Re: [uml-devel] [RFC] PATCH 3/4 - Time virtualization : PTRACE_SYSCALL_MASK

[Jeff Dike]

On Wed, Apr 26, 2006 at 05:47:54PM +0200, Blaisorblade wrote:
> If we can do without MASK_STRICT_VERIFY, that works fully, and
> anyway it's simpler - however, say, when running strace -e read,tee
> (sys_tee will soon be added, it seems) this call would fail, while it
> would be desirable to have it work as strace -e read. 
> 
> MASK_STRICT_VERIFY isn't necessarily the best solution, but if
> userspace must search the maximum allowed syscall by multiple
> attempts, we've still a bad API.
> 
> Probably, a better option (_instead_ of MASK_STRICT_VERIFY) would be
> to return somewhere an "extended error code" saying which is the
> last allowed syscall or (better) which is the first syscall which
> failed. I.e. if there is strace -e read,splice,tee and nor splice nor
> tee are supported, then this value would be __NR_splice and strace (or
> any app) could then decide what to do.

Why not just zero out the bits that the kernel knows about?  Then, if
we return -EINVAL, the process just looks at the remaining bits that
are set to see what system calls the kernel didn't know about.

				Jeff
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/