[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <26dd45f4-8fb7-476f-b8ec-9cb685a138ef@wanadoo.fr>
Date: Fri, 8 Dec 2023 18:45:50 +0100
From: Christophe JAILLET <christophe.jaillet@...adoo.fr>
To: yang.guang5@....com.cn, harry.wentland@....com
Cc: jiang.xuexin@....com.cn, chen.haonan2@....com.cn,
cgel.zte@...il.com, sunpeng.li@....com, rodrigo.siqueira@....com,
alexander.deucher@....com, christian.koenig@....com,
xinhui.pan@....com, airlied@...il.com, daniel@...ll.ch,
wenjing.liu@....com, jun.lei@....com, qingqing.zhuo@....com,
alvin.lee2@....com, samson.tam@....com, aric.cyr@....com,
chiawen.huang@....com, gabe.teeger@....com,
amd-gfx@...ts.freedesktop.org, dri-devel@...ts.freedesktop.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH linux-next] drm/amd/display: replace kzalloc and memcpy
with kmemdup
Le 08/12/2023 à 03:44, yang.guang5@....com.cn a écrit :
> From: Yang Guang <yang.guang5@....com.cn>
>
> Convert kzalloc/memcpy operations to memdup makes for
> cleaner code and avoids memcpy() failures
Hi,
usually, function's names are written with () in commit description.
(i.e. kzalloc()/memcpy()).
memdup should be kmemdup().
Finally the proposed change does not avoid memcpy() failures. Should it
fail (what does it mean in this context?), kmemdup() would behave
exactly the same.
>
> Signed-off-by: Chen Haonan <chen.haonan2@....com.cn>
> ---
> drivers/gpu/drm/amd/display/dc/core/dc.c | 10 ++++------
> 1 file changed, 4 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c
> index 76b47f178127..867e1a0fdef6 100644
> --- a/drivers/gpu/drm/amd/display/dc/core/dc.c
> +++ b/drivers/gpu/drm/amd/display/dc/core/dc.c
> @@ -2264,12 +2264,10 @@ struct dc_state *dc_copy_state(struct dc_state *src_ctx)
>
> #ifdef CONFIG_DRM_AMD_DC_FP
> if (new_ctx->bw_ctx.dml2) {
> - dml2 = kzalloc(sizeof(struct dml2_context), GFP_KERNEL);
> - if (!dml2)
> - return NULL;
> -
> - memcpy(dml2, src_ctx->bw_ctx.dml2, sizeof(struct dml2_context));
> - new_ctx->bw_ctx.dml2 = dml2;
> + dml2 = kmemdup(src_ctx->bw_ctx.dml2, sizeof(struct dml2_context), GFP_KERNEL);
sizeof(struct dml2_context) could be sizeof(*dlm2) to be less verbose.
CJ
> + if (!dml2)
> + return NULL;
> + new_ctx->bw_ctx.dml2 = dml2;
> }
> #endif
>
Powered by blists - more mailing lists