On Fri, 21 Apr 2000, Stephen C. Tweedie wrote:
> Hi,
>
> On Fri, Apr 21, 2000 at 04:30:28PM +0530, Amit S. Kale wrote:
> >
> > Reading a raw device starting from end returns 0.
> > If you do a 'dd' to a raw device without specifying count of blocks, it writes
> > to the end. After this write returns 0 which causes dd to retry the write ( Not
> > sure why dd retries write ) Thus dd falls into an infinite loop.
> >
> > This does not happen on a block device as writing starting from end returns
> > ENOSPC
> >
> > Should this be considered as a bug in dd or incorrect behavior of raw device
> > driver?
>
> I don't know. POSIX certainly doesn't specify exactly what to do in
> this case. I never use "dd" on raw devices, as (at least on Linux) it
> does not align its buffers correctly,
Other unixes use bounce buffers so alignment is not required.
dd in RH6.2 apparently uses page aligned buffers.
> but "lmdd" does not have the
> problem you describe: it terminates correctly after a short write.
>
> Do you know what other Unixen return in this case?
I checked behaviors on solaris. Both dd and raw devices behave differently
there.
1. dd on solaris terminates if number of bytes written are less then
request. The shown by dd is is error 'unexpected short write'
This happens when a write starts before end of a device and extends beyond it.
dd on linux is certainly buggy.
2. raw device returns ENXIO when writing to a raw device starting from end of
the device.
-- Amit Kale Veritas Software ( http://www.veritas.com )- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sun Apr 23 2000 - 21:00:18 EST