Skip to contentSkip to conversation

Alpha compile problem solved by Andrea (pte_alloc)

DISCUSSION Magnus Naeslund(f) 5 participants 12 messages Apr 27, 2001 – May 1, 2001 Alpha port

lore
Share & tools
lore
#0 lore raw

Hello yesterday i installed redhat6.2 on our little alpha server over here. It's an Ruffian EV56 system, and a hand upgraded redhat to be able to cope with 2.4.

I got an compile error that told me that pte_alloc was declared wrong in some files..
Then in the back of my mind i figured that Andrea does a lot of alpha work, so i grepped after pte_alloc in his patches.

I found:
ftp://ftp.kernel.org/pub/linux/kernel/people/andrea/patches/v2.4/2.4.3/alpha -numa-2

Now my kernel compiled, and it works great. (Thanks Andrea :)) Just a little gotcha if anyone gets this problem (now it's in the mail archives, where i didnt find it).

Andrea:
Is that patch harmless, or is it experimental?
Is there any other patches you recommend me to apply to my kernel?

Regards

Magnus Naeslund

#1 lore raw
Quoted 13 lines · Magnus Naeslund(f)

Hello yesterday i installed redhat6.2 on our little alpha server over here. It's an Ruffian EV56 system, and a hand upgraded redhat to be able to cope with 2.4.

I got an compile error that told me that pte_alloc was declared wrong in some files..
Then in the back of my mind i figured that Andrea does a lot of alpha work, so i grepped after pte_alloc in his patches.

I found:
ftp://ftp.kernel.org/pub/linux/kernel/people/andrea/patches/v2.4/2.4.3/alpha -numa-2

Quoted 6 lines

Now my kernel compiled, and it works great. (Thanks Andrea :)) Just a little gotcha if anyone gets this problem (now it's in the mail archives, where i didnt find it).

Andrea:
Is that patch harmless, or is it experimental?

The patch is ready for production. I just submitted it two times to Linus but no luck so far. However alternate patches are been merged in Linus's tree recently and they fix the compile problems at least.

Quoted 1 line

Is there any other patches you recommend me to apply to my kernel?

specifically for the alpha (but of course ok for x86 kernels too) in order against pre7:

	ftp://ftp.us.kernel.org/pub/linux/kernel/people/andrea/kernels/v2.4/2.4.4pre7aa1/00_alpha-numa-6
	ftp://ftp.us.kernel.org/pub/linux/kernel/people/andrea/kernels/v2.4/2.4.4pre7aa1/00_numa-sched-5
	ftp://ftp.us.kernel.org/pub/linux/kernel/people/andrea/kernels/v2.4/2.4.4pre7aa1/00_alpha-tlb-page-sym-1
	ftp://ftp.us.kernel.org/pub/linux/kernel/people/andrea/kernels/v2.4/2.4.4pre7aa1/00_softirq-SMP-fixes-2
	ftp://ftp.us.kernel.org/pub/linux/kernel/people/andrea/kernels/v2.4/2.4.4pre7aa1/00_rwsem-10

Andrea

#2 lore raw
In reply to Magnus Naeslund(f)

From: "Andrea Arcangeli" <[email protected]>

[snip]

Quoted 6 lines
Quoted 1 line

Is there any other patches you recommend me to apply to my kernel?

specifically for the alpha (but of course ok for x86 kernels too) in order against pre7:

ftp://ftp.us.kernel.org/pub/linux/kernel/people/andrea/kernels/v2.4/2.4.4pre 7aa1/00_alpha-numa-6

Quoted 1 line

ftp://ftp.us.kernel.org/pub/linux/kernel/people/andrea/kernels/v2.4/2.4.4pre 7aa1/00_numa-sched-5

Quoted 1 line

ftp://ftp.us.kernel.org/pub/linux/kernel/people/andrea/kernels/v2.4/2.4.4pre 7aa1/00_alpha-tlb-page-sym-1

Quoted 1 line

ftp://ftp.us.kernel.org/pub/linux/kernel/people/andrea/kernels/v2.4/2.4.4pre 7aa1/00_softirq-SMP-fixes-2

Quoted 1 line

ftp://ftp.us.kernel.org/pub/linux/kernel/people/andrea/kernels/v2.4/2.4.4pre 7aa1/00_rwsem-10

Quoted 3 lines

Andrea

Thanks...
Will apply them..

Magnus

#3 lore raw
In reply to Magnus Naeslund(f)

Do you know if anyone has fixed the lazy vmalloc code? I know of as of early 2.4 it was broken on alpha. At the time I noticed it I didn't have time to persue it, but before I forget to even put in a bug report I thought I'd ask if you know anything about it?

Eric

#4 lore raw
In reply to Magnus Naeslund(f)
Quoted 5 lines · Eric W. Biederman

Do you know if anyone has fixed the lazy vmalloc code? I know of as of early 2.4 it was broken on alpha. At the time I noticed it I didn't have time to persue it, but before I forget to even put in a bug report I thought I'd ask if you know anything about it?

On alpha it's racy if you set CONFIG_ALPHA_LARGE_VMALLOC y (so don't do that as you don't need it). As long as you use only 1 entry of the pgd for the whole vmalloc space (CONFIG_ALPHA_LARGE_VMALLOC n) alpha is safe.

OTOH x86 is racy and there's no workaround available at the moment.

Andrea

#5 lore raw
Quoted 4 lines

On alpha it's racy if you set CONFIG_ALPHA_LARGE_VMALLOC y (so don't do that as you don't need it). As long as you use only 1 entry of the pgd for the whole vmalloc space (CONFIG_ALPHA_LARGE_VMALLOC n) alpha is safe.

Its racy for all cases on the Alpha because the exception table fixes are not done.

Quoted 1 line

OTOH x86 is racy and there's no workaround available at the moment.

-ac fixes all known problems there

Alan

#6 lore raw
In reply to Andrea Arcangeli
Quoted 7 lines · Alan Cox
Quoted 4 lines

On alpha it's racy if you set CONFIG_ALPHA_LARGE_VMALLOC y (so don't do that as you don't need it). As long as you use only 1 entry of the pgd for the whole vmalloc space (CONFIG_ALPHA_LARGE_VMALLOC n) alpha is safe.

Its racy for all cases on the Alpha because the exception table fixes are not done.

you're talking about the module races, I was only talking only about vmalloc lazy pgd mapping, they're different things even if they are both related to the page fault hanlder.

I don't use modules on the alpha so...

Quoted 3 lines
Quoted 1 line

OTOH x86 is racy and there's no workaround available at the moment.

-ac fixes all known problems there

I will check that shortly, thanks. (so far all the fixes I seen floating around for such races were wrong)

Andrea

#7 lore raw
In reply to Andrea Arcangeli
Quoted 10 lines · Andrea Arcangeli
Quoted 7 lines · Alan Cox
Quoted 4 lines

On alpha it's racy if you set CONFIG_ALPHA_LARGE_VMALLOC y (so don't do that as you don't need it). As long as you use only 1 entry of the pgd for the whole vmalloc space (CONFIG_ALPHA_LARGE_VMALLOC n) alpha is safe.

Its racy for all cases on the Alpha because the exception table fixes are not done.

you're talking about the module races, I was only talking only about

here the fix for your module race (still untested though):

arch/alpha/mm/extable.c +9 -4
diff -urN 2.4.4/arch/alpha/mm/extable.c alpha-modrace/arch/alpha/mm/extable.c
@@ -45,20 +45,25 @@
 	/* There is only the kernel to search.  */
 	ret = search_one_table(__start___ex_table, __stop___ex_table - 1,
 			       addr - gp);
-	if (ret) return ret;
 #else
+	unsigned long flags;
 	/* The kernel is the last "module" -- no need to treat it special. */
 	struct module *mp;
+
+	ret = 0;
+	spin_lock_irqsave(&modlist_lock, flags);
 	for (mp = module_list; mp ; mp = mp->next) {
-		if (!mp->ex_table_start)
+		if (!mp->ex_table_start || !(mp->flags&(MOD_RUNNING|MOD_INITIALIZING)))
 			continue;
 		ret = search_one_table(mp->ex_table_start,
 				       mp->ex_table_end - 1, addr - mp->gp);
-		if (ret) return ret;
+		if (ret)
+			break;
 	}
+	spin_unlock_irqrestore(&modlist_lock, flags);
 #endif
 
-	return 0;
+	return ret;
 }
 
 unsigned

For the large-vmalloc races I'd take a very lazy approch:

arch/alpha/config.in renamed from arch/alpha/config.in.~1~ +9 -7
@@ -211,13 +211,15 @@
 
 # The machine must be able to support more than 8GB physical memory
 # before large vmalloc might even pretend to be an issue.
-if [ "$CONFIG_ALPHA_GENERIC" = "y" -o "$CONFIG_ALPHA_DP264" = "y" \
-	-o "$CONFIG_ALPHA_WILDFIRE" = "y" -o "$CONFIG_ALPHA_TITAN" = "y" ]
-then
-	bool 'Large VMALLOC support' CONFIG_ALPHA_LARGE_VMALLOC
-else
-	define_bool CONFIG_ALPHA_LARGE_VMALLOC n
-fi
+#if [ "$CONFIG_ALPHA_GENERIC" = "y" -o "$CONFIG_ALPHA_DP264" = "y" \
+#	-o "$CONFIG_ALPHA_WILDFIRE" = "y" -o "$CONFIG_ALPHA_TITAN" = "y" ]
+#then
+#	bool 'Large VMALLOC support' CONFIG_ALPHA_LARGE_VMALLOC
+#else
+#	define_bool CONFIG_ALPHA_LARGE_VMALLOC n
+#fi
+# LARGE_VMALLOC is racy, if you *really* need it then fix it first
+define_bool CONFIG_ALPHA_LARGE_VMALLOC n
 
 source drivers/pci/Config.in
 

I mean: I certainly don't need it, not even on the 256G boxes, the non LARGE_VMALLOC is simpler and _faster_ (it drops a branch from the page fault handler fast path) and so I'd prefer to spend my time on other things than fixing LARGE_VMALLOC races, but still the above will avoid people to get bitten by such race until somebody fixes it. If anybody has a rasonable example for which I may need more than 8giga of kernel vmalloc memory then I can change my mind of course.

Andrea

#8 lore raw
In reply to Andrea Arcangeli

Hi,

Quoted 3 lines · Alan Cox
Quoted 1 line

OTOH x86 is racy and there's no workaround available at the moment.

-ac fixes all known problems there

Is there some place from where one can download all the patches in -ac kernels as separate patches, not just one monster patch (same way Andrea is doing)?

I assume you are maintaining them as separate patches anyway in order to be able to feed them to Linus.

Quoted 1 line

Alan

                                                                  -o)
    Hubert Mantel              Goodbye, dots...                   /\\
                                                                 _\_v
#9 lore raw
Quoted 2 lines

I assume you are maintaining them as separate patches anyway in order to be able to feed them to Linus.

Nope - the dependancies between them are too complex

#10 lore raw
In reply to Magnus Naeslund(f)
Quoted 11 lines · Andrea Arcangeli
Quoted 5 lines · Eric W. Biederman

Do you know if anyone has fixed the lazy vmalloc code? I know of as of early 2.4 it was broken on alpha. At the time I noticed it I didn't have time to persue it, but before I forget to even put in a bug report I thought I'd ask if you know anything about it?

On alpha it's racy if you set CONFIG_ALPHA_LARGE_VMALLOC y (so don't do that as you don't need it). As long as you use only 1 entry of the pgd for the whole vmalloc space (CONFIG_ALPHA_LARGE_VMALLOC n) alpha is safe.

Hmm. I was having problems reproducible with
CONFIG_ALPHA_LARGE_VMALLOC n.

Enabling the large vmalloc was my work around, because the large vmalloc whet back to the prelazy allocation code.

I was getting repeatable problems inside of an mtd driver. The problem I had was entries failed to propagate across different tasks. I think it was something like the first pgd was lazily allocated and not propagated.

I don't have a SRM on my 264 alpha so alpha (for reference on which code paths were followed.

Quoted 2 lines

OTOH x86 is racy and there's no workaround available at the moment.

GH

Well racy is easier to work with than just plain non-functional.

Eric

#11 lore raw
In reply to Magnus Naeslund(f)
Quoted 5 lines · Eric W. Biederman

Hmm. I was having problems reproducible with
CONFIG_ALPHA_LARGE_VMALLOC n.

Enabling the large vmalloc was my work around, because the large vmalloc whet back to the prelazy allocation code.

I don't have a clue about your problems but certainly the CONFIG_ALPHA_LARGE_VMALLOC n is not racy while the CONFIG_ALPHA_LARGE_VMALLOC y is racy.

Quoted 1 line

problem I had was entries failed to propagate across different tasks.

With CONFIG_ALPHA_LARGE_VMALLOC n the entry is propagated before starting using the new pgd so it cannot race, there's no special page fault case for that beacuse you will never get a page fault because of an unmapped pgd entry in the vmalloc space in first place.

Andrea