From mboxrd@z Thu Jan 1 00:00:00 1970 From: Aleksa Sarai Subject: Re: [PATCH v12 01/12] lib: introduce copy_struct_{to,from}_user helpers Date: Wed, 11 Sep 2019 20:37:30 +1000 Message-ID: <20190911103730.72unmfp7lsvvafxo@yavin> References: <20190904201933.10736-1-cyphar@cyphar.com> <20190904201933.10736-2-cyphar@cyphar.com> <20190905073205.GY2332@hirez.programming.kicks-ass.net> <20190905092622.tlb6nn3uisssdfbu@yavin.dot.cyphar.com> <20190905094305.GJ2349@hirez.programming.kicks-ass.net> <20190905105749.GW2386@hirez.programming.kicks-ass.net> Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="===============3127660236630967693==" Return-path: In-Reply-To: <20190905105749.GW2386@hirez.programming.kicks-ass.net> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: "linux-arm-kernel" Errors-To: linux-arm-kernel-bounces+linux-arm-kernel=m.gmane.org@lists.infradead.org To: Peter Zijlstra Cc: linux-ia64@vger.kernel.org, linux-sh@vger.kernel.org, Alexander Shishkin , Rasmus Villemoes , Alexei Starovoitov , linux-kernel@vger.kernel.org, David Howells , linux-kselftest@vger.kernel.org, sparclinux@vger.kernel.org, Jiri Olsa , linux-arch@vger.kernel.org, linux-s390@vger.kernel.org, Tycho Andersen , Aleksa Sarai , Shuah Khan , Ingo Molnar , linux-arm-kernel@lists.infradead.org, linux-mips@vger.kernel.org, linux-xtensa@linux-xtensa.org, Kees Cook , Arnd Bergmann , Jann Horn , linuxppc-dev@lists.ozlabs.org, linux-m68k@lists.linux-m68k.org, Al Viro , Andy Lutomirski , Shuah Khan List-Id: linux-api@vger.kernel.org --===============3127660236630967693== Content-Type: multipart/signed; micalg=pgp-sha256; protocol="application/pgp-signature"; boundary="xpghzs3x7rsccuuz" Content-Disposition: inline --xpghzs3x7rsccuuz Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On 2019-09-05, Peter Zijlstra wrote: > On Thu, Sep 05, 2019 at 11:43:05AM +0200, Peter Zijlstra wrote: > > On Thu, Sep 05, 2019 at 07:26:22PM +1000, Aleksa Sarai wrote: > > > On 2019-09-05, Peter Zijlstra wrote: > > > > On Thu, Sep 05, 2019 at 06:19:22AM +1000, Aleksa Sarai wrote: > > > > > +/** > > > > > + * copy_struct_to_user: copy a struct to user space > > > > > + * @dst: Destination address, in user space. > > > > > + * @usize: Size of @dst struct. > > > > > + * @src: Source address, in kernel space. > > > > > + * @ksize: Size of @src struct. > > > > > + * > > > > > + * Copies a struct from kernel space to user space, in a way tha= t guarantees > > > > > + * backwards-compatibility for struct syscall arguments (as long= as future > > > > > + * struct extensions are made such that all new fields are *appe= nded* to the > > > > > + * old struct, and zeroed-out new fields have the same meaning a= s the old > > > > > + * struct). > > > > > + * > > > > > + * @ksize is just sizeof(*dst), and @usize should've been passed= by user space. > > > > > + * The recommended usage is something like the following: > > > > > + * > > > > > + * SYSCALL_DEFINE2(foobar, struct foo __user *, uarg, size_t, = usize) > > > > > + * { > > > > > + * int err; > > > > > + * struct foo karg =3D {}; > > > > > + * > > > > > + * // do something with karg > > > > > + * > > > > > + * err =3D copy_struct_to_user(uarg, usize, &karg, sizeof(k= arg)); > > > > > + * if (err) > > > > > + * return err; > > > > > + * > > > > > + * // ... > > > > > + * } > > > > > + * > > > > > + * There are three cases to consider: > > > > > + * * If @usize =3D=3D @ksize, then it's copied verbatim. > > > > > + * * If @usize < @ksize, then kernel space is "returning" a new= er struct to an > > > > > + * older user space. In order to avoid user space getting inc= omplete > > > > > + * information (new fields might be important), all trailing = bytes in @src > > > > > + * (@ksize - @usize) must be zerored > > > >=20 > > > > s/zerored/zero/, right? > > >=20 > > > It should've been "zeroed". > >=20 > > That reads wrong to me; that way it reads like this function must take > > that action and zero out the 'rest'; which is just wrong. > >=20 > > This function must verify those bytes are zero, not make them zero. > >=20 > > > > > , otherwise -EFBIG is re= turned. > > > >=20 > > > > 'Funny' that, copy_struct_from_user() below seems to use E2BIG. > > >=20 > > > This is a copy of the semantics that sched_[sg]etattr(2) uses -- E2BI= G for > > > a "too big" struct passed to the kernel, and EFBIG for a "too big" > > > struct passed to user-space. I would personally have preferred EMSGSI= ZE > > > instead of EFBIG, but felt using the existing error codes would be le= ss > > > confusing. > >=20 > > Sadly a recent commit: > >=20 > > 1251201c0d34 ("sched/core: Fix uclamp ABI bug, clean up and robustify= sched_read_attr() ABI logic and code") > >=20 > > Made the situation even 'worse'. >=20 > And thinking more about things; I'm not convinced the above patch is > actually right. >=20 > Do we really want to simply truncate all the attributes of the task? >=20 > And should we not at least set sched_flags when there are non-default > clamp values applied? >=20 > See; that is I think the primary bug that had chrt failing; we tried to > publish the default clamp values as !0. I just saw this patch in -rc8 -- should I even attempt to port sched_getattr(2) to copy_struct_to_user()? I agree that publishing a default non-zero value is a mistake -- once you do that, old user space will either get confused or lose information. --=20 Aleksa Sarai Senior Software Engineer (Containers) SUSE Linux GmbH --xpghzs3x7rsccuuz Content-Type: application/pgp-signature; name="signature.asc" -----BEGIN PGP SIGNATURE----- iHUEABYIAB0WIQSxZm6dtfE8gxLLfYqdlLljIbnQEgUCXXjOZwAKCRCdlLljIbnQ EgeZAP9UC+Kf1AuY3XgSz6a5avhF1Eskr6DzrSR4wx0T62dnoQD9GcCXU0oVrERB 0xz5K9MrU1nBr6ERqmBwygo/DVsTwAk= =oep8 -----END PGP SIGNATURE----- --xpghzs3x7rsccuuz-- --===============3127660236630967693== Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel --===============3127660236630967693==--