[LTP] [PATCH 5/5] syscalls/posix_fadvise0: Start using new library.
Tue Nov 6 16:54:53 CET 2018
> The only thing that should matter in the context of this test is
> whether or not posix_fadvise() is supported by the kernel.
> For some reason that is not clear to me (are we tinyfying the kernel syscall
> by syscall now?), the support for fadvise in the kernel can be configured out
> with CONFIG_ADVISE_SYSCALLS=n. So IMO, this test really needs a
> runtime check, not a build time check like above.
Yes, I would expect that we would get ENOSYS in a case that kernel is
build without the support for the call. And we don't have to check in
the setup, we can do tst_brk(TCONF, ...) in the actuall test as well.
More information about the ltp