include/tst_test_macros.h: Add TST_EXP_{PASS, FD}_SILENT macros
PATCH v3 1/3 include/tst_test_macros.h: Add TST_EXP_{PASS, FD}_SILENT macros
1) TST_EXP_{PASS,FD}_SILENT don't report TPASS when SCALL succeeds.
2) TST_EXP_{PASS,FD} calls TST_EXP_{PASS,FD}_SILENT and reports TPASS when SCALL succeeds.
--- include/tst_test_macros.h | 25 ++++++++++++++++++------- 1 file changed, 18 insertions(+), 7 deletions(-)
include/tst_test_macros.h +18 -7
@@ -46,7 +46,7 @@ extern void *TST_RET_PTR;tst_res_ (__FILE__, __LINE__,RES , \ TST_FMT_(TST_2_(dummy, ##__VA_ARGS__, SCALL) FMT, __VA_ARGS__), PAR) -#define TST_EXP_FD (SCALL , ...) \ +#define TST_EXP_FD_SILENT (SCALL , ...) \ do { \ TEST(SCALL); \ \
@@ -64,14 +64,20 @@ extern void *TST_RET_PTR;break ; \ } \ \ - TST_MSGP_(TPASS, " returned fd %ld", TST_RET, \ -#SCALL , ##__VA_ARGS__); \ - \TST_PASS =1 ; \ \ } while (0) -#define TST_EXP_PASS (SCALL , ...) \ +#define TST_EXP_FD (SCALL , ...) \ + do { \ + TST_EXP_FD_SILENT(SCALL, __VA_ARGS__); \ + \ + if (TST_PASS) \ + TST_MSGP_(TPASS, " returned fd %ld", TST_RET, \ + #SCALL, ##__VA_ARGS__); \ + } while (0) + +#define TST_EXP_PASS_SILENT (SCALL , ...) \ do { \ TEST(SCALL); \ \
@@ -89,12 +95,17 @@ extern void *TST_RET_PTR;break ; \ } \ \ - TST_MSG_(TPASS, " passed", #SCALL, ##__VA_ARGS__); \ - \TST_PASS =1 ; \ \ } while (0) +#define TST_EXP_PASS (SCALL , ...) \ + do { \ + TST_EXP_PASS_SILENT(SCALL, __VA_ARGS__); \ + \ + if (TST_PASS) \ + TST_MSG_(TPASS, " passed", #SCALL, ##__VA_ARGS__); \ + } while (0) \#define TST_EXP_FAIL (SCALL ,ERRNO , ...) \ do { \
Signature · 1 line
2.21.0
PATCH v3 1/3 include/tst_test_macros.h: Add TST_EXP_{PASS, FD}_SILENT macros
Hi Cyril,
If you have no objection, I will push the patchset after the new release. :-)
Best Regards?
Xiao Yang
Quoted 67 lines · Xiao Yang
1) TST_EXP_{PASS,FD}_SILENT don't report TPASS when SCALL succeeds.
2) TST_EXP_{PASS,FD} calls TST_EXP_{PASS,FD}_SILENT and reports TPASS when SCALL succeeds.Signed-off-by Xiao Yang <[email protected]>--- include/tst_test_macros.h | 25 ++++++++++++++++++------- 1 file changed, 18 insertions(+), 7 deletions(-)include/tst_test_macros.h +18
-7@@ -46,7 +46,7 @@ extern void *TST_RET_PTR;tst_res_ (__FILE__, __LINE__,RES , \ TST_FMT_(TST_2_(dummy, ##__VA_ARGS__, SCALL) FMT, __VA_ARGS__), PAR) -#define TST_EXP_FD (SCALL , ...) \ +#define TST_EXP_FD_SILENT (SCALL , ...) \ do { \ TEST(SCALL); \ \@@ -64,14 +64,20 @@ extern void *TST_RET_PTR;break ; \ } \ \ - TST_MSGP_(TPASS, " returned fd %ld", TST_RET, \ -#SCALL , ##__VA_ARGS__); \ - \TST_PASS =1 ; \ \ } while (0) -#define TST_EXP_PASS (SCALL , ...) \ +#define TST_EXP_FD (SCALL , ...) \ + do { \ + TST_EXP_FD_SILENT(SCALL, __VA_ARGS__); \ + \ + if (TST_PASS) \ + TST_MSGP_(TPASS, " returned fd %ld", TST_RET, \ + #SCALL, ##__VA_ARGS__); \ + } while (0) + +#define TST_EXP_PASS_SILENT (SCALL , ...) \ do { \ TEST(SCALL); \ \@@ -89,12 +95,17 @@ extern void *TST_RET_PTR;break ; \ } \ \ - TST_MSG_(TPASS, " passed", #SCALL, ##__VA_ARGS__); \ - \TST_PASS =1 ; \ \ } while (0) +#define TST_EXP_PASS (SCALL , ...) \ + do { \ + TST_EXP_PASS_SILENT(SCALL, __VA_ARGS__); \ + \ + if (TST_PASS) \ + TST_MSG_(TPASS, " passed", #SCALL, ##__VA_ARGS__); \ + } while (0) \#define TST_EXP_FAIL (SCALL ,ERRNO , ...) \ do { \
PATCH v3 2/3 syscalls/open01.c: Take use of TST_EXP_FD_SILENT
In reply to Xiao YangCurrent TST_EXP_FD macro reports the double passed for one subtest, for example:
--------------------------------------------
./open01
tst_test.c:1261: TINFO: Timeout per run is 0h 05m 00s open01.c:48: TPASS: open() with sticky bit returned fd 3 open01.c:59: TPASS: sticky bit is set as expected open01.c:48: TPASS: open() with sirectory bit returned fd 3 open01.c:59: TPASS: sirectory bit is set as expected
Summary:
passed 4 failed 0 broken 0 skipped 0
warnings 0
--------------------------------------------
It is just a minor cleanup rather than a fix.
--- testcases/kernel/syscalls/open/open01.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
testcases/kernel/syscalls/open/open01.c +1 -1
@@ -45,7 +45,7 @@ static void verify_open(unsigned int n)struct tcase *tc = &tcases[n];struct stat buf; -TST_EXP_FD (open (tc->filename, tc->flag, tc->mode), +TST_EXP_FD_SILENT (open (tc->filename, tc->flag, tc->mode),"open() with %s" , tc->desc);if (!TST_PASS )return ;
Signature · 1 line
2.21.0
PATCH v3 3/3 syscalls: Take use of TST_EXP_PASS_SILENT
In reply to Xiao YangCurrent TST_EXP_PASS macro reports the double passed for
one subtest, for example:
--------------------------------------------
tst_test.c:1261: TINFO: Timeout per run is 0h 05m 00s access02.c:62: TPASS: access(file_f, F_OK) as root passed access02.c:141: TPASS: access(file_f, F_OK) as root behaviour is correct. access02.c:62: TPASS: access(file_f, F_OK) as nobody passed access02.c:141: TPASS: access(file_f, F_OK) as nobody behaviour is correct. access02.c:62: TPASS: access(file_r, R_OK) as root passed
...
Summary:
passed 32 failed 0 broken 0 skipped 0
warnings 0
--------------------------------------------
It is just a minor cleanup rather than a fix.
--- testcases/kernel/syscalls/access/access02.c | 2 +- testcases/kernel/syscalls/bind/bind04.c | 2 +- testcases/kernel/syscalls/bind/bind05.c | 2 +- testcases/kernel/syscalls/brk/brk01.c | 6 +++++- testcases/kernel/syscalls/capget/capget01.c | 4 +++- 5 files changed, 11 insertions(+), 5 deletions(-)
testcases/kernel/syscalls/access/access02.c +1 -1
@@ -59,7 +59,7 @@ static void access_test(struct tcase *tc, const char *user)struct stat stat_buf;char command[64 ]; -TST_EXP_PASS (access (tc->pathname, tc->mode), +TST_EXP_PASS_SILENT (access (tc->pathname, tc->mode),"access(%s, %s) as %s" , tc->pathname, tc->name, user);if (!TST_PASS )
testcases/kernel/syscalls/bind/bind04.c +1 -1
@@ -118,7 +118,7 @@ static void test_bind(unsigned int n) listen_sock =SAFE_SOCKET (tc->address->sa_family, tc->type, tc->protocol); -TST_EXP_PASS (bind (listen_sock, tc->address, tc->addrlen),"bind()" ); +TST_EXP_PASS_SILENT (bind (listen_sock, tc->address, tc->addrlen),"bind()" );if (!TST_PASS ) {SAFE_CLOSE (listen_sock);
testcases/kernel/syscalls/bind/bind05.c +1 -1
@@ -131,7 +131,7 @@ static void test_bind(unsigned int n)tst_res (TINFO ,"Testing %s" , tc->description); sock =SAFE_SOCKET (tc->address->sa_family, tc->type, tc->protocol); -TST_EXP_PASS (bind (sock, tc->address, tc->addrlen),"bind()" ); +TST_EXP_PASS_SILENT (bind (sock, tc->address, tc->addrlen),"bind()" );if (!TST_PASS ) {SAFE_CLOSE (sock);
testcases/kernel/syscalls/brk/brk01.c +5 -1
@@ -31,7 +31,9 @@ void verify_brk(void)break ; } -TST_EXP_PASS (brk ((void *)new_brk),"brk()" ); +TST_EXP_PASS_SILENT (brk ((void *)new_brk),"brk()" ); +if (!TST_PASS ) +return ; cur_brk = (uintptr_t )sbrk (0 );
@@ -46,6 +48,8 @@ void verify_brk(void)if (i %3 ==0 ) *((char *)cur_brk) =0 ; } + +tst_res (TPASS ,"brk() works fine" ); }static struct tst_test test = {
testcases/kernel/syscalls/capget/capget01.c +3 -1
@@ -32,8 +32,10 @@ static void verify_capget(unsigned int n) hdr->version = tc->version; hdr->pid = pid; -TST_EXP_PASS (tst_syscall (__NR_capget, hdr, data), +TST_EXP_PASS_SILENT (tst_syscall (__NR_capget, hdr, data),"capget() with %s" , tc->message); +if (!TST_PASS ) +return ;if (data[0 ].effective &1 <<CAP_NET_RAW )tst_res (TFAIL ,"capget() gets CAP_NET_RAW unexpectedly in pE" );
Signature · 1 line
2.21.0
PATCH v3 3/3 syscalls: Take use of TST_EXP_PASS_SILENT
Hi!
Quoted 6 lines
testcases/kernel/syscalls/capget/capget01.c +0
-0@@ -32,8 +32,10 @@ static void verify_capget(unsigned int n) hdr->version = tc->version;
testcases/kernel/syscalls/capget/capget01.c static void verify_capget(unsigned int n) · line 41
Quoted 10 lines
testcases/kernel/syscalls/capget/capget01.c +3
-1hdr->pid = pid; -TST_EXP_PASS (tst_syscall (__NR_capget, hdr, data), +TST_EXP_PASS_SILENT (tst_syscall (__NR_capget, hdr, data),"capget() with %s" , tc->message); +if (!TST_PASS ) +return ;if (data[0 ].effective &1 <<CAP_NET_RAW )tst_res (TFAIL ,"capget() gets CAP_NET_RAW unexpectedly in pE" );
I do not agree with the change in the capget01 here since there are really two testcases there and the test was producing the same amount of TPASS messages before the change to the TST_EXP_PASS() as well.
Other than that the patchset is fine.
So with the capget change removed you can add my:
Signature · 2 lines
Cyril Hrubis [email protected]
PATCH v3 3/3 syscalls: Take use of TST_EXP_PASS_SILENT
Quoted 15 lines · Cyril Hrubis
Hi!
Quoted 14 lines
testcases/kernel/syscalls/capget/capget01.c +3
-1@@ -32,8 +32,10 @@ static void verify_capget(unsigned int n) hdr->version = tc->version; hdr->pid = pid; -TST_EXP_PASS (tst_syscall (__NR_capget, hdr, data), +TST_EXP_PASS_SILENT (tst_syscall (__NR_capget, hdr, data),"capget() with %s" , tc->message); +if (!TST_PASS ) +return ;
Quoted 5 lines
testcases/kernel/syscalls/capget/capget01.c static void verify_capget(unsigned int n) · line 41
Quoted 2 lines
testcases/kernel/syscalls/capget/capget01.c +0
-0if (data[0 ].effective&1 <<CAP_NET_RAW )tst_res (TFAIL ,"capget() gets CAP_NET_RAW unexpectedly in pE" );I do not agree with the change in the capget01 here since there are really two testcases there and the test was producing the same amount of TPASS messages before the change to the
TST_EXP_PASS()as well.
Hi Cyril,
Running capget01 with TST_EXP_PASS() shows:
--------------------------------------------------------
capget01.c:35: TPASS: capget() with LINUX_CAPABILITY_VERSION_1 passed capget01.c:41: TPASS: capget() doesn't get CAP_NET_RAW as expected in PE capget01.c:35: TPASS: capget() with LINUX_CAPABILITY_VERSION_2 passed capget01.c:41: TPASS: capget() doesn't get CAP_NET_RAW as expected in PE capget01.c:35: TPASS: capget() with LINUX_CAPABILITY_VERSION_3 passed capget01.c:41: TPASS: capget() doesn't get CAP_NET_RAW as expected in PE
Summary:
passed 6
--------------------------------------------------------
Running capget01 with TST_EXP_PASS_SILENT() shows:
--------------------------------------------------------
capget01.c:43: TPASS: capget() doesn't get CAP_NET_RAW as expected in PE capget01.c:43: TPASS: capget() doesn't get CAP_NET_RAW as expected in PE capget01.c:43: TPASS: capget() doesn't get CAP_NET_RAW as expected in PE
Summary:
passed 3
--------------------------------------------------------
If you want to keep two TPASS for a subtest, how about merging two TPASS into one? like this:
------------------------------------------------------------------------------------------
capget01.c:43: TPASS: capget() with LINUX_CAPABILITY_VERSION_1 doesn't get CAP_NET_RAW as expected in PE
capget01.c:43: TPASS: capget() with LINUX_CAPABILITY_VERSION_2 doesn't get CAP_NET_RAW as expected in PE
capget01.c:43: TPASS: capget() with LINUX_CAPABILITY_VERSION_3 doesn't get CAP_NET_RAW as expected in PE
Summary:
passed 3
------------------------------------------------------------------------------------------
BTW, I just want to avoid many TPASS for a subtest.
Best Regards,
Xiao Yang
Quoted 6 lines
Other than that the patchset is fine.
So with the capget change removed you can add my:
Reviewed-by Cyril Hrubis<[email protected]>
PATCH v3 3/3 syscalls: Take use of TST_EXP_PASS_SILENT
Hi!
Quoted 1 line
BTW, I just want to avoid many TPASS for a subtest.
Why? The capget01 clearly does two different tests per iteration.
There is absolutely no correlation between the number of TPASS test produces and the number of iteration test does.
Signature · 2 lines
Cyril Hrubis [email protected]
PATCH v3 3/3 syscalls: Take use of TST_EXP_PASS_SILENT
Quoted 6 lines · Cyril Hrubis
Hi!
Quoted 1 line
BTW, I just want to avoid many TPASS for a subtest.
Why? The capget01 clearly does two different tests per iteration.
There is absolutely no correlation between the number of TPASS test produces and the number of iteration test does.
Hi Cyril,
I pushed the patchset without capget01 for now. But I think there is a whole test per iteration, like three steps:
1) Drop CAP_NET_RAW.
2) Get effective by specifying kernel revision.
3) Check that the effective doesn't include CAP_NET_RAW. Is it necessary for per steps to report TPASS? :-)
Best Regards,
Xiao Yang