From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:57861) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1bP4IK-0004MC-DN for qemu-devel@nongnu.org; Mon, 18 Jul 2016 04:52:57 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1bP4IH-0008Dc-16 for qemu-devel@nongnu.org; Mon, 18 Jul 2016 04:52:56 -0400 Received: from 6.mo178.mail-out.ovh.net ([46.105.53.132]:39934) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1bP4IG-0008DS-N2 for qemu-devel@nongnu.org; Mon, 18 Jul 2016 04:52:52 -0400 Received: from player169.ha.ovh.net (b9.ovh.net [213.186.33.59]) by mo178.mail-out.ovh.net (Postfix) with ESMTP id 0822F10105C4 for ; Mon, 18 Jul 2016 10:52:50 +0200 (CEST) Date: Mon, 18 Jul 2016 10:52:39 +0200 From: Greg Kurz Message-ID: <20160718105239.646accea@bahia.lan> In-Reply-To: <20160718011725.GE16769@voom.fritz.box> References: <1468483025-1084-1-git-send-email-david@gibson.dropbear.id.au> <1468483025-1084-3-git-send-email-david@gibson.dropbear.id.au> <20160714115945.GQ14615@voom.fritz.box> <20160716001156.227ab4c9@bahia.lan> <20160718011725.GE16769@voom.fritz.box> MIME-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; boundary="Sig_/EfD6Ivz7CvzIh4wQo50Y5O2"; protocol="application/pgp-signature" Subject: Re: [Qemu-devel] [RFC 2/2] linux-user: Fix cpu_index generation List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: David Gibson Cc: Bharata B Rao , Peter Maydell , Riku Voipio , QEMU Developers , Igor Mammedov --Sig_/EfD6Ivz7CvzIh4wQo50Y5O2 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: quoted-printable On Mon, 18 Jul 2016 11:17:25 +1000 David Gibson wrote: > On Sat, Jul 16, 2016 at 12:11:56AM +0200, Greg Kurz wrote: > > On Thu, 14 Jul 2016 21:59:45 +1000 > > David Gibson wrote: > > =20 > > > On Thu, Jul 14, 2016 at 03:50:56PM +0530, Bharata B Rao wrote: =20 > > > > On Thu, Jul 14, 2016 at 3:24 PM, Peter Maydell wrote: =20 > > > > > On 14 July 2016 at 08:57, David Gibson wrote: =20 > > > > >> With CONFIG_USER_ONLY, generation of cpu_index values is done di= fferently > > > > >> than for full system targets. This method turns out to be broke= n, since > > > > >> it can fairly easily result in duplicate cpu_index values for > > > > >> simultaneously active cpus (i.e. threads in the emulated process= ). > > > > >> > > > > >> Consider this sequence: > > > > >> Create thread 1 > > > > >> Create thread 2 > > > > >> Exit thread 1 > > > > >> Create thread 3 > > > > >> > > > > >> With the current logic thread 1 will get cpu_index 1, thread 2 w= ill get > > > > >> cpu_index 2 and thread 3 will also get cpu_index 2 (because ther= e are 2 > > > > >> threads in the cpus list at the point of its creation). > > > > >> > > > > >> We mostly get away with this because cpu_index values aren't tha= t important > > > > >> for userspace emulation. Still, it can't be good, so this patch= fixes it > > > > >> by making CONFIG_USER_ONLY use the same bitmap based allocation = that full > > > > >> system targets already use. > > > > >> > > > > >> Signed-off-by: David Gibson > > > > >> --- > > > > >> exec.c | 19 ------------------- > > > > >> 1 file changed, 19 deletions(-) > > > > >> > > > > >> diff --git a/exec.c b/exec.c > > > > >> index 011babd..e410dab 100644 > > > > >> --- a/exec.c > > > > >> +++ b/exec.c > > > > >> @@ -596,7 +596,6 @@ AddressSpace *cpu_get_address_space(CPUState= *cpu, int asidx) > > > > >> } > > > > >> #endif > > > > >> > > > > >> -#ifndef CONFIG_USER_ONLY > > > > >> static DECLARE_BITMAP(cpu_index_map, MAX_CPUMASK_BITS); > > > > >> > > > > >> static int cpu_get_free_index(Error **errp) > > > > >> @@ -617,24 +616,6 @@ static void cpu_release_index(CPUState *cpu) > > > > >> { > > > > >> bitmap_clear(cpu_index_map, cpu->cpu_index, 1); > > > > >> } > > > > >> -#else > > > > >> - > > > > >> -static int cpu_get_free_index(Error **errp) > > > > >> -{ > > > > >> - CPUState *some_cpu; > > > > >> - int cpu_index =3D 0; > > > > >> - > > > > >> - CPU_FOREACH(some_cpu) { > > > > >> - cpu_index++; > > > > >> - } > > > > >> - return cpu_index; > > > > >> -} > > > > >> - > > > > >> -static void cpu_release_index(CPUState *cpu) > > > > >> -{ > > > > >> - return; > > > > >> -} > > > > >> -#endif =20 > > > > > > > > > > Won't this change impose a maximum limit of 256 simultaneous > > > > > threads? That seems a little low for comfort. =20 > > > >=20 > > > > This was the reason why the bitmap logic wasn't applied to > > > > CONFIG_USER_ONLY when it was introduced. > > > >=20 > > > > https://lists.gnu.org/archive/html/qemu-devel/2015-05/msg01980.html= =20 > > >=20 > > > Ah.. good point. > > >=20 > > > Hrm, ok, my next idea would be to just (globally) sequentially > > > allocate cpu_index values for CONFIG_USER, and never try to re-use > > > them. Does that seem reasonable? > > > =20 > >=20 > > Isn't it only deferring the problem to later ? =20 >=20 > You mean that we could get duplicate indexes after the value wraps > around? >=20 Yes. > I suppose, but duplicates after spawning 4 billion threads seems like > a substantial improvement over duplicates after spawning 3 in the > wrong order.. >=20 Agreed. It takes ~ 20 seconds to user QEMU to spawn 10000 threads on my palmetto box, so the wrap around would occur after ~ 100 days. :) > > Maybe it is possible to define MAX_CPUMASK_BITS to a much higher > > value fo CONFIG_USER only instead ? =20 >=20 > Perhaps. It does mean carrying around a huge bitmap, though. >=20 > Another option is to remove cpu_index entirely for the user only > case. I have some patches for this, which are very ugly but it's > possible they can be cleaned up to something reasonable (the biggest > chunk is moving a bunch of ARM stuff under #ifndef CONFIG_USER_ONLY > for what I think are registers that aren't accessible in user mode). >=20 >=20 > > > > But then we didn't have actual removal, but we do now. =20 > > >=20 > > > You mean patch 1/2 in this set? Or something else? > > >=20 > > > Even so, 256 does seem a bit low for a number of simultaneously active > > > threads - there are some bug hairy multi-threaded programs out there. > > > =20 > > =20 >=20 >=20 >=20 --Sig_/EfD6Ivz7CvzIh4wQo50Y5O2 Content-Type: application/pgp-signature Content-Description: OpenPGP digital signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iEYEARECAAYFAleMmNcACgkQAvw66wEB28KClwCgm7qIG6aQyVM7cwPN5I655uHw AcYAoJuynabqTWdUk91w2fzpvWuMIwK7 =NPNS -----END PGP SIGNATURE----- --Sig_/EfD6Ivz7CvzIh4wQo50Y5O2--