This is the mail archive of the glibc-bugs@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[Bug libc/21026] [MIPS] readahead syscall is broken on n64


https://sourceware.org/bugzilla/show_bug.cgi?id=21026

--- Comment #5 from Dmitry V. Levin <ldv at sourceware dot org> ---
(In reply to joseph@codesourcery.com from comment #4)
> sysdeps/unix/sysv/linux/wordsize-64 cannot be used for n64 because it 
> implies struct stat = struct stat64, which is not the case for n64.

Anyway, readahead should be added, it's a straightforward fix.

> I suppose https://sourceware.org/ml/libc-alpha/2016-09/msg00527.html needs 
> to be reviewed (with a careful check against the kernel ABI for all 
> supported glibc configurations).

I'm afraid this is too late for 2.25, and the bug is quite user visible: it
breaks strace test suite.

> Unfortunately I don't see a way for the 
> glibc testsuite to verify correctness here (by passing function arguments 
> that would produce a different return from the syscall if the arguments 
> are passed incorrectly).

Yes, readahead is not a regular syscall that would produce different return
values.  Anyway, one can use strace, it's better than nothing.  For most of
architectures supported by glibc it's verified to be correct wrt decoding of
readahead syscall.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]