2021-09-15 14:19:59

by Peter Zijlstra

[permalink] [raw]
Subject: [PATCH 00/20] futex: splitup and waitv syscall

Hi,

Neither Thomas nor myself much liked that futex2.c nor do we think that CONFIG_
symbol makes much sense.

However, futex.c is a wee bit long and splitting it up makes sense. So I've
taken the liberty to replace your initial patch with 15 of my own and then
rebased the remaining patches on top of that.

Thomas, does something like this work for you?


2021-09-15 15:15:43

by André Almeida

[permalink] [raw]
Subject: Re: [PATCH 00/20] futex: splitup and waitv syscall

Hi Peter,

Às 11:07 de 15/09/21, Peter Zijlstra escreveu:
> Hi,
>
> Neither Thomas nor myself much liked that futex2.c nor do we think that CONFIG_
> symbol makes much sense.
>
> However, futex.c is a wee bit long and splitting it up makes sense. So I've
> taken the liberty to replace your initial patch with 15 of my own and then
> rebased the remaining patches on top of that.
>

Thank you for doing that. futex.c is indeed too big and I believe that
this really make the code easier to read, and the way you organized
makes more sense than the way I did. Maybe even robust could have a
separated file?

Thanks,
André

2021-09-15 18:26:29

by André Almeida

[permalink] [raw]
Subject: Re: [PATCH 00/20] futex: splitup and waitv syscall

Às 11:07 de 15/09/21, Peter Zijlstra escreveu:
> Hi,
>
> Neither Thomas nor myself much liked that futex2.c nor do we think that CONFIG_
> symbol makes much sense.
>
> However, futex.c is a wee bit long and splitting it up makes sense. So I've
> taken the liberty to replace your initial patch with 15 of my own and then
> rebased the remaining patches on top of that.
>
> Thomas, does something like this work for you?
>
I tested this patchset:

- On top of a full distro
- Using futex kselftests
- Running glibc's nptl tests, all 387 passed

It seems that things are working as expected. You can add my

Reviewed-by: André Almeida <[email protected]>

for patches 01 - 15.