Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.33 as permitted sender) client-ip=23.128.96.33;
MIME-Version: 1.0
References: <20230913152238.905247-1-mszeredi@redhat.com> <20230913152238.905247-3-mszeredi@redhat.com>
 <44631c05-6b8a-42dc-b37e-df6776baa5d4@app.fastmail.com> <20230925-total-debatten-2a1f839fde5a@brauner>
In-Reply-To: <20230925-total-debatten-2a1f839fde5a@brauner>
From:   Miklos Szeredi <miklos@szeredi.hu>
Date:   Mon, 25 Sep 2023 15:13:36 +0200
Message-ID: <CAJfpegvUCoKebYS=_3eZtCH49nObotuWc=_khFcHshKjRG8h6Q@mail.gmail.com>
Subject: Re: [RFC PATCH 2/3] add statmnt(2) syscall
To:     Christian Brauner <brauner@kernel.org>
Cc:     Arnd Bergmann <arnd@arndb.de>,
        Miklos Szeredi <mszeredi@redhat.com>,
        linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org,
        linux-api@vger.kernel.org, linux-man@vger.kernel.org,
        linux-security-module@vger.kernel.org, Karel Zak <kzak@redhat.com>,
        Ian Kent <raven@themaw.net>,
        David Howells <dhowells@redhat.com>,
        Linus Torvalds <torvalds@linux-foundation.org>,
        Alexander Viro <viro@zeniv.linux.org.uk>,
        Christian Brauner <christian@brauner.io>,
        Amir Goldstein <amir73il@gmail.com>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Mon, 25 Sept 2023 at 15:04, Christian Brauner <brauner@kernel.org> wrote:
>
> On Mon, Sep 25, 2023 at 02:57:31PM +0200, Arnd Bergmann wrote:
> > On Wed, Sep 13, 2023, at 17:22, Miklos Szeredi wrote:
> >
> > >  asmlinkage long sys_fstatfs64(unsigned int fd, size_t sz,
> > >                             struct statfs64 __user *buf);
> > > +asmlinkage long sys_statmnt(u64 mnt_id, u64 mask,
> > > +                       struct statmnt __user *buf, size_t bufsize,
> > > +                       unsigned int flags);
> >
> > This definition is problematic on 32-bit architectures for two
> > reasons:
> >
> > - 64-bit register arguments are passed in pairs of registers
> >   on two architectures, so anything passing those needs to
> >   have a separate entry point for compat syscalls on 64-bit
> >   architectures. I would suggest also using the same one on
> >   32-bit ones, so you don't rely on the compiler splitting
> >   up the long arguments into pairs.
> >
> > - There is a limit of six argument registers for system call
> >   entry points, but with two pairs and three single registers
> >   you end up with seven of them.
> >
> > The listmnt syscall in patch 3 also has the first problem,
> > but not the second.
>
> Both fields could also just be moved into the struct itself just like we
> did for clone3() and others.

Let's not mix in and out args, please.

How about passing u64 *?

Thanks,
Miklos