From: roland <for_spam@gm...> - 2004-03-03 23:19:30
no - this doesnt help fixing the "hang at NET:..." bug.
but the mail from jonathan boler gave the essential hint and i tried further.
conclusion: this bug indeed seems to be related to cpu frequency!
so - what did i do ?
i did an
"echo 0%40%50%performance >/proc/cpufreq"
to reduce the speed of my 2.6 ghz p4 (40% minimum, 50% maximum)
THE UML STARTS FINE AND THE HANG AT IPV4 SETUP IS GONE!
so - this is the workaround.
now we need a patch ;)
do we need the so called "rtc fix" ?
does it exist already for 2.6 kernel ?
----- Original Message -----
From: "BlaisorBlade" <blaisorblade_spam@...>
Sent: Tuesday, March 02, 2004 8:05 PM
Subject: Re: [uml-user] Hang at ipv4 setup
Alle 11:07, lunedì 1 marzo 2004, Jonathan Boler ha scritto:
> I was just wondering if there has there been any progress in fixing the
> hang at NET: Registered protocol family 2 bug ?
> Is there a patch available ?
Yes. At least, it should work (that is a fix for what we think is another
symptom of the your bug). Get these two patches and apply them with patch -l
-p1 (it is a little mangled for spaces by the mailer). So, please report even
if you get success. If you have any problem, try using only the first one.
Paolo Giarrusso, aka Blaisorblade
Linux registered user n. 292729
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
User-mode-linux-user mailing list
From: Jeff Dike <jdike@ad...> - 2004-03-04 02:49:34
> i did an "modprobe p4-clockmod" and then "echo 0%40%50%performance >/
> proc/cpufreq" to reduce the speed of my 2.6 ghz p4 (40% minimum, 50%
> THE UML STARTS FINE AND THE HANG AT IPV4 SETUP IS GONE!
> so - this is the workaround.
> now we need a patch ;)
> do we need the so called "rtc fix" ? does it exist already for 2.6
> kernel ?
2.4 and 2.6 are identical as far as the rtc stuff goes. So, there's something
else wrong. I'm wondering whether this is somehow related to the rdtsc
problem that came up recently.
From: Christopher S. Aker <caker@th...> - 2004-03-04 04:56:37
Here is my original post with patch.
Some of it is already in the 2.6um tree. Just the very first patch in the patch to
time.c is necessary. The lines are offset from 2.6 in this 2.4 patch, but it
should apply to 2.6um easy enough...
diff -Naur orig/arch/um/kernel/time.c mine/arch/um/kernel/time.c
--- orig/arch/um/kernel/time.c 2003-12-18 02:30:20.000000000 -0500
+++ mine/arch/um/kernel/time.c 2003-12-18 02:32:38.000000000 -0500
@@ -79,10 +79,11 @@
-static long get_host_hz(void)
+static unsigned long long get_host_hz(void)
char mhzline, *end;
- int ret, mult, mhz, rest, len;
+ int ret, mult, rest, len;
+ unsigned long long mhz;
ret = cpu_feature("cpu MHz", mhzline,
sizeof(mhzline) / sizeof(mhzline));