Messages in this thread | | | Date | Thu, 31 May 2007 23:51:26 +0530 | From | "Satyam Sharma" <> | Subject | Re: [RFC] LZO de/compression support - take 6 |
| |
On 5/31/07, Nitin Gupta <nitingupta910@gmail.com> wrote: > [...] > The author (Markus Oberhumer) of LZO provided these comments for this patch: > > --- > I've only briefly looked over it, but it's obvious that your version does not > work on architechtures which do not allow unaligned access (arm, mips, ...). > > As for further quality assurance, your version should generate byte-identical > object code to LZO 2.02 when using the same compiler & flags. So you could do > some semi-automatic testing by compiling with -ffunction-sections and use > "objcopy --only-section .text.XXX" to compare the md5sum of all generated > functions. This also works fine with crosscompilers. > > Finally I'm not too happy that you renamed the functions and #defines like > LZO1X_WORKMEM_SIZE - please stay compatible with the official library version.
As suggested by Johannes earlier, it'd be great if you could submit the various changes (as per Changelog) as _individual patches_ on the original userspace code. That would be easier for others to review, and there's lesser chances of bugs / issues leaking in that way.
As for "byte-identical object code", I definitely do *not* think it is necessarily a requirement / good idea. As long as all the changes you make are reviewed individually / closely by people here on this list, there's very low chances of any bugs creeping in. [ F.e. I see nothing wrong in removing the usage of "register" -- that could clearly lead to different object code, but with no bugs introduced. ] - 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/
| |