Date: Sun, 19 Apr 1998 13:00:06 +1000 From: Bruce Evans <bde@zeta.org.au> To: cvs-all@FreeBSD.ORG, cvs-committers@FreeBSD.ORG, cvs-lib@FreeBSD.ORG, cvs-sys@FreeBSD.ORG, des@FreeBSD.ORG Subject: Re: cvs commit: src/lib/libc/sys lseek.2 src/sys/kern vfs_syscalls.c Message-ID: <199804190300.NAA15945@godzilla.zeta.org.au>
next in thread | raw e-mail | index | archive | help
>des 1998/04/18 12:24:45 PDT
>
> Modified files:
> lib/libc/sys lseek.2
> Log:
> Return EINVAL and do not changefile pointer if resulting offset is negative.
> PR: kern/6184
>
> Revision Changes Path
> 1.7 +2 -2 src/lib/libc/sys/lseek.2
>
> Modified files:
> sys/kern vfs_syscalls.c
> Log:
> Return EINVAL and do not change file pointer if resulting offset is negative.
> PR: kern/6184
This is wrong. Please back it out. It breaks at least the ability to
seek to high addresses in /dev/[k]mem on 64 bit systems. Not long ago
when off_t was 32 bits, it was normal to seek to "negative" addresses in
/dev/kmem, so portable programs should be able to handle such addresses
at least if they actually request them. libkvm/kvm_proc.c:kvm_uread()
shows how to handle them:
errno = 0;
if (lseek(fd, (off_t)uva, 0) == -1 && errno != 0) {
/* Error. */
...
}
/* No error. */
...
This allows seeking to (off_t)-1 provided the system supports it.
I replied to the PR, but for some reason the followup didn't get attached
to the PR.
Bruce
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199804190300.NAA15945>