice(pipefd[0], NULL, outfd, NULL, bytes,
> > -		       SPLICE_F_MOVE | SPLICE_F_MORE);
> > +		bytes = splice(pipefd[0], NULL, outfd, NULL,
> > bytes,
> > +			       SPLICE_F_MOVE | SPLICE_F_MORE);
> > +		if (bytes <= 0)
> > +			break;
> 
> Please check [1]: that's not enough: it is important to print the
> error
> to be able to find out what went wrong in case of error. Also, the
> two
> splice() should return the same value, and if "bytes" is 0 here,
> that's
> not normal.
> 
> [1]
> https://lore.kernel.org/mptcp/20260130-mptcp-sft-no-ignore-splice-err-v1-1-af0a458b6ec8@kernel.org
> 
> >  	}
> >  
> >  	close(pipefd[0]);
> >  	close(pipefd[1]);
> >  
> > -	return 0;
> > +	return bytes < 0 ? bytes : 0;
> 
> The returned value is supposed to be positive (that's the exit
> returned
> value). Also, I don't think you are supposed to have bytes > 0 here,
> no?
> 
> >  }
> >  
> >  static int copyfd_io_splice(int infd, int peerfd, int outfd,
> > unsigned int size,
> 
> Cheers,
> Matt

Re: [PATCH mptcp-next] Squash to "selftests: mptcp: add splice io mode"

[MPTCP CI]

Hi Geliang,

Thank you for your modifications, that's great!

Our CI did some validations and here is its report:

- KVM Validation: normal (except selftest_mptcp_join): Success! ✅
- KVM Validation: normal (only selftest_mptcp_join): Success! ✅
- KVM Validation: debug (except selftest_mptcp_join): Unstable: 2 failed test(s): packetdrill_dss packetdrill_fastopen 🔴
- KVM Validation: debug (only selftest_mptcp_join): Success! ✅
- KVM Validation: btf-normal (only bpftest_all): Success! ✅
- KVM Validation: btf-debug (only bpftest_all): Success! ✅
- Task: https://github.com/multipath-tcp/mptcp_net-next/actions/runs/21513386911

Initiator: Patchew Applier
Commits: https://github.com/multipath-tcp/mptcp_net-next/commits/d48f3ead61fc
Patchwork: https://patchwork.kernel.org/project/mptcp/list/?series=1048930


If there are some issues, you can reproduce them using the same environment as
the one used by the CI thanks to a docker image, e.g.:

    $ cd [kernel source code]
    $ docker run -v "${PWD}:${PWD}:rw" -w "${PWD}" --privileged --rm -it \
        --pull always mptcp/mptcp-upstream-virtme-docker:latest \
        auto-normal

For more details:

    https://github.com/multipath-tcp/mptcp-upstream-virtme-docker


Please note that despite all the efforts that have been already done to have a
stable tests suite when executed on a public CI like here, it is possible some
reported issues are not due to your modifications. Still, do not hesitate to
help us improve that ;-)

Cheers,
MPTCP GH Action bot
Bot operated by Matthieu Baerts (NGI0 Core)