Re: regression in ktime.h circa 3.16.0-rc5+ breaks lirc irsend, bad commit 166afb64511

[John Stultz <john.stultz@linaro.org>]

On Fri, May 1, 2015 at 3:02 AM, Trevor Cordes <trevor@tecnopolis.ca> wrote:
> May  1 04:46:12 piles kernel: [  356.838029] JDB: ktime_to_us: -20157485 -> divns 18446744073689394 != old method: -20158
>
> The last 2-3 or 3 groups of output I could produce on demand by stopping mythbackend and running:
> systemctl restart lircd.service ; irsend SEND_ONCE dct700 info
>
> Subsequent irsends don't trigger the bug, since (as I found out a
> while ago) by that point lircd is "hung", at least for a long while.
> Hey!  Maybe lircd is then hung for 18446744073689394 us or ns :-)
> If this result is used as a delay timer, the negative would produce
> 0 delay, and the + number the "hang".  I calculate that hang is 584
> years?  :-)
>
> So it looks like maybe my theory wasn't so wacky: we're dealing
> with a caller passing negative numbers (or 32/64 weirdness).  Very
> strange as it seems the caller *wants* (or is happy with) negative
> numbers!

So yea. I see the issue now. The ktime_divns() function uses do_div
which is an unsigned divide. So negative numbers look like large
numbers, and it doesn't do sign-extension, so when it divides we lose
the sign bit.

ktimes are normally positive values, so this probably isn't something
its run-across often, thus why you've just noticed. It does make me
suspicious that the driver is doing something funny still, but I've
gone ahead and written up a patch to fix this since this is clearly an
issue here.

I'll send it out here shortly. If you could give it a spin at your
leisure, and if it works give me a Tested-by: tag I'd appreciate it!

Great work again on chasing this down, and thanks for helping with
debugging and validating the fix!
-john