Re: xsrc/50429: Xserver fails to start on Ultra-5 andNetBSD/sparc64 7.0

[Michael <macallan%netbsd.org@localhost>]

The following reply was made to PR xsrc/50429; it has been noted by GNATS.

From: Michael <macallan%netbsd.org@localhost>
To: gnats-bugs%NetBSD.org@localhost
Cc: 
Subject: Re: xsrc/50429: Xserver fails to start on Ultra-5 andNetBSD/sparc64
  7.0
Date: Fri, 20 Nov 2015 01:16:32 -0500

 On my U60 I get this, turns out I was wrong about the Rage II not
 having the MMIO aperture:
 
 machfb0 at pci0 dev 5 function 0: vendor 1002 product 4754 (rev. 0x9a)
 machfb0: using MMIO aperture
 machfb0: 16 MB aperture at 0x02000000, 4 KB registers at 0x03000000
 machfb0: 128 KB ROM at 0x02020000
 machfb0: 2048 KB SGRAM 62.999 MHz, maximum RAMDAC clock 170 MHz
 machfb0: initializing the DSP
 machfb0: initial resolution 1440x900 at 8 bpp
 wsdisplay0 at machfb0 kbdmux 1: console (default, vt100 emulation)
 
     Base address register at 0x10
       type: 32-bit nonprefetchable memory
       base: 0x02000000, not sized
     Base address register at 0x14
       type: i/o
       base: 0x00000000, not sized
     Base address register at 0x18
       type: 32-bit nonprefetchable memory
       base: 0x03000000, not sized
 
 [  1169.766] (II) MACH64(0): Using 8 MB linear aperture at 0x02800000.
 [  1169.766] (!!) MACH64(0): Virtual resolutions will be limited to 2047 kB
  due to linear aperture size and/or placement of hardware cursor image area.
 [  1169.767] (II) MACH64(0): Using Block 0 MMIO aperture at 0x03000400.
 [  1169.767] (II) MACH64(0): Using Block 1 MMIO aperture at 0x03000000.
 [  1169.770] (II) MACH64(0): MMIO write caching enabled.
 [  1169.770] (--) MACH64(0): 2048 kB of SGRAM (1:1) detected (using 2047 kB).
 
 Looking through the driver source, in your case it should see a NULL in
 the 0x18 BAR, therefore ignore it, and try to mmap() the MMIO registers
 through the first aperture, at the last page of the first 8MB. I can't
 really see a reason why that would fail.
 Please sprinkle xf86Msg() over atividmem.c::ATIMapApertures() to see
 what exactly it's trying to map - apparently the first two calls to
 pci_device_map_range() succeed and only the last one, for &pATI->pMMIO,
 fails.