Re: A patch for linux 2.1.127

Re: A patch for linux 2.1.127

[Steven N. Hirsch]

On Sun, 8 Nov 1998, H.J. Lu wrote:

> Hi,
> 
> Here is a patch for Linux 2.1.127. Now I am running Linux 2.1.127
> compiled with egcs 1.1.1 prerelease using -O6.
> 

HJ,

You will note that the schedule() changes completely broke lockd, and that
your linux-2.1.1xx.diff patch no longer results in a working RPC system.

Tried to E-Mail you, but your mailer bounces my mail.

Steve

Re: A patch for linux 2.1.127

[Michael Harnois]

HJ's patch works here, with egcs-1.1a. The no-boot problem with
optimizations greater than -O2 is gone.

--
Michael D. Harnois, Redeemer Lutheran Church, Washburn, IA 
mharnois@willinet.net                      aa0bt@aa0bt.ampr.org 
 If they can get you asking the wrong questions, 
 they don't have to worry about the answers.
      -- Thomas Pynchon

A patch for linux 2.1.127

[H.J. Lu]

Hi,

Here is a patch for Linux 2.1.127. Now I am running Linux 2.1.127
compiled with egcs 1.1.1 prerelease using -O6.

I don't know whom to blame. I have to say both. As you can see,
calibrate_tsc is just asm code written in the asm statement. It
does't make much sense to let compiler choose registers in that
case since most of registers are used in the asm statement and
there are only a few free registers left. Why not we just finish
the register allocation ourselves. The C compiler cannot do any
much better than it.

The problem starts when egcs 1.1.1 inlines calibrate_tsc. For

	: "r" (5 * 1000020/HZ)

It puts 5 * 1000020/HZ in eax. It is overwritten by

	"inb  $0x61, %%al\n\t" /* Read port */

If egcs does a better job, it should know eax will be clobbered in
the asm statement and choose an unclobbered regiister instead. The
interesting thing is it only happens when calibrate_tsc is inlined.
Maybe the constraints in the asm statement are not 100% correct.

Anyway, I am enclosing the relevant part here. I hope someone can
give a better answer than mine.

Thanks.


-- 
H.J. Lu (hjl@gnu.org)
----
Index: arch/i386/kernel/time.c
===================================================================
RCS file: /home/work/cvs/linux/linux/arch/i386/kernel/time.c,v
retrieving revision 1.1.1.14
diff -u -p -r1.1.1.14 time.c
--- time.c	1998/10/24 13:53:40	1.1.1.14
+++ time.c	1998/11/08 19:14:04
@@ -566,8 +566,8 @@ __initfunc(static unsigned long calibrat
                "divl %%ecx\n\t" /* eax= 2^32 / (1 * TSC counts per microsecond) */
 	       /* Return eax for the use of fast_gettimeoffset */
                "movl %%eax, %0\n\t"
-               : "=r" (retval)
-               : "r" (5 * 1000020/HZ)
+               : "=S" (retval)
+               : "D" (5 * 1000020/HZ)
                : /* we clobber: */ "ax", "bx", "cx", "dx", "cc", "memory");
        return retval;
 }
----
__initfunc(static unsigned long calibrate_tsc(void))
{
       unsigned long retval;

       __asm__( /* set the Gate high, program CTC channel 2 for mode 0
		 * (interrupt on terminal count mode), binary count, 
		 * load 5 * LATCH count, (LSB and MSB)
		 * to begin countdown, read the TSC and busy wait.
		 * BTW LATCH is calculated in timex.h from the HZ value
		 */

	       /* Set the Gate high, disable speaker */
	       "inb  $0x61, %%al\n\t" /* Read port */
	       "andb $0xfd, %%al\n\t" /* Turn off speaker Data */
	       "orb  $0x01, %%al\n\t" /* Set Gate high */
	       "outb %%al, $0x61\n\t" /* Write port */

	       /* Now let's take care of CTC channel 2 */
	       "movb $0xb0, %%al\n\t" /* binary, mode 0, LSB/MSB, ch 2*/
	       "outb %%al, $0x43\n\t" /* Write to CTC command port */
	       "movb $0x0c, %%al\n\t"
	       "outb %%al, $0x42\n\t" /* LSB of count */
	       "movb $0xe9, %%al\n\t"
	       "outb %%al, $0x42\n\t" /* MSB of count */

               /* Read the TSC; counting has just started */
               "rdtsc\n\t"
               /* Move the value for safe-keeping. */
               "movl %%eax, %%ebx\n\t"
               "movl %%edx, %%ecx\n\t"

	       /* Busy wait. Only 50 ms wasted at boot time. */
               "0: inb $0x61, %%al\n\t" /* Read Speaker Output Port */
	       "testb $0x20, %%al\n\t" /* Check CTC channel 2 output (bit 5) */
               "jz 0b\n\t"

               /* And read the TSC.  5 jiffies (50.00077ms) have elapsed. */
               "rdtsc\n\t"

               /* Great.  So far so good.  Store last TSC reading in
                * last_tsc_low (only 32 lsb bits needed) */
               "movl %%eax, last_tsc_low\n\t"
               /* And now calculate the difference between the readings. */
               "subl %%ebx, %%eax\n\t"
               "sbbl %%ecx, %%edx\n\t" /* 64-bit subtract */
	       /* but probably edx = 0 at this point (see below). */
               /* Now we have 5 * (TSC counts per jiffy) in eax.  We want
                * to calculate TSC->microsecond conversion factor. */

               /* Note that edx (high 32-bits of difference) will now be 
                * zero iff CPU clock speed is less than 85 GHz.  Moore's
                * law says that this is likely to be true for the next
                * 12 years or so.  You will have to change this code to
                * do a real 64-by-64 divide before that time's up. */
               "movl %%eax, %%ecx\n\t"
               "xorl %%eax, %%eax\n\t"
               "movl %1, %%edx\n\t"
               "divl %%ecx\n\t" /* eax= 2^32 / (1 * TSC counts per microsecond) */
	       /* Return eax for the use of fast_gettimeoffset */
               "movl %%eax, %0\n\t"
               : "=S" (retval)
               : "D" (5 * 1000020/HZ)
               : /* we clobber: */ "ax", "bx", "cx", "dx", "cc", "memory");
       return retval;
}

__initfunc(void time_init(void))
{
	xtime.tv_sec = get_cmos_time();
	xtime.tv_usec = 0;

/*
 * If we have APM enabled or the CPU clock speed is variable
 * (CPU stops clock on HLT or slows clock to save power)
 * then the TSC timestamps may diverge by up to 1 jiffy from
 * 'real time' but nothing will break.
 * The most frequent case is that the CPU is "woken" from a halt
 * state by the timer interrupt itself, so we get 0 error. In the
 * rare cases where a driver would "wake" the CPU and request a
 * timestamp, the maximum error is < 1 jiffy. But timestamps are
 * still perfectly ordered.
 * Note that the TSC counter will be reset if APM suspends
 * to disk; this won't break the kernel, though, 'cuz we're
 * smart.  See devices/char/apm_bios.c.
 */
	if (boot_cpu_data.x86_capability & X86_FEATURE_TSC) {
		do_gettimeoffset = do_fast_gettimeoffset;
		do_get_fast_time = do_gettimeofday;
		irq0.handler = pentium_timer_interrupt;
		fast_gettimeoffset_quotient = calibrate_tsc();
		
		/* report CPU clock rate in Hz.
		 * The formula is (10^6 * 2^32) / (2^32 * 1 / (clocks/us)) =
		 * clock/second. Our precision is about 100 ppm.
		 */
		{	unsigned long eax=0, edx=1000000;
			__asm__("divl %2"
	       		:"=a" (cpu_hz), "=d" (edx)
               		:"r" (fast_gettimeoffset_quotient),
                	"0" (eax), "1" (edx));
			printk("Detected %ld Hz processor.\n", cpu_hz);
		}
	}
	setup_x86_irq(0, &irq0);
}

egcs on Mipsel?

[Greg]

Hello all,
	I just got ahold of a Cobalt Qube and being the harbinger of pain
that I am, started madly re-building RedHat 5.2 .src.rpms with the
eventual intentions of getting X running on the machine. In the process, I
have run up against a road block.
	Many packages require egcs to compile. I'm not having any luck
getting egcs to compile for the mipsel architecture. In fact, it is dying
in the "configure" stage. Am I barking up the wrong tree here? Does anyone
have an egcs package or set of patches that will allow it to be compiled
on a mipsel?

[root@acct egcs-1.1b]# ./configure
Configuring for a mipsel-unknown-linux-gnu host.
Created "Makefile" in /home/redhat/BUILD/egcs-1.1b using "mt-frag"
Configuration mipsel-unknown-linux-gnu not supported
Configure in /home/redhat/BUILD/egcs-1.1b/gcc failed, exiting.

BTW: If anyone wants access to the box for testing patches, please let
     me know. I'll be happy to create an account for you...

egcs on Mipsel?

[Gavin Romig-Koch]

Greg writes:
 > 	I just got ahold of a Cobalt Qube and being the harbinger of pain
 > that I am, started madly re-building RedHat 5.2 .src.rpms with the
 > eventual intentions of getting X running on the machine. In the process, I
 > have run up against a road block.
 > ...
 >
 > [root@acct egcs-1.1b]# ./configure
 > Configuring for a mipsel-unknown-linux-gnu host.
 > Created "Makefile" in /home/redhat/BUILD/egcs-1.1b using "mt-frag"
 > Configuration mipsel-unknown-linux-gnu not supported
 > Configure in /home/redhat/BUILD/egcs-1.1b/gcc failed, exiting.
 > 
 > BTW: If anyone wants access to the box for testing patches, please let
 >      me know. I'll be happy to create an account for you...

Assuming no one more expert on this has answered you already:

Looking through egcs/gcc/configure, it looks like egcs hasn't
been ported to this target (mipsel-unknown-linux-gnu) yet.  
It has been ported to both other mipsel targets, and other
linux-gnu targets.  

Depending on how adventurous you are you could, try fixing this
yourself, contact others who might have already fixed this (like
Cobalt, or the linux-mips people) in some other version of gcc,
or wait for someone else to do it.

I've done some other mips mini-ports, I'm not really an expert
on it, and I don't have lots of spare time, but if you want to
try fixing this yourself, I help if I can.


                                               -gavin...

Re: EGCS on Mipsel?

[Greg]

On Tue, 1 Dec 1998, Robert Lipe wrote:

> > 	I'm heading up a development project to port RedHat 5.2 to the
> > Cobalt Qube/RAQ servers. These run the Mipsel architecture. We're about
> > 50% through the project, but a significant amount of the SRPMS are
> > requiring egcs to build. We do NOT have a working egcs RPM at this time.
> >
> > Has anyone been successful at getting egcs to compile and work on a Cobalt
> > Qube/RAQ product?
> 
> I've never seen a Linux/MIPS system, but I've used the GNU tools on
> embedded mipsel targets for about 6 years.  For a long time, I suspected
> I was the *only* user of little endian MIPS targets. :-) On the premise
> that you must have had some sort of GCC-like substance ported to this
> system in order to get this far, it should be "merely" a matter of
> backstitching the configuration from that version stuff into the current
> version and getting it committed.
> 
> If you can grant clock cycles, disk space, and remote telnet/rlogin/ssh
> access to an appropriate system (esp. if CVS access to the EGCS tree is
> available) I can try to help you make it go.
> 
> Privately, we can work out the details like *which* mips processor is in
> use, last version of EGCS/GCC known to work, where your current compiler
> came from, etc...

An account on the machine will be forthcoming in a private E-mail. :)

As for the information, the machine is a Cobalt Qube designed by Coablt
Microserver http://www.cobaltmicro.com . The CPU identifies itself as;

[root@acct /proc]# more cpuinfo
cpu                     : MIPS
cpu model               : Nevada V1.0
system type             : Cobalt Microserver 27
BogoMIPS                : 149.50
byteorder               : little endian
unaligned accesses      : 6977910
wait instruction        : yes
microsecond timers      : yes
extra interrupt vector  : yes
hardware watchpoint     : no

Here's a list of RPM packages that are installed;

setup-1.7-3C7
filesystem-1.3.1-2
basesystem-4.9-1
glibc-2.0.4-C4
MAKEDEV-2.2-11
SysVinit-2.71-3
termcap-9.12.6-6
libtermcap-2.0.8-6
grep-2.0-6
bash-1.14.7-2
fileutils-3.16-3
info-3.9-5
textutils-1.22-2
CGI.pm-2.36-1
GD-1.15-1
GIFgraph-0.90-4
apache-admin-1.2.5-7
apache-1.2.5-6
bdflush-1.5-6
bind-4.9.5p1-3
bind-utils-4.9.5p1-3
binutils-2.8.1-1C1
bison-1.25-2
bootpc-061-3
cgiwrap-1.0-1
cpio-2.4.2-5
cracklib-2.5-2
cracklib-dicts-2.5-2
crontabs-1.5-2C1
dev-2.5.2-1
dhcpcd-0.65-1
dhcpd-5.16-4C1
diffutils-2.7-7
discuss-1.0-17
docorg-1.1-3
libtiff-3.4-4
file-3.22-6
findutils-4.1-13
finger-0.10-2
ftp-0.10-1
gawk-3.0.2-2
gcc-2.7.2-C1
gcc-c++-2.7.2-C1
gcc-objc-2.7.2-C1
getty_ps-2.0.7j-2
glibc-debug-2.0.4-C4
glibc-devel-2.0.4-C4
glibc-profile-2.0.4-C4
glimpse-4.0-4
gzip-1.2.4-9
hdparm-3.1-3
imap-4.1.BETA-6C1
initscripts-cobalt-1.0-4
kernel-2.0.33-C12
kernel-headers-2.0.33-C12
less-321-4
libtermcap-devel-2.0.8-6
logrotate-2.3-4
m4-1.4-7
mailx-5.5.kw-7
gdbm-1.7.3-17
mktemp-1.4-2
mount-2.7f-1C1
ncurses-1.9.9e-5C1
ncurses-devel-1.9.9e-5C1
net-tools-1.33-2
netatalk-1.4b2-1C5
netkit-base-0.10-2
nfs-server-2.2beta16-8
nfs-server-clients-2.2beta16-8
pam-0.57-3
panel-utils-1.0-21
rpm-devel-2.5.5-5.2
ircii-4.4-4
Net-Ping-2.02-1
portmap-4.0-4
procmail-3.10-11
procps-1.2.2-2
psmisc-16-1
quota-1.55-5C1
cobalt-release-1.0-1
flex-2.5.4a-3
samba-1.9.18p2-C6
sed-2.05-7
qpopper-2.3-1C1
sendmail-8.8.7-11
sendmail-cf-8.8.7-10
sendmail-doc-8.8.7-10
sh-utils-1.16-6C1
special-sauce-2.5-2
strace-3.1-4
sysklogd-1.3-16
tar-1.11.8-13C1
tcp_wrappers-7.5-2
telnet-0.10-1
texinfo-3.9-5
time-1.7-2
tmpwatch-1.2-2
ui-2.3-31
upgrade-1.0-7
util-linux-2.7-5C2
vim-4.5-3
vixie-cron-3.0.1-15
which-1.0-6
wu-ftpd-2.4.2b15-2
web-stat-1.0-4
gpm-devel-1.13-6
libpng-devel-1.0.1-5
libpng-1.0.1-5
fetchmail-4.4.4-2
zlib-1.1.2-2
zlib-devel-1.1.2-2
slang-0.99.38-4
slang-devel-0.99.38-4
gpm-1.13-6
ssh-1.2.25-1us
ssh-clients-1.2.25-1us
ssh-server-1.2.25-1us
ssh-extras-1.2.25-1us
patch-2.5-2
pine-3.96-8
ncftp-2.4.3-3
make-3.76.1-3
pwdb-0.54-8
gettext-0.10.35-3
popt-1.1.1-2
gdbm-devel-1.7.3-17
rpm-2.5.5-5.2
libelf-0.6.4-2
shadow-utils-980403-4
e2fsprogs-1.12-2
e2fsprogs-devel-1.12-2
libtiff-devel-3.4-4
perl-5.004_04-1
screen-3.7.4-3
autoconf-2.12-6
automake-1.3-2
byacc-1.9-8
readline-2.2.1-1
readline-devel-2.2.1-1
unzip-5.31-3
jed-0.98.4-6

Re: EGCS on Mipsel?

[Robert Lipe]

> 	I'm heading up a development project to port RedHat 5.2 to the
> Cobalt Qube/RAQ servers. These run the Mipsel architecture. We're about
> 50% through the project, but a significant amount of the SRPMS are
> requiring egcs to build. We do NOT have a working egcs RPM at this time.
>
> Has anyone been successful at getting egcs to compile and work on a Cobalt
> Qube/RAQ product?

I've never seen a Linux/MIPS system, but I've used the GNU tools on
embedded mipsel targets for about 6 years.  For a long time, I suspected
I was the *only* user of little endian MIPS targets. :-) On the premise
that you must have had some sort of GCC-like substance ported to this
system in order to get this far, it should be "merely" a matter of
backstitching the configuration from that version stuff into the current
version and getting it committed.

If you can grant clock cycles, disk space, and remote telnet/rlogin/ssh
access to an appropriate system (esp. if CVS access to the EGCS tree is
available) I can try to help you make it go.

Privately, we can work out the details like *which* mips processor is in
use, last version of EGCS/GCC known to work, where your current compiler
came from, etc...

RJL

EGCS on Mipsel?

[Greg]

Hello all,
	I'm heading up a development project to port RedHat 5.2 to the
Cobalt Qube/RAQ servers. These run the Mipsel architecture. We're about
50% through the project, but a significant amount of the SRPMS are
requiring egcs to build. We do NOT have a working egcs RPM at this time.

Has anyone been successful at getting egcs to compile and work on a Cobalt
Qube/RAQ product?

See message below..

From damin@intel.cleveland.lug.net Mon Nov 30 12:57:18 1998
Date: Mon, 30 Nov 1998 12:57:17 -0500 (EST)
From: Greg <damin@intel.cleveland.lug.net>
To: linux-mips@fnet.fr
Subject: RedHat 5.2 on Cobalt Qube/RAQ
X-Status: 

Howdy all,
	The Cleveland Linux Users's Group is exactly 50% finished with our
rebuild of RedHat 5.2 RPMS on the Cobalt Qube / RAQ platform. We are
significantly ahead of schedule with our efforts, which can be accessed
at;

ftp://intel.cleveland.lug.net/pub/Mipsel

	Of the 432 SRPMS for RedHat 5.2, we have successfully rebuilt 216
of them. This equates to some 280 binary RPMS that are available for your
pleasure on our FTP archive. It would be helpful if someone would consider
mirroring our archive in the event our machine should become unavailable.
	Still don't have a working EGCS RPM, and many of the SRPMS listed
below are Intel specific and most likely will require extensive porting
(if even needed) to work on the Mipsel platform. The usefulness of some of
them is without a doubt in question, but our object is to make the job
easier for RedHat and other Mipsel users to get and install current
software... So what the hell.. maybe there IS some value, somewhere, for
Zgv running on a Cobalt RAQ! ;)
	In any case, here is a list of our Phase 2 SRPMS that will require
some tweaking to build. Once we have worked through these for the next
couple of weeks, we'll re-build a majority of the packages to include more
current compiler tools etc..

Stuff that isn't done yet..

AfterStep-1.5-0.7.src.rpm
AfterStep-APPS-1.5-0.3.src.rpm
AnotherLevel-0.7.3-1.src.rpm
ElectricFence-2.0.5-11.src.rpm
ImageMagick-4.1.0-1.src.rpm
MAKEDEV-2.3.1-7.src.rpm
SVGATextMode-1.8-2.src.rpm
SysVinit-2.74-5.src.rpm
WindowMaker-0.20.1-3.src.rpm
X11R6-contrib-3.3.2-3.src.rpm
XFree86-3.3.2.3-25.src.rpm
XFree86-ISO8859-2-1.0-1.src.rpm
XFree86-ISO8859-9-2.1.2-1.src.rpm
Xaw3d-1.3-17.src.rpm
Xconfigurator-3.82-1.src.rpm
aboot-0.5-7.src.rpm
acm-4.7-8.src.rpm
anonftp-2.6-1.src.rpm
aout-libs-1.4-9.src.rpm
awesfx-0.4.2-1.src.rpm
bin86-0.4-5.src.rpm
blt-2.4f-3.src.rpm
bzip2-0.9.0b-2.src.rpm
chkconfig-0.9.6-1.src.rpm
clock-1.1-4.src.rpm
cmu-snmp-3.5-3.src.rpm
control-panel-3.7-7.src.rpm
cxhextris-1.0-12.src.rpm
dip-3.3.7o-12.src.rpm
dosemu-0.98.1-2.src.rpm
ee-0.3-8.src.rpm
egcs-1.0.3a-14.src.rpm
elftoaout-2.2-1.src.rpm
elm-2.4.25-14.src.rpm
emacs-20.3-3.src.rpm
exmh-2.0.2-4.src.rpm
faces-1.6.1-11.src.rpm
fdisk-2.5-1.src.rpm
flying-6.20-9.src.rpm
fvwm-1.24r-15.src.rpm
fvwm2-2.0.46-11.src.rpm
gcc-2.7.2.3-14.src.rpm
gd-1.3-3.src.rpm
gdb-4.17.0.4-3.src.rpm
gecko-1.6-0.src.rpm
ghostscript-4.03-1.src.rpm
gimp-1.0.1-2.src.rpm
gimp-manual-1.0.0-1.src.rpm
glibc-2.0.7-29.src.rpm
glint-2.6.2-1.src.rpm
gnome-core-0.20.1-2.src.rpm
gnome-libs-0.20-3.src.rpm
gnome-linuxconf-0.14-4rh.src.rpm
gnuplot-3.6-beta347.src.rpm
gperf-2.7-2.src.rpm
groff-1.11a-6.src.rpm
guavac-1.2-2.src.rpm
gv-3.5.8-5.src.rpm
howto-5.2-2.src.rpm
ical-2.2-7.src.rpm
imap-4.4-2.src.rpm
inn-1.7.2-14.src.rpm
isapnptools-1.15a-3.src.rpm
ispell-3.1.20-11.src.rpm
kaffe-1.0.b2-2.src.rpm
kbdconfig-1.8.3-1.src.rpm
kernel-2.0.36-0.7.src.rpm
kernel-alpha-2.0.35-5.src.rpm
kernel-sparc-2.0.35-11.src.rpm
kernelcfg-0.5-3.src.rpm
kterm-6.2.0-3.src.rpm
ld.so-1.9.5-8.src.rpm
libc-5.3.12-27.src.rpm
libg++-2.7.2.8-9.src.rpm
lilo-0.20-2.src.rpm
linuxconf-1.12r5-6rh.src.rpm
lsof-4.37-1.src.rpm
ltrace-0.3.4-1.src.rpm
lynx-2.8.1-5.src.rpm
man-1.5f-1.src.rpm
maplay-1.2-8.src.rpm
mars-nwe-0.99pl10-1.src.rpm
mc-4.1.35-2.src.rpm
mgetty-1.1.14-5.src.rpm
minlabel-1.2-3.src.rpm
mkbootdisk-1.1-2.src.rpm
mkinitrd-1.8-3.src.rpm
mod_perl-1.15-3.src.rpm
mod_php-2.0.1-5.src.rpm
mod_php3-3.0.5-2.src.rpm
modutils-2.1.85-9.src.rpm
moonclock-1.0-13.src.rpm
mouseconfig-3.1.3-1.src.rpm
mtools-3.9.1-2.src.rpm
multimedia-2.1-12.src.rpm
mxp-1.0-9.src.rpm
ncompress-4.2.4-11.src.rpm
ncpfs-2.2.0-1.src.rpm
ncurses-4.2-10.src.rpm
ncurses3-1.9.9e-7.src.rpm
netcfg-2.19-5.src.rpm
netscape-4.07-1.src.rpm
newt-0.30-2.src.rpm
nfs-server-2.2beta37-1.src.rpm
pam-0.64-3.src.rpm
passwd-0.50-11.src.rpm
perl-5.004m4-1.src.rpm
perl-MD5-1.7-3.src.rpm
pine-4.04-2.src.rpm
playmidi-2.4-3.src.rpm
pmake-1.0-10.src.rpm
popt-1.1.1-2.src.rpm
portmap-4.0-12.src.rpm
postgresql-6.3.2-10.src.rpm
ppp-2.3.5-1.src.rpm
printtool-3.29-3.src.rpm
procinfo-14-2.src.rpm
procmail-3.10-13.src.rpm
procps-1.2.9-2.src.rpm
psacct-6.3-5.src.rpm
psmisc-17-3.src.rpm
pwdb-0.55-1.src.rpm
python-1.5.1-5.src.rpm
pythonlib-1.22-2.src.rpm
quickstrip-1.1-4.src.rpm
quota-1.55-9.src.rpm
raidtools-0.50beta10-2.src.rpm
rcs-5.7-7.src.rpm
rdate-0.960923-5.src.rpm
rdist-1.0-9.src.rpm
redhat-release-5.2-1.src.rpm
rhbackup-0.2-1.src.rpm
rhl-alpha-install-addend-en-5.2-1.src.rpm
rhl-install-guide-en-5.2-1.src.rpm
rhmask-1.0-3.src.rpm
rhs-hwdiag-0.30-1.src.rpm
rhs-printfilters-1.46-3.src.rpm
rhsound-1.7-3.src.rpm
rootfiles-5.2-2.src.rpm
routed-0.10-10.src.rpm
rsh-0.10-4.src.rpm
rsync-2.1.1-1.src.rpm
rusers-0.10-8.src.rpm
rwall-0.10-7.src.rpm
rwho-0.10-8.src.rpm
rxvt-2.4.7-2.src.rpm
sash-2.1-1.src.rpm
setserial-2.14-4.src.rpm
setuptool-1.0-1.src.rpm
sgml-tools-1.0.7-1.src.rpm
sh-utils-1.16-14.src.rpm
shapecfg-2.0.36-1.src.rpm
sharutils-4.2-10.src.rpm
silo-0.8-2.src.rpm
smbfs-2.0.1-4.src.rpm
sndconfig-0.27-1.src.rpm
spice-2g6-9.src.rpm
spider-1.0-7.src.rpm
strace-3.1-11.src.rpm
svgalib-1.3.0-2.src.rpm
tcltk-8.0.3-20.src.rpm
tcsh-6.07.09-1.src.rpm
telnet-0.10-5.src.rpm
termfiles_sparc-1.1-1.src.rpm
tetex-0.9-6.src.rpm
timeconfig-2.5-1.src.rpm
transfig-3.2.1-1.src.rpm
usermode-1.4.3-1.src.rpm
usernet-1.0.8-1.src.rpm
util-linux-2.8-11.src.rpm
uucp-1.06.1-16.src.rpm
vga_cardgames-1.3.1-9.src.rpm
vga_gamespack-1.3-9.src.rpm
wmakerconf-1.1.1-3.src.rpm
wu-ftpd-2.4.2b18-2.src.rpm
x3270-3.1.1.6-3.src.rpm
xanim-27070-1.src.rpm
xbanner-1.31-3.src.rpm
xbill-2.0-4.src.rpm
xbl-1.0h-3.src.rpm
xboing-2.4-5.src.rpm
xcpustate-2.5-3.src.rpm
xdaliclock-2.10-3.src.rpm
xdemineur-1.1-9.src.rpm
xearth-1.0-10.src.rpm
xevil-1.5-7.src.rpm
xfig-3.2.2-1.src.rpm
xfishtank-2.0-10.src.rpm
xfm-1.3.2-10.src.rpm
xgalaga-1.6c-8.src.rpm
xgammon-0.98-12.src.rpm
xgopher-1.3.3-6.src.rpm
xjewel-1.6-9.src.rpm
xlander-1.2-9.src.rpm
xlispstat-3.52.5-1.src.rpm
xloadimage-4.1-10.src.rpm
xmailbox-2.5-5.src.rpm
xmorph-1996.07.12-5.src.rpm
xosview-1.6.1.a-4.src.rpm
xpaint-2.4.9-6.src.rpm
xpat2-1.04-8.src.rpm
xpdf-0.7a-1.src.rpm
xpilot-3.6.2-4.src.rpm
xpm-3.4j-3.src.rpm
xpuzzles-5.4.1-4.src.rpm
xrn-9.01-1.src.rpm
xsnow-1.40-7.src.rpm
xsysinfo-1.6-3.src.rpm
xterm-color-1.1-9.src.rpm
xtoolwait-1.1-3.src.rpm
xtrojka-1.2.3-4.src.rpm
xv-3.10a-13.src.rpm
xwpick-2.20-9.src.rpm
xxgdb-1.12-6.src.rpm
ytalk-3.0.3-6.src.rpm
zgv-3.0-5.src.rpm

egcs on mipsel?

[Greg]

In case my last message didn't get through...

I'm not having luck getting egcs-1.1b to build on a Cobalt Qube, which
uses a Little Endian MIPS processor. Is this an impossible task? Or is it
something that is in the works? Or has anyone done it?

I'd like to see patches, rpms, or comments on this.

Or.. is this just a dead end trying to get this to work?