* [PATCH] test/statx: verify against statx(2) on all archs
@ 2020-07-09 21:34 Tobias Klauser
2020-07-10 13:55 ` Jens Axboe
0 siblings, 1 reply; 4+ messages in thread
From: Tobias Klauser @ 2020-07-09 21:34 UTC (permalink / raw)
To: io-uring
Use __NR_statx in do_statx and unconditionally use it to check the
result on all architectures, not just x86_64. This relies on the
fact that __NR_statx should be defined if struct statx and STATX_ALL are
available as well.
Don't fail the test if the statx syscall returns EOPNOTSUPP though.
Signed-off-by: Tobias Klauser <[email protected]>
---
test/statx.c | 27 ++++++++++-----------------
1 file changed, 10 insertions(+), 17 deletions(-)
diff --git a/test/statx.c b/test/statx.c
index a397593f7c11..d354d19fb095 100644
--- a/test/statx.c
+++ b/test/statx.c
@@ -15,13 +15,11 @@
#include "liburing.h"
-#if defined(__x86_64)
static int do_statx(int dfd, const char *path, int flags, unsigned mask,
struct statx *statxbuf)
{
- return syscall(332, dfd, path, flags, mask, statxbuf);
+ return syscall(__NR_statx, dfd, path, flags, mask, statxbuf);
}
-#endif
static int create_file(const char *file, size_t size)
{
@@ -42,14 +40,16 @@ static int create_file(const char *file, size_t size)
return ret != size;
}
+static int statx_syscall_supported(void)
+{
+ return errno == EOPNOTSUPP ? 0 : -1;
+}
+
static int test_statx(struct io_uring *ring, const char *path)
{
struct io_uring_cqe *cqe;
struct io_uring_sqe *sqe;
- struct statx x1;
-#if defined(__x86_64)
- struct statx x2;
-#endif
+ struct statx x1, x2;
int ret;
sqe = io_uring_get_sqe(ring);
@@ -74,15 +74,13 @@ static int test_statx(struct io_uring *ring, const char *path)
io_uring_cqe_seen(ring, cqe);
if (ret)
return ret;
-#if defined(__x86_64)
ret = do_statx(-1, path, 0, STATX_ALL, &x2);
if (ret < 0)
- return -1;
+ return statx_syscall_supported();
if (memcmp(&x1, &x2, sizeof(x1))) {
fprintf(stderr, "Miscompare between io_uring and statx\n");
goto err;
}
-#endif
return 0;
err:
return -1;
@@ -92,10 +90,7 @@ static int test_statx_fd(struct io_uring *ring, const char *path)
{
struct io_uring_cqe *cqe;
struct io_uring_sqe *sqe;
- struct statx x1;
-#if defined(__x86_64)
- struct statx x2;
-#endif
+ struct statx x1, x2;
int ret, fd;
fd = open(path, O_RDONLY);
@@ -128,16 +123,14 @@ static int test_statx_fd(struct io_uring *ring, const char *path)
io_uring_cqe_seen(ring, cqe);
if (ret)
return ret;
-#if defined(__x86_64)
memset(&x2, 0, sizeof(x2));
ret = do_statx(fd, "", AT_EMPTY_PATH, STATX_ALL, &x2);
if (ret < 0)
- return -1;
+ return statx_syscall_supported();
if (memcmp(&x1, &x2, sizeof(x1))) {
fprintf(stderr, "Miscompare between io_uring and statx\n");
goto err;
}
-#endif
return 0;
err:
return -1;
--
2.27.0
^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [PATCH] test/statx: verify against statx(2) on all archs
2020-07-09 21:34 [PATCH] test/statx: verify against statx(2) on all archs Tobias Klauser
@ 2020-07-10 13:55 ` Jens Axboe
2020-07-10 14:25 ` Tobias Klauser
0 siblings, 1 reply; 4+ messages in thread
From: Jens Axboe @ 2020-07-10 13:55 UTC (permalink / raw)
To: Tobias Klauser, io-uring
On 7/9/20 3:34 PM, Tobias Klauser wrote:
> Use __NR_statx in do_statx and unconditionally use it to check the
> result on all architectures, not just x86_64. This relies on the
> fact that __NR_statx should be defined if struct statx and STATX_ALL are
> available as well.
>
> Don't fail the test if the statx syscall returns EOPNOTSUPP though.
Applied, thanks.
--
Jens Axboe
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] test/statx: verify against statx(2) on all archs
2020-07-10 13:55 ` Jens Axboe
@ 2020-07-10 14:25 ` Tobias Klauser
2020-07-10 14:27 ` Jens Axboe
0 siblings, 1 reply; 4+ messages in thread
From: Tobias Klauser @ 2020-07-10 14:25 UTC (permalink / raw)
To: Jens Axboe; +Cc: io-uring
On 2020-07-10 at 15:55:24 +0200, Jens Axboe <[email protected]> wrote:
> On 7/9/20 3:34 PM, Tobias Klauser wrote:
> > Use __NR_statx in do_statx and unconditionally use it to check the
> > result on all architectures, not just x86_64. This relies on the
> > fact that __NR_statx should be defined if struct statx and STATX_ALL are
> > available as well.
> >
> > Don't fail the test if the statx syscall returns EOPNOTSUPP though.
>
> Applied, thanks.
And thanks for your follow-up fix! Looking at it I noticed that
statx_syscall_supported introduced by this change should check for
ENOSYS, not EOPNOTSUPP. Will send another follow-up.
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] test/statx: verify against statx(2) on all archs
2020-07-10 14:25 ` Tobias Klauser
@ 2020-07-10 14:27 ` Jens Axboe
0 siblings, 0 replies; 4+ messages in thread
From: Jens Axboe @ 2020-07-10 14:27 UTC (permalink / raw)
To: Tobias Klauser; +Cc: io-uring
On 7/10/20 8:25 AM, Tobias Klauser wrote:
> On 2020-07-10 at 15:55:24 +0200, Jens Axboe <[email protected]> wrote:
>> On 7/9/20 3:34 PM, Tobias Klauser wrote:
>>> Use __NR_statx in do_statx and unconditionally use it to check the
>>> result on all architectures, not just x86_64. This relies on the
>>> fact that __NR_statx should be defined if struct statx and STATX_ALL are
>>> available as well.
>>>
>>> Don't fail the test if the statx syscall returns EOPNOTSUPP though.
>>
>> Applied, thanks.
>
> And thanks for your follow-up fix! Looking at it I noticed that
> statx_syscall_supported introduced by this change should check for
> ENOSYS, not EOPNOTSUPP. Will send another follow-up.
Ugh yeah, I missed that in yours. I'll apply the followup.
--
Jens Axboe
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2020-07-10 14:27 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2020-07-09 21:34 [PATCH] test/statx: verify against statx(2) on all archs Tobias Klauser
2020-07-10 13:55 ` Jens Axboe
2020-07-10 14:25 ` Tobias Klauser
2020-07-10 14:27 ` Jens Axboe
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox