On 9/5/22 15:05, Kevin Brodsky wrote:
On 02/09/2022 14:29, carsten.haitzler@foss.arm.com wrote:
From: Carsten Haitzler carsten.haitzler@foss.arm.com
Signed-off-by: Carsten Haitzler carsten.haitzler@foss.arm.com
include/lapi/keyctl.h | 10 +++++----- runtest/morello_transitional_extended | 9 +++++++++ 2 files changed, 14 insertions(+), 5 deletions(-)
diff --git a/include/lapi/keyctl.h b/include/lapi/keyctl.h index 3be782494..9c847a429 100644 --- a/include/lapi/keyctl.h +++ b/include/lapi/keyctl.h @@ -42,13 +42,13 @@ static inline key_serial_t request_key(const char *type, static inline long keyctl(int cmd, ...) { va_list va; - unsigned long arg2, arg3, arg4, arg5; + uintptr_t arg2, arg3, arg4, arg5; va_start(va, cmd); - arg2 = va_arg(va, unsigned long); - arg3 = va_arg(va, unsigned long); - arg4 = va_arg(va, unsigned long); - arg5 = va_arg(va, unsigned long); + arg2 = va_arg(va, uintptr_t); + arg3 = va_arg(va, uintptr_t); + arg4 = va_arg(va, uintptr_t); + arg5 = va_arg(va, uintptr_t);
The changes make sense to me but I wonder if they shouldn't be in a separate patch?
i like to keep the test and the code that fixes it together myself... in this case it's just adding the tests to the list to say "these work now". :) personal preference. in patch collection as it's a tiny patch - so it's pretty obvious which thing is which... :)
va_end(va); return tst_syscall(__NR_keyctl, cmd, arg2, arg3, arg4, arg5); diff --git a/runtest/morello_transitional_extended b/runtest/morello_transitional_extended index a31f0884c..cdd494694 100644 --- a/runtest/morello_transitional_extended +++ b/runtest/morello_transitional_extended @@ -17,3 +17,12 @@ semget06 semget06 semop01 semop01 semop02 semop02 semop03 semop03
+keyctl01 keyctl01 +keyctl02 keyctl02 +keyctl03 keyctl03 +keyctl04 keyctl04 +keyctl05 keyctl05 +keyctl06 keyctl06 +keyctl07 keyctl07 +keyctl08 keyctl08
I'm wondering why we're not including keyctl09, is it because it is only run if CONFIG_USER_DECRYPTED_DATA is set and we are not setting it? If it's just skipped then we could include it anyway, but no strong opinion either way.
correct. we don't have enough kernel config to run it - but it's not a "failing test" due to purecap so i chose not to add it.