Re: 2.6.16.6 ( - 2.6.16.11 ) compile failure on an alpha

[Brian Uhrain <buhrain@rosettastone.com>]

[-- Attachment #1: Type: text/plain, Size: 609 bytes --]

Hi,

Tomi Lapinlampi wrote:
> When trying to compile 2.6.16.11 on an alpha I encountered a compile
> error in arch/alpha/kernel/setup.c. I found that the problem appears with
> all kernels from 2.6.16.6 up to 2.6.16.11.
> 2.6.16.5 compiles fine, just like all older 2.6.16 -releases.
> I did not test this with 2.6.17-rc -kernels.

The problem is that the version of the patch meant for the 2.6.17-rc 
kernels was merged into 2.6.16.5.  'for_each_possible_cpu' does not 
exist in the 2.6.16.x tree.  Attached is a patch that replaces the usage 
of for_each_possible_cpu with the 2.6.16 equivalent.

  - Brian

[-- Attachment #2: 2.6.16.11-alpha-compile-fix.patch --]
[-- Type: text/plain, Size: 723 bytes --]

Signed-off-by: Brian Uhrain <buhrain@rosettastone.com>

---
 arch/alpha/kernel/setup.c |   12 +++++++-----
 1 file changed, 7 insertions(+), 5 deletions(-)

--- linux-2.6.16.11.orig/arch/alpha/kernel/setup.c	2006-04-25 11:21:03.000000000 +0100
+++ linux-2.6.16.11/arch/alpha/kernel/setup.c	2006-04-25 11:22:56.557266608 +0100
@@ -483,11 +483,13 @@ register_cpus(void)
 {
 	int i;
 
-	for_each_possible_cpu(i) {
-		struct cpu *p = kzalloc(sizeof(*p), GFP_KERNEL);
-		if (!p)
-			return -ENOMEM;
-		register_cpu(p, i, NULL);
+	for (i = 0; i < NR_CPUS; i++) {
+		if (cpu_possible(i)) {
+			struct cpu *p = kzalloc(sizeof(*p), GFP_KERNEL);
+			if (!p)
+				return -ENOMEM;
+			register_cpu(p, i, NULL);
+		}
 	}
 	return 0;
 }