Messages in this thread | | | Date | Mon, 3 Nov 2014 23:26:38 +0100 | Subject | Re: bisected regression: qla2xxx endianness on sparc64 | From | Geert Uytterhoeven <> |
| |
On Sun, Nov 2, 2014 at 7:10 PM, Meelis Roos <mroos@linux.ee> wrote: > Between 3.17 and 3.18-rc2, qla2xxx is broken on my sparc64 machines. It > fails to boot (hangs in firmware rings init). > > This is the result of bisect: > > 98aee70d19a7e3203649fa2078464e4f402a0ad8 is the first bad commit > commit 98aee70d19a7e3203649fa2078464e4f402a0ad8 > Author: Joe Carnuccio <joe.carnuccio@qlogic.com> > Date: Thu Sep 25 05:16:38 2014 -0400 > > qla2xxx: Add endianizer to max_payload_size modifier. > > Signed-off-by: Joe Carnuccio <joe.carnuccio@qlogic.com> > Signed-off-by: Saurav Kashyap <saurav.kashyap@qlogic.com> > Signed-off-by: Christoph Hellwig <hch@lst.de> > > :040000 040000 041c1a26dc2a7988900acc982e3bd65d3cf7e751 > 9700d09c3226fc352d44352f22b84f1be632d324 M s > > This may not be the only problem - when bisecting, I also came to commits > that got past this step but hang after about 165 seconds of uptime while > running userspace startup scripts. But let that be another issue at the > moment.
Switching from uint16_t to __le16 but _removing_ cpu_to_le16() operations looks indeed very fishy.
Gr{oetje,eeting}s,
Geert
-- Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org
In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that. -- Linus Torvalds
| |