Re: [Dri-devel] Two Linux framebuffer patches, Radeon and Rage128

2003-08-14 Thread David Dawes
On Sun, Aug 10, 2003 at 08:57:43PM -0700, Jon Smirl wrote: As far as I know linux/pci_ids.h is not being generated, it is hand built. That probably explains why it had about 15 entries in it that were wrong for the Radeon/Rage128. Same is true for xf86PciInfo.h (being hand generated), although

Re: [Dri-devel] Two Linux framebuffer patches, Radeon and Rage128

2003-08-14 Thread Jon Smirl
Update of Rage128 and Radeon PCI IDs patch. Adds a few more Radeon IDs from XFree code. Removes switch in computing name and uses table instead. In the kernel DEVICE_NAME_SIZE is currently defined as 50 but most of the pretty names are around 100 chars in length. Any objections to raising

Re: [Dri-devel] Two Linux framebuffer patches, Radeon and Rage128

2003-08-14 Thread Jon Smirl
As far as I know linux/pci_ids.h is not being generated, it is hand built. That probably explains why it had about 15 entries in it that were wrong for the Radeon/Rage128. My understanding is that drivers/pci/pci.ids is semi-hand built. It is partially generated from http://pciids.sf.net/ and

Re: [Dri-devel] Two Linux framebuffer patches, Radeon and Rage128

2003-08-11 Thread Ian Romanick
Jon Smirl wrote: These add all known Radeon and Rage128 PCI IDs to their respective framebuffer drivers. It also updates linux/pci_ids.h with these IDs. Both drivers will display pretty_name if CONFIG_PCI_NAMES is enabled, otherwise a name is generated. Please check over the chip family

[Dri-devel] Two Linux framebuffer patches, Radeon and Rage128

2003-08-10 Thread Jon Smirl
These add all known Radeon and Rage128 PCI IDs to their respective framebuffer drivers. It also updates linux/pci_ids.h with these IDs. Both drivers will display pretty_name if CONFIG_PCI_NAMES is enabled, otherwise a name is generated. Please check over the chip family definitions in the