Re: [RFC] LZO de/compression support - take 7

["Nitin Gupta" <nitingupta910@gmail.com>]

On 6/6/07, Nitin Gupta <nitingupta910@gmail.com> wrote:
> On 6/6/07, Christoph Hellwig <hch@infradead.org> wrote:
> > On Wed, Jun 06, 2007 at 06:04:17PM +0530, Nitin Gupta wrote:
> > > Hi,
> > >
> > > This is kernel port of LZO1X-1 compressor and LZO1X decompressor (safe
> > > version only).
> > >
> > > * Changes since 'take 6' (Full Changelog after this):
> > > 1) Fixed unaligned memory access problems as pointed out by author --
> > > LZO_UNALIGNED_OK_{2,4} collapsed to UNALIGNED_OK which is #defined for
> > > archs that allow unaligned access (see lib/lzo1x/Makefile). For other
> > > archs, we fall down to byte-by-byte access avoiding any unaligned
> > > access. In original code, LZO_UNALIGNED_OK_{2,4} resp. describe if we
> > > can do unaligned access for ushort and uint32.
> >
> > Please use get_unaligned() for any unaligned access in kernel space.
> >
> >
>
> Done.

Just as a side note: I think retaining code to do byte-by-byte access
also is worth it, such as:
+ #ifdef UNALIGNED_OK
+		if (get_unaligned((const unsigned short *)m_pos) ==
+				get_unaligned((const unsigned short *)ip)) {
+#else
+		if (m_pos[0] != ip[0] || m_pos[1] != ip[1]) {
+#endif

Since on archs where unaligned access is slower than byte-by-byte
access (e.g. where unaligned access dirty work is done in s/w), we do
not want to use get_unaligned() - instead simply byte-by-byte access
is better.

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