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: Fri, 6 Sep 2019 10:09:08 +1000 Message-ID: <20190906000908.xpvkuhun7v6onp6w@yavin.dot.cyphar.com> References: <20190904201933.10736-1-cyphar@cyphar.com> <20190904201933.10736-2-cyphar@cyphar.com> <20190905180750.GQ1131@ZenIV.linux.org.uk> <20190905230003.bek7vqdvruzi4ybx@yavin.dot.cyphar.com> <20190905234944.GT1131@ZenIV.linux.org.uk> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha256; protocol="application/pgp-signature"; boundary="45flctx23d2mffx4" Return-path: Content-Disposition: inline In-Reply-To: <20190905234944.GT1131@ZenIV.linux.org.uk> Sender: linux-kernel-owner@vger.kernel.org To: Al Viro Cc: Jeff Layton , "J. Bruce Fields" , Arnd Bergmann , David Howells , Shuah Khan , Shuah Khan , Ingo Molnar , Peter Zijlstra , Christian Brauner , Rasmus Villemoes , Eric Biederman , Andy Lutomirski , Andrew Morton , Alexei Starovoitov , Kees Cook , Jann Horn , Tycho Andersen , David Drysdale , Chanho Min , Oleg Nesterov , Alexander List-Id: linux-api@vger.kernel.org --45flctx23d2mffx4 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On 2019-09-06, Al Viro wrote: > On Fri, Sep 06, 2019 at 09:00:03AM +1000, Aleksa Sarai wrote: > > > > + return -EFAULT; > > > > + } > > > > + /* Copy the interoperable parts of the struct. */ > > > > + if (__copy_to_user(dst, src, size)) > > > > + return -EFAULT; > > >=20 > > > Why not simply clear_user() and copy_to_user()? > >=20 > > I'm not sure I understand what you mean -- are you asking why we need to > > do memchr_inv(src + size, 0, rest) earlier? >=20 > I'm asking why bother with __ and separate access_ok(). Ah right, it was a dumb "optimisation" (since we need to do access_ok() anyway since we should early -EFAULT in that case). I've dropped the __ usages in my working copy. > > > if ((unsigned long)addr & 1) { > > > u8 v; > > > if (get_user(v, (__u8 __user *)addr)) > > > return -EFAULT; > > > if (v) > > > return -E2BIG; > > > addr++; > > > } > > > if ((unsigned long)addr & 2) { > > > u16 v; > > > if (get_user(v, (__u16 __user *)addr)) > > > return -EFAULT; > > > if (v) > > > return -E2BIG; > > > addr +=3D2; > > > } > > > if ((unsigned long)addr & 4) { > > > u32 v; > > > if (get_user(v, (__u32 __user *)addr)) > > > return -EFAULT; > > > if (v) > > > return -E2BIG; > > > } > > > >=20 > Actually, this is a dumb way to do it - page size on anything > is going to be a multiple of 8, so you could just as well > read 8 bytes from an address aligned down. Then mask the > bytes you don't want to check out and see if there's anything > left. >=20 > You can have readability boundaries inside a page - it's either > the entire page (let alone a single word) being readable, or > it's EFAULT for all parts. >=20 > > > would be saner, and things like x86 could trivially add an > > > asm variant - it's not hard. Incidentally, memchr_inv() is > > > an overkill in this case... > >=20 > > Why is memchr_inv() overkill? >=20 > Look at its implementation; you only care if there are > non-zeroes, you don't give a damn where in the buffer > the first one would be. All you need is the same logics > as in "from userland" case > if (!count) > return true; > offset =3D (unsigned long)from & 7 > p =3D (u64 *)(from - offset); > v =3D *p++; > if (offset) { // unaligned > count +=3D offset; > v &=3D ~aligned_byte_mask(offset); // see strnlen_user.c > } > while (count > 8) { > if (v) > return false; > v =3D *p++; > count -=3D 8; > } > if (count !=3D 8) > v &=3D aligned_byte_mask(count); > return v =3D=3D 0; >=20 > All there is to it... Alright, will do (for some reason I hadn't made the connection that memchr_inv() is doing effectively the same word-by-word comparison but also detecting where the first byte is). --=20 Aleksa Sarai Senior Software Engineer (Containers) SUSE Linux GmbH --45flctx23d2mffx4 Content-Type: application/pgp-signature; name="signature.asc" -----BEGIN PGP SIGNATURE----- iHUEABYIAB0WIQSxZm6dtfE8gxLLfYqdlLljIbnQEgUCXXGjoAAKCRCdlLljIbnQ EpC9AP0R1Y7fvOkhCrlqhEeSXH2/w/eSafFO51uuSnY7m3dVegEAm16vVXT68ypo Z7fWiISgwHeOk0U5O9VS4cZGMgtS3ws= =nF2N -----END PGP SIGNATURE----- --45flctx23d2mffx4--