Hi Michael,
Sorry, it took a bit longer than expected.
Sometimes FOSDEM is not good for the progress of free software ;-)
On Thu, Jan 31, 2013 at 1:14 AM, Michael Schmitz <schmitzmic@xxxxxxxxx> wrote:
This rather lengthy list of patches updates Atari EtherNAT and EtherNEC
network support, plus adds USB driver support for the ISP1160 chip found on
EtherNAT and NetUSBee adapters.
Network stuff:
1 declare original EtherNEC driver obsolete
2 declare original EtherNAT driver obsolete
3 use CONFIG_ATARI in smc91x.h
4 change number of interrupt sources to include EtherNAT interrupts
5 fix EtherNAT interrupt number, conditionalize platform device register
6 add platform support for EtherNEC
7 EtherNAT ethernet support using new driver - smc91x.c
8 EtherNEC ethernet support using new driver - ne.c
9 IRQ: add handle_polled_irq for timer based software interrupts
10 use dedicated irq_chip for timer D interrupts
11 add interrupt chip definition for EtherNAT
12 remove unnecessary cruft from Atari interrupt setup
USB related:
13 implement 16 bit access macros for ROM port ISA adapters
14 Implement ndelay based on the existing udelay code
15 add platform device definitions for ISP1160 drivers
16 Add ISP1160 USB host controller support for EtherNAT/NetUSBee
17 NetUSBee tweaks for ISP1160 USB HCD driver
Here are my general comments for this series:
- Please run scripts/checkpatch.pl on your patches. 12 out of 17 have style
problems. Some of them you do want to fix for your next submission.
- You can use the --compose option of "git send-email" to write a separate
introductory email (i.e. the "[PATCH 0/17]" one).
- I don't know what went wrong with "git send-email", but the actual subjects
of the emails ended up in the bodies, too.
- For upstream submission, several of the patches should be folded into
older commits. As you based your series on m68k-queue, you can run
git rebase -i v3.8-rc5 [top]
to strip out the unrelated commits and fold/reorder/fix/clean up the rest.
- As you sent your patches on Jan 31, I assume m68k-queue was based on
v3.8-rc5 at the time you created the patches
- If your commits are on the current branch, the "top" is optional.
Else it would be your queue branch (e.g. for me that's "m68k-queue",
"for-linus", "for-next", or a more specific "for-vX.Y").
You may well decide to leave out #9 - use of EtherNEC and NetUSBee is then possible
only with the 'noirqdebug' kernel option.
Yeah, this is something that should be sorted out with Thomas Gleixner
on lkml (CC added
for path 09/17).
With the present code, the timer D interrupt frequency is approximately 100 Hz. On fast
hardware, it should probably be doubled (change the timer D counter value fron 254 to 123
to achieve that. I had the timer rate adjustable via kernel options in an earlier
incarnation pf the ethernet drivers - whoever feels like it can revive that bit please).
Perhaps you can derive it from CPU frequency, i.e. loops_per_jiffy?
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@xxxxxxxxxxxxxx
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
--
To unsubscribe from this list: send the line "unsubscribe linux-m68k" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
