Re: [PATCH v4 1/2] clk: hisilicon: Use correct return value about hisi_reset_init()

From: Tiezhu Yang
Date: Fri May 29 2020 - 06:20:29 EST


On 05/29/2020 05:36 PM, Stephen Boyd wrote:
Quoting Tiezhu Yang (2020-05-28 23:44:20)
On 05/29/2020 12:31 PM, Stephen Boyd wrote:
Quoting Tiezhu Yang (2020-05-28 21:02:05)
I think you didn't understand my question. I'm asking where is this
patch applied to the kernel and what commit is it? I don't see it in the
clk tree.
Sorry for that, actually I do not quite understand what you mean.

In my opinion, after the following commit, when devm_ioremap_resource()
is called in hisi_reset_init(), hisi_reset_init() still returns NULL and
it only returns
-ENOMEM when call hisi_reset_init() failed, I think it may returns
-EINVAL, -EBUSY
or -ENOMEM if failed, this is what I want to fix.

"reset: hisilicon: fix potential NULL pointer dereference"
https://git.kernel.org/pub/scm/linux/kernel/git/clk/linux.git/commit/drivers/clk/hisilicon/reset.c?h=clk-next&id=e9a2310fb689151166df7fd9971093362d34bd79

This commit doesn't change the value that is returned by
hisi_reset_init() on an error. It still returns NULL when an error
happens.

Yes, I agree, but after this commit e9a2310fb689 ("reset:
hisilicon: fix potential NULL pointer dereference"), the
return value of hisi_reset_init() is not so correct because
it replaces devm_ioremap() with devm_ioremap_resource().

Do you think the code of this patch is OK but the "Fixes:" commit
is not accurate? If so, can we remove the "Fixes:" tag?