Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   =?utf-8?B?QmrDtnJuIFTDtnBlbA==?= <bjorn@kernel.org>
To:     Alexandre Ghiti <alex@ghiti.fr>,
        Alexandre Ghiti <alexghiti@rivosinc.com>,
        Paul Walmsley <paul.walmsley@sifive.com>,
        Palmer Dabbelt <palmer@dabbelt.com>,
        Albert Ou <aou@eecs.berkeley.edu>,
        Alan Kao <alankao@andestech.com>,
        linux-riscv@lists.infradead.org, linux-kernel@vger.kernel.org
Cc:     Bo YU <tsu.yubo@gmail.com>, Aurelien Jarno <aurelien@aurel32.net>
Subject: Re: [PATCH -fixes] riscv: uaccess: Return the number of bytes
 effectively copied
In-Reply-To: <dff5c443-bebf-9bd9-af1c-85ebfb7c2aec@ghiti.fr>
References: <20230811110304.1613032-1-alexghiti@rivosinc.com>
 <dff5c443-bebf-9bd9-af1c-85ebfb7c2aec@ghiti.fr>
Date:   Fri, 11 Aug 2023 14:33:53 +0200
Message-ID: <87r0o9ahce.fsf@all.your.base.are.belong.to.us>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

Alexandre Ghiti <alex@ghiti.fr> writes:

> On 11/08/2023 13:03, Alexandre Ghiti wrote:
>> It was reported that the riscv kernel hangs while executing the test
>> in [1].
>>
>> Indeed, the test hangs when trying to write a buffer to a file. The
>> problem is that the riscv implementation of raw_copy_from_user() does not
>> return the number of bytes written when an exception happens and is fixed
>> up.
>
>
> I'll respin another version as the changelog and the title are=20
> incorrect: the uaccess routines should not return the number of bytes=20
> copied but actually the number of bytes not copied (this is what this=20
> patch implements).
>
> I'll wait for feedbacks before doing so!

Yikes! Nice catch.

Functions like fault_in_readable() will fail horribly w/o this. I wonder
why we haven't seen this problem more?

Feel free to add my RB to your next spin!

Reviewed-by: Bj=C3=B6rn T=C3=B6pel <bjorn@rivosinc.com>