tools/nolibc: add pipe(), pipe2() and their testcase
Subsystems (2)
v2 cover tools/nolibc: add pipe(), pipe2() and their testcase
Hi, Willy, Thomas,
Thanks to your advice and I really learned a lot from it.
V2 now uses pipe2() to wrap pipe(), and fixes the strcmp issue in test case.
Best regards,
Yuan Tan
Yuan Tan (2):
tools/nolibc: add pipe() and pipe2() support
selftests/nolibc: add testcase for pipe
tools/include/nolibc/sys.h | 24 ++++++++++++++ tools/testing/selftests/nolibc/nolibc-test.c | 35 ++++++++++++++++++++ 2 files changed, 59 insertions(+)
Signature · 1 line
2.34.1
PATCH v2 1/2 tools/nolibc: add pipe() and pipe2() support
According to manual page [1], posix spec [2] and source code like arch/mips/kernel/syscall.c, for historic reasons, the sys_pipe() syscall on some architectures has an unusual calling convention. It returns results in two registers which means there is no need for it to do verify the validity of a userspace pointer argument. Historically that used to be expensive in Linux. These days the performance advantage is negligible.
Nolibc doesn't support the unusual calling convention above, luckily Linux provides a generic sys_pipe2() with an additional flags argument from 2.6.27. If flags is 0, then pipe2() is the same as pipe(). So here we use sys_pipe2() to implement the pipe().
pipe2() is also provided to allow users to use flags argument on demand.
[1]: https://man7.org/linux/man-pages/man2/pipe.2.html
[2]: https://pubs.opengroup.org/onlinepubs/9699919799/functions/pipe.html
--- tools/include/nolibc/sys.h | 24 ++++++++++++++++++++++++ 1 file changed, 24 insertions(+)
tools/include/nolibc/sys.h +24 -0
@@ -752,6 +752,30 @@ int open(const char *path, int flags, ...) } +/* +* int pipe2(int pipefd[2], int flags); +* int pipe(int pipefd[2]); +*/ + +static __attribute__((unused)) +int sys_pipe2 (int pipefd[2 ],int flags) +{ +return my_syscall2 (__NR_pipe, pipefd, flags); +} + +static __attribute__((unused)) +int pipe2 (int pipefd[2 ],int flags) +{ +return __sysret (sys_pipe2 (pipefd, flags)); +} + +static __attribute__((unused)) +int pipe (int pipefd[2 ]) +{ +pipe2 (pipefd,0 ); +} + +/* * int prctl(int option, unsigned long arg2, unsigned long arg3, * unsigned long arg4, unsigned long arg5);
Signature · 1 line
2.34.1
Re: [PATCH v2 1/2] tools/nolibc: add pipe() and pipe2() support
Quoted 33 lines · Yuan Tan
According to manual page [1], posix spec [2] and source code like
arch/mips/kernel/syscall.c, for historic reasons, thesys_pipe()syscall on some architectures has an unusual calling convention. It returns results in two registers which means there is no need for it to do verify the validity of a userspace pointer argument. Historically that used to be expensive in Linux. These days the performance advantage is negligible.Nolibc doesn't support the unusual calling convention above, luckily Linux provides a generic
sys_pipe2()with an additional flags argument from 2.6.27. If flags is 0, thenpipe2()is the same aspipe(). So here we usesys_pipe2()to implement thepipe().
pipe2()is also provided to allow users to use flags argument on demand.[1]: https://man7.org/linux/man-pages/man2/pipe.2.html
[2]: https://pubs.opengroup.org/onlinepubs/9699919799/functions/pipe.htmlSuggested-by Zhangjin Wu <[email protected]>Signed-off-by Yuan Tan <[email protected]>--- tools/include/nolibc/sys.h | 24 ++++++++++++++++++++++++ 1 file changed, 24 insertions(+)tools/include/nolibc/sys.h +0
-0@@ -752,6 +752,30 @@ int open(const char *path, int flags, ...) }
tools/include/nolibc/sys.h int open(const char *path, int flags, ...) · line 764
Quoted 10 lines
tools/include/nolibc/sys.h +10
-0+/* +* int pipe2(int pipefd[2], int flags); +* int pipe(int pipefd[2]); +*/ + +static __attribute__((unused)) +int sys_pipe2 (int pipefd[2 ],int flags) +{ +return my_syscall2 (__NR_pipe, pipefd, flags); +}
Should be __NR_pipe2.
Quoted 19 lines
tools/include/nolibc/sys.h +13
-0+static __attribute__((unused)) +int pipe2 (int pipefd[2 ],int flags) +{ +return __sysret (sys_pipe2 (pipefd, flags)); +} + +static __attribute__((unused)) +int pipe (int pipefd[2 ]) +{ +pipe2 (pipefd,0 ); +} + +/* * int prctl(int option, unsigned long arg2, unsigned long arg3, * unsigned long arg4, unsigned long arg5); Signature · 1 line
2.34.1
Re: [PATCH v2 1/2] tools/nolibc: add pipe() and pipe2() support
In reply to Yuan Tantools/include/nolibc/sys.h int open(const char *path, int flags, ...) · line 775
Quoted 6 lines · Yuan Tan
[..]
tools/include/nolibc/sys.h +4
-0+static __attribute__((unused)) +int pipe (int pipefd[2 ]) +{ +pipe2 (pipefd,0 );
This is missing a "return".
Quoted 9 lines
tools/include/nolibc/sys.h +3
-0+} + +/* * int prctl(int option, unsigned long arg2, unsigned long arg3, * unsigned long arg4, unsigned long arg5); Signature · 1 line
2.34.1
PATCH v2 2/2 selftests/nolibc: add testcase for pipe
In reply to Yuan TanAdd a testcase of pipe that child process sends message to parent process.
Here we use memcmp() to avoid the output buffer issue.
--- tools/testing/selftests/nolibc/nolibc-test.c | 35 ++++++++++++++++++++ 1 file changed, 35 insertions(+)
tools/testing/selftests/nolibc/nolibc-test.c +35 -0
@@ -767,6 +767,41 @@ int test_mmap_munmap(void)return ret; } +int test_pipe (void ) +{ +const char *const msg ="hello, nolibc" ; +int pipefd[2 ]; +char buf[32 ]; +pid_t pid; +ssize_t len; + +if (pipe (pipefd) ==-1 ) +return 1 ; + + pid =fork (); + +switch (pid) { +case -1 : +return 1 ; + +case 0 : +close (pipefd[0 ]); +write (pipefd[1 ], msg,strlen (msg)); +close (pipefd[1 ]); +exit (EXIT_SUCCESS ); + +default : +close (pipefd[1 ]); + len =read (pipefd[0 ], buf,sizeof (buf)); +close (pipefd[0 ]); +waitpid (pid,NULL ,0 ); + +if (len !=strlen (msg)) +return 1 ; +return !!memcmp (buf, msg, len); + } +} +/* Run syscall tests between IDs <min> and <max>. * Return 0 on success, non-zero on failure.
Signature · 1 line
2.34.1
Re: [PATCH v2 2/2] selftests/nolibc: add testcase for pipe
Quoted 2 lines · Yuan Tan
Add a testcase of pipe that child process sends message to parent process.
Thinking about it some more:
What's the advantage of going via a child process? The pipe should work the same within the same process.
Quoted 1 line
Here we use
memcmp()to avoid the output buffer issue.
This sentence is meaningless without the background from v1. You can drop it.
Quoted 13 lines
Suggested-by Thomas Weißschuh <[email protected]>Suggested-by Willy Tarreau <[email protected]>Signed-off-by Yuan Tan <[email protected]>--- tools/testing/selftests/nolibc/nolibc-test.c | 35 ++++++++++++++++++++ 1 file changed, 35 insertions(+)tools/testing/selftests/nolibc/nolibc-test.c +0
-0
tools/testing/selftests/nolibc/nolibc-test.c int test_mmap_munmap(void) · line 770
Quoted 4 lines
tools/testing/selftests/nolibc/nolibc-test.c +1
-0@@ -767,6 +767,41 @@ int test_mmap_munmap(void)return ret; } +int test_pipe (void )
Should be static and actually get called :-)
Quoted 40 lines
tools/testing/selftests/nolibc/nolibc-test.c +34
-0+{ +const char *const msg ="hello, nolibc" ; +int pipefd[2 ]; +char buf[32 ]; +pid_t pid; +ssize_t len; + +if (pipe (pipefd) ==-1 ) +return 1 ; + + pid =fork (); + +switch (pid) { +case -1 : +return 1 ; + +case 0 : +close (pipefd[0 ]); +write (pipefd[1 ], msg,strlen (msg)); +close (pipefd[1 ]); +exit (EXIT_SUCCESS ); + +default : +close (pipefd[1 ]); + len =read (pipefd[0 ], buf,sizeof (buf)); +close (pipefd[0 ]); +waitpid (pid,NULL ,0 ); + +if (len !=strlen (msg)) +return 1 ; +return !!memcmp (buf, msg, len); + } +} +/* Run syscall tests between IDs <min> and <max>. * Return 0 on success, non-zero on failure. Signature · 1 line
2.34.1
Re: [PATCH v2 2/2] selftests/nolibc: add testcase for pipe
Hi Thomas,
Quoted 9 lines · Thomas Weißschuh
Quoted 2 lines · Yuan Tan
Add a testcase of pipe that child process sends message to parent process.
Thinking about it some more:
What's the advantage of going via a child process? The pipe should work the same within the same process.
The pipe is commonly used for process communication, and I think as a test case it is supposed to cover the most common scenarios.
Quoted 5 lines
Quoted 1 line
Here we use
memcmp()to avoid the output buffer issue.This sentence is meaningless without the background from v1. You can drop it.
Got it.
Quoted 61 lines
Quoted 15 lines
Suggested-by Thomas Weißschuh <[email protected]>Suggested-by Willy Tarreau <[email protected]>Signed-off-by Yuan Tan <[email protected]>---
tools/testing/selftests/nolibc/nolibc-test.c | 35 ++++++++++++++++++++1 file changed, 35 insertions(+)tools/testing/selftests/nolibc/nolibc-test.c +0
-0@@ -767,6 +767,41 @@ int test_mmap_munmap(void) return ret; }Quoted 2 lines
+inttest_pipe(void)Should be static and actually get called :-)
Quoted 40 lines
tools/testing/selftests/nolibc/nolibc-test.c +34
-0+{ + const char *const msg = "hello, nolibc"; + int pipefd[2]; + char buf[32]; + pid_t pid; + ssize_t len; + + if (pipe(pipefd) == -1) + return 1; + + pid = fork(); + + switch (pid) { + case -1: + return 1; + + case 0: + close(pipefd[0]); + write(pipefd[1], msg, strlen(msg)); + close(pipefd[1]); + exit(EXIT_SUCCESS); + + default: + close(pipefd[1]); + len = read(pipefd[0], buf, sizeof(buf)); + close(pipefd[0]); + waitpid(pid, NULL, 0); + + if (len != strlen(msg)) + return 1; + return !!memcmp(buf, msg, len); + } +} +
/* Run syscall tests between IDs <min> and <max>.
* Return 0 on success, non-zero on failure.Signature · 1 line
2.34.1
Re: [PATCH v2 2/2] selftests/nolibc: add testcase for pipe
Quoted 5 lines · Yuan Tan
Quoted 4 lines · Thomas Weißschuh
Quoted 2 lines · Yuan Tan
Add a testcase of pipe that child process sends message to parent process.
Quoted 8 lines
Quoted 5 lines
Thinking about it some more:
What's the advantage of going via a child process? The pipe should work the same within the same process.
The pipe is commonly used for process communication, and I think as a test case it is supposed to cover the most common scenarios.
The testcase is supposed to cover the code of nolibc.
It should be the *minimal* amount of code to be reasonable sure that the code in nolibc does the correct thing.
If pipe() returns a value that behaves like a pipe I see no reason to doubt it will also survive fork().
Validating that would mean testing the kernel and not nolibc. For the kernel there are different testsuites.
Less code means less work for everyone involved, now and in the future.
Re: [PATCH v2 2/2] selftests/nolibc: add testcase for pipe
Hi Thomas,
Quoted 24 lines · Thomas Weißschuh
Quoted 8 lines · Yuan Tan
Quoted 7 lines · Thomas Weißschuh
Quoted 3 lines · Yuan Tan
Add a testcase of pipe that child process sends message to parent
process.Thinking about it some more:
Quoted 7 lines
Quoted 3 lines
What's the advantage of going via a child process? The pipe should work the same within the same process.
The pipe is commonly used for process communication, and I think as a
test case it is supposed to cover the most common scenarios.The testcase is supposed to cover the code of nolibc.
It should be the *minimal* amount of code to be reasonable sure that the
code in nolibc does the correct thing.
Ifpipe()returns a value that behaves like a pipe I see no reason to doubt it will also survivefork().
Quoted 6 lines
Validating that would mean testing the kernel and not nolibc. For the kernel there are different testsuites.
Less code means less work for everyone involved, now and in the future.
It's a good point and I never thought about this aspect.
I wonder whether the code below is enough?
static int test_pipe(void)
{
int pipefd[2];
if (pipe(pipefd) == -1)
return 1;
close(pipefd[0]);
close(pipefd[1]);
return 0;
}And I forgot to add this line:
CASE_TEST(pipe); EXPECT_SYSZR(1, test_pipe()); break;
I will add it in next patch.
Re: [PATCH v2 2/2] selftests/nolibc: add testcase for pipe
Quoted 38 lines · Yuan Tan
Hi Thomas,
Quoted 24 lines · Thomas Weißschuh
Quoted 8 lines · Yuan Tan
Quoted 7 lines · Thomas Weißschuh
Quoted 3 lines · Yuan Tan
Add a testcase of pipe that child process sends message to parent
process.Thinking about it some more:
Quoted 7 lines
Quoted 3 lines
What's the advantage of going via a child process? The pipe should work the same within the same process.
The pipe is commonly used for process communication, and I think as a
test case it is supposed to cover the most common scenarios.The testcase is supposed to cover the code of nolibc.
It should be the *minimal* amount of code to be reasonable sure that the
code in nolibc does the correct thing.
Ifpipe()returns a value that behaves like a pipe I see no reason to doubt it will also survivefork().Quoted 6 lines
Validating that would mean testing the kernel and not nolibc. For the kernel there are different testsuites.
Less code means less work for everyone involved, now and in the future.
It's a good point and I never thought about this aspect.
I wonder whether the code below is enough?
static int
test_pipe(void)
Quoted 11 lines
{
int pipefd[2];
if (pipe(pipefd) == -1)
return 1;
close(pipefd[0]);
close(pipefd[1]);
return 0;
}That is very barebones.
If accidentally a wrong syscall number was used and the used syscall would not take any arguments this test would still succeed.
Keeping the write-read-cycle from the previous revision would test that nicely. Essentially the same code as before but without the fork().
Quoted 7 lines
And I forgot to add this line:
CASE_TEST(pipe); EXPECT_SYSZR(1, test_pipe()); break;I will add it in next patch.
Re: [PATCH v2 2/2] selftests/nolibc: add testcase for pipe
Hi Thomas,
Quoted 54 lines · Thomas Weißschuh
Quoted 44 lines · Yuan Tan
Hi Thomas,
Quoted 27 lines · Thomas Weißschuh
Quoted 8 lines · Yuan Tan
Quoted 7 lines · Thomas Weißschuh
Quoted 3 lines · Yuan Tan
Add a testcase of pipe that child process sends message to parent
process.Thinking about it some more:
Quoted 8 lines
Quoted 3 lines
What's the advantage of going via a child process? The pipe should work the same within the same process.
The pipe is commonly used for process communication, and I think as
a
test case it is supposed to cover the most common scenarios.The testcase is supposed to cover the code of nolibc. It should be the *minimal* amount of code to be reasonable sure that
the
code in nolibc does the correct thing.
Ifpipe()returns a value that behaves like a pipe I see no reason to
doubt it will also survivefork().Quoted 6 lines
Validating that would mean testing the kernel and not nolibc. For the kernel there are different testsuites.
Less code means less work for everyone involved, now and in the future.
It's a good point and I never thought about this aspect.
I wonder whether the code below is enough?
static int
test_pipe(void)
{
int pipefd[2];Quoted 8 lines
if (pipe(pipefd) == -1)
return 1;close(pipefd[0]);
close(pipefd[1]);return 0;
}
Quoted 17 lines
That is very barebones.
If accidentally a wrong syscall number was used and the used syscall would not take any arguments this test would still succeed.
Keeping the write-read-cycle from the previous revision would test that
nicely. Essentially the same code as before but without thefork().Quoted 7 lines
And I forgot to add this line:
CASE_TEST(pipe);
EXPECT_SYSZR(1, test_pipe()); break;I will add it in next patch.
In the situation you described, that is indeed the case.
Would this be fine?
static int test_pipe(void)
{
const char *const msg = "hello, nolibc";
int pipefd[2];
char buf[32];
ssize_t len;
if (pipe(pipefd) == -1)
return 1;
write(pipefd[1], msg, strlen(msg));
close(pipefd[1]);
len = read(pipefd[0], buf, sizeof(buf));
close(pipefd[0]);
if (len != strlen(msg))
return 1;
return !!memcmp(buf, msg, len);
}Re: [PATCH v2 2/2] selftests/nolibc: add testcase for pipe
Quoted 91 lines · Yuan Tan
Hi Thomas,
Quoted 54 lines · Thomas Weißschuh
Quoted 44 lines · Yuan Tan
Hi Thomas,
Quoted 27 lines · Thomas Weißschuh
Quoted 8 lines · Yuan Tan
Quoted 7 lines · Thomas Weißschuh
Quoted 3 lines · Yuan Tan
Add a testcase of pipe that child process sends message to parent
process.Thinking about it some more:
Quoted 8 lines
Quoted 3 lines
What's the advantage of going via a child process? The pipe should work the same within the same process.
The pipe is commonly used for process communication, and I think as
a
test case it is supposed to cover the most common scenarios.The testcase is supposed to cover the code of nolibc. It should be the *minimal* amount of code to be reasonable sure that
the
code in nolibc does the correct thing.
Ifpipe()returns a value that behaves like a pipe I see no reason to
doubt it will also survivefork().Quoted 6 lines
Validating that would mean testing the kernel and not nolibc. For the kernel there are different testsuites.
Less code means less work for everyone involved, now and in the future.
It's a good point and I never thought about this aspect.
I wonder whether the code below is enough?
static int
test_pipe(void)
{
int pipefd[2];Quoted 8 lines
if (pipe(pipefd) == -1)
return 1;close(pipefd[0]);
close(pipefd[1]);return 0;
}Quoted 17 lines
That is very barebones.
If accidentally a wrong syscall number was used and the used syscall would not take any arguments this test would still succeed.
Keeping the write-read-cycle from the previous revision would test that
nicely. Essentially the same code as before but without thefork().Quoted 7 lines
And I forgot to add this line:
CASE_TEST(pipe);
EXPECT_SYSZR(1, test_pipe()); break;I will add it in next patch.
In the situation you described, that is indeed the case.
Would this be fine?
static int
test_pipe(void){ const char *const msg = "hello, nolibc"; int pipefd[2]; char buf[32]; ssize_t len; if (pipe(pipefd) == -1) return 1; write(pipefd[1], msg, strlen(msg)); close(pipefd[1]);
Quoted 8 lines
len = read(pipefd[0], buf, sizeof(buf)); close(pipefd[0]); if (len != strlen(msg)) return 1; return !!memcmp(buf, msg, len); }
Looks good!
The return value of write() could also be validated but given we validate the return value from read() it shouldn't make a difference.
(Also the manual manipulation of "buf" is gone that necessitated the check in v1 of the series)
Re: [PATCH v2 2/2] selftests/nolibc: add testcase for pipe
Hi Thomas,
Quoted 108 lines · Thomas Weißschuh
Quoted 101 lines · Yuan Tan
Hi Thomas,
Quoted 64 lines · Thomas Weißschuh
Quoted 48 lines · Yuan Tan
Hi Thomas,
Quoted 30 lines · Thomas Weißschuh
Quoted 13 lines · Yuan Tan
Quoted 11 lines · Thomas Weißschuh
Quoted 4 lines · Yuan Tan
Add a testcase of pipe that child process sends message to
parent
process.Thinking about it some more:
What's the advantage of going via a child process? The pipe should work the same within the same process.
Quoted 5 lines
The pipe is commonly used for process communication, and I think as
a
test case it is supposed to cover the most common scenarios.The testcase is supposed to cover the code of nolibc. It should be the *minimal* amount of code to be reasonable sure
that
the
code in nolibc does the correct thing.
Ifpipe()returns a value that behaves like a pipe I see no reason to
doubt it will also survivefork().Quoted 7 lines
Validating that would mean testing the kernel and not nolibc. For the kernel there are different testsuites.
Less code means less work for everyone involved, now and in the
future.It's a good point and I never thought about this aspect.
I wonder whether the code below is enough?
static int
test_pipe(void)
{
int pipefd[2];Quoted 8 lines
if (pipe(pipefd) == -1)
return 1;close(pipefd[0]);
close(pipefd[1]);return 0;
}That is very barebones.
If accidentally a wrong syscall number was used and the used syscall
would not take any arguments this test would still succeed.Quoted 14 lines
Keeping the write-read-cycle from the previous revision would test
that
nicely. Essentially the same code as before but without thefork().Quoted 7 lines
And I forgot to add this line:
CASE_TEST(pipe);
EXPECT_SYSZR(1, test_pipe()); break;I will add it in next patch.
In the situation you described, that is indeed the case.
Would this be fine?
static int
test_pipe(void)
{
const char *const msg = "hello, nolibc";
int pipefd[2];
char buf[32];
ssize_t len;if (pipe(pipefd) == -1)
return 1;write(pipefd[1], msg, strlen(msg));
close(pipefd[1]);
len =read(pipefd[0], buf, sizeof(buf));
close(pipefd[0]);Quoted 5 lines
if (len != strlen(msg))
return 1;return !!memcmp(buf, msg, len);
}
Quoted 8 lines
Looks good!
The return value of
write()could also be validated but given we validate the return value fromread()it shouldn't make a difference.(Also the manual manipulation of "buf" is gone that necessitated the check in v1 of the series)
I am sorry that I didn't catch your last sentence.
Did you mean this piece of code does not need any further modifications right?
Re: [PATCH v2 2/2] selftests/nolibc: add testcase for pipe
Aug 1, 2023 14:23:22 Yuan Tan <[email protected]>:
Quoted 28 lines
Quoted 28 lines
Quoted 17 lines
Would this be fine?
static int
test_pipe(void)
{
const char *const msg = "hello, nolibc";
int pipefd[2];
char buf[32];
ssize_t len;if (pipe(pipefd) == -1)
return 1;write(pipefd[1], msg, strlen(msg));
close(pipefd[1]);
len =read(pipefd[0], buf, sizeof(buf));
close(pipefd[0]);Quoted 5 lines
if (len != strlen(msg))
return 1;return !!memcmp(buf, msg, len);
}Looks good!
The return value of
write()could also be validated but given we validate the return value fromread()it shouldn't make a difference.
Quoted 8 lines
Quoted 3 lines
(Also the manual manipulation of "buf" is gone that necessitated the check in v1 of the series)
I am sorry that I didn't catch your last sentence.
Did you mean this piece of code does not need any further modifications right?
Yes, for me this is great!
Sorry for being unclear.
Hi Yuan!
Quoted 4 lines · Yuan Tan
Thanks to your advice and I really learned a lot from it.
V2 now uses
pipe2()to wrappipe(), and fixes the strcmp issue in test case.
Thanks!
I have some more comments for the patches.
For new revisions you should also link to the old revision.
Quoted 14 lines
Best regards,
Yuan TanYuan Tan (2):
tools/nolibc: addpipe()andpipe2()support
selftests/nolibc: add testcase for pipetools/include/nolibc/sys.h | 24 ++++++++++++++ tools/testing/selftests/nolibc/nolibc-test.c | 35 ++++++++++++++++++++ 2 files changed, 59 insertions(+)Signature · 1 line
2.34.1
Hi Thomas,
Quoted 9 lines · Thomas Weißschuh
Hi Yuan!
Quoted 5 lines · Yuan Tan
Thanks to your advice and I really learned a lot from it.
V2 now uses
pipe2()to wrappipe(), and fixes the strcmp issue in test
case.
Quoted 6 lines
Thanks!
I have some more comments for the patches.
For new revisions you should also link to the old revision.
I sincerely appreciate your valuable advice. As a newcomer, I will certainly take note of it next time. :)
Quoted 16 lines
Quoted 15 lines
Best regards,
Yuan TanYuan Tan (2):
tools/nolibc: addpipe()andpipe2()support
selftests/nolibc: add testcase for pipetools/include/nolibc/sys.h | 24 ++++++++++++++
tools/testing/selftests/nolibc/nolibc-test.c| 35
++++++++++++++++++++2 files changed, 59 insertions(+)Signature · 1 line
2.34.1