On 23/04/2024 23:51, Beata Michalska wrote:
On Tue, Apr 23, 2024 at 10:38:26AM +0100, Kevin Brodsky wrote:
The mremap06 test relies on being able to move a mapping into the unmapped space of an existing reservation (second call to mremap() in do_test()). This is explicitly disallowed in PCuABI (for all syscalls that can create mappings), and it is unclear if the original bug can be reproduced if remapping to another adddress, so let's just skip the test in PCuABI.
Signed-off-by: Kevin Brodsky kevin.brodsky@arm.com
testcases/kernel/syscalls/mremap/mremap06.c | 4 ++++ 1 file changed, 4 insertions(+)
diff --git a/testcases/kernel/syscalls/mremap/mremap06.c b/testcases/kernel/syscalls/mremap/mremap06.c index a19262750772..478840165cdc 100644 --- a/testcases/kernel/syscalls/mremap/mremap06.c +++ b/testcases/kernel/syscalls/mremap/mremap06.c @@ -96,6 +96,10 @@ static void setup(void) { int ret, i; +#ifdef __CHERI_PURE_CAPABILITY__
- tst_brk(TCONF, "Cannot test mremap() into unmapped space in purecap: not supported");
+#endif
Just thinking at loud, but maybe we could temporarily disable the test for PCuABI in related runtest files : runtest/morello_transitional & runtest/syscalls_morello_purecap_skip I'm kinda hesitating disabling the test if there is a chance it could be tweaked to work with PCuABI.
Sure, I was hesitating between the two options. It doesn't make much difference in practice. Should I send a v2 then?
Kevin
BR Beata
page_size = getpagesize(); mmap_size = (NUM_PAGES+1) * page_size; mremap_size = NUM_PAGES * page_size; -- 2.43.0
linux-morello-ltp mailing list -- linux-morello-ltp@op-lists.linaro.org To unsubscribe send an email to linux-morello-ltp-leave@op-lists.linaro.org