Skip to contentSkip to conversation

stumped with APM suspend/resume problem going from 2.4.5 -> 2.4.17

DISCUSSION Matt 2 participants 2 messages Feb 6, 2002 1 message not in the archive

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

Matt

author
#1 lore raw

disabling both CONFIG_SMP and CONFIG_X86_UP_APIC fixed the suspend/resume problem. I haven't confirmed whether or not CONFIG_X86_UP_APIC alone brings back the problem -- will let you know.

the dmi settings are:

DMI 2.3 present.
DMI table at 0x000E0010.

/proc/cpuinfo comes up as:

processor : 0
vendor_id : GenuineIntel
cpu family : 6
model : 8
model name : Pentium III (Coppermine)
stepping : 6
cpu MHz : 696.980
cache size : 256 KB
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 2
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 sep mtrr pge mca cmov
pat pse36 mmx fxsr sse
bogomips : 1389.36

thanks for all your help.
-Matt

Quoted 40 lines · Mikael Pettersson
Quoted 15 lines · Matt

I have a Thinkpad T20 and have been using a hand-built 2.4.5 kernel for some time. Recently, ever since I got a USB scanner and camera, I've been experiencing spurious machine hangs. I read on one of the lists that there was a bug in the USB core that sounded related so I tried to upgrade.

Everything works ok with the stock 2.4.17 kernel (except the addition of the pcmcia-cs pacakge version 3.1.31) I built except that "apm -s" and then resume causes the machine to hang. On resume, even outside of X, the screen blanks or appears but doesn't respond to keyboard input. No oops, no messages, no nothing. The only choice I have at that point is a hard reset.

I just tried the 2.5.3 beta kernel and had the same problem.

Quoted 8 lines

Please let me know if there's anything else I can do to help diagnose the problem and/or workaroound it. I'll include my .config file in case that helps.

Thanks and keep up the great work.
-Matt
...
CONFIG_SMP=y

CONFIG_SMP could actually trigger the problem. Since about 2.4.10, the kernel will try to enable and use the processor's "local APIC" (a built-in interrupt controller) if CONFIG_SMP _or_ CONFIG_X86_UP_APIC is enabled. Unfortunately, some machines have crap BIOSen that break this feature. The Dell Inspiron is the standard example of this, but the T20 might also be affected.

Does /proc/cpuinfo list the "apic" feature flag?

If so, does disabling CONFIG_SMP and running a normal UP kernel eliminate the hangs?

If a plain UP kernel works, does UP with CONFIG_X86_UP_APIC reintroduce the hangs?

Quoted 7 lines

If all this is true, then it's a safe bet that your BIOSs (actually the "SMM" part of it) can't handle an enabled local APIC. I'm working on a patch for this: could you edit arch/i386/kernel/dmi_scan.c, change the #define dmi_printk(x) to use printk instead of being empty, reboot, and send me the DMI strings printed during boot?

/Mikael

_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com

#2 lore raw

Re: stumped with APM suspend/resume problem going from 2.4.5 ->

Quoted 3 lines

disabling both CONFIG_SMP and CONFIG_X86_UP_APIC fixed the suspend/resume problem. I haven't confirmed whether or not CONFIG_X86_UP_APIC alone brings back the problem -- will let you know.

On my thinkpad enabling CONFIG_SMP with or without UP_APIC breaks APM support at the resume stage. When I did some digging its deadlocking against itself somehow with the cli/sti calls in the apm resume code path. Remove those and it works

I didn't have time to get to the bottom of the problem