[I do not subscribe to debian-mips; please include debian-x when replying.]
Ryan Murray reported to me that a 4.1 compile on mips fails.
<neuro> Overfiend: make[8]: Entering directory `/home/rmurray/build/xfree86-4.1.0/build-tree/xc/programs/Xserver/hw/xfree86/drivers/newport'
<neuro> Overfiend: newport_driver.c:250: macro `RANGE' used with only 4 args
The problem here is that a patch from David S. Miller (see
debian/patches/100* in the source package) added a 5th argument to the
RANGE() macro. This fifth argument is a PCI domain.
The definition of the macro is in
xc/programs/Xserver/hw/xfree86/common/xf86str.h.
Ryan also says that machines with newport cards won't have PCI devices. I
have the shallowest possible understanding of this code so I am not sure
what happens on machines that use this macro with no PCI bus. Perhaps
whatever bus happens to be available is assigned bus 0. The existing code
in the newport driver describes what it is doing as a "hack".
244 /* This is a hack because don't have the RAC info(and don't want it).
245 * Set it as an ISA entity to get the entity field set up right.
246 */
247 entity = xf86ClaimIsaSlot(drv, 0, dev, TRUE);
248 base = (NEWPORT_BASE_ADDR0 + busID * NEWPORT_BASE_OFFSET);
249 RANGE(range[0], base, base + sizeof(NewportRegs),\
250 ResExcMemBlock);
251 pScrn = xf86ConfigIsaEntity(pScrn, 0, entity, NULL, range, \
252 NULL, NULL, NULL, NULL);
I do not have access, let alone physical access, to any mips boxen so I am
going to rely on one of Debian's MIPS porters to resolve this problem. I'm
CC'ing David S. Miller as well, in case he has any insights he'd like to
share.
You can get the last 4.1.0 prerelease .debs from the URL in my .signature.
Thanks for your assistance.
--
G. Branden Robinson | No math genius, eh? Then perhaps you
Debian GNU/Linux | could explain to me where you got
branden@debian.org | these... PENROSE TILES!
http://people.debian.org/~branden/ | -- Stephen R. Notley
Attachment:
pgpXCbe6djQRG.pgp
Description: PGP signature