AGPGART for VIA vt8235, kernel 2.4.21-pre2

Here is a patch that enables AGPGART for the VIA vt8235 chipset on P4X400 motherboards. Rather trivial, adds the pci id, then a few lines in the agp sources to get it identified. I've only been able to test this on my computer, where it works.

Rich

--BEGIN--

diff -ur linux/drivers/char/agp/agp.h linux-2.4.21-pre2/drivers/char/agp/agp.h

drivers/char/agp/agp.h 2003-01-02 14:35:02.000000000 -0500 renamed from drivers/char/agp/agp.h 2003-01-02 14:36:15.000000000 -0500 +3 -0

@@ -157,6 +157,9 @@
 
 #define PGE_EMPTY(p) (!(p) || (p) == (unsigned long) agp_bridge.scratch_page)
 
+#ifndef PCI_DEVICE_ID_VIA_8235_0
+#define PCI_DEVICE_ID_VIA_8235_0       0x3168
+#endif
 #ifndef PCI_DEVICE_ID_VIA_82C691_0
 #define PCI_DEVICE_ID_VIA_82C691_0      0x0691
 #endif

diff -ur linux/drivers/char/agp/agpgart_be.c linux-2.4.21-pre2/drivers/char/agp/agpgart_be.c

drivers/char/agp/agpgart_be.c 2003-01-02 14:35:02.000000000 -0500 renamed from drivers/char/agp/agpgart_be.c 2003-01-02 14:36:15.000000000 -0500 +6 -0

@@ -4640,6 +4640,12 @@
 #endif /* CONFIG_AGP_SIS */
 
 #ifdef CONFIG_AGP_VIA
+  { PCI_DEVICE_ID_VIA_8235_0,
+     PCI_VENDOR_ID_VIA,
+     VIA_P4X400,
+     "Via",
+     "P4X400",
+     via_generic_setup },
   { PCI_DEVICE_ID_VIA_8501_0,
      PCI_VENDOR_ID_VIA,
      VIA_MVP4,

diff -ur linux/drivers/pci/pci.ids linux-2.4.21-pre2/drivers/pci/pci.ids

drivers/pci/pci.ids 2003-01-02 14:35:02.000000000 -0500 renamed from drivers/pci/pci.ids 2003-01-02 14:36:16.000000000 -0500 +1 -0

@@ -2751,6 +2751,7 @@
   3147  VT8233A ISA Bridge
   3148  P4M266 Host Bridge
   3156  P/KN266 Host Bridge
+  3168  VT8235 [P4X400 AGP]
   3177  VT8233A ISA Bridge
      1458 5001 GA-7VAX Mainboard
   3189  VT8377 [KT400 AGP] Host Bridge

diff -ur linux/include/linux/agp_backend.h linux-2.4.21-pre2/include/linux/agp_backend.h

include/linux/agp_backend.h 2003-01-02 14:35:02.000000000 -0500 renamed from include/linux/agp_backend.h 2003-01-02 14:36:17.000000000 -0500 +1 -2

@@ -54,6 +54,7 @@
   INTEL_I850,
   INTEL_I860,
   VIA_GENERIC,
+  VIA_P4X400,
   VIA_VP3,
   VIA_MVP3,
   VIA_MVP4,
 
--END--
 
 
-

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

The vt8235 is the southbridge chip and that the AGP bridge is located in the northbridge, which most likely has a different number.

Quoted 3 lines

The vt8235 is the southbridge chip and that the AGP bridge is located in the northbridge, which most likely has a different number.

Sorry about that, thought the AGP was in the southbridge chip. I see however that someone else has submitted a patch with the correct naming conventions (It does seem to change between P4X333 and P4X400 however).

In fact, P4X333 defines a chipset rather than a chip. According to ..
http://www.viatech.com/en/apollo/p4x333.jsp and
http://www.viatech.com/en/apollo/p4x400.jsp ,
Just to confirm, device 3168 is the host bridge in lspci output right? And this does all work when you run a DRI application ?

Actually, the P4X400 won't work in current situations. I'll bet if you put an AGPx8 card in there, it'll do the same "change into AGP3.0 mode" trick that the KT400 does. I'll move onto that after I finish up the KT400 GART driver.

What exactly do you mean by 'it' ? The PCI device id ?

If they share the same device ID, then Bernhards patch is really no better...

#define PCI_DEVICE_ID_VIA_P4X333 0x3168

In fact, P4X333 defines a chipset rather than a chip. According to ..
http://www.viatech.com/en/apollo/p4x333.jsp and
http://www.viatech.com/en/apollo/p4x400.jsp , the northbridge is a VT8754 in both models, so the correct define would seem to be PCI_DEVICE_ID_VIA_8754

Just to confirm, device 3168 is the host bridge in lspci output right? And this does all work when you run a DRI application ?