warnings report on builder wrong?
Joel Sherrill
joel at rtems.org
Mon Feb 13 23:33:13 UTC 2023
With fresh tools on CentOS 7, I have 5 warnings in beaglebonewhite. Log
attached.
Sorry it took so long. I ran out of disk space, then got distracted.
I have no idea why the rtems-bsp-builder isn't tripping these unless it
isn't actually hitting the default build settings. But these warnings
looked pretty generic at first glance.
--joel
On Mon, Feb 13, 2023 at 12:22 AM Joel Sherrill <joel at rtems.org> wrote:
>
>
> On Sun, Feb 12, 2023, 8:01 PM Karel Gardas <karel at functional.vision>
> wrote:
>
>>
>> Hello,
>>
>> looking into the last bsp-builder report here:
>> https://lists.rtems.org/pipermail/build/2023-February/041799.html
>>
>> and I'm surprised by the fact that beaglebonewhile shows 0 warnings
>> while I remember well to fix 3 warnings this weekend.
>>
>
> I read this at first.as saying you fixed 3 and they were still present in
> the build@ logs. That could happen if you got unlucky timing on the
> commit versus when the entire sweep started. The sweeper updates the git
> source for RTEMS.and the rsb before beginning a sweep. So a xo.mut after
> that initial point could cause that.
>
> I checked other hosts and Ubuntu, FreeBSD 12, and FreeBSD 13 all have no
> warnings for any beagle variant. I wondered if the git update had an error
> and missed a fix or commit.
>
> I'll look more in the morning but I'm stumped right now. Maybe someone has
> an insight I don't at the moment.
>
>>
>> I even rerun build just a minute ago and they are still here in full
>> glory:
>>
>> In file included from ../../../bsps/arm/beagle/start/bspstart.c:41:
>> ../../../bsps/arm/beagle/start/bspstart.c: In function
>> 'beagle_get_node_unit':
>> ../../../bsps/arm/beagle/start/bspstart.c:124:26: warning: array
>> subscript has type 'char' [-Wchar-subscripts]
>> 124 | if (!isdigit(prop[i]))
>> | ~~~~^~~
>>
>>
>>
>> In file included from ../../../bsps/arm/beagle/pwmss/pwmss.c:38:
>> ../../../bsps/arm/beagle/include/bsp.h: In function
>> 'pwmss_module_clk_config':
>> ../../../bsps/arm/beagle/include/bsp.h:66:16: warning: 'clkctrl' may be
>> used uninitialized [-Wmaybe-uninitialized]
>> 66 | #define REG(x)(*((volatile uint32_t *)(x)))
>> | ^
>> ../../../bsps/arm/beagle/pwmss/pwmss.c:59:14: note: 'clkctrl' was
>> declared here
>> 59 | uint32_t clkctrl;
>> | ^~~~~~~
>>
>>
>>
>> So, why is that builder does not report those? This is default
>> beaglebonewhite configuration so it should be pretty close to what
>> builder builds. Also the toolchain is 6/rtems-arm build last week...
>>
>> Thanks,
>> Karel
>> _______________________________________________
>> devel mailing list
>> devel at rtems.org
>> http://lists.rtems.org/mailman/listinfo/devel
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rtems.org/pipermail/devel/attachments/20230213/e313bbc2/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: b-arm-beaglebonewhite.log.xz
Type: application/octet-stream
Size: 30948 bytes
Desc: not available
URL: <http://lists.rtems.org/pipermail/devel/attachments/20230213/e313bbc2/attachment-0001.obj>
More information about the devel
mailing list