Re: ni_usb: bug when setting kernel timer

My change to the kernel version requirement was incorrect. Kernel 4.14.0
has unsigned long as the timer_list function argument type.

https://elixir.bootlin.com/linux/v4.14/source/include/linux/timer.h#L20
https://sourceforge.net/p/linux-gpib/code/1752/tree//trunk/linux-gpib-kernel/compat/include/linux/timer.h#l29

I'll look more into the redefinition error. Comparing 4.13, 4.14, and 4.15
declarations of setup_timer it looks like there were a few changes between
each so I maybe just particularly unlucky in kernel versions.

https://elixir.bootlin.com/linux/v4.13/source/include/linux/timer.h#L154
https://elixir.bootlin.com/linux/v4.14/source/include/linux/timer.h#L175
https://elixir.bootlin.com/linux/v4.15/source/include/linux/timer.h#L131

Regards,
Derek

On Mon, May 28, 2018 at 7:49 PM, Derek Kozel <[email protected]> wrote:

> Hello,
>
> I've just had a chance to try the newly separated source trees and
> encountered an error with the setup_timer compat header. The make log is
> attached as linux-gpib-r1752-setup_timer-error.txt.
>
> I'm running Ubuntu 16.04 with kernel 4.14.0-041400-generic. I have SVN rev
> 1752.
>
> As the error was setup_timer being redefined I changed the kernel version
> check to < 4.14 and attempted to recompile. Many drivers succeeded, but
> then I encountered an incompatible pointer type error. The log for that is
> also attached, I reran make after the first error to abbreviate the output.
> My guess is that I'll have to try understanding the code a bit more to fix
> this error, but I'm also slightly surprised it hasn't been encountered by
> anyone else. Can anyone shed light on it?
>
> Many thanks,
> Derek
>
> On Tue, May 8, 2018 at 3:41 PM, Frank Mori Hess <[email protected]> wrote:
>
>> On Mon, May 7, 2018 at 10:46 AM, Stefan Mahr <[email protected]>
>> wrote:
>> > Hi,
>> >
>> > There's a timer initialisation bug in ni_usb driver, especially
>> > for NI_USB_B. The timer is setup after first use of mod_timer
>> > (ni_usb_b_read_serial_number->usb_send_bulk_msg), so kernel throws a
>> > bug in timer.c: BUG_ON(!timer->function);
>> >
>> > Attached patch should fix this issue.
>> >
>>
>> Thanks, I'm going to try implementing timer_setup using setup_timer in
>> a compatibility header later today, to get rid of needing to #ifdef
>> every timer setup.  I'll incorporate your change into that.
>>
>>
>> --
>> Frank
>>
>> ------------------------------------------------------------
>> ------------------
>> Check out the vibrant tech community on one of the world's most
>> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
>> _______________________________________________
>> Linux-gpib-general mailing list
>> [email protected]
>> https://lists.sourceforge.net/lists/listinfo/linux-gpib-general
>>
>
>

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot

_______________________________________________
Linux-gpib-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-gpib-general