Skip to contentSkip to conversation

mm/vmalloc: Extend find_vmap_lowest_match_check with extra arguments

PATCH Superseded Song Liu 4 participants 10 messages Aug 31, 2022 – Sep 4, 2022

Subsystems (2)

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore

Song Liu

author
#0 lore raw

PATCH 1/1 mm/vmalloc: Extend find_vmap_lowest_match_check with extra arguments

find_vmap_lowest_match() is now able to handle different roots. Make similar changes to find_vmap_lowest_match_check() and find_vmap_lowest_linear_match() to handle different trees.

Fixes f9863be49312 ("mm/vmalloc: extend __alloc_vmap_area() with extra arguments")
Cc Uladzislau Rezki (Sony) <[email protected]>
Cc Baoquan He <[email protected]>
Cc Andrew Morton <[email protected]>
---
 mm/vmalloc.c | 13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)
mm/vmalloc.c +7 -6
index dd6cdb201195..088b421601c4 100644
@@ -1300,12 +1300,12 @@ find_vmap_lowest_match(struct rb_root *root, unsigned long size,
 #include <linux/random.h>
 
 static struct vmap_area *
-find_vmap_lowest_linear_match(unsigned long size,
+find_vmap_lowest_linear_match(struct list_head *head, unsigned long size,
 	unsigned long align, unsigned long vstart)
 {
 	struct vmap_area *va;
 
-	list_for_each_entry(va, &free_vmap_area_list, list) {
+	list_for_each_entry(va, head, list) {
 		if (!is_within_this_va(va, size, align, vstart))
 			continue;
 
@@ -1316,7 +1316,8 @@ find_vmap_lowest_linear_match(unsigned long size,
 }
 
 static void
-find_vmap_lowest_match_check(unsigned long size, unsigned long align)
+find_vmap_lowest_match_check(struct rb_root *root, struct list_head *head,
+			     unsigned long size, unsigned long align)
 {
 	struct vmap_area *va_1, *va_2;
 	unsigned long vstart;
@@ -1325,8 +1326,8 @@ find_vmap_lowest_match_check(unsigned long size, unsigned long align)
 	get_random_bytes(&rnd, sizeof(rnd));
 	vstart = VMALLOC_START + rnd;
 
-	va_1 = find_vmap_lowest_match(size, align, vstart, false);
-	va_2 = find_vmap_lowest_linear_match(size, align, vstart);
+	va_1 = find_vmap_lowest_match(root, size, align, vstart, false);
+	va_2 = find_vmap_lowest_linear_match(head, size, align, vstart);
 
 	if (va_1 != va_2)
 		pr_emerg("not lowest: t: 0x%p, l: 0x%p, v: 0x%lx\n",
@@ -1513,7 +1514,7 @@ __alloc_vmap_area(struct rb_root *root, struct list_head *head,
 		return vend;
 
 #if DEBUG_AUGMENT_LOWEST_MATCH_CHECK
-	find_vmap_lowest_match_check(size, align);
+	find_vmap_lowest_match_check(root, head, size, align);
 #endif
 
 	return nva_start_addr;
Signature · 1 line
2.30.2
Reviewed-by on v1 1/1
#1 lore raw
Quoted 3 lines · Song Liu

find_vmap_lowest_match() is now able to handle different roots. Make similar changes to find_vmap_lowest_match_check() and find_vmap_lowest_linear_match() to handle different trees.

Good catch. Guess usually nobody eables DEBUG_AUGMENT_LOWEST_MATCH_CHECK to trigger compilation of the code.

Quoted 62 lines
Fixes f9863be49312 ("mm/vmalloc: extend __alloc_vmap_area() with extra arguments")
Cc Uladzislau Rezki (Sony) <[email protected]>
Cc Baoquan He <[email protected]>
Cc Andrew Morton <[email protected]>
---
 mm/vmalloc.c | 13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)
mm/vmalloc.c +7 -6
index dd6cdb201195..088b421601c4 100644
@@ -1300,12 +1300,12 @@ find_vmap_lowest_match(struct rb_root *root, unsigned long size,
 #include <linux/random.h>
 
 static struct vmap_area *
-find_vmap_lowest_linear_match(unsigned long size,
+find_vmap_lowest_linear_match(struct list_head *head, unsigned long size,
 	unsigned long align, unsigned long vstart)
 {
 	struct vmap_area *va;
 
-	list_for_each_entry(va, &free_vmap_area_list, list) {
+	list_for_each_entry(va, head, list) {
 		if (!is_within_this_va(va, size, align, vstart))
 			continue;
 
@@ -1316,7 +1316,8 @@ find_vmap_lowest_linear_match(unsigned long size,
 }
 
 static void
-find_vmap_lowest_match_check(unsigned long size, unsigned long align)
+find_vmap_lowest_match_check(struct rb_root *root, struct list_head *head,
+			     unsigned long size, unsigned long align)
 {
 	struct vmap_area *va_1, *va_2;
 	unsigned long vstart;
@@ -1325,8 +1326,8 @@ find_vmap_lowest_match_check(unsigned long size, unsigned long align)
 	get_random_bytes(&rnd, sizeof(rnd));
 	vstart = VMALLOC_START + rnd;
 
-	va_1 = find_vmap_lowest_match(size, align, vstart, false);
-	va_2 = find_vmap_lowest_linear_match(size, align, vstart);
+	va_1 = find_vmap_lowest_match(root, size, align, vstart, false);
+	va_2 = find_vmap_lowest_linear_match(head, size, align, vstart);
 
 	if (va_1 != va_2)
 		pr_emerg("not lowest: t: 0x%p, l: 0x%p, v: 0x%lx\n",
@@ -1513,7 +1514,7 @@ __alloc_vmap_area(struct rb_root *root, struct list_head *head,
 		return vend;
 
 #if DEBUG_AUGMENT_LOWEST_MATCH_CHECK
-	find_vmap_lowest_match_check(size, align);
+	find_vmap_lowest_match_check(root, head, size, align);
 #endif
 
 	return nva_start_addr;
Signature · 1 line
2.30.2
maintainer Reviewed-by
#2 lore raw
Quoted 10 lines · Song Liu
Quoted 3 lines

find_vmap_lowest_match() is now able to handle different roots. Make similar changes to find_vmap_lowest_match_check() and find_vmap_lowest_linear_match() to handle different trees.

Good catch. Guess usually nobody eables DEBUG_AUGMENT_LOWEST_MATCH_CHECK to trigger compilation of the code.

I do it.

--
Uladzislau Rezki

maintainer
#3 lore raw
In reply to Song Liu
Quoted 3 lines · On Tue, 30 Aug 2022 22:27:34 -0700 Song Liu <[email protected]> wrote:

find_vmap_lowest_match() is now able to handle different roots. Make similar changes to find_vmap_lowest_match_check() and find_vmap_lowest_linear_match() to handle different trees.

What are the runtime effects of this change?

author
#4 lore raw
Quoted 8 lines · Andrew Morton
Quoted 3 lines · On Tue, 30 Aug 2022 22:27:34 -0700 Song Liu <[email protected]> wrote:

find_vmap_lowest_match() is now able to handle different roots. Make similar changes to find_vmap_lowest_match_check() and find_vmap_lowest_linear_match() to handle different trees.

What are the runtime effects of this change?

The code is gated by DEBUG_AUGMENT_LOWEST_MATCH_CHECK. It is only compiled when the developer enables it explicitly. Therefore, there isn't
any runtime effect.

Thanks,
Song

maintainer
#5 lore raw

This is only for debug purpose.

Even without this patch, the debug path would work correctly. The difference is just only in
whether roots are hardcoded or passed over function paramter.

--
Uladzislau Rezki

Quoted 18 lines · Song Liu
Quoted 8 lines · Andrew Morton
Quoted 3 lines · On Tue, 30 Aug 2022 22:27:34 -0700 Song Liu <[email protected]> wrote:

find_vmap_lowest_match() is now able to handle different roots. Make similar changes to find_vmap_lowest_match_check() and find_vmap_lowest_linear_match() to handle different trees.

What are the runtime effects of this change?

The code is gated by DEBUG_AUGMENT_LOWEST_MATCH_CHECK. It is only compiled when the developer enables it explicitly. Therefore, there isn't
any runtime effect.

Thanks,
Song

Signature · 1 line
Uladzislau Rezki
#6 lore raw
Quoted 5 lines · Uladzislau Rezki

This is only for debug purpose.

Even without this patch, the debug path would work correctly. The difference is just only in
whether roots are hardcoded or passed over function paramter.

Calling find_vmap_lowest_match() inside find_vmap_lowest_match_check() will fail compilation because the function interface has been changed.

Quoted 26 lines · Song Liu
Quoted 18 lines · Andrew Morton
Quoted 8 lines · On Tue, 30 Aug 2022 22:27:34 -0700 Song Liu <[email protected]> wrote:
Quoted 4 lines

find_vmap_lowest_match() is now able to handle different roots. Make similar changes to find_vmap_lowest_match_check() and find_vmap_lowest_linear_match() to handle different trees.

What are the runtime effects of this change?

The code is gated by DEBUG_AUGMENT_LOWEST_MATCH_CHECK. It is only compiled when the developer enables it explicitly. Therefore, there isn't
any runtime effect.

Thanks,
Song

Signature · 1 line
Uladzislau Rezki
maintainer
#7 lore raw
Quoted 7 lines
Quoted 3 lines

Even without this patch, the debug path would work correctly. The difference is just only in
whether roots are hardcoded or passed over function paramter.

Calling find_vmap_lowest_match() inside find_vmap_lowest_match_check() will fail compilation because the function interface has been changed.

Ah. That makes sense, though the commit message has to reflect it. So it is only about compilation error if debug is ON.

Signature · 1 line
Uladzislau Rezki
#8 lore raw
Quoted 9 lines · Uladzislau Rezki
Quoted 7 lines
Quoted 3 lines

Even without this patch, the debug path would work correctly. The difference is just only in
whether roots are hardcoded or passed over function paramter.

Calling find_vmap_lowest_match() inside find_vmap_lowest_match_check() will fail compilation because the function interface has been changed.

Ah. That makes sense, though the commit message has to reflect it. So it is only about compilation error if debug is ON.

Indeed, the current patch log sounds like an improvement or normal change. In fact it's a code fix.

maintainer
#9 lore raw
Quoted 5 lines · Uladzislau Rezki
Quoted 4 lines
Quoted 4 lines
Quoted 3 lines

Even without this patch, the debug path would work correctly. The difference is just only in
whether roots are hardcoded or passed over function paramter.

Quoted 9 lines
Quoted 5 lines
Quoted 3 lines

Calling find_vmap_lowest_match() inside find_vmap_lowest_match_check() will fail compilation because the function interface has been changed.

Ah. That makes sense, though the commit message has to reflect it. So it is only about compilation error if debug is ON.

Indeed, the current patch log sounds like an improvement or normal change. In fact it's a code fix.

Then i think it is worth to mention about this in the commit message. At least i have missed the main point of this change looking at the commit message.

Song Liu, Could you please upload a v2 of it stating exactly what it fixes?

<snip>

urezki@pc638:~/data/raid0/coding/linux-next.git$ git diff
mm/vmalloc.c +1 -1
index e68c0081e861..7552f1f8350e 100644
@@ -713,7 +713,7 @@ EXPORT_SYMBOL(vmalloc_to_pfn);
 /*** Global kva allocator ***/
 
 #define DEBUG_AUGMENT_PROPAGATE_CHECK 0
-#define DEBUG_AUGMENT_LOWEST_MATCH_CHECK 0
+#define DEBUG_AUGMENT_LOWEST_MATCH_CHECK 1
 
 
 static DEFINE_SPINLOCK(vmap_area_lock);
urezki@pc638:~/data/raid0/coding/linux-next.git$ make -j64 bzImage
  DESCEND objtool
  CALL    scripts/checksyscalls.sh
  CHK     include/generated/compile.h
  CC      mm/vmalloc.o
mm/vmalloc.c: In function ‘find_vmap_lowest_match_check’:
mm/vmalloc.c:1328:32: warning: passing argument 1 of ‘find_vmap_lowest_match’ makes pointer from integer without a cast [-Wint-conversion]
 1328 |  va_1 = find_vmap_lowest_match(size, align, vstart, false);
      |                                ^~~~
      |                                |
      |                                long unsigned int
mm/vmalloc.c:1236:40: note: expected ‘struct rb_root *’ but argument is of type ‘long unsigned int’
 1236 | find_vmap_lowest_match(struct rb_root *root, unsigned long size,
      |                        ~~~~~~~~~~~~~~~~^~~~
mm/vmalloc.c:1328:9: error: too few arguments to function ‘find_vmap_lowest_match’
 1328 |  va_1 = find_vmap_lowest_match(size, align, vstart, false);
      |         ^~~~~~~~~~~~~~~~~~~~~~
mm/vmalloc.c:1236:1: note: declared here
 1236 | find_vmap_lowest_match(struct rb_root *root, unsigned long size,
      | ^~~~~~~~~~~~~~~~~~~~~~
make[1]: *** [scripts/Makefile.build:250: mm/vmalloc.o] Error 1
make[1]: *** Waiting for unfinished jobs....
make: *** [Makefile:2003: mm] Error 2
make: *** Waiting for unfinished jobs....
urezki@pc638:~/data/raid0/coding/linux-next.git$
<snip>

Thank you in advance!

--
Uladzislau Rezki