Re: [PATCH] unpack_sha1_header(): detect malformed object header

From: Junio C Hamano <gitster@pobox.com>
Date: Sep 26, 2016, 16:15 UTC
Message-ID: <xmqqfuomvdqe.fsf@gitster.mtv.corp.google.com>
In-Reply-To: <20160926140309.l2h4b65gpqyutepn@sigill.intra.peff.net>

Jeff King <peff@peff.net> writes:

Show 10 quoted lines

> This part I don't understand, though. We clearly need to look for the
> NUL. But why do we need to look for the space? The loop in
> parse_sha1_header() can easily detect this as it looks for the end of
> the type name (and if it hits the end-of-string, can bail as in your
> original patch).
> I.e., the root of the problem is that we pass parse_sha1_header() a the
> "ptr" half of a ptr/len buffer, and it has no idea how much we read.
> But once we get it that information (either by passing the length, or by
> ensuring that the buffer is NUL-terminated, it should be easy for it to
> do the right thing.

Yup.

Show 5 quoted lines

> Anyway, here's my ptr/len version (which passes the length back out of
> unpack_sha1_header via an in/out pointer). After thinking on it, though,
> I'm of the opinion that we're better off just ensuring that "hdr" is
> NUL-terminated. We end up assuming that anyway later, since we have to
> know how much of the header buffer was consumed by parsing.

I'd agree, not because I didn't first go in this <ptr,len> route myself, but because the attached change does look quite invasive. Also, I think it is OK to ask unpack_*_header() to fail if what it turns can no way be a header, e.g. lacks NUL termination.

> Do note the final call below in the streaming loose-open code, which
> exhibits that, but also seems to call parse_sha1_header() without
> checking its return value. I think that needs fixed regardless of the
> approach.

Good that your attempt to signature-changing change caught it. I'll take a further look.

Thanks.