Re: [PATCH 1/5] Linux: Add tables with system call numbers

[Florian Weimer <fw@deneb.enyo.de>]

* Florian Weimer:

> * Siddhesh Poyarekar:
>
>> On 31/12/19 9:34 pm, Florian Weimer wrote:
>>> I'll see what I can do about this.  It will require quite a bit
>>> re-testing.
>>> 
>>
>> Thanks, the freeze (I'm evaluating and writing that email as I write
>> this) will be slushy at least till the end of this week.  At worst (i.e.
>> if you can't make it by the end of the week), we can conclude the review
>> and you can commit it right after release.
>>
>>> No, there are system call tables which are shared across tuples,
>>> e.g. little-endian and big-endian variants.  These also get updated
>>> multiple times.
>>> 
>>> The only alternative would be to lock a different file, perhaps
>>> sysdeps/unix/sysv/linux/Makefile.  I don't know the impact yet on the
>>> execution time.
>>
>> OK, we can hold off on that, the current version of this bit is fine.
>
> Here's a new version of the patch, with the file comment and the
> atomic file replacement.
>
> Initial testing looks good, but build-many-glibcs.py is still running.

A glibcs step for build-many-glibcs.py passed, and the .so* files are
unchanged.