[Bug libc/23951] New: pread64(fd, buf, (size_t)-1, offset) do not works as expected, using 32bit glibc-2.27 under 64bit-linux-4.4

["wangfangpeng1 at huawei dot com" <sourceware-bugzilla at sourceware dot org>]

https://sourceware.org/bugzilla/show_bug.cgi?id=23951

            Bug ID: 23951
           Summary: pread64(fd, buf, (size_t)-1, offset) do not works as
                    expected, using 32bit glibc-2.27 under 64bit-linux-4.4
           Product: glibc
           Version: 2.27
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: libc
          Assignee: unassigned at sourceware dot org
          Reporter: wangfangpeng1 at huawei dot com
                CC: drepper.fsp at gmail dot com
  Target Milestone: ---

Hi 

    I got the wrong ret value and errno, when call pread64(fd, buf, (size_t)-1,
offset) using 32bit glibc-2.27, under 64bit-linux-4.4; but it works well for
64bit glibc-2.27 and 64bit linux-4.4(this bug can be found in arm64,x86_64… ).

root:/tmp # ./p-x86-32
pread64.ret = 0, errno = 0
root:/tmp # ./p-x86-64
pead64.ret = -1, errno = 14

    Furthermore, I find that in 32bit glibc-2.27, sizeof(size_t)==4, but in the
64bit linux kernel sizeof(size_t)==8, and 0xffffffff is a reasonable count for
64bit linux kernel.

    So, do you think it’s a bug for glibc to be fixed up? 

    By the way glibc-2.20 have no such a problem.

    Looking forward to your early reply…


Glibc-2.27: sysdeps/unix/sysv/linux/pread64.c
ssize_t
__libc_pread64 (int fd, void *buf, size_t count, off64_t offset)
{
  return SYSCALL_CANCEL (pread64, fd, buf, count, SYSCALL_LL64_PRW (offset));
}

Kernel-4.4: fs/read_write.c
SYSCALL_DEFINE4(pread64, unsigned int, fd, char __user *, buf,
                            size_t, count, loff_t, pos)
{
         struct fd f;
         ssize_t ret = -EBADF;

         if (pos < 0)
                   return -EINVAL;

         f = fdget(fd);
         if (f.file) {
                   ret = -ESPIPE;
                   if (f.file->f_mode & FMODE_PREAD)
                            ret = vfs_read(f.file, buf, count, &pos);
                   fdput(f);
         }

         return ret;
}

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