On Tue, 20 Feb 2024, Maciej Wieczor-Retman wrote:
On 2024-02-20 at 15:45:23 +0200, Ilpo Järvinen wrote:
On Tue, 20 Feb 2024, Maciej Wieczor-Retman wrote:
Ctrl-c handler isn't aware of what test is currently running. Because of that it executes all cleanups even if they aren't necessary. Since the ctrl-c handler uses the sigaction system no parameters can be passed to it as function arguments.
Add a global variable to make ctrl-c handler aware of the currently run test and only execute the correct cleanup callback.
Signed-off-by: Maciej Wieczor-Retman maciej.wieczor-retman@intel.com
tools/testing/selftests/resctrl/resctrl.h | 2 ++ .../testing/selftests/resctrl/resctrl_tests.c | 20 +++++++++---------- tools/testing/selftests/resctrl/resctrl_val.c | 2 +- 3 files changed, 13 insertions(+), 11 deletions(-)
diff --git a/tools/testing/selftests/resctrl/resctrl.h b/tools/testing/selftests/resctrl/resctrl.h index 0f49df4961ea..79b45cbeb628 100644 --- a/tools/testing/selftests/resctrl/resctrl.h +++ b/tools/testing/selftests/resctrl/resctrl.h @@ -128,6 +128,8 @@ extern pid_t bm_pid, ppid; extern char llc_occup_path[1024]; +extern struct resctrl_test current_test;
Why this is not just a pointer?
I tried making this as a pointer but the 'test' in test_prepare() is of type 'const struct resctrl_test *' and there are warnings about dropping the const modifier.
Why cannot the pointer be const too? The signal handler is not supposed to modify the contents of the resctrl_test struct.
There are two types of constness in C. One (the most commonly used one) relates to immutability of the contents of the struct the pointer (or char *) points to while the other enforces the pointer itself to remain immutable. Usually, the former is what is useful and it's what you get when you naturally write "const struct".
- current_test = *test;
I'd prefer to keep this internal to signal handling functions so that either the struct resctrl_test or just the cleanup handler is passed to signal_handler_register().
Okay, would moving this assignment to signal_handler_register() be okay then?
Yes, that's what I'm after here. Lets keep it internal to the signal handling code.