Skip to contentSkip to conversation

selftests/bpf: test_progs: don't leak fd in bpf

PATCH Applied Brian Vazquez 5 participants 12 messages Oct 1, 2019 – Oct 2, 2019

Subsystems (5)
lore
Share & tools
lore
#0 lore raw

cover selftests/bpf: test_progs: don't leak fd in bpf

This patch series fixes some fd leaks in tcp_rtt and test_sockopt_inherit bpf prof_tests.

Brian Vazquez (2):
selftests/bpf: test_progs: don't leak server_fd in tcp_rtt selftests/bpf: test_progs: don't leak server_fd in
test_sockopt_inherit

 tools/testing/selftests/bpf/prog_tests/sockopt_inherit.c | 2 +-
 tools/testing/selftests/bpf/prog_tests/tcp_rtt.c         | 3 ++-
 2 files changed, 3 insertions(+), 2 deletions(-)
Signature · 1 line
2.23.0.444.g18eeb5a265-goog
#1 lore raw

PATCH 1/2 selftests/bpf: test_progs: don't leak server_fd in tcp_rtt

server_fd needs to be closed if pthread can't be created.

Fixes 8a03222f508b ("selftests/bpf: test_progs: fix client/server race

in tcp_rtt")
Cc: Stanislav Fomichev <[email protected]>

---
 tools/testing/selftests/bpf/prog_tests/tcp_rtt.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
tools/testing/selftests/bpf/prog_tests/tcp_rtt.c +2 -1
index a82da555b1b02..f4cd60d6fba2e 100644
@@ -260,13 +260,14 @@ void test_tcp_rtt(void)
 
 	if (CHECK_FAIL(pthread_create(&tid, NULL, server_thread,
 				      (void *)&server_fd)))
-		goto close_cgroup_fd;
+		goto close_server_fd;
 
 	pthread_mutex_lock(&server_started_mtx);
 	pthread_cond_wait(&server_started, &server_started_mtx);
 	pthread_mutex_unlock(&server_started_mtx);
 
 	CHECK_FAIL(run_test(cgroup_fd, server_fd));
+close_server_fd:
 	close(server_fd);
 close_cgroup_fd:
 	close(cgroup_fd);
Signature · 1 line
2.23.0.444.g18eeb5a265-goog
Acked-by on v1 1/2
#2 lore raw

Re: [PATCH bpf 1/2] selftests/bpf: test_progs: don't leak server_fd in tcp_rtt

Quoted 5 lines · Brian Vazquez

server_fd needs to be closed if pthread can't be created.

Fixes 8a03222f508b ("selftests/bpf: test_progs: fix client/server race

in tcp_rtt")

Fixes tag hast to be on single line, no wrapping. Besides that:

Quoted 30 lines
Cc Stanislav Fomichev <[email protected]>
---
 tools/testing/selftests/bpf/prog_tests/tcp_rtt.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
tools/testing/selftests/bpf/prog_tests/tcp_rtt.c +2 -2
index a82da555b1b02..f4cd60d6fba2e 100644
@@ -260,13 +260,14 @@ void test_tcp_rtt(void)
 
        if (CHECK_FAIL(pthread_create(&tid, NULL, server_thread,
                                      (void *)&server_fd)))
-               goto close_cgroup_fd;
+               goto close_server_fd;
 
        pthread_mutex_lock(&server_started_mtx);
        pthread_cond_wait(&server_started, &server_started_mtx);
        pthread_mutex_unlock(&server_started_mtx);
 
        CHECK_FAIL(run_test(cgroup_fd, server_fd));
+close_server_fd:
        close(server_fd);
 close_cgroup_fd:
        close(cgroup_fd);
--

2.23.0.444.g18eeb5a265-goog

#3 lore raw

PATCH 2/2 selftests/bpf: test_progs: don't leak server_fd in test_sockopt_inherit

In reply to Brian Vazquez

server_fd needs to be close if pthread can't be created.

Fixes e3e02e1d9c24 ("selftests/bpf: test_progs: convert test_sockopt_inherit")
Cc Stanislav Fomichev <[email protected]>
---
 tools/testing/selftests/bpf/prog_tests/sockopt_inherit.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/testing/selftests/bpf/prog_tests/sockopt_inherit.c +1 -1
index 6cbeea7b4bf16..8547ecbdc61ff 100644
@@ -195,7 +195,7 @@ static void run_test(int cgroup_fd)
 
 	if (CHECK_FAIL(pthread_create(&tid, NULL, server_thread,
 				      (void *)&server_fd)))
-		goto close_bpf_object;
+		goto close_server_fd;
 
 	pthread_mutex_lock(&server_started_mtx);
 	pthread_cond_wait(&server_started, &server_started_mtx);
Signature · 1 line
2.23.0.444.g18eeb5a265-goog
Acked-by on v1 2/2
#4 lore raw

Re: [PATCH bpf 2/2] selftests/bpf: test_progs: don't leak server_fd in test_sockopt_inherit

Quoted 1 line · Brian Vazquez

I don't think there is a need to add "test_progs:" to subject, " test_sockopt_inherit" is specific enough ;)

Quoted 1 line

server_fd needs to be close if pthread can't be created.

typo: closed

Quoted 5 lines
Fixes e3e02e1d9c24 ("selftests/bpf: test_progs: convert test_sockopt_inherit")
Cc Stanislav Fomichev <[email protected]>

---

Quoted 19 lines
 tools/testing/selftests/bpf/prog_tests/sockopt_inherit.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/testing/selftests/bpf/prog_tests/sockopt_inherit.c +1 -2
index 6cbeea7b4bf16..8547ecbdc61ff 100644
@@ -195,7 +195,7 @@ static void run_test(int cgroup_fd)
 
        if (CHECK_FAIL(pthread_create(&tid, NULL, server_thread,
                                      (void *)&server_fd)))
-               goto close_bpf_object;
+               goto close_server_fd;
 
        pthread_mutex_lock(&server_started_mtx);
        pthread_cond_wait(&server_started, &server_started_mtx);
--

2.23.0.444.g18eeb5a265-goog

#5 lore raw

Re: [PATCH bpf 2/2] selftests/bpf: test_progs: don't leak server_fd in test_sockopt_inherit

Thanks for reviewing the patches Andrii!

Although Daniel fixed them and applied them correctly.

Quoted 38 lines · Andrii Nakryiko
Quoted 1 line · Brian Vazquez

I don't think there is a need to add "test_progs:" to subject, " test_sockopt_inherit" is specific enough ;)

Quoted 1 line

server_fd needs to be close if pthread can't be created.

typo: closed

Quoted 5 lines
Fixes e3e02e1d9c24 ("selftests/bpf: test_progs: convert test_sockopt_inherit")
Cc Stanislav Fomichev <[email protected]>

---

Quoted 19 lines
 tools/testing/selftests/bpf/prog_tests/sockopt_inherit.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/testing/selftests/bpf/prog_tests/sockopt_inherit.c +1 -2
index 6cbeea7b4bf16..8547ecbdc61ff 100644
@@ -195,7 +195,7 @@ static void run_test(int cgroup_fd)
 
        if (CHECK_FAIL(pthread_create(&tid, NULL, server_thread,
                                      (void *)&server_fd)))
-               goto close_bpf_object;
+               goto close_server_fd;
 
        pthread_mutex_lock(&server_started_mtx);
        pthread_cond_wait(&server_started, &server_started_mtx);
--

2.23.0.444.g18eeb5a265-goog

maintainer
#6 lore raw

Re: [PATCH bpf 2/2] selftests/bpf: test_progs: don't leak server_fd in test_sockopt_inherit

Quoted 3 lines · Brian Vazquez

Thanks for reviewing the patches Andrii!

Although Daniel fixed them and applied them correctly.

After last kernel/maintainer summit at LPC, I reworked all my patchwork scripts [0] which I use for bpf trees in order to further reduce manual work and add more sanity checks at the same time. Therefore, the broken Fixes: tag was a good test-case. ;-)

Thanks,
Daniel

[0] https://git.kernel.org/pub/scm/linux/kernel/git/dborkman/pw.git/

Quoted 22 lines
Quoted 20 lines · Andrii Nakryiko
Quoted 1 line · Brian Vazquez

I don't think there is a need to add "test_progs:" to subject, " test_sockopt_inherit" is specific enough ;)

Quoted 1 line

server_fd needs to be close if pthread can't be created.

typo: closed

Quoted 4 lines
Fixes e3e02e1d9c24 ("selftests/bpf: test_progs: convert test_sockopt_inherit")
Cc Stanislav Fomichev <[email protected]>

---

Quoted 2 lines
 tools/testing/selftests/bpf/prog_tests/sockopt_inherit.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
#7 lore raw

Re: [PATCH bpf 2/2] selftests/bpf: test_progs: don't leak server_fd in test_sockopt_inherit

Quoted 9 lines · Daniel Borkmann
Quoted 3 lines · Brian Vazquez

Thanks for reviewing the patches Andrii!

Although Daniel fixed them and applied them correctly.

After last kernel/maintainer summit at LPC, I reworked all my patchwork scripts [0] which I use for bpf trees in order to further reduce manual work and add more sanity checks at the same time. Therefore, the broken Fixes: tag was a good test-case. ;-)

Do you scripts also capitalize first word after libbpf: prefix? Is that intentional? Is that a recommended subject casing:

"libbpf: Do awesome stuff" vs "libbpf: do awesome stuff"?

Quoted 28 lines

Thanks,
Daniel

[0] https://git.kernel.org/pub/scm/linux/kernel/git/dborkman/pw.git/

Quoted 22 lines
Quoted 20 lines · Andrii Nakryiko
Quoted 1 line · Brian Vazquez

I don't think there is a need to add "test_progs:" to subject, " test_sockopt_inherit" is specific enough ;)

Quoted 1 line

server_fd needs to be close if pthread can't be created.

typo: closed

Quoted 4 lines
Fixes e3e02e1d9c24 ("selftests/bpf: test_progs: convert test_sockopt_inherit")
Cc Stanislav Fomichev <[email protected]>

---

Quoted 2 lines
 tools/testing/selftests/bpf/prog_tests/sockopt_inherit.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
maintainer
#8 lore raw

Re: [PATCH bpf 2/2] selftests/bpf: test_progs: don't leak server_fd in test_sockopt_inherit

Quoted 6 lines · Andrii Nakryiko
Quoted 5 lines · Daniel Borkmann
Quoted 3 lines · Brian Vazquez

Thanks for reviewing the patches Andrii!

Although Daniel fixed them and applied them correctly.

Quoted 8 lines
Quoted 3 lines

After last kernel/maintainer summit at LPC, I reworked all my patchwork scripts [0] which I use for bpf trees in order to further reduce manual work and add more sanity checks at the same time. Therefore, the broken Fixes: tag was a good test-case. ;-)

Do you scripts also capitalize first word after libbpf: prefix? Is that intentional? Is that a recommended subject casing:

"libbpf: Do awesome stuff" vs "libbpf: do awesome stuff"?

Right now we have a bit of a mix on that regard, and basically what the pw-apply script from [0] is doing, is the following to provide some more context:

- Pulls the series mbox specified by series id from patchwork, dumps all necessary information about the series, e.g. whether it's complete and all patches are present, etc.
- Pushes the mbox through mb2q which is a script that x86 maintainers and few others use for their patch management and spills out a new mbox. This is effectively 'normalizing' the patches from the mbox to bring in some more consistency, meaning it adds Link: tags to every patch based on the message id and checks whether the necessary mailing list aka bpf was in Cc, so we always have lore BPF archive links, sorts tags so they all have a consistent order, it allows to propagate Acked-by, Reviewed-by, Tested-by tags from cover letter into the individual patches, it also capitalizes the first word after the subsystem prefix.
- It applies and merges the resulting mbox, and performs additional checks for the newly added commit range, that is, it checks whether Fixes tags are correctly formatted, whether the commit exists at all in the tree or whether subject / sha is wrong, and throws warnings to me so I can fix them up if needed or toss out the series again worst case, as well as checks whether SOB from the patch authors is present and matches their name.
- It allows to set the patches from the series into accepted state in patchwork.

So overall less manual work / checks than what used to be before while improving / ensuring more consistency in the commits at the same time. If you have further suggestions / improvements / patches to pw.git, happy to hear. :)

Thanks,
Daniel

Quoted 24 lines
Quoted 24 lines

[0] https://git.kernel.org/pub/scm/linux/kernel/git/dborkman/pw.git/

Quoted 22 lines
Quoted 20 lines · Andrii Nakryiko
Quoted 1 line · Brian Vazquez

I don't think there is a need to add "test_progs:" to subject, " test_sockopt_inherit" is specific enough ;)

Quoted 1 line

server_fd needs to be close if pthread can't be created.

typo: closed

Quoted 4 lines
Fixes e3e02e1d9c24 ("selftests/bpf: test_progs: convert test_sockopt_inherit")
Cc Stanislav Fomichev <[email protected]>

---

Quoted 2 lines
 tools/testing/selftests/bpf/prog_tests/sockopt_inherit.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
#9 lore raw

Re: [PATCH bpf 2/2] selftests/bpf: test_progs: don't leak server_fd in test_sockopt_inherit

Quoted 43 lines · Daniel Borkmann
Quoted 6 lines · Andrii Nakryiko
Quoted 5 lines · Daniel Borkmann
Quoted 3 lines · Brian Vazquez

Thanks for reviewing the patches Andrii!

Although Daniel fixed them and applied them correctly.

Quoted 8 lines
Quoted 3 lines

After last kernel/maintainer summit at LPC, I reworked all my patchwork scripts [0] which I use for bpf trees in order to further reduce manual work and add more sanity checks at the same time. Therefore, the broken Fixes: tag was a good test-case. ;-)

Do you scripts also capitalize first word after libbpf: prefix? Is that intentional? Is that a recommended subject casing:

"libbpf: Do awesome stuff" vs "libbpf: do awesome stuff"?

Right now we have a bit of a mix on that regard, and basically what the pw-apply script from [0] is doing, is the following to provide some more context:

- Pulls the series mbox specified by series id from patchwork, dumps all necessary information about the series, e.g. whether it's complete and all patches are present, etc.
- Pushes the mbox through mb2q which is a script that x86 maintainers and few others use for their patch management and spills out a new mbox. This is effectively 'normalizing' the patches from the mbox to bring in some more consistency, meaning it adds Link: tags to every patch based on the message id and checks whether the necessary mailing list aka bpf was in Cc, so we always have lore BPF archive links, sorts tags so they all have a consistent order, it allows to propagate Acked-by, Reviewed-by, Tested-by tags from cover letter into the individual patches, it also capitalizes the first word after the subsystem prefix.
- It applies and merges the resulting mbox, and performs additional checks for the newly added commit range, that is, it checks whether Fixes tags are correctly formatted, whether the commit exists at all in the tree or whether subject / sha is wrong, and throws warnings to me so I can fix them up if needed or toss out the series again worst case, as well as checks whether SOB from the patch authors is present and matches their name.
- It allows to set the patches from the series into accepted state in patchwork.

Quoted 5 lines

So overall less manual work / checks than what used to be before while improving / ensuring more consistency in the commits at the same time. If you have further suggestions / improvements / patches to pw.git, happy to hear. :)

"libbpf: Captilized subj" looks weird, but I can live with that. I'll post subsequent patches with that casing. I'm glad a lot of that stuff is semi-automated, it's terrible to have to always check all that manually :)

Quoted 27 lines

Thanks,
Daniel

Quoted 24 lines
Quoted 24 lines

[0] https://git.kernel.org/pub/scm/linux/kernel/git/dborkman/pw.git/

Quoted 22 lines
Quoted 20 lines · Andrii Nakryiko
Quoted 1 line · Brian Vazquez

I don't think there is a need to add "test_progs:" to subject, " test_sockopt_inherit" is specific enough ;)

Quoted 1 line

server_fd needs to be close if pthread can't be created.

typo: closed

Quoted 4 lines
Fixes e3e02e1d9c24 ("selftests/bpf: test_progs: convert test_sockopt_inherit")
Cc Stanislav Fomichev <[email protected]>

---

Quoted 2 lines
 tools/testing/selftests/bpf/prog_tests/sockopt_inherit.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
Reviewed-by on the v1 series
#10 lore raw
In reply to Brian Vazquez
Quoted 2 lines · Brian Vazquez

This patch series fixes some fd leaks in tcp_rtt and test_sockopt_inherit bpf prof_tests.

Thanks! For the series:

Quoted 12 lines

Brian Vazquez (2):
selftests/bpf: test_progs: don't leak server_fd in tcp_rtt selftests/bpf: test_progs: don't leak server_fd in
test_sockopt_inherit

 tools/testing/selftests/bpf/prog_tests/sockopt_inherit.c | 2 +-
 tools/testing/selftests/bpf/prog_tests/tcp_rtt.c         | 3 ++-
 2 files changed, 3 insertions(+), 2 deletions(-)
Signature · 1 line
2.23.0.444.g18eeb5a265-goog
maintainer Applied
#11 lore raw
In reply to Brian Vazquez
Quoted 11 lines · Brian Vazquez

This patch series fixes some fd leaks in tcp_rtt and test_sockopt_inherit bpf prof_tests.

Brian Vazquez (2):
selftests/bpf: test_progs: don't leak server_fd in tcp_rtt selftests/bpf: test_progs: don't leak server_fd in
test_sockopt_inherit

 tools/testing/selftests/bpf/prog_tests/sockopt_inherit.c | 2 +-
 tools/testing/selftests/bpf/prog_tests/tcp_rtt.c         | 3 ++-
 2 files changed, 3 insertions(+), 2 deletions(-)

Applied, thanks!