drivers/soc/mediatek/mtk-cmdq-helper.c | 241 ++++++++++++------------- 1 file changed, 112 insertions(+), 129 deletions(-)
This series performs various cleanups to the MediaTek CMDQ Helper lib, reducing code duplication and enhancing human readability. This also avoids double initialization struct cmdq_instruction as, in some cases, it was stack-initialized to zero and then overwritten completely anyway a bit later. I'd expect compilers to be somehow smart about that, but still, while at it ... why not :-) Tested on MT8192 Asurada, MT8195 Tomato Chromebooks. AngeloGioacchino Del Regno (3): soc: mediatek: mtk-cmdq: Move mask build and append to function soc: mediatek: mtk-cmdq: Mark very unlikely branches as such soc: mediatek: mtk-cmdq: Move cmdq_instruction init to declaration drivers/soc/mediatek/mtk-cmdq-helper.c | 241 ++++++++++++------------- 1 file changed, 112 insertions(+), 129 deletions(-) -- 2.46.0
On Wed, 18 Sep 2024 12:06:17 +0200, AngeloGioacchino Del Regno wrote:
> This series performs various cleanups to the MediaTek CMDQ Helper lib,
> reducing code duplication and enhancing human readability.
>
> This also avoids double initialization struct cmdq_instruction as,
> in some cases, it was stack-initialized to zero and then overwritten
> completely anyway a bit later.
> I'd expect compilers to be somehow smart about that, but still, while
> at it ... why not :-)
>
> [...]
Applied to v6.12-next/soc, thanks!
[1/3] soc: mediatek: mtk-cmdq: Move mask build and append to function
https://git.kernel.org/mediatek/c/2400e830
[2/3] soc: mediatek: mtk-cmdq: Mark very unlikely branches as such
https://git.kernel.org/mediatek/c/21ab3dae
[3/3] soc: mediatek: mtk-cmdq: Move cmdq_instruction init to declaration
https://git.kernel.org/mediatek/c/66705b89
Cheers,
Angelo
On 02/10/2024 11:08, AngeloGioacchino Del Regno wrote: > On Wed, 18 Sep 2024 12:06:17 +0200, AngeloGioacchino Del Regno wrote: >> This series performs various cleanups to the MediaTek CMDQ Helper lib, >> reducing code duplication and enhancing human readability. >> >> This also avoids double initialization struct cmdq_instruction as, >> in some cases, it was stack-initialized to zero and then overwritten >> completely anyway a bit later. >> I'd expect compilers to be somehow smart about that, but still, while >> at it ... why not :-) >> >> [...] > > Applied to v6.12-next/soc, thanks! > > [1/3] soc: mediatek: mtk-cmdq: Move mask build and append to function > https://git.kernel.org/mediatek/c/2400e830 > [2/3] soc: mediatek: mtk-cmdq: Mark very unlikely branches as such > https://git.kernel.org/mediatek/c/21ab3dae > [3/3] soc: mediatek: mtk-cmdq: Move cmdq_instruction init to declaration > https://git.kernel.org/mediatek/c/66705b89 > You probably oversaw the sparse warning email on 3/3? As I oversaw that you already merged this. Regards, Matthias
Il 02/10/24 15:00, Matthias Brugger ha scritto: > > > On 02/10/2024 11:08, AngeloGioacchino Del Regno wrote: >> On Wed, 18 Sep 2024 12:06:17 +0200, AngeloGioacchino Del Regno wrote: >>> This series performs various cleanups to the MediaTek CMDQ Helper lib, >>> reducing code duplication and enhancing human readability. >>> >>> This also avoids double initialization struct cmdq_instruction as, >>> in some cases, it was stack-initialized to zero and then overwritten >>> completely anyway a bit later. >>> I'd expect compilers to be somehow smart about that, but still, while >>> at it ... why not :-) >>> >>> [...] >> >> Applied to v6.12-next/soc, thanks! >> >> [1/3] soc: mediatek: mtk-cmdq: Move mask build and append to function >> https://git.kernel.org/mediatek/c/2400e830 >> [2/3] soc: mediatek: mtk-cmdq: Mark very unlikely branches as such >> https://git.kernel.org/mediatek/c/21ab3dae >> [3/3] soc: mediatek: mtk-cmdq: Move cmdq_instruction init to declaration >> https://git.kernel.org/mediatek/c/66705b89 >> > > You probably oversaw the sparse warning email on 3/3? > > As I oversaw that you already merged this. > Yeah, I did. There's no problem anyway, as we can always replace the commits without noise, I haven't pushed anything to -next :-) I plan to do that (drop patch 2/3, delete `.mask = 0` on 3/3) soon, at max next week anyway. Cheers, Angelo
On 03/10/2024 10:08, AngeloGioacchino Del Regno wrote: > Il 02/10/24 15:00, Matthias Brugger ha scritto: >> >> >> On 02/10/2024 11:08, AngeloGioacchino Del Regno wrote: >>> On Wed, 18 Sep 2024 12:06:17 +0200, AngeloGioacchino Del Regno wrote: >>>> This series performs various cleanups to the MediaTek CMDQ Helper lib, >>>> reducing code duplication and enhancing human readability. >>>> >>>> This also avoids double initialization struct cmdq_instruction as, >>>> in some cases, it was stack-initialized to zero and then overwritten >>>> completely anyway a bit later. >>>> I'd expect compilers to be somehow smart about that, but still, while >>>> at it ... why not :-) >>>> >>>> [...] >>> >>> Applied to v6.12-next/soc, thanks! >>> >>> [1/3] soc: mediatek: mtk-cmdq: Move mask build and append to function >>> https://git.kernel.org/mediatek/c/2400e830 >>> [2/3] soc: mediatek: mtk-cmdq: Mark very unlikely branches as such >>> https://git.kernel.org/mediatek/c/21ab3dae >>> [3/3] soc: mediatek: mtk-cmdq: Move cmdq_instruction init to declaration >>> https://git.kernel.org/mediatek/c/66705b89 >>> >> >> You probably oversaw the sparse warning email on 3/3? >> >> As I oversaw that you already merged this. >> > > Yeah, I did. There's no problem anyway, as we can always replace the commits > without noise, I haven't pushed anything to -next :-) > > I plan to do that (drop patch 2/3, delete `.mask = 0` on 3/3) soon, at max > next week anyway. > Perfect, before you do any locally please pull from remote as I did some changes, not sure if you saw my message on IRC. Regarding -next: we don't need to keep commit history in linux-next neither, to even it got pushed to -next, we can drop patches in the queue with any problem. Have a nice afternoon :) Matthias
Il 07/10/24 16:33, Matthias Brugger ha scritto: > > > On 03/10/2024 10:08, AngeloGioacchino Del Regno wrote: >> Il 02/10/24 15:00, Matthias Brugger ha scritto: >>> >>> >>> On 02/10/2024 11:08, AngeloGioacchino Del Regno wrote: >>>> On Wed, 18 Sep 2024 12:06:17 +0200, AngeloGioacchino Del Regno wrote: >>>>> This series performs various cleanups to the MediaTek CMDQ Helper lib, >>>>> reducing code duplication and enhancing human readability. >>>>> >>>>> This also avoids double initialization struct cmdq_instruction as, >>>>> in some cases, it was stack-initialized to zero and then overwritten >>>>> completely anyway a bit later. >>>>> I'd expect compilers to be somehow smart about that, but still, while >>>>> at it ... why not :-) >>>>> >>>>> [...] >>>> >>>> Applied to v6.12-next/soc, thanks! >>>> >>>> [1/3] soc: mediatek: mtk-cmdq: Move mask build and append to function >>>> https://git.kernel.org/mediatek/c/2400e830 >>>> [2/3] soc: mediatek: mtk-cmdq: Mark very unlikely branches as such >>>> https://git.kernel.org/mediatek/c/21ab3dae >>>> [3/3] soc: mediatek: mtk-cmdq: Move cmdq_instruction init to declaration >>>> https://git.kernel.org/mediatek/c/66705b89 >>>> >>> >>> You probably oversaw the sparse warning email on 3/3? >>> >>> As I oversaw that you already merged this. >>> >> >> Yeah, I did. There's no problem anyway, as we can always replace the commits >> without noise, I haven't pushed anything to -next :-) >> >> I plan to do that (drop patch 2/3, delete `.mask = 0` on 3/3) soon, at max >> next week anyway. >> > > Perfect, before you do any locally please pull from remote as I did some changes, > not sure if you saw my message on IRC. > > Regarding -next: we don't need to keep commit history in linux-next neither, to > even it got pushed to -next, we can drop patches in the queue with any problem. My day was pretty full, I didn't. Thanks for shooting the email :-D > > Have a nice afternoon :) You too! Angelo
© 2016 - 2024 Red Hat, Inc.