Longhaul - call suspend(PMSG_FREEZE) before and resume() after

I posted this to Dave Jones, but:

So I'm posting this patch here too. I'm not subscribed to this list. If You wish email me.
Minor change: preempt_disabled() goto_sleep_pci() and wakeup_pci() causes "sheduling while atomic". So in this patch below it isn't preempt_disabled(). But I think that this should be protected in some way.

How it works:
1. Call suspend(PMSG_FREEZE) for each block device.
2. Call suspend(PMSG_FREEZE) for each PCI device.
3. Change CPU frequency.
4. Call resume() for each PCI device.
5. Call resume() for each block device.

Result: No more broken DMA transfers caused by frequency change.

diff -uprN -X linux-2.6.16.18-vanilla/Documentation/dontdiff linux-2.6.16.18-vanilla/arch/i386/kernel/cpu/cpufreq/Kconfig linux-2.6.16.18/arch/i386/kernel/cpu/cpufreq/Kconfig
@@ -203,7 +203,6 @@ config X86_LONGRUN
 config X86_LONGHAUL
 	tristate "VIA Cyrix III Longhaul"
 	select CPU_FREQ_TABLE
-	depends on BROKEN
 	help
 	  This adds the CPUFreq driver for VIA Samuel/CyrixIII, 
 	  VIA Cyrix Samuel/C3, VIA Cyrix Ezra and VIA Cyrix Ezra-T 

diff -uprN -X linux-2.6.16.18-vanilla/Documentation/dontdiff linux-2.6.16.18-vanilla/block/genhd.c linux-2.6.16.18/block/genhd.c
@@ -16,9 +16,9 @@
 #include <linux/kobj_map.h>
 #include <linux/buffer_head.h>
 
-static struct subsystem block_subsys;
+struct subsystem block_subsys;
 
-static DECLARE_MUTEX(block_subsys_sem);
+DECLARE_MUTEX(block_subsys_sem);
 
 /*
  * Can be deleted altogether. Later.

@@ -511,8 +511,10 @@ static struct kset_uevent_ops block_ueve
 };
 
 /* declare block_subsys. */
-static decl_subsys(block, &ktype_block, &block_uevent_ops);
+decl_subsys(block, &ktype_block, &block_uevent_ops);
 
+EXPORT_SYMBOL(block_subsys);
+EXPORT_SYMBOL(block_subsys_sem);
 
 /*
  * aggregate disk stat collector.  Uses the same stats that the sysfs

----------------------------------------------------------------------
Tysiace zdjec swietnych samochodow! >>> http://link.interia.pl/f1944

Hi!

Well, suspend/resume is normally called during system suspend/resume, so it has 'interesting' locking requirements.

Result: system hang if userspace tries to do request at the same time -- PCI drivers probably were not designed for _that_... but we probably want to fix them, anyway, so this is probably ok (but expect to do some driver debugging).

But you should really add that preempt_disable and not try this on smp system...
Pavel

Thanks for all the (sleeping) penguins.