Date:   Fri, 17 Dec 2021 16:14:25 +1100
From:   Herbert Xu <herbert@gondor.apana.org.au>
To:     Kai Ye <yekai13@huawei.com>
Cc:     linux-crypto@vger.kernel.org, linux-kernel@vger.kernel.org,
        wangzhou1@hisilicon.com
Subject: Re: [PATCH v3] crypto: hisilicon - replace 'smp_processor_id' with
 the raw version of the macro
Message-ID: <20211217051425.GA20625@gondor.apana.org.au>
References: <20211206105109.21847-1-yekai13@huawei.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20211206105109.21847-1-yekai13@huawei.com>
User-Agent: Mutt/1.10.1 (2018-07-13)
Precedence: bulk

On Mon, Dec 06, 2021 at 06:51:09PM +0800, Kai Ye wrote:
> smp_processor_id() is unsafe if it's used in a preemption-on critical
> section. It will cause the call trace when the preemption-on and sets the
> CONFIG_DEBUG_PREEMPT. So replace 'smp_processor_id' with the raw version of
> the marco in preemptible to avoid the following call trace:

What you wrote here makes zero sense.  If smp_processor_id is
unsafe then the raw version is even more unsafe.

Please provide the actual justification as to why it is *safe*.

Even better, put the justification in the patch as a comment
instead of the changelog.

Thanks,
-- 
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt