Skip to contentSkip to conversation

alpha: fix vmalloc breakage

PATCH In review Ivan Kokshaysky 2 participants 3 messages Jan 7, 2009 – Jan 12, 2009

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 alpha: fix vmalloc breakage

On alpha, we have to map some stuff in the VMALLOC space very early in the boot process (to make SRM console callbacks work and so on, see arch/alpha/mm/init.c). For old VM allocator, we just manually placed a vm_struct onto the global vmlist and this worked for ages.
Unfortunately, the new allocator isn't aware of this, so it constantly tries to allocate the VM space which is already in use, making vmalloc on alpha defunct.

This patch forces KVA to import vmlist entries on init.

---
 mm/vmalloc.c |   13 +++++++++++++
 1 files changed, 13 insertions(+), 0 deletions(-)
mm/vmalloc.c +13 -0
index c5db9a7..0dd9c2e 100644
@@ -24,6 +24,7 @@
 #include <linux/rbtree.h>
 #include <linux/radix-tree.h>
 #include <linux/rcupdate.h>
+#include <linux/bootmem.h>
 
 #include <asm/atomic.h>
 #include <asm/uaccess.h>
@@ -984,6 +985,8 @@ EXPORT_SYMBOL(vm_map_ram);
 
 void __init vmalloc_init(void)
 {
+	struct vmap_area *va;
+	struct vm_struct *tmp;
 	int i;
 
 	for_each_possible_cpu(i) {
@@ -996,6 +999,16 @@ void __init vmalloc_init(void)
 		vbq->nr_dirty = 0;
 	}
 
+	/* Import existing vmlist entries. */
+	for (tmp = vmlist; tmp; tmp = tmp->next) {
+		va = alloc_bootmem(sizeof(struct vmap_area));
+		if (!va)
+			break;
+		va->flags = tmp->flags | VM_VM_AREA;
+		va->va_start = (unsigned long)tmp->addr;
+		va->va_end = va->va_start + tmp->size;
+		__insert_vmap_area(va);
+	}
 	vmap_initialized = true;
 }
 
#1 lore raw
Quoted 38 lines · Ivan Kokshaysky

On alpha, we have to map some stuff in the VMALLOC space very early in the boot process (to make SRM console callbacks work and so on, see arch/alpha/mm/init.c). For old VM allocator, we just manually placed a vm_struct onto the global vmlist and this worked for ages.
Unfortunately, the new allocator isn't aware of this, so it constantly tries to allocate the VM space which is already in use, making vmalloc on alpha defunct.

This patch forces KVA to import vmlist entries on init.

---
 mm/vmalloc.c |   13 +++++++++++++
 1 files changed, 13 insertions(+), 0 deletions(-)
mm/vmalloc.c +3 -0
index c5db9a7..0dd9c2e 100644
@@ -24,6 +24,7 @@
 #include <linux/rbtree.h>
 #include <linux/radix-tree.h>
 #include <linux/rcupdate.h>
+#include <linux/bootmem.h>
 
 #include <asm/atomic.h>
 #include <asm/uaccess.h>
@@ -984,6 +985,8 @@ EXPORT_SYMBOL(vm_map_ram);
 
 void __init vmalloc_init(void)
 {
+	struct vmap_area *va;
+	struct vm_struct *tmp;
 	int i;
 
 	for_each_possible_cpu(i) {

mm/vmalloc.c void __init vmalloc_init(void) · line 1006

Quoted 8 lines
mm/vmalloc.c +5 -0
@@ -996,6 +999,16 @@ void __init vmalloc_init(void)
 		vbq->nr_dirty = 0;
 	}
 
+	/* Import existing vmlist entries. */
+	for (tmp = vmlist; tmp; tmp = tmp->next) {
+		va = alloc_bootmem(sizeof(struct vmap_area));
+		if (!va)
+			break;

You don't need this check, alloc_bootmem() panics when it fails.

Hannes

#2 lore raw
Quoted 5 lines · Johannes Weiner

mm/vmalloc.c void __init vmalloc_init(void) · line 1006

Quoted 3 lines
mm/vmalloc.c +3 -0
+		va = alloc_bootmem(sizeof(struct vmap_area));
+		if (!va)
+			break;

You don't need this check, alloc_bootmem() panics when it fails.

Ah, indeed. Thanks.

Ivan.