Re: [OE-core] [wrynose][PATCH] minicom: fix baud rate setting with glibc 2.42 and newer

["Yoann Congal" <yoann.congal@smile.fr>]

On Mon Oct 19, 2026 at 2:09 PM CEST, Ayoub Zaki wrote:
> Hi Yoann,
> On 10/19/26 10:16, Yoann Congal wrote:
>> Le lun. 19 oct. 2026 à 10:04, Ayoub Zaki <ayoub.zaki@embetrix.com 
>> <mailto:ayoub.zaki@embetrix.com>> a écrit :
>> 
>>     Hi Yoann,
>>     On 10/19/26 08:43, Yoann Congal wrote:
>>      > On Sat Oct 17, 2026 at 1:18 PM CEST, Ayoub Zaki via
>>     lists.openembedded.org <http://lists.openembedded.org> wrote:
>>      >> Backport the upstream baud rate fix to minicom 2.10 in wrynose.
>>      >>
>>      >> glibc 2.42 added support for arbitrary termios baud rates on
>>     Linux and
>>      >> changed the Bxxx speed constants from bitmasks to integer baud
>>     rates.
>>      >> The macros combine these values directly with c_cflag, resulting in
>>      >> incorrect baud rates and garbled UART output. Use the libc functions
>>      >> instead.
>>      >>
>>      >> Signed-off-by: Ayoub Zaki <ayoub.zaki@embetrix.com
>>     <mailto:ayoub.zaki@embetrix.com>>
>>      >> ---
>>      >>   ...ove-cfset-i-o-speed-macros-for-glibc.patch | 41 +++++++++++
>>     ++++++++
>>      >>   meta/recipes-extended/minicom/minicom_2.10.bb <http://
>>     minicom_2.10.bb> |  4 +-
>>      >>   2 files changed, 44 insertions(+), 1 deletion(-)
>>      >>   create mode 100644 meta/recipes-extended/minicom/files/0001-
>>     src-sysdep.h-remove-cfset-i-o-speed-macros-for-glibc.patch
>>      >
>>      > Hello,
>>      >
>>      > I think this patch is breaking tests:
>>      > [...]
>>      >
>>      >
>>      > Can you look into this please?
>> 
>>     Thanks for reporting this:
>>     The patch exposes a hardcoded minicom/ directory assumption in the
>>     devtool selftests:
>>     https://github.com/openembedded/openembedded-core/blob/master/meta/lib/
>>     oeqa/selftest/cases/devtool.py#L1360 <https://github.com/
>>     openembedded/openembedded-core/blob/master/meta/lib/oeqa/selftest/
>>     cases/devtool.py#L1360>.
>> 
>>     devtool puts generated patches in the existing files/ directory but the
>>     tests stage and clean up minicom/ causing both failures.
>> 
>>     Could we fix those selftest assumptions separately?
>> 
>> 
>> Yes, the test must be updated to match the modified recipe. Can you send 
>> a v2 as a series, with the test fix included?
>> 
>
> The same directory assumptions affect several other devtool selftests. 
> I’ll send separate fixes for master and wrynose to address this more 
> generally.
>
> Once the selftest fix is applied to wrynose, the original minicom patch 
> can be applied unchanged.

Sounds good!
Please ping this patch when all that happened.

Thanks!
-- 
Yoann Congal
Smile ECS