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;
MIME-Version: 1.0
References: <20230623142054.3945121-1-kyletso@google.com> <2023062359-moody-bunch-3011@gregkh>
In-Reply-To: <2023062359-moody-bunch-3011@gregkh>
From:   Kyle Tso <kyletso@google.com>
Date:   Fri, 23 Jun 2023 22:51:41 +0800
Message-ID: <CAGZ6i=1a+fz0p2pkxv3zGKgDgcaTtGQgoo-zfU-SYhq+rVsJJA@mail.gmail.com>
Subject: Re: [PATCH] sysfs: Replace %p with %pK in the warning of sysfs_emit*
To:     Greg KH <gregkh@linuxfoundation.org>
Cc:     rafael@kernel.org, badhri@google.com, linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Fri, Jun 23, 2023 at 10:30=E2=80=AFPM Greg KH <gregkh@linuxfoundation.or=
g> wrote:
>
> On Fri, Jun 23, 2023 at 10:20:54PM +0800, Kyle Tso wrote:
> > According to Documentation/core-api/printk-formats.rst, kernel pointers
> > need to be printed with %pK format specifier to respect kptr_restrict i=
n
> > sysctl.
> >
> > Also replace the function names in the strings with %s and __func__ as
> > checkpatch.pl suggested.
> >
> > Signed-off-by: Kyle Tso <kyletso@google.com>
> > ---
> >  fs/sysfs/file.c | 4 ++--
> >  1 file changed, 2 insertions(+), 2 deletions(-)
> >
> > diff --git a/fs/sysfs/file.c b/fs/sysfs/file.c
> > index a12ac0356c69..56712f0886ef 100644
> > --- a/fs/sysfs/file.c
> > +++ b/fs/sysfs/file.c
> > @@ -732,7 +732,7 @@ int sysfs_emit(char *buf, const char *fmt, ...)
> >       int len;
> >
> >       if (WARN(!buf || offset_in_page(buf),
> > -              "invalid sysfs_emit: buf:%p\n", buf))
> > +              "invalid %s: buf:%pK\n", __func__, buf))
>
> Wait, no, these are not going out in sysfs files, as they are still
> going to dmesg, right?
>
> So I think this line in the documentation matters:
>         This modifier is *only* intended when producing content of a file=
 read by
>         userspace from e.g. procfs or sysfs, not for dmesg. Please refer =
to the
>         section about %p above for discussion about how to manage hashing=
 pointers
>         in printk().
>
> so %p should still be ok here, the hashed value will end up in the
> kernerl log, which is what we want, and the WARN() traceback will show
> the problem properly, right?
>
> >               return 0;
> >
> >       va_start(args, fmt);
> > @@ -760,7 +760,7 @@ int sysfs_emit_at(char *buf, int at, const char *fm=
t, ...)
> >       int len;
> >
> >       if (WARN(!buf || offset_in_page(buf) || at < 0 || at >=3D PAGE_SI=
ZE,
> > -              "invalid sysfs_emit_at: buf:%p at:%d\n", buf, at))
> > +              "invalid %s: buf:%pK at:%d\n", __func__, buf, at))
>
>
> Same here, %p should be correct.
>
> Sorry for missing this before.
>
> greg k-h

Yes, you are right. Sorry for the noise.

Kyle