On 2010.03.11 08:31:57 +0100, Miguel Ojeda wrote:
quoted text
> On Wed, Mar 10, 2010 at 11:09 PM, Miguel Ojeda
> <miguel.ojeda.sandonis@gmail.com> wrote:
> > Hi,
> >
> > The commit 5877960869333e42ebeb733e8d9d5630ff96d350 (included since 2.6.32.4) crashes (locks up) the 82915G/GV/910GL Controller when intel-agp.c tries to access nonexistent GTT entries at:
> >
> > -               for (i = intel_private.gtt_entries; i < current_size->num_entries; i++) {
> > +               for (i = intel_private.gtt_entries; i < intel_private.gtt_total_size; i++) {
> >
> > Rationale: I915 (gma900) has 128 MB of video memory (maximum), as per intel.com ( http://www.intel.com/support/graphics/intel915g/sb/CS-012579.htm ) and lscpi:

I think that page is wrong, and http://www.intel.com/design/chipsets/datashts/301467.htm
has info that 256K is for GTT bar, so max video memory size is 256M. On my 915G
board, I can choose 128M/256M in BIOS setup.

quoted text
> >
> > 00:02.0 VGA compatible controller: Intel Corporation 82915G/GV/910GL Integrated Graphics Controller (rev 04) (prog-if 00 [VGA controller])
> >        Subsystem: Intel Corporation Device 4147
> >        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
> >        Status: Cap+ 66MHz- UDF- FastB2B+ ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
> >        Latency: 0
> >        Interrupt: pin A routed to IRQ 11
> >        Region 0: Memory at ff480000 (32-bit, non-prefetchable) [size=512K]
> >        Region 1: I/O ports at ec00 [size=8]
> >        Region 2: Memory at d8000000 (32-bit, prefetchable) [size=128M]
> >        Region 3: Memory at ff440000 (32-bit, non-prefetchable) [size=256K]

This also tells bar 3 for GTT has 256K.

quoted text
> >        Capabilities: <access denied>
> >
> >
> > AFAIK, that implies that its gtt_total_size (in pages) should be 32K (as num_entries showed before the commit) instead of 64K.
> >
> > Note: The IS_I915 macro includes 945; however, only GMA900 (I915) had 128 MB as the maximum AFAIK. Therefore, I divided the IS_I915 macro. I do not know about the "E7221" (please check).
> >
> > How to reproduce: Access kernel.org in iceweasel (Debian Lenny) and the X server will crash. Sometimes, the kernel freezes.
> >

I can't produce this on my 915G board with 128M or 256M memory config. Could you
paste dmesg in your failure or just hang?

quoted text
> > Please review. The fix should be applied to stable series, as well as 2.6.33 and 2.6.34-rc1.
> >
> > Signed-off-by: Miguel Ojeda <miguel.ojeda.sandonis@gmail.com>
> > ---
> > --- linux-2.6.32.stable/drivers/char/agp/intel-agp.c.old        2010-03-10 15:32:36.000000000 +0100
> > +++ linux-2.6.32.stable/drivers/char/agp/intel-agp.c    2010-03-10 22:38:23.000000000 +0100
> > @@ -65,11 +65,11 @@
> >  #define PCI_DEVICE_ID_INTEL_IGDNG_MC2_HB    0x006a
> >  #define PCI_DEVICE_ID_INTEL_IGDNG_M_IG     0x0046
> >
> > -/* cover 915 and 945 variants */
> >  #define IS_I915 (agp_bridge->dev->device == PCI_DEVICE_ID_INTEL_E7221_HB || \
> >                 agp_bridge->dev->device == PCI_DEVICE_ID_INTEL_82915G_HB || \
> > -                agp_bridge->dev->device == PCI_DEVICE_ID_INTEL_82915GM_HB || \
> > -                agp_bridge->dev->device == PCI_DEVICE_ID_INTEL_82945G_HB || \
> > +                agp_bridge->dev->device == PCI_DEVICE_ID_INTEL_82915GM_HB)
> > +
> > +#define IS_I945 (agp_bridge->dev->device == PCI_DEVICE_ID_INTEL_82945G_HB || \
> >                 agp_bridge->dev->device == PCI_DEVICE_ID_INTEL_82945GM_HB || \
> >                 agp_bridge->dev->device == PCI_DEVICE_ID_INTEL_82945GME_HB)
> >
> > @@ -724,14 +724,14 @@ static void intel_i830_init_gtt_entries(
> >                        break;
> >                case I915_GMCH_GMS_STOLEN_48M:
> >                        /* Check it's really I915G */
> > -                       if (IS_I915 || IS_I965 || IS_G33 || IS_G4X)
> > +                       if (IS_I915 || IS_I945 || IS_I965 || IS_G33 || IS_G4X)
> >                                gtt_entries = MB(48) - KB(size);
> >                        else
> >                                gtt_entries = 0;
> >                        break;
> >                case I915_GMCH_GMS_STOLEN_64M:
> >                        /* Check it's really I915G */
> > -                       if (IS_I915 || IS_I965 || IS_G33 || IS_G4X)
> > +                       if (IS_I915 || IS_I945 || IS_I965 || IS_G33 || IS_G4X)
> >                                gtt_entries = MB(64) - KB(size);
> >                        else
> >                                gtt_entries = 0;
> > @@ -1305,6 +1305,8 @@ static int intel_i915_create_gatt_table(
> >
> >        if (IS_G33)
> >            gtt_map_size = 1024 * 1024; /* 1M on G33 */
> > +       else if (IS_I915)
> > +           gtt_map_size = 128 * 1024; /* 128K on I915 */
> >        intel_private.gtt = ioremap(temp2, gtt_map_size);
> >        if (!intel_private.gtt)
> >                return -ENOMEM;
> >
> >
> >
> 
> Cc'ing the original committers.

-- 
Open Source Technology Center, Intel ltd.

$gpg --keyserver wwwkeys.pgp.net --recv-keys 4D781827