Skip to contentSkip to conversation

staging: vc04_services: Convert kmap() to kmap_local_page()

PATCH In review Fabio M. De Francesco 3 participants 4 messages Mar 30, 2022 – Apr 1, 2022

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

PATCH 1/1 staging: vc04_services: Convert kmap() to kmap_local_page()

The use of kmap() is being deprecated in favor of kmap_local_page() where it is feasible. In file interface/vchiq_arm/vchiq_arm.c, function free_pagelist() calls kmap() / kunmap() from two places.

With kmap_local_page(), the mapping is per thread, CPU local and not globally visible. Therefore, free_pagelist() is a function where the use of kmap_local_page() in place of kmap() is correctly suited.

Convert to kmap_local_page() but, instead of open coding it, use the memcpy_to_page() helper.

---
 .../vc04_services/interface/vchiq_arm/vchiq_arm.c   | 13 +++++--------
 1 file changed, 5 insertions(+), 8 deletions(-)
drivers/staging/vc04_services/interface/vchiq_arm/vchiq_arm.c +5 -8
index f0bfacfdea80..efb1383b5218 100644
@@ -431,21 +431,18 @@ free_pagelist(struct vchiq_pagelist_info *pagelistinfo,
 			if (head_bytes > actual)
 				head_bytes = actual;
 
-			memcpy((char *)kmap(pages[0]) +
+			memcpy_to_page(pages[0],
 				pagelist->offset,
 				fragments,
 				head_bytes);
-			kunmap(pages[0]);
 		}
 		if ((actual >= 0) && (head_bytes < actual) &&
-		    (tail_bytes != 0)) {
-			memcpy((char *)kmap(pages[num_pages - 1]) +
-				((pagelist->offset + actual) &
-				(PAGE_SIZE - 1) & ~(g_cache_line_size - 1)),
+		    (tail_bytes != 0))
+			memcpy_to_page(pages[num_pages - 1],
+				(pagelist->offset + actual) &
+				(PAGE_SIZE - 1) & ~(g_cache_line_size - 1),
 				fragments + g_cache_line_size,
 				tail_bytes);
-			kunmap(pages[num_pages - 1]);
-		}
 
 		down(&g_free_fragments_mutex);
 		*(char **)fragments = g_free_fragments;
Signature · 1 line
2.34.1
#1 lore raw
Quoted 50 lines · On mercoledì 30 marzo 2022 21:14:14 CEST Fabio M. De Francesco wrote:

The use of kmap() is being deprecated in favor of kmap_local_page() where it is feasible. In file interface/vchiq_arm/vchiq_arm.c, function free_pagelist() calls kmap() / kunmap() from two places.

With kmap_local_page(), the mapping is per thread, CPU local and not globally visible. Therefore, free_pagelist() is a function where the use of kmap_local_page() in place of kmap() is correctly suited.

Convert to kmap_local_page() but, instead of open coding it, use the memcpy_to_page() helper.

---
 .../vc04_services/interface/vchiq_arm/vchiq_arm.c   | 13 +++++--------
 1 file changed, 5 insertions(+), 8 deletions(-)
drivers/staging/vc04_services/interface/vchiq_arm/vchiq_arm.c +5 -8
index f0bfacfdea80..efb1383b5218 100644
@@ -431,21 +431,18 @@ free_pagelist(struct vchiq_pagelist_info *pagelistinfo,
 			if (head_bytes > actual)
 				head_bytes = actual;
 
-			memcpy((char *)kmap(pages[0]) +
+			memcpy_to_page(pages[0],
 				pagelist->offset,
 				fragments,
 				head_bytes);
-			kunmap(pages[0]);
 		}
 		if ((actual >= 0) && (head_bytes < actual) &&
-		    (tail_bytes != 0)) {
-			memcpy((char *)kmap(pages[num_pages - 1]) +
-				((pagelist->offset + actual) &
-				(PAGE_SIZE - 1) & ~(g_cache_line_size - 1)),
+		    (tail_bytes != 0))
+			memcpy_to_page(pages[num_pages - 1],
+				(pagelist->offset + actual) &
+				(PAGE_SIZE - 1) & ~(g_cache_line_size - 1),
 				fragments + g_cache_line_size,
 				tail_bytes);
-			kunmap(pages[num_pages - 1]);
-		}
 
 		down(&g_free_fragments_mutex);
 		*(char **)fragments = g_free_fragments;
Signature · 1 line
2.34.1

Hi Greg,

I've just received a message from you that says that a patch that I sent on March 31 has been applied to staging testing. I know that you usually apply patches in first come first served fashion (FIFO), therefore I wonder why this patch has not yet been applied.

Please don't misunderstand me: I have no hurry. I'm asking only because I suspect that this patch, sent on March 30th) could have been overlooked since it has the very identical subject of another patch that I sent on the same day (or the day before, I'm not sure about it now) and which has already been applied. Therefore, they may appear to be the same patch, because the only difference is that the drivers are different.

Thanks,

Fabio

#2 lore raw
Quoted 58 lines · Fabio M. De Francesco
Quoted 50 lines · On mercoledì 30 marzo 2022 21:14:14 CEST Fabio M. De Francesco wrote:

The use of kmap() is being deprecated in favor of kmap_local_page() where it is feasible. In file interface/vchiq_arm/vchiq_arm.c, function free_pagelist() calls kmap() / kunmap() from two places.

With kmap_local_page(), the mapping is per thread, CPU local and not globally visible. Therefore, free_pagelist() is a function where the use of kmap_local_page() in place of kmap() is correctly suited.

Convert to kmap_local_page() but, instead of open coding it, use the memcpy_to_page() helper.

---
 .../vc04_services/interface/vchiq_arm/vchiq_arm.c   | 13 +++++--------
 1 file changed, 5 insertions(+), 8 deletions(-)
drivers/staging/vc04_services/interface/vchiq_arm/vchiq_arm.c +5 -8
index f0bfacfdea80..efb1383b5218 100644
@@ -431,21 +431,18 @@ free_pagelist(struct vchiq_pagelist_info *pagelistinfo,
 			if (head_bytes > actual)
 				head_bytes = actual;
 
-			memcpy((char *)kmap(pages[0]) +
+			memcpy_to_page(pages[0],
 				pagelist->offset,
 				fragments,
 				head_bytes);
-			kunmap(pages[0]);
 		}
 		if ((actual >= 0) && (head_bytes < actual) &&
-		    (tail_bytes != 0)) {
-			memcpy((char *)kmap(pages[num_pages - 1]) +
-				((pagelist->offset + actual) &
-				(PAGE_SIZE - 1) & ~(g_cache_line_size - 1)),
+		    (tail_bytes != 0))
+			memcpy_to_page(pages[num_pages - 1],
+				(pagelist->offset + actual) &
+				(PAGE_SIZE - 1) & ~(g_cache_line_size - 1),
 				fragments + g_cache_line_size,
 				tail_bytes);
-			kunmap(pages[num_pages - 1]);
-		}
 
 		down(&g_free_fragments_mutex);
 		*(char **)fragments = g_free_fragments;
Signature · 1 line
2.34.1

Hi Greg,

I've just received a message from you that says that a patch that I sent on March 31 has been applied to staging testing. I know that you usually apply patches in first come first served fashion (FIFO), therefore I wonder why this patch has not yet been applied.

Quoted 6 lines

Please don't misunderstand me: I have no hurry. I'm asking only because I suspect that this patch, sent on March 30th) could have been overlooked since it has the very identical subject of another patch that I sent on the same day (or the day before, I'm not sure about it now) and which has already been applied. Therefore, they may appear to be the same patch, because the only difference is that the drivers are different.

I wanted to give others the chance to review this before applying it :)

Reviewed-by on v1 1/1
#3 lore raw
In reply to Fabio M. De Francesco
Quoted 4 lines · Fabio M. De Francesco

The use of kmap() is being deprecated in favor of kmap_local_page() where it is feasible. In file interface/vchiq_arm/vchiq_arm.c, function free_pagelist() calls kmap() / kunmap() from two places.

Quoted 8 lines

With kmap_local_page(), the mapping is per thread, CPU local and not globally visible. Therefore, free_pagelist() is a function where the use of kmap_local_page() in place of kmap() is correctly suited.

Convert to kmap_local_page() but, instead of open coding it, use the memcpy_to_page() helper.

LGTM:

Quoted 38 lines
---
 .../vc04_services/interface/vchiq_arm/vchiq_arm.c   | 13 +++++--------
 1 file changed, 5 insertions(+), 8 deletions(-)
drivers/staging/vc04_services/interface/vchiq_arm/vchiq_arm.c +5 -8
index f0bfacfdea80..efb1383b5218 100644
@@ -431,21 +431,18 @@ free_pagelist(struct vchiq_pagelist_info *pagelistinfo,
 			if (head_bytes > actual)
 				head_bytes = actual;
 
-			memcpy((char *)kmap(pages[0]) +
+			memcpy_to_page(pages[0],
 				pagelist->offset,
 				fragments,
 				head_bytes);
-			kunmap(pages[0]);
 		}
 		if ((actual >= 0) && (head_bytes < actual) &&
-		    (tail_bytes != 0)) {
-			memcpy((char *)kmap(pages[num_pages - 1]) +
-				((pagelist->offset + actual) &
-				(PAGE_SIZE - 1) & ~(g_cache_line_size - 1)),
+		    (tail_bytes != 0))
+			memcpy_to_page(pages[num_pages - 1],
+				(pagelist->offset + actual) &
+				(PAGE_SIZE - 1) & ~(g_cache_line_size - 1),
 				fragments + g_cache_line_size,
 				tail_bytes);
-			kunmap(pages[num_pages - 1]);
-		}
 
 		down(&g_free_fragments_mutex);
 		*(char **)fragments = g_free_fragments;
Signature · 1 line
2.34.1