Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Message-ID: <1531339833.26425.1.camel@surriel.com>
Subject: Re: Lazy FPU restoration / moving kernel_fpu_end() to context switch
From:   Rik van Riel <riel@surriel.com>
To:     Sebastian Andrzej Siewior <bigeasy@linutronix.de>,
        "Jason A. Donenfeld" <Jason@zx2c4.com>
Cc:     Andrew Lutomirski <luto@kernel.org>, dave.hansen@linux.intel.com,
        LKML <linux-kernel@vger.kernel.org>, X86 ML <x86@kernel.org>
Date:   Wed, 11 Jul 2018 16:10:33 -0400
In-Reply-To: <20180711162850.52jmzsuwegpk7rag@breakpoint.cc>
References: <CAHmME9pBqGhCjdwx64GxYTKWiMkDNY3v2gnVL_Xm2q=3guOAsQ@mail.gmail.com>
         <CALCETrU+2mBPDfkBz1i_GT1EOJau+mzj4yOK8N0UnT2pGjiUWQ@mail.gmail.com>
         <CAHmME9qhmSVNBq_GnkYrGU-PPEX9Fk0yiGktsCeEXo+LTB9KPg@mail.gmail.com>
         <20180711162850.52jmzsuwegpk7rag@breakpoint.cc>
Content-Type: multipart/signed; micalg="pgp-sha256";
        protocol="application/pgp-signature"; boundary="=-PBWDaPbCpr3to4z8rhDe"
Mime-Version: 1.0
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk


--=-PBWDaPbCpr3to4z8rhDe
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Wed, 2018-07-11 at 18:28 +0200, Sebastian Andrzej Siewior wrote:
> On 2018-06-15 22:33:47 [+0200], Jason A. Donenfeld wrote:
> > On Fri, Jun 15, 2018 at 8:32 PM Andy Lutomirski <luto@kernel.org>
> > wrote:
> > > quite in the form you imagined.  The idea that we've tossed
> > > around is
> > > to restore FPU state on return to user mode.  Roughly, we'd
> > > introduce
> > > a new thread flag TIF_FPU_UNLOADED (name TBD).
> > > prepare_exit_to_usermode() would notice this flag, copy the
> > > fpstate to
> > > fpregs, and clear the flag.  (Or maybe exit_to_usermode_loop() --
> > > No
> > > one has quite thought it through, but I think it should be
> > > outside the
> > > loop.)  We'd update all the FPU accessors to understand the flag.
> >=20
> > Yes! This is exactly what I was thinking. Then those calls to
> > begin()
> > and end() could be placed as close to the actual FPU usage as
> > possible.
>=20
> I was thinking about this myself. Did anyone try to hack something in
> the meantime? I might want to look into this, too :)

I have implemented this before, back before the
big rewrite of the syscall entry and exit code.

It seemed to work fine.

--=20
All Rights Reversed.
--=-PBWDaPbCpr3to4z8rhDe
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: This is a digitally signed message part
Content-Transfer-Encoding: 7bit

-----BEGIN PGP SIGNATURE-----

iQEzBAABCAAdFiEEKR73pCCtJ5Xj3yADznnekoTE3oMFAltGZDkACgkQznnekoTE
3oNjyQgAwGynix8qb8Oi/WeLSl9f9nUgh2s2mqE2a56rq6xBoRopdm/zYttDypQI
CggihoFE+M2gFjfarlSFGaJiKv/VwASvLcVjd5olmBTJ/qgC6Nt1icNAOKkAJMe8
VTl/wcN9M2J86dx7NJPDBLg77oEzL15erc5wOQ4mihxE/bg4lI8sjjo8EWYOv3BY
4VUdSnHGm8XMKZuHj96mltdowedGkTfL0Md2Tns56XsCLxqW4ec1dGJggb55qvnr
IqaBjWNEGZi80nK4HP6YVcoUL6dDcjqfhlQOfRp2luK6vHb4a7YBEdujK9O4JtaV
7AGvolsr9NfhK+kAxT+7Z9EG9YmaAQ==
=pYzj
-----END PGP SIGNATURE-----

--=-PBWDaPbCpr3to4z8rhDe--