Hi, Just now as I was typing an email in Sylpheed, my system froze up, became entirely unresponsive: and I had to employ the reset button (which of course would have been difficult remotely). This is the second time it's happened, and I'm alarmed, because this is exactly what I'm trying to avoid by converting to Linux. I would like some ideas on how to troubleshoot this problem. I don't really trust Sylpheed, as it has numerous quirks to it, but I'm told an email client could not likely cause a complete system failure. So I don't know what to suspect. Have I got some bad hardware? I watched a DVD last night, which may be related, I don't know. Here are my specs: Basics: Iwill KK266plus M/B running 266 FSB AMD Thunderbird 1.4GHz BIOS version is (Award) v6.00PG, dated 2001 southbridge is labeled "VIA VT82C686B" (the infamous "686" bug: is this the problem?) north bridge is covered by an IWill heat sync, but I assume it is the VI half-gig' of PC133 cas2 SDRAM Other hardware includes: ASUS v7700 GTS 32-meg GeForce2 3 Western Digital HDs, and an I/O Magic MagicDVD 8x (secondary slave) SoundBlaster Live! Platinum (with LiveDrive) My OS is: SuSE Linux 7.1 Pro kernel 2.4.16 (built from source) Elightenment 0.16.5 If Linux is really as stable is it's cracked up to be, then this is probably a hardware bug. Any suggestions? And if it is the motherboard, what can I do about that? I have been unsuccessful getting IWill to cooperate with me concerning other problems I've had. Stumped, --Jason V. C.
On Monday 14 January 2002 20.19, Jason A.Van Cleve wrote:
If Linux is really as stable is it's cracked up to be, then this is probably a hardware bug.
99.999% of all lockups I've seen or heard about have been related to one of three things: X (by far the majority), over heating and faulty memory. //Anders
Jason A.Van Cleve wrote:
Hi,
Just now as I was typing an email in Sylpheed, my system froze up, became entirely unresponsive: and I had to employ the reset button (which of course would have been difficult remotely). This is the second time it's happened, and I'm alarmed, because this is exactly what I'm trying to avoid by converting to Linux.
I would like some ideas on how to troubleshoot this problem. I don't really trust Sylpheed, as it has numerous quirks to it, but I'm told an email client could not likely cause a complete system failure. So I don't know what to suspect. Have I got some bad hardware? I watched a DVD last night, which may be related, I don't know. Here are my specs:
I'm running Sylpheed here under fvwm2, and it has been very solid. -- $|=1;while(1){print pack("h*",'75861647f302d4560275f6272797f3');sleep(1); for(1..16){for(8,32,8,7){print chr($_);}select(undef,undef,undef,.05);}}
Hi If You have had the same kind of problem with other OS, maybe it's time to check timings & other stuff in BIOS... Too fast memory timings do that kind of problems... Jaska. Viestissä Maanantai 14. Tammikuuta 2002 22:35, zentara kirjoitti:
Jason A.Van Cleve wrote:
Hi,
Just now as I was typing an email in Sylpheed, my system froze up, became entirely unresponsive: and I had to employ the reset button (which of course would have been difficult remotely). This is the second time it's happened, and I'm alarmed, because this is exactly what I'm trying to avoid by converting to Linux.
I would like some ideas on how to troubleshoot this problem. I don't really trust Sylpheed, as it has numerous quirks to it, but I'm told an email client could not likely cause a complete system failure. So I don't know what to suspect. Have I got some bad hardware? I watched a DVD last night, which may be related, I don't know. Here are my specs:
I'm running Sylpheed here under fvwm2, and it has been very solid.
Windows crashes too, of course, but never in this way, and never in the middle of a simple text operation. Would a memory timing error allow the mouse cursor to continue moving but nothing else? That was the case last time.
--Jason
On Mon, 14 Jan 2002 22:43:15 +0200, Jaakko Tamminen
Hi
If You have had the same kind of problem with other OS, maybe it's time to check timings & other stuff in BIOS... Too fast memory timings do that kind of problems...
Jaska.
Viestissä Maanantai 14. Tammikuuta 2002 22:35, zentara kirjoitti:
Jason A.Van Cleve wrote:
Hi,
Just now as I was typing an email in Sylpheed, my system froze up, became entirely unresponsive: and I had to employ the reset button (which of course would have been difficult remotely). This is the second time it's happened, and I'm alarmed, because this is exactly what I'm trying to avoid by converting to Linux.
I would like some ideas on how to troubleshoot this problem. I don't really trust Sylpheed, as it has numerous quirks to it, but I'm told an email client could not likely cause a complete system failure. So I don't know what to suspect. Have I got some bad hardware? I watched a DVD last night, which may be related, I don't know. Here are my specs:
I'm running Sylpheed here under fvwm2, and it has been very solid.
-- To unsubscribe send e-mail to suse-linux-e-unsubscribe@suse.com For additional commands send e-mail to suse-linux-e-help@suse.com Also check the FAQ at http://www.suse.com/support/faq and the archives at http://lists.suse.com
Windows crashes too, of course, but never in this way, and never in the middle of a simple text operation. Would a memory timing error allow the mouse cursor to continue moving but nothing else? That was the case last time.
If the mouse pointer continues to move then you don't have a hard crash! Sounds more like an X problem, or possibly a bug in the application which is causing the keyboard to get blocked. When it crashes, do the Caps and Num lock LEDs still work? Does Ctrl-Alt-F1 take you to a text console? If not you might need to hook up another box, either by a network or by a serial tty, in order to track down what's going wrong. -- 8:06am up 34 days, 16:15, 2 users, load average: 0.36, 0.16, 0.11
On Tue, 15 Jan 2002 08:10:32 +0000, Derek Fountain
If the mouse pointer continues to move then you don't have a hard crash!
I suppose you're right. I just meant that figuratively. My bad.
When it crashes, do the Caps and Num lock LEDs still work? Does Ctrl-Alt-F1
No, that's the strange part. The mouse cursor moves, but that was it. No mouse-clicks, no cap's lock/num' lock, no ctrl-alt-F1, no three-fingered salute. Just mouse motion. I'm having other problems with my IDE because of that VIA chipset problem (known as the "686 bug"), but that doesn't seem to fit here, does it? I'm going to try installing Xfree86 4.1.0 and see if that helps. --Jason Van Cleve
Hi, If you use the nvidia drivers (and not nv which is installed by default), I've seen these random freezes because of AGP 4x enabled in the bios, if you use the nvidia drivers (and not nv which is installed by default). Try disabling it then HTH, Matt On Tuesday 15 January 2002 02:19, Jason A.Van Cleve wrote:
Hi, [...] Other hardware includes:
ASUS v7700 GTS 32-meg GeForce2 [...]
On Tue, 15 Jan 2002 17:23:05 +0700, "Matt T."
Hi,
If you use the nvidia drivers (and not nv which is installed by default), I've seen these random freezes because of AGP 4x enabled in the bios, if you use the nvidia drivers (and not nv which is installed by default). Try disabling it then
HTH, Matt
On Tuesday 15 January 2002 02:19, Jason A.Van Cleve wrote:
Hi, [...] Other hardware includes:
ASUS v7700 GTS 32-meg GeForce2 [...]
-- To unsubscribe send e-mail to suse-linux-e-unsubscribe@suse.com For additional commands send e-mail to suse-linux-e-help@suse.com Also check the FAQ at http://www.suse.com/support/faq and the archives at http://lists.suse.com
On Tue, 15 Jan 2002 17:23:05 +0700, "Matt T."
If you use the nvidia drivers (and not nv which is installed by default), I've seen these random freezes because of AGP 4x enabled in the bios, if you
Whoa, sorry about the empty message there, I fat-fingered the send button. I do use that Nvidia driver, and I'd hate to revert back to nv, because the new driver's done wonderful things for me: E starts much faster now, for instance. Maybe Xfree86 4.1.0 will fix some things. --Jason V. C.
Hi Jason, You can keep the nvidia driver, just disable AGP4x in your bios. Your system will fall back to AGP 2x, I guess you will hardly see a difference. Since I disabled AGP4 x, I had not one single freeze anymore. HTH, Matt On Wednesday 16 January 2002 12:55, Jason A.Van Cleve wrote:
On Tue, 15 Jan 2002 17:23:05 +0700, "Matt T."
mentioned: If you use the nvidia drivers (and not nv which is installed by default), I've seen these random freezes because of AGP 4x enabled in the bios, if you
Whoa, sorry about the empty message there, I fat-fingered the send button. I do use that Nvidia driver, and I'd hate to revert back to nv, because the new driver's done wonderful things for me: E starts much faster now, for instance.
Maybe Xfree86 4.1.0 will fix some things.
--Jason V. C.
to revive this old thread, news from the gentoo folks which is, um, frightening if the machine in question is an athlon: http://features.linuxtoday.com/news_story.php3?ltsn=2002-01-21-001-20-NW-KN i'm running an athlon; fwiw, lilo in suse-7.3 chokes on the mem=nopentium option. -- dep There is sobbing of the strong, And a pall upon the land; But the People in their weeping Bare the iron hand; Beware the People weeping When they bare the iron hand.
On Mon, 21 Jan 2002, dep wrote:
to revive this old thread, news from the gentoo folks which is, um, frightening if the machine in question is an athlon:
http://features.linuxtoday.com/news_story.php3?ltsn=2002-01-21-001-20-NW-KN
i'm running an athlon; fwiw, lilo in suse-7.3 chokes on the mem=nopentium option.
Interesting. Would this bug also happen on a Cyrix M2 CPU? I been getting a crash every other week and haven't been able to track it down. Maybe if I add mem=nopentium that would help. Thanks! Christopher Reimer
I'm running a Duron 800mhz box w/ an Asus A7V mb..works fine for me. * dep (dep@drippingwithirony.com) [020121 18:45]: ->to revive this old thread, news from the gentoo folks which is, um, ->frightening if the machine in question is an athlon: -> ->http://features.linuxtoday.com/news_story.php3?ltsn=2002-01-21-001-20-NW-KN -> ->i'm running an athlon; fwiw, lilo in suse-7.3 chokes on the ->mem=nopentium option. ->-- ->dep -> ->There is sobbing of the strong, ->And a pall upon the land; ->But the People in their weeping ->Bare the iron hand; ->Beware the ->People weeping ->When they bare the iron hand. -> ->-- ->To unsubscribe send e-mail to suse-linux-e-unsubscribe@suse.com ->For additional commands send e-mail to suse-linux-e-help@suse.com ->Also check the FAQ at http://www.suse.com/support/faq and the ->archives at http://lists.suse.com -> -----=====-----=====-----=====-----=====----- Ben Rosenberg mailto:ben@whack.org -----=====-----=====-----=====-----=====----- I'm out of my mind, but feel free to leave a message...
On Monday 21 January 2002 22:39, Ben Rosenberg wrote: | I'm running a Duron 800mhz box w/ an Asus A7V mb..works fine for | me. i'm on an athlon 2.1g with a gigabyte mb; there have been a few otherwise inexplicable events, all involving lockups. and i'm running an agp vid card with its various bells and whistles. indeed, i begin to wonder if this might have had something to do with the weird installation problems i encountered when using the graphical install. -- dep There is sobbing of the strong, And a pall upon the land; But the People in their weeping Bare the iron hand; Beware the People weeping When they bare the iron hand.
On Monday 21 January 2002 22:56, dep wrote: | i'm on an athlon 2.1g On Monday 21 January 2002 22:56, dep *meant*: i'm on an athlon 1.2g -- dep There is sobbing of the strong, And a pall upon the land; But the People in their weeping Bare the iron hand; Beware the People weeping When they bare the iron hand.
*shrug* I just used YaST to enter the parameter and it worked... lilo.conf looks like this.... -- append="mem=nopentium" boot=/dev/hda lba32 vga=791 message=/boot/message menu-scheme=Wg:kw:Wg:Wg read-only prompt timeout=80 -- I'm not sure why yours won't take the parameter... I used the GUI installer to update this box from 7.1 --> 7.3 * dep (dep@drippingwithirony.com) [020121 19:58]: ->On Monday 21 January 2002 22:39, Ben Rosenberg wrote: ->| I'm running a Duron 800mhz box w/ an Asus A7V mb..works fine for ->| me. -> ->i'm on an athlon 2.1g with a gigabyte mb; there have been a few ->otherwise inexplicable events, all involving lockups. and i'm running ->an agp vid card with its various bells and whistles. indeed, i begin ->to wonder if this might have had something to do with the weird ->installation problems i encountered when using the graphical install. -----=====-----=====-----=====-----=====----- Ben Rosenberg mailto:ben@whack.org -----=====-----=====-----=====-----=====----- I'm out of my mind, but feel free to leave a message...
On Monday 21 January 2002 23:35, Ben Rosenberg wrote: | *shrug* I just used YaST to enter the parameter and it worked... | | lilo.conf looks like this.... yup. just did some screwing around with the append line -- already had ide-scsi and so on -- and did seem to get it to work. (there is something else flaky going on -- my /etc/lilo.conf had come to contain *nothing* except failsafe and memtest, which is very weird. fixed it, but i'd love to know how it got broken.) while i'm here, maybe someone can explain the open gl handling in 7.3. i just got and built xscreensaver 4.0, which built okay except for the gl stuff, which didn't build at all for lack of glu.h. i installed the mesa stuff -- god, i wish the install routine were a little friendlier when it comes to development packages -- and now, going into the gl screensavers and trying to build 'em, i get a world of hurt, to wit: gcc -Wall -Wstrict-prototypes -Wnested-externs -Wno-format -std=c89 -U__STRICT_ANSI__ -L/usr/lib -o rubik rubik.o screenhack-gl.o xlock-gl.o fps.o ../xlockmore.o ../../utils/resources.o ../../utils/visual.o ../../utils/visual-gl.o ../../utils/usleep.o ../../utils/yarandom.o ../../utils/hsv.o ../../utils/colors.o -L/usr/X11R6/lib -lpthread -lSM -lICE -lXt -lX11 -lXmu -lXext -lm rubik.o: In function `pickcolor': /home/dep/download/xscreensaver-4.00/hacks/glx/rubik.c:480: undefined reference to `glMaterialfv' /home/dep/download/xscreensaver-4.00/hacks/glx/rubik.c:482: undefined reference to `glMaterialfv' rubik.o: In function `draw_cubit': /home/dep/download/xscreensaver-4.00/hacks/glx/rubik.c:500: undefined reference to `glNewList' /home/dep/download/xscreensaver-4.00/hacks/glx/rubik.c:501: undefined reference to `glBegin' [several pages of errors deleted] /home/dep/download/xscreensaver-4.00/hacks/glx/fps.c:153: undefined reference to `glMatrixMode' /home/dep/download/xscreensaver-4.00/hacks/glx/fps.c:154: undefined reference to `glPopMatrix' /home/dep/download/xscreensaver-4.00/hacks/glx/fps.c:158: undefined reference to `glPopAttrib' collect2: ld returned 1 exit status make: *** [rubik] Error 1 dep@depoffice:~/download/xscreensaver-4.00/hacks/glx > anybody know what package is needed in place of the mesa stuff, which seems not to be the answer? -- dep There is sobbing of the strong, And a pall upon the land; But the People in their weeping Bare the iron hand; Beware the People weeping When they bare the iron hand.
On Tuesday 22 January 2002 14.15, dep wrote:
gcc -Wall -Wstrict-prototypes -Wnested-externs -Wno-format -std=c89 -U__STRICT_ANSI__ -L/usr/lib -o rubik rubik.o screenhack-gl.o xlock-gl.o fps.o ../xlockmore.o ../../utils/resources.o ../../utils/visual.o ../../utils/visual-gl.o ../../utils/usleep.o ../../utils/yarandom.o ../../utils/hsv.o ../../utils/colors.o -L/usr/X11R6/lib -lpthread -lSM -lICE -lXt -lX11 -lXmu -lXext -lm
I don't see any -lGL or -lMesaGL here. Did you rerun configure after installing mesa-devel? If you did, try ./configure --with-gl. Worked For Me(tm) :)
rubik.o: In function `pickcolor': /home/dep/download/xscreensaver-4.00/hacks/glx/rubik.c:480: undefined
the linker can't find GL-functions, but without the -lGL or -lMesaGL it's not going to. regards Anders
On Tuesday 22 January 2002 08:30, Anders Johansson wrote: | I don't see any -lGL or -lMesaGL here. Did you rerun configure | after installing mesa-devel? If you did, try ./configure --with-gl. | Worked For Me(tm) :) yeah, it was installed. in the time since i sent the note i got and built mesa-4.01, which solved that problem. now it complains of a lack of libxml, which puzzles me -- i have both libxml and libxml2, including devel packages for both, installed. running ldconfig didn't cure this (and yeah, i ran make distclean and made sure that config.cache had gone to the bit bucket). | the linker can't find GL-functions, but without the -lGL or | -lMesaGL it's not going to. this raises another question. if memory serves, suse has something that is supposed to do better 3d than mesa for hardware that supports it, which mine does. but i do not remember where or what it is, or, more important, where the project is located so i can get the newest of it. or is it mesa that is supposed to be quicker? -- dep There is sobbing of the strong, And a pall upon the land; But the People in their weeping Bare the iron hand; Beware the People weeping When they bare the iron hand.
On Tuesday 22 January 2002 14.58, dep wrote:
this raises another question. if memory serves, suse has something that is supposed to do better 3d than mesa for hardware that supports it, which mine does. but i do not remember where or what it is, or, more important, where the project is located so i can get the newest of it. or is it mesa that is supposed to be quicker?
In my case it compiles against Mesa but the dynamic linker picks up nvidia's GL libs, so I get hardware acceleration. //Anders
----- Original Message -----
From: "dep"
On Monday 21 January 2002 22:39, Ben Rosenberg wrote: | I'm running a Duron 800mhz box w/ an Asus A7V mb..works fine for | me.
i'm on an athlon 2.1g with a gigabyte mb; there have been a few otherwise inexplicable events, all involving lockups. and i'm running an agp vid card with its various bells and whistles. indeed, i begin to wonder if this might have had something to do with the weird installation problems i encountered when using the graphical install. -- dep
This could be similar to the problem I had installing. I tried the SuSE update unsuccessfully twice. I got it by using the "safe" install which disables idedma and apic. Even after the install I would lock up from time to time until I disabled idedma for good. That's funny since I never had problems using it before with 7.2 and I've never had lockups until I installed 7.3. I've still got an old 600mhz Athlon. Most of my video problems went away when I installed nvidia's drivers. John
There was a previous message thread which discussed this. To state it rather vaguely,
the handling of dma was changed between the kernels in 7.2 and 7.3 as I remember.
01/22/02 08:43:16 AM, "John Scott"
----- Original Message ----- From: "dep"
To: "Ben Rosenberg" ; "SLE" Sent: Tuesday, January 22, 2002 4:56 AM Subject: Re: [SLE] SuSE Crashes Hard On Monday 21 January 2002 22:39, Ben Rosenberg wrote: | I'm running a Duron 800mhz box w/ an Asus A7V mb..works fine for | me.
i'm on an athlon 2.1g with a gigabyte mb; there have been a few otherwise inexplicable events, all involving lockups. and i'm running an agp vid card with its various bells and whistles. indeed, i begin to wonder if this might have had something to do with the weird installation problems i encountered when using the graphical install. -- dep
This could be similar to the problem I had installing. I tried the SuSE update unsuccessfully twice. I got it by using the "safe" install which disables idedma and apic. Even after the install I would lock up from time to time until I disabled idedma for good. That's funny since I never had problems using it before with 7.2 and I've never had lockups until I installed 7.3. I've still got an old 600mhz Athlon. Most of my video problems went away when I installed nvidia's drivers.
John
-- To unsubscribe send e-mail to suse-linux-e-unsubscribe@suse.com For additional commands send e-mail to suse-linux-e-help@suse.com Also check the FAQ at http://www.suse.com/support/faq and the archives at http://lists.suse.com
1/22/02 2:42:40 AM, dep
to revive this old thread, news from the gentoo folks which is, um, frightening if the machine in question is an athlon:
http://features.linuxtoday.com/news_story.php3?ltsn=2002-01-21-001-20-NW-KN
i'm running an athlon; fwiw, lilo in suse-7.3 chokes on the mem=nopentium option. -- dep
Well, I for one have been hit badly by the AMD/AGP problem and was one of those on the NVidia
forums that maintained there was a problem with this combination, as this seemed to be the most
common denominator of those who suffered hard lockups.
Fortunately I managed to refrain from launching into a tirade against all things nVidia, as they
appear to be the innocent party on this one!
I've suffered repeated hard crashes against any driver newer than 1251 or any kernel beyond
2.4.4 (so if one of these clauses is true -> CRASH!).
I put the mem=nopentium option in lilo.conf but haven't yet rebooted so what you say about 7.3
choking on this is quite worrying.
I run 7.2...can anyone vouch as to whether this nopentium directive is safe?
Regards,
Tim Harrell
On Tue, 2002-01-22 at 06:56, Tim Harrell wrote:
1/22/02 2:42:40 AM, dep
wrote: to revive this old thread, news from the gentoo folks which is, um, frightening if the machine in question is an athlon:
http://features.linuxtoday.com/news_story.php3?ltsn=2002-01-21-001-20-NW-KN
i'm running an athlon; fwiw, lilo in suse-7.3 chokes on the mem=nopentium option. -- dep
Well, I for one have been hit badly by the AMD/AGP problem and was one of those on the NVidia forums that maintained there was a problem with this combination, as this seemed to be the most common denominator of those who suffered hard lockups. Fortunately I managed to refrain from launching into a tirade against all things nVidia, as they appear to be the innocent party on this one!
I've suffered repeated hard crashes against any driver newer than 1251 or any kernel beyond 2.4.4 (so if one of these clauses is true -> CRASH!).
I put the mem=nopentium option in lilo.conf but haven't yet rebooted so what you say about 7.3 choking on this is quite worrying.
I run 7.2...can anyone vouch as to whether this nopentium directive is safe?
Regards,
Tim Harrell
Tim, I have SuSE 7.2 Pro with the upgraded 2.4.16 kernel from Mantal's "next" directory on the SuSE FTP site (/pub/people/mantel/next). I think he's got some 2.4.17 stuff in there now, though. I also have been a victim of this bug, having a 1.33Ghz Athlon and an nVidia GeForce2MX card... at least I think I have. I'll know before the end of February though, as I'm going to be doing some testing on this to see if the extent is as bad as everyone says it is. My append line in /etc/lilo.conf looks like this: append = "idebus=66 hdc=ide-scsi mem=nopentium" and if I do a "dmesg | grep pemtium" I get: Kernel command line: auto BOOT_IMAGE=linux ro root=302 BOOT_FILE=/boot/vmlinuz.srh idebus=66 hdc=ide-scsi mem=nopentium So it looks like it's working... Have a great new week! -Steven
1/22/02 12:36:52 PM, Steven Hatfield
On Tue, 2002-01-22 at 06:56, Tim Harrell wrote:
I put the mem=nopentium option in lilo.conf but haven't yet rebooted so what you say about 7.3 choking on this is quite worrying.
I run 7.2...can anyone vouch as to whether this nopentium directive is safe?
Regards,
Tim Harrell
Tim, I have SuSE 7.2 Pro with the upgraded 2.4.16 kernel from Mantal's "next" directory on the SuSE FTP site (/pub/people/mantel/next). I think he's got some 2.4.17 stuff in there now, though. I also have been a victim of this bug, having a 1.33Ghz Athlon and an nVidia GeForce2MX card... at least I think I have. I'll know before the end of February though, as I'm going to be doing some testing on this to see if the extent is as bad as everyone says it is.
My append line in /etc/lilo.conf looks like this:
append = "idebus=66 hdc=ide-scsi mem=nopentium"
and if I do a "dmesg | grep pemtium" I get:
Kernel command line: auto BOOT_IMAGE=linux ro root=302 BOOT_FILE=/boot/vmlinuz.srh idebus=66 hdc=ide-scsi mem=nopentium
So it looks like it's working...
Have a great new week! -Steven
Hi Steven,
I just saw Ben's reply, which was delayed for some reason in my mail pickup, and I'll try this
nopentium option in the next boot. I loaded the SuSE shrinkwrapped 2.4.16 last week and I
wanted to see how that fared with the 2313 drivers out of curiosity, so far it's holding up pretty well.
2.4.16 looks like the best thing off the 2.4 tree so far.
If the option in lilo didn't work, I would have just gone into the sources and compiled it out
manually. I was just being cautious and hoping to reduce downtime (this box also serves a router
for my modest home LAN).
Regards,
Tim Harrell
participants (13)
-
Anders Johansson
-
Ben Rosenberg
-
Christopher D. Reimer
-
dep
-
Derek Fountain
-
Jaakko Tamminen
-
James Bliss
-
Jason A.Van Cleve
-
John Scott
-
Matt T.
-
Steven Hatfield
-
Tim Harrell
-
zentara