Message ID | 20230309103346.5574-1-pvorel@suse.cz |
---|---|
State | Accepted |
Headers | show |
Series | [1/1] endian_switch01.c: Remove useless TST_NO_DEFAULT_MAIN | expand |
Hi Jan, out of curiosity, what is main4() used for? Kind regards, Petr
On Thu, Mar 9, 2023 at 11:44 AM Petr Vorel <pvorel@suse.cz> wrote: > > Hi Jan, > > out of curiosity, what is main4() used for? It's from 086c14f7a465 ("The following hack fixes the "endian_switch01.c:115: warning: ‘main’ takes only zero or two arguments" warning. Signed-off-by: CAI Qian <caiqian@cclom.cn>.") but I'm not sure it is still used these days. As you found my misplaced TST_NO_DEFAULT_MAIN, that seems to confirm we can do without it. I'll have a look on a ppc system. > > Kind regards, > Petr >
On Thu, Mar 9, 2023 at 12:17 PM Jan Stancek <jstancek@redhat.com> wrote: > > On Thu, Mar 9, 2023 at 11:44 AM Petr Vorel <pvorel@suse.cz> wrote: > > > > Hi Jan, > > > > out of curiosity, what is main4() used for? > > It's from 086c14f7a465 ("The following hack fixes the > "endian_switch01.c:115: warning: ‘main’ takes only zero or two > arguments" warning. Signed-off-by: CAI Qian <caiqian@cclom.cn>.") > but I'm not sure it is still used these days. > > As you found my misplaced TST_NO_DEFAULT_MAIN, that seems to confirm > we can do without it. I'll have a look on a ppc system. So I agree with your patch here. And then I'd suggest we follow it with one that replaces main4 with @@ -42,6 +42,9 @@ void check_le_switch_supported(void) exit(errno); } + if (!(getauxval(AT_HWCAP) & PPC_FEATURE_TRUE_LE)) + tst_brk(TCONF, "Processor does not support little-endian mode"); + What do you think?
> On Thu, Mar 9, 2023 at 12:17 PM Jan Stancek <jstancek@redhat.com> wrote: > > On Thu, Mar 9, 2023 at 11:44 AM Petr Vorel <pvorel@suse.cz> wrote: > > > Hi Jan, > > > out of curiosity, what is main4() used for? > > It's from 086c14f7a465 ("The following hack fixes the > > "endian_switch01.c:115: warning: ‘main’ takes only zero or two > > arguments" warning. Signed-off-by: CAI Qian <caiqian@cclom.cn>.") > > but I'm not sure it is still used these days. +1 > > As you found my misplaced TST_NO_DEFAULT_MAIN, that seems to confirm > > we can do without it. I'll have a look on a ppc system. > So I agree with your patch here. And then I'd suggest we follow it Thank you, I merged this patch. > with one that replaces main4 with > @@ -42,6 +42,9 @@ void check_le_switch_supported(void) > exit(errno); > } > + if (!(getauxval(AT_HWCAP) & PPC_FEATURE_TRUE_LE)) > + tst_brk(TCONF, "Processor does not support little-endian mode"); > + > What do you think? Indeed, that's looks to me better. You can add my ack to the patch. But please test it. Kind regards, Petr
diff --git a/testcases/kernel/syscalls/switch/endian_switch01.c b/testcases/kernel/syscalls/switch/endian_switch01.c index bee35184a..eae193fc1 100644 --- a/testcases/kernel/syscalls/switch/endian_switch01.c +++ b/testcases/kernel/syscalls/switch/endian_switch01.c @@ -24,8 +24,6 @@ # define PPC_FEATURE_TRUE_LE 0x00000002 # endif -# define TST_NO_DEFAULT_MAIN - /* * Make minimal call to 0x1ebe. If we get ENOSYS then syscall is not * available, likely because of:
TST_NO_DEFAULT_MAIN must be defined before including tst_test.h. Therefore test could have been compiled although there was no main() defined. Fixes: 68fb61225 ("endian_switch01: rewrite to newlib") Signed-off-by: Petr Vorel <pvorel@suse.cz> --- testcases/kernel/syscalls/switch/endian_switch01.c | 2 -- 1 file changed, 2 deletions(-)