Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 04 Oct 2008 14:31:06 -0400
From:      Robert Noland <rnoland@FreeBSD.org>
To:        "Paul B. Mahol" <onemda@gmail.com>
Cc:        freebsd-x11 <freebsd-x11@freebsd.org>
Subject:   Re: drm MSI support
Message-ID:  <1223145067.1619.46.camel@wombat.2hip.net>
In-Reply-To: <3a142e750810041119wc63cdf7u85149c88c5182696@mail.gmail.com>
References:  <1223134762.1619.32.camel@wombat.2hip.net> <1223138743.1619.35.camel@wombat.2hip.net> <3a142e750810041119wc63cdf7u85149c88c5182696@mail.gmail.com>

next in thread | previous in thread | raw e-mail | index | archive | help

--=-v/ztb80ezidumJoLFhTu
Content-Type: text/plain
Content-Transfer-Encoding: quoted-printable

On Sat, 2008-10-04 at 20:19 +0200, Paul B. Mahol wrote:
> On 10/4/08, Robert Noland <rnoland@2hip.net> wrote:
> > On Sat, 2008-10-04 at 11:39 -0400, Robert Noland wrote:
> >> I would like to get some initial feedback on this patch.  It enables M=
SI
> >> on capable cards.  (I'm using it on my i965gm)  The guys at Intel have
> >> done this in the Linux driver.  I have implemented it in our
> >> infrastructure code which means that we will enable it on any hardware
> >> that reports capable.  The Intel guys have explicitly blacklisted a
> >> couple of 945 class devices, stating that they report capable but don'=
t
> >> appear to work.  While I would like to get verification that these chi=
ps
> >> don't work, don't be surprised if it doesn't work if you have chips 27=
72
> >> or 27a2.  I have not implemented the blacklist bits yet.
> >>
> >> The Linux crew has never enabled MSI on radeon hardware, though severa=
l
> >> devices appear to report capable.  When I inquired as to why they hadn=
't
> >> done so, they just stated that they hadn't gotten around to it.  I'm
> >> particularly interested to see how this works on radeons.
> >>
> >> In order to use this patch, you need to be running CURRENT as I had to
> >> make a few other changes to vgapci to get this working. You can see if
> >> you hardware is capable by running pciconf -lvc.  You will see somethi=
ng
> >> like this:
> >
> > CURRENT needs to be at least:
> > r183194 | rnoland | 2008-09-19 15:11:35 -0400 (Fri, 19 Sep 2008) | 6
> > lines
> >
> > pci_setup_intr() will only enable MSI/MSI-X for direct children.  Add
> > methods
> > to vga_pci.c to request on behalf of it's children.  This causes vgapci
> > to show
> > up as the interrupt owner in vmstat -i, rather than the child device.
> >
> > Approved by:    jhb(mentor)
> >
> > robert.
> >
> >> vgapci0@pci0:0:2:0:     class=3D0x030000 card=3D0x01f91028 chip=3D0x2a=
028086
> >> rev=3D0x0c hdr=3D0x00
> >>     vendor     =3D 'Intel Corporation'
> >>     device     =3D 'Mobile 965 Express Integrated Graphics Controller'
> >>     class      =3D display
> >>     subclass   =3D VGA
> >>     cap 05[90] =3D MSI supports 1 message
> >>     cap 01[d0] =3D powerspec 3  supports D0 D3  current D0
> >>
> >> When drm loads it will also report that it has enabled MSI.
> >>
> >> Please send me reports of what chips do/don't work.
>=20
> vgapci0@pci0:0:2:0:     class=3D0x030000 card=3D0x30a2103c chip=3D0x27a28=
086
> rev=3D0x03 hdr=3D0x00
>     vendor     =3D 'Intel Corporation'
>     device     =3D 'Mobile 945GM/GU Express Integrated Graphics Controlle=
r'
>     class      =3D display
>     subclass   =3D VGA
>     cap 05[90] =3D MSI supports 1 message
>     cap 01[d0] =3D powerspec 2  supports D0 D3  current D0
>=20
>=20
> drm0: <Intel i945GM> on vgapci0
> info: [drm] MSI enabled 1 message(s)
> vgapci0: child drm0 requested pci_enable_busmaster
> info: [drm] AGP at 0xe0000000 256MB
> info: [drm] Initialized i915 1.6.0 20080730
> drm0: [ITHREAD]
> vgapci0: child drm0 requested pci_disable_busmaster
> info: [drm] MSI released
> drm0: detached
>=20
> It decrease framerate of glxgears 10x.
> stellarium performance is even worse (worse than using software
> rendering when DRI is disabled).

I expect that you aren't getting interrupts, vmstat -ia will tell you.
Yours is one of the cards that Intel blacklisted though, so it's not
entirely unexpected.

robert.

> _______________________________________________
> freebsd-x11@freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-x11
> To unsubscribe, send any mail to "freebsd-x11-unsubscribe@freebsd.org"

--=-v/ztb80ezidumJoLFhTu
Content-Type: application/pgp-signature; name=signature.asc
Content-Description: This is a digitally signed message part

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.9 (FreeBSD)

iEYEABECAAYFAkjntmoACgkQM4TrQ4qfROM/TwCePXrrdFtdrU1DqujQtIwKjFo6
gfEAnjBnXfqJB1lVzT7R1qpVmpVp1i+I
=cYb1
-----END PGP SIGNATURE-----

--=-v/ztb80ezidumJoLFhTu--




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1223145067.1619.46.camel>