Skip to contentSkip to conversation

lib/test_bpf: Call page_address() on page acquired with GFP_KERNEL flag

PATCH Applied Sumitra Sharma 4 participants 5 messages Jun 23, 2023 – Jun 29, 2023

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

PATCH v4 1/1 lib/test_bpf: Call page_address() on page acquired with GFP_KERNEL flag

generate_test_data() acquires a page with alloc_page(GFP_KERNEL). The GFP_KERNEL is typical for kernel-internal allocations. The caller requires ZONE_NORMAL or a lower zone for direct access.

Therefore the page cannot come from ZONE_HIGHMEM. Thus there's no need to map it with kmap().

Also, the kmap() is being deprecated in favor of kmap_local_page() [1].

Hence, use a plain page_address() directly.

Since the page passed to the page_address() is not from the highmem zone, the page_address() function will always return a valid kernel virtual address and will not return NULL. Hence, remove the check 'if (!ptr)'.

Remove the unused variable 'ptr' and label 'err_free_page'.

[1]: https://lore.kernel.org/all/[email protected]/

---

- Link to v1: https://lore.kernel.org/bpf/[email protected]/T/
- Link to v2: https://lore.kernel.org/all/[email protected]/
- Link to v3: https://lore.kernel.org/all/[email protected]/

Changes in v4:

- Remove unused label 'err_free_page'.

Changes in v3:
Noted by: Fabio M. De Francesco<[email protected]>
- Remove the check 'if (!ptr)'.
- Remove the unused variable 'ptr'.
- Change the commit message.

Changes in v2:
Noted by: Fabio M. De Francesco<[email protected]>
- Remove the kmap() call and call page_address() instead.
- Change the commit subject and message.

 lib/test_bpf.c | 10 +---------
 1 file changed, 1 insertion(+), 9 deletions(-)
lib/test_bpf.c +1 -9
index ade9ac672adb..a5cc5f9fc4e8 100644
@@ -14381,25 +14381,17 @@ static void *generate_test_data(struct bpf_test *test, int sub)
 		 * single fragment to the skb, filled with
 		 * test->frag_data.
 		 */
-		void *ptr;
-
 		page = alloc_page(GFP_KERNEL);
 
 		if (!page)
 			goto err_kfree_skb;
 
-		ptr = kmap(page);
-		if (!ptr)
-			goto err_free_page;
-		memcpy(ptr, test->frag_data, MAX_DATA);
-		kunmap(page);
+		memcpy(page_address(page), test->frag_data, MAX_DATA);
 		skb_add_rx_frag(skb, 0, page, 0, MAX_DATA, MAX_DATA);
 	}
 
 	return skb;
 
-err_free_page:
-	__free_page(page);
 err_kfree_skb:
 	kfree_skb(skb);
 	return NULL;
Signature · 1 line
2.25.1
Reviewed-by on v4 1/1
#1 lore raw
Quoted 16 lines · On venerdì 23 giugno 2023 17:16:44 CEST Sumitra Sharma wrote:

generate_test_data() acquires a page with alloc_page(GFP_KERNEL). The GFP_KERNEL is typical for kernel-internal allocations. The caller requires ZONE_NORMAL or a lower zone for direct access.

Therefore the page cannot come from ZONE_HIGHMEM. Thus there's no need to map it with kmap().

Also, the kmap() is being deprecated in favor of kmap_local_page() [1].

Hence, use a plain page_address() directly.

Since the page passed to the page_address() is not from the highmem zone, the page_address() function will always return a valid kernel virtual address and will not return NULL. Hence, remove the check 'if (!ptr)'.

Quoted 5 lines

Remove the unused variable 'ptr' and label 'err_free_page'.

[1]: https://lore.kernel.org/all/[email protected]/

I can't see no more things that prevent my tag, then it is...

Thanks for working on it for all the time it took to get to v4 and for listening to what you've been suggested to change.

Well done :-)

Fabio

Quoted 61 lines

---

- Link to v1:
https://lore.kernel.org/bpf/[email protected]/T/ - Link to
v2: https://lore.kernel.org/all/[email protected]/ - Link to v3: https://lore.kernel.org/all/[email protected]/

Changes in v4:

Closes:
https://lore.kernel.org/oe-kbuild-all/[email protected]/ - Remove unused label 'err_free_page'.

Changes in v3:
Noted by: Fabio M. De Francesco<[email protected]>
- Remove the check 'if (!ptr)'.
- Remove the unused variable 'ptr'.
- Change the commit message.

Changes in v2:
Noted by: Fabio M. De Francesco<[email protected]>
- Remove the kmap() call and call page_address() instead.
- Change the commit subject and message.

 lib/test_bpf.c | 10 +---------
 1 file changed, 1 insertion(+), 9 deletions(-)
lib/test_bpf.c +1 -10
index ade9ac672adb..a5cc5f9fc4e8 100644
@@ -14381,25 +14381,17 @@ static void *generate_test_data(struct bpf_test
 *test, int sub) * single fragment to the skb, filled with
 		 * test->frag_data.
 		 */
-		void *ptr;
-
 		page = alloc_page(GFP_KERNEL);
 
 		if (!page)
 			goto err_kfree_skb;
 
-		ptr = kmap(page);
-		if (!ptr)
-			goto err_free_page;
-		memcpy(ptr, test->frag_data, MAX_DATA);
-		kunmap(page);
+		memcpy(page_address(page), test->frag_data, MAX_DATA);
 		skb_add_rx_frag(skb, 0, page, 0, MAX_DATA, MAX_DATA);
 	}
 
 	return skb;
 
-err_free_page:
-	__free_page(page);
 err_kfree_skb:
 	kfree_skb(skb);
 	return NULL;
--

2.25.1

#2 lore raw
Quoted 23 lines · Fabio M. De Francesco
Quoted 16 lines · On venerdì 23 giugno 2023 17:16:44 CEST Sumitra Sharma wrote:

generate_test_data() acquires a page with alloc_page(GFP_KERNEL). The GFP_KERNEL is typical for kernel-internal allocations. The caller requires ZONE_NORMAL or a lower zone for direct access.

Therefore the page cannot come from ZONE_HIGHMEM. Thus there's no need to map it with kmap().

Also, the kmap() is being deprecated in favor of kmap_local_page() [1].

Hence, use a plain page_address() directly.

Since the page passed to the page_address() is not from the highmem zone, the page_address() function will always return a valid kernel virtual address and will not return NULL. Hence, remove the check 'if (!ptr)'.

Quoted 5 lines

Remove the unused variable 'ptr' and label 'err_free_page'.

[1]: https://lore.kernel.org/all/[email protected]/

Quoted 8 lines

I can't see no more things that prevent my tag, then it is...

Thanks for working on it for all the time it took to get to v4 and for listening to what you've been suggested to change.

Well done :-)

Thank you, Fabio and Ira, for your tags and kind words :)

Thanks & regards
Sumitra

Quoted 68 lines

Fabio

Quoted 61 lines

---

- Link to v1:
https://lore.kernel.org/bpf/[email protected]/T/ - Link to
v2: https://lore.kernel.org/all/[email protected]/ - Link to v3: https://lore.kernel.org/all/[email protected]/

Changes in v4:

Closes:
https://lore.kernel.org/oe-kbuild-all/[email protected]/ - Remove unused label 'err_free_page'.

Changes in v3:
Noted by: Fabio M. De Francesco<[email protected]>
- Remove the check 'if (!ptr)'.
- Remove the unused variable 'ptr'.
- Change the commit message.

Changes in v2:
Noted by: Fabio M. De Francesco<[email protected]>
- Remove the kmap() call and call page_address() instead.
- Change the commit subject and message.

 lib/test_bpf.c | 10 +---------
 1 file changed, 1 insertion(+), 9 deletions(-)
lib/test_bpf.c +1 -10
index ade9ac672adb..a5cc5f9fc4e8 100644
@@ -14381,25 +14381,17 @@ static void *generate_test_data(struct bpf_test
 *test, int sub) * single fragment to the skb, filled with
 		 * test->frag_data.
 		 */
-		void *ptr;
-
 		page = alloc_page(GFP_KERNEL);
 
 		if (!page)
 			goto err_kfree_skb;
 
-		ptr = kmap(page);
-		if (!ptr)
-			goto err_free_page;
-		memcpy(ptr, test->frag_data, MAX_DATA);
-		kunmap(page);
+		memcpy(page_address(page), test->frag_data, MAX_DATA);
 		skb_add_rx_frag(skb, 0, page, 0, MAX_DATA, MAX_DATA);
 	}
 
 	return skb;
 
-err_free_page:
-	__free_page(page);
 err_kfree_skb:
 	kfree_skb(skb);
 	return NULL;
--

2.25.1

Reviewed-by on v4 1/1
#3 lore raw
In reply to Sumitra Sharma
Quoted 16 lines · Sumitra Sharma

generate_test_data() acquires a page with alloc_page(GFP_KERNEL). The GFP_KERNEL is typical for kernel-internal allocations. The caller requires ZONE_NORMAL or a lower zone for direct access.

Therefore the page cannot come from ZONE_HIGHMEM. Thus there's no need to map it with kmap().

Also, the kmap() is being deprecated in favor of kmap_local_page() [1].

Hence, use a plain page_address() directly.

Since the page passed to the page_address() is not from the highmem zone, the page_address() function will always return a valid kernel virtual address and will not return NULL. Hence, remove the check 'if (!ptr)'.

Quoted 6 lines

Remove the unused variable 'ptr' and label 'err_free_page'.

[1]: https://lore.kernel.org/all/[email protected]/

Just saw the report from kernel bot...

Better on this version than I saw before! :-D

Quoted 59 lines

---

- Link to v1: https://lore.kernel.org/bpf/[email protected]/T/
- Link to v2: https://lore.kernel.org/all/[email protected]/
- Link to v3: https://lore.kernel.org/all/[email protected]/

Changes in v4:

- Remove unused label 'err_free_page'.

Changes in v3:
Noted by: Fabio M. De Francesco<[email protected]>
- Remove the check 'if (!ptr)'.
- Remove the unused variable 'ptr'.
- Change the commit message.

Changes in v2:
Noted by: Fabio M. De Francesco<[email protected]>
- Remove the kmap() call and call page_address() instead.
- Change the commit subject and message.

 lib/test_bpf.c | 10 +---------
 1 file changed, 1 insertion(+), 9 deletions(-)
lib/test_bpf.c +1 -9
index ade9ac672adb..a5cc5f9fc4e8 100644
@@ -14381,25 +14381,17 @@ static void *generate_test_data(struct bpf_test *test, int sub)
 		 * single fragment to the skb, filled with
 		 * test->frag_data.
 		 */
-		void *ptr;
-
 		page = alloc_page(GFP_KERNEL);
 
 		if (!page)
 			goto err_kfree_skb;
 
-		ptr = kmap(page);
-		if (!ptr)
-			goto err_free_page;
-		memcpy(ptr, test->frag_data, MAX_DATA);
-		kunmap(page);
+		memcpy(page_address(page), test->frag_data, MAX_DATA);
 		skb_add_rx_frag(skb, 0, page, 0, MAX_DATA, MAX_DATA);
 	}
 
 	return skb;
 
-err_free_page:
-	__free_page(page);
 err_kfree_skb:
 	kfree_skb(skb);
 	return NULL;
Signature · 1 line
2.25.1
#4 lore raw
In reply to Sumitra Sharma
This patch was applied to bpf/bpf-next.git (master) by Daniel Borkmann <[email protected]>:

Hello:

This patch was applied to bpf/bpf-next.git (master) by Daniel Borkmann <[email protected]>:

Quoted 8 lines · On Fri, 23 Jun 2023 08:16:44 -0700 you wrote:

generate_test_data() acquires a page with alloc_page(GFP_KERNEL). The GFP_KERNEL is typical for kernel-internal allocations. The caller requires ZONE_NORMAL or a lower zone for direct access.

Therefore the page cannot come from ZONE_HIGHMEM. Thus there's no need to map it with kmap().

[...]

Here is the summary with links:
- [v4] lib/test_bpf: Call page_address() on page acquired with GFP_KERNEL flag

    https://git.kernel.org/bpf/bpf-next/c/da1a055d01ed

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html