Subject: Re: bankeddev for Retina?
To: David C Greene <DCG9367@tntech.edu>
From: None <email@example.com>
Date: 09/21/1994 23:03:08
> I wanted to try running MW's X11R6 server for the Retina, but
> have to recompile the kernel for BANKEDDEVPAGER support.
> I did this and the compile stopped because it could not find
I don't know whether the functionality is still in the 1.0 kernel
(given that they probably changed the vm-system since 0.9, the
necessary code is probably gone in vm/*[ch]. There is a bankeddev.h in
incoming on ftp.eunet.ch, if you really want to try it.
I'd like to clearly warn people with Z2 Retinas: this server will run
*dog* slow, I don't have any code in it to reduce trashing caused by
excess bank switching like I did in the R5 server. For banked chipsets,
you really should go for an X-server that does bank-switching in user-mode,
not my (sun-based) server that assumes a contigous linearly mapped
If the picasso grf driver should get ready in time, the server should run
on that board. Note that there is a critical error in the (at least 0.9)
grfioctl.h file: GRFIOCBITBLT should be:
#define GRFIOCBITBLT _IOW('G', 57, struct grf_bitblt)
it was defined as _IOR before, which caused the kernel not to copy the argument
into kernel space, and thus the functions dealing with it would have had to
deal with completely random data... They did not, because the second bug,
in grf.c, prevented any bitblt-handler to be called at all...:
+ case GRFIOCBITBLT:
return grfdev[gp->g_type].gd_mode (gp, GM_GRFIOCTL, cmd, data);
/* check to see whether it's a command recognized by the
view code if the unit is 0 XXX */
1.0 maintainers: please check that these bugs are gone there too...
EUnet Switzerland Tel: +41 1 291 45 80 Markus Wild
Zweierstrasse 35 Hotline: +41 1 291 45 60 firstname.lastname@example.org
CH-8004 Zuerich Fax: +41 1 291 46 42 S=mw;P=EUnet;A=EUnet;C=CH
>Solaris 2 is not an upgrade from Solaris 1. They just want you to THINK it is.