Skip to contentSkip to conversation

powerpc/64s: Clarify copy_and_flush() cache sync loop comment

PATCH Superseded Nikhil Kumar Singh 6 participants 15 messages Jul 1 – Aug 6 PowerPC

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 powerpc/64s: Clarify copy_and_flush() cache sync loop comment

The value loaded into r0 in copy_and_flush() represents the number of 8-byte words processed between cache synchronization operations.

The existing comment refers to cache line size, which can make it appear that the value is a cache line size in bytes rather than a loop count. Clarify the comment to explain that the loop processes 8 words (64 bytes) per cache synchronization iteration, and that increasing the value would skip cache maintenance for intermediate cache lines.

This is a comment-only change with no functional impact.

---
 arch/powerpc/kernel/head_64.S | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)
arch/powerpc/kernel/head_64.S +8 -8
index 63432a33ec49..e21c2bce8f7e 100644
@@ -713,14 +713,14 @@ p_end: .8byte _end - copy_to_here
 _GLOBAL(copy_and_flush)
 	addi	r5,r5,-8
 	addi	r6,r6,-8
-4:	li	r0,8			/* Use the smallest common	*/
-					/* denominator cache line	*/
-					/* size.  This results in	*/
-					/* extra cache line flushes	*/
-					/* but operation is correct.	*/
-					/* Can't get cache line size	*/
-					/* from NACA as it is being	*/
-					/* moved too.			*/
+4:	li	r0,8			/* r0 is the number of 8-byte words       */
+					/* to copy per cache sync iteration.      */
+					/* 8 words * 8 bytes = 64 bytes. 64B is   */
+					/* the current default cache line size.   */
+					/* This is a loop count, not a byte       */
+					/* count. Increasing it will skip         */
+					/* dcbst/icbi for lines in between and    */
+					/* leave stale instructions in icache.    */
 
 	mtctr	r0			/* put # words/line in ctr	*/
 3:	addi	r6,r6,8			/* copy a cache line		*/
Signature · 1 line
2.43.5
Reviewed-by on v1 1/1
#1 lore raw
Quoted 30 lines · Nikhil Kumar Singh

The value loaded into r0 in copy_and_flush() represents the number of 8-byte words processed between cache synchronization operations.

The existing comment refers to cache line size, which can make it appear that the value is a cache line size in bytes rather than a loop count. Clarify the comment to explain that the loop processes 8 words (64 bytes) per cache synchronization iteration, and that increasing the value would skip cache maintenance for intermediate cache lines.

This is a comment-only change with no functional impact.

---
  arch/powerpc/kernel/head_64.S | 16 ++++++++--------
  1 file changed, 8 insertions(+), 8 deletions(-)
arch/powerpc/kernel/head_64.S +0 -6
index 63432a33ec49..e21c2bce8f7e 100644
@@ -713,14 +713,14 @@ p_end: .8byte _end - copy_to_here
  _GLOBAL(copy_and_flush)
  	addi	r5,r5,-8
  	addi	r6,r6,-8
-4:	li	r0,8			/* Use the smallest common	*/
-					/* denominator cache line	*/
-					/* size.  This results in	*/
-					/* extra cache line flushes	*/
-					/* but operation is correct.	*/
-					/* Can't get cache line size	*/

arch/powerpc/kernel/head_64.S p_end: .8byte _end - copy_to_here · line 723

Quoted 10 lines
arch/powerpc/kernel/head_64.S +8 -2
-					/* from NACA as it is being	*/
-					/* moved too.			*/
+4:	li	r0,8			/* r0 is the number of 8-byte words       */
+					/* to copy per cache sync iteration.      */
+					/* 8 words * 8 bytes = 64 bytes. 64B is   */
+					/* the current default cache line size.   */
+					/* This is a loop count, not a byte       */
+					/* count. Increasing it will skip         */
+					/* dcbst/icbi for lines in between and    */
+					/* leave stale instructions in icache.    */

Looks good to me, previous comment stating cache line size may make it look like r0 is number of bytes being copied which isn't the case when the code is read. Hence:

Thanks,
- Aditya G

Quoted 3 lines
arch/powerpc/kernel/head_64.S +0 -0
  	mtctr	r0			/* put # words/line in ctr	*/
  3:	addi	r6,r6,8			/* copy a cache line		*/
#2 lore raw
In reply to Nikhil Kumar Singh
Quoted 31 lines · Nikhil Kumar Singh

The value loaded into r0 in copy_and_flush() represents the number of 8-byte words processed between cache synchronization operations.

The existing comment refers to cache line size, which can make it appear that the value is a cache line size in bytes rather than a loop count. Clarify the comment to explain that the loop processes 8 words (64 bytes) per cache synchronization iteration, and that increasing the value would skip cache maintenance for intermediate cache lines.

This is a comment-only change with no functional impact.

---
 arch/powerpc/kernel/head_64.S | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)
arch/powerpc/kernel/head_64.S +0 -7
index 63432a33ec49..e21c2bce8f7e 100644
@@ -713,14 +713,14 @@ p_end: .8byte _end - copy_to_here
 _GLOBAL(copy_and_flush)
 	addi	r5,r5,-8
 	addi	r6,r6,-8
-4:	li	r0,8			/* Use the smallest common	*/
-					/* denominator cache line	*/
-					/* size.  This results in	*/
-					/* extra cache line flushes	*/
-					/* but operation is correct.	*/
-					/* Can't get cache line size	*/
-					/* from NACA as it is being	*/

arch/powerpc/kernel/head_64.S p_end: .8byte _end - copy_to_here · line 724

Quoted 10 lines
arch/powerpc/kernel/head_64.S +8 -1
-					/* moved too.			*/
+4:	li	r0,8			/* r0 is the number of 8-byte words       */
+					/* to copy per cache sync iteration.      */
+					/* 8 words * 8 bytes = 64 bytes. 64B is   */
+					/* the current default cache line size.   */
+					/* This is a loop count, not a byte       */
+					/* count. Increasing it will skip         */
+					/* dcbst/icbi for lines in between and    */
+					/* leave stale instructions in icache.    */
 

You dropped 2 pieces of information here, That we can't use NACA and that we are doing an extra cache line flush(atleast in case of P10).

Also use tab instead of spaces before commment close per line "*/" can you keep the last character aligned to rest of the file. It looks little nicer.

Regards,
Mukesh

Quoted 5 lines
arch/powerpc/kernel/head_64.S +0 -0
 	mtctr	r0			/* put # words/line in ctr	*/
 3:	addi	r6,r6,8			/* copy a cache line		*/
Signature · 1 line
2.43.5
#3 lore raw

PATCH v2 1/1 powerpc/64s: Clarify copy_and_flush() cache sync loop comment

In reply to Nikhil Kumar Singh

The value loaded into r0 in copy_and_flush() represents the number of 8-byte words processed between cache synchronization operations.

The existing comment refers to cache line size, which can make it appear that the value is a cache line size in bytes rather than a loop count. Clarify the comment to explain that the loop processes 8 words (64 bytes) per cache synchronization iteration, and that increasing the value would skip cache maintenance for intermediate cache lines.

This is a comment-only change with no functional impact.

---
 arch/powerpc/kernel/head_64.S | 20 ++++++++++++--------
 1 file changed, 12 insertions(+), 8 deletions(-)
arch/powerpc/kernel/head_64.S +12 -8
index 63432a33ec49..a54f6f979173 100644
@@ -713,14 +713,18 @@ p_end: .8byte _end - copy_to_here
 _GLOBAL(copy_and_flush)
 	addi	r5,r5,-8
 	addi	r6,r6,-8
-4:	li	r0,8			/* Use the smallest common	*/
-					/* denominator cache line	*/
-					/* size.  This results in	*/
-					/* extra cache line flushes	*/
-					/* but operation is correct.	*/
-					/* Can't get cache line size	*/
-					/* from NACA as it is being	*/
-					/* moved too.			*/
+4:	li	r0,8			/* r0 is the number of 8-byte words	*/
+					/* to copy per cache sync iteration.	*/
+					/* 8 words * 8 bytes = 64 bytes. 64B is	*/
+					/* the current default cache line size.	*/
+					/* This is a loop count, not a byte	*/
+					/* count. Increasing it may skip	*/
+					/* dcbst/icbi for lines in between and	*/
+					/* leave stale instructions in icache.	*/
+					/* This results in extra cache line	*/
+					/* flushes but operation is correct.	*/
+					/* Can't get cache line size from NACA	*/
+					/* as it is being moved too.		*/
 
 	mtctr	r0			/* put # words/line in ctr	*/
 3:	addi	r6,r6,8			/* copy a cache line		*/
Signature · 1 line
2.43.5
#4 lore raw
Quoted 34 lines · Nikhil Kumar Singh

The value loaded into r0 in copy_and_flush() represents the number of 8-byte words processed between cache synchronization operations.

The existing comment refers to cache line size, which can make it appear that the value is a cache line size in bytes rather than a loop count. Clarify the comment to explain that the loop processes 8 words (64 bytes) per cache synchronization iteration, and that increasing the value would skip cache maintenance for intermediate cache lines.

This is a comment-only change with no functional impact.

---
 arch/powerpc/kernel/head_64.S | 20 ++++++++++++--------
 1 file changed, 12 insertions(+), 8 deletions(-)
arch/powerpc/kernel/head_64.S +2 -8
index 63432a33ec49..a54f6f979173 100644
@@ -713,14 +713,18 @@ p_end: .8byte _end - copy_to_here
 _GLOBAL(copy_and_flush)
 	addi	r5,r5,-8
 	addi	r6,r6,-8
-4:	li	r0,8			/* Use the smallest common	*/
-					/* denominator cache line	*/
-					/* size.  This results in	*/
-					/* extra cache line flushes	*/
-					/* but operation is correct.	*/
-					/* Can't get cache line size	*/
-					/* from NACA as it is being	*/
-					/* moved too.			*/
+4:	li	r0,8			/* r0 is the number of 8-byte words	*/
+					/* to copy per cache sync iteration.	*/

arch/powerpc/kernel/head_64.S p_end: .8byte _end - copy_to_here · line 727

Quoted 10 lines
arch/powerpc/kernel/head_64.S +10 -0
+					/* 8 words * 8 bytes = 64 bytes. 64B is	*/
+					/* the current default cache line size.	*/
+					/* This is a loop count, not a byte	*/
+					/* count. Increasing it may skip	*/
+					/* dcbst/icbi for lines in between and	*/
+					/* leave stale instructions in icache.	*/
+					/* This results in extra cache line	*/
+					/* flushes but operation is correct.	*/
+					/* Can't get cache line size from NACA	*/
+					/* as it is being moved too.		*/

The comment looks fine. But the alignement of the comment is still different from earlier. Keep the alignment same as earlier, it looks a little nicer.

Regards,
Mukesh

Quoted 6 lines
arch/powerpc/kernel/head_64.S +0 -0
 	mtctr	r0			/* put # words/line in ctr	*/
 3:	addi	r6,r6,8			/* copy a cache line		*/
Signature · 1 line
2.43.5
#5 lore raw

Hi Mukesh,

Thanks for reviewing the patch and for the suggestion.
I'd prefer to keep the current alignment as I find it a bit more readable.

Thanks,
~ Nikhil

Quoted 57 lines · Mukesh Kumar Chaurasiya
Quoted 34 lines · Nikhil Kumar Singh

The value loaded into r0 in copy_and_flush() represents the number of 8-byte words processed between cache synchronization operations.

The existing comment refers to cache line size, which can make it appear that the value is a cache line size in bytes rather than a loop count. Clarify the comment to explain that the loop processes 8 words (64 bytes) per cache synchronization iteration, and that increasing the value would skip cache maintenance for intermediate cache lines.

This is a comment-only change with no functional impact.

---
  arch/powerpc/kernel/head_64.S | 20 ++++++++++++--------
  1 file changed, 12 insertions(+), 8 deletions(-)
arch/powerpc/kernel/head_64.S +2 -8
index 63432a33ec49..a54f6f979173 100644
@@ -713,14 +713,18 @@ p_end: .8byte _end - copy_to_here
  _GLOBAL(copy_and_flush)
  	addi	r5,r5,-8
  	addi	r6,r6,-8
-4:	li	r0,8			/* Use the smallest common	*/
-					/* denominator cache line	*/
-					/* size.  This results in	*/
-					/* extra cache line flushes	*/
-					/* but operation is correct.	*/
-					/* Can't get cache line size	*/
-					/* from NACA as it is being	*/
-					/* moved too.			*/
+4:	li	r0,8			/* r0 is the number of 8-byte words	*/
+					/* to copy per cache sync iteration.	*/

arch/powerpc/kernel/head_64.S p_end: .8byte _end - copy_to_here · line 727

Quoted 10 lines
arch/powerpc/kernel/head_64.S +10 -0
+					/* 8 words * 8 bytes = 64 bytes. 64B is	*/
+					/* the current default cache line size.	*/
+					/* This is a loop count, not a byte	*/
+					/* count. Increasing it may skip	*/
+					/* dcbst/icbi for lines in between and	*/
+					/* leave stale instructions in icache.	*/
+					/* This results in extra cache line	*/
+					/* flushes but operation is correct.	*/
+					/* Can't get cache line size from NACA	*/
+					/* as it is being moved too.		*/

The comment looks fine. But the alignement of the comment is still different from earlier. Keep the alignment same as earlier, it looks a little nicer.

Regards,
Mukesh

Quoted 6 lines
arch/powerpc/kernel/head_64.S +0 -0
  	mtctr	r0			/* put # words/line in ctr	*/
  3:	addi	r6,r6,8			/* copy a cache line		*/
Signature · 1 line
2.43.5
#6 lore raw
Quoted 8 lines · Nikhil Kumar Singh

Hi Mukesh,

Thanks for reviewing the patch and for the suggestion.
I'd prefer to keep the current alignment as I find it a bit more readable.

Thanks,
~ Nikhil

There is a reason we keep the character limit to 80 chars. Please go through these coding guidelines.

https://docs.kernel.org/process/coding-style.html

Regards,
Mukesh

Quoted 58 lines · Mukesh Kumar Chaurasiya
Quoted 56 lines · Nikhil Kumar Singh
Quoted 34 lines

The value loaded into r0 in copy_and_flush() represents the number of 8-byte words processed between cache synchronization operations.

The existing comment refers to cache line size, which can make it appear that the value is a cache line size in bytes rather than a loop count. Clarify the comment to explain that the loop processes 8 words (64 bytes) per cache synchronization iteration, and that increasing the value would skip cache maintenance for intermediate cache lines.

This is a comment-only change with no functional impact.

---
  arch/powerpc/kernel/head_64.S | 20 ++++++++++++--------
  1 file changed, 12 insertions(+), 8 deletions(-)
arch/powerpc/kernel/head_64.S +2 -8
index 63432a33ec49..a54f6f979173 100644
@@ -713,14 +713,18 @@ p_end: .8byte _end - copy_to_here
  _GLOBAL(copy_and_flush)
  	addi	r5,r5,-8
  	addi	r6,r6,-8
-4:	li	r0,8			/* Use the smallest common	*/
-					/* denominator cache line	*/
-					/* size.  This results in	*/
-					/* extra cache line flushes	*/
-					/* but operation is correct.	*/
-					/* Can't get cache line size	*/
-					/* from NACA as it is being	*/
-					/* moved too.			*/
+4:	li	r0,8			/* r0 is the number of 8-byte words	*/
+					/* to copy per cache sync iteration.	*/

arch/powerpc/kernel/head_64.S p_end: .8byte _end - copy_to_here · line 727

Quoted 10 lines
arch/powerpc/kernel/head_64.S +10 -0
+					/* 8 words * 8 bytes = 64 bytes. 64B is	*/
+					/* the current default cache line size.	*/
+					/* This is a loop count, not a byte	*/
+					/* count. Increasing it may skip	*/
+					/* dcbst/icbi for lines in between and	*/
+					/* leave stale instructions in icache.	*/
+					/* This results in extra cache line	*/
+					/* flushes but operation is correct.	*/
+					/* Can't get cache line size from NACA	*/
+					/* as it is being moved too.		*/

The comment looks fine. But the alignement of the comment is still different from earlier. Keep the alignment same as earlier, it looks a little nicer.

Regards,
Mukesh

Quoted 5 lines
arch/powerpc/kernel/head_64.S +0 -0
  	mtctr	r0			/* put # words/line in ctr	*/
  3:	addi	r6,r6,8			/* copy a cache line		*/
Signature · 1 line
2.43.5
#7 lore raw

Hi Mukesh,

Thanks for the clarification, sorry I missed that part. I'll update the comment to follow the 80-column limit and keep the alignment consistent in the next revision.

Thanks,
~ Nikhil

Quoted 74 lines · Mukesh Kumar Chaurasiya
Quoted 8 lines · Nikhil Kumar Singh

Hi Mukesh,

Thanks for reviewing the patch and for the suggestion.
I'd prefer to keep the current alignment as I find it a bit more readable.

Thanks,
~ Nikhil

There is a reason we keep the character limit to 80 chars. Please go through these coding guidelines.

https://docs.kernel.org/process/coding-style.html

Regards,
Mukesh

Quoted 57 lines · Mukesh Kumar Chaurasiya
Quoted 56 lines · Nikhil Kumar Singh
Quoted 34 lines

The value loaded into r0 in copy_and_flush() represents the number of 8-byte words processed between cache synchronization operations.

The existing comment refers to cache line size, which can make it appear that the value is a cache line size in bytes rather than a loop count. Clarify the comment to explain that the loop processes 8 words (64 bytes) per cache synchronization iteration, and that increasing the value would skip cache maintenance for intermediate cache lines.

This is a comment-only change with no functional impact.

---
   arch/powerpc/kernel/head_64.S | 20 ++++++++++++--------
   1 file changed, 12 insertions(+), 8 deletions(-)
arch/powerpc/kernel/head_64.S +2 -8
index 63432a33ec49..a54f6f979173 100644
@@ -713,14 +713,18 @@ p_end: .8byte _end - copy_to_here
   _GLOBAL(copy_and_flush)
   	addi	r5,r5,-8
   	addi	r6,r6,-8
-4:	li	r0,8			/* Use the smallest common	*/
-					/* denominator cache line	*/
-					/* size.  This results in	*/
-					/* extra cache line flushes	*/
-					/* but operation is correct.	*/
-					/* Can't get cache line size	*/
-					/* from NACA as it is being	*/
-					/* moved too.			*/
+4:	li	r0,8			/* r0 is the number of 8-byte words	*/
+					/* to copy per cache sync iteration.	*/

arch/powerpc/kernel/head_64.S p_end: .8byte _end - copy_to_here · line 727

Quoted 10 lines
arch/powerpc/kernel/head_64.S +10 -0
+					/* 8 words * 8 bytes = 64 bytes. 64B is	*/
+					/* the current default cache line size.	*/
+					/* This is a loop count, not a byte	*/
+					/* count. Increasing it may skip	*/
+					/* dcbst/icbi for lines in between and	*/
+					/* leave stale instructions in icache.	*/
+					/* This results in extra cache line	*/
+					/* flushes but operation is correct.	*/
+					/* Can't get cache line size from NACA	*/
+					/* as it is being moved too.		*/

The comment looks fine. But the alignement of the comment is still different from earlier. Keep the alignment same as earlier, it looks a little nicer.

Regards,
Mukesh

Quoted 5 lines
arch/powerpc/kernel/head_64.S +0 -0
   	mtctr	r0			/* put # words/line in ctr	*/
   3:	addi	r6,r6,8			/* copy a cache line		*/
Signature · 1 line
2.43.5
#8 lore raw
In reply to Nikhil Kumar Singh

Hi Mukesh,

Ideally, checkpatch should take care of this, I see no checkpatch errors.

```
./scripts/checkpatch.pl v2-0001-powerpc-64s-Clarify-copy_and_flush-cache-sync-loo.patch
total: 0 errors, 0 warnings, 26 lines checked

v2-0001-powerpc-64s-Clarify-copy_and_flush-cache-sync-loo.patch has no obvious style problems and is ready for submission.
```

Thanks
~ Nikhil

Quoted 65 lines · Nikhil Kumar Singh

Hi Mukesh,

Thanks for reviewing the patch and for the suggestion.
I'd prefer to keep the current alignment as I find it a bit more readable.

Thanks,
~ Nikhil

Quoted 56 lines · Mukesh Kumar Chaurasiya
Quoted 36 lines · Nikhil Kumar Singh

The value loaded into r0 in copy_and_flush() represents the number of 8-byte words processed between cache synchronization operations.

The existing comment refers to cache line size, which can make it appear that the value is a cache line size in bytes rather than a loop count. Clarify the comment to explain that the loop processes 8 words (64 bytes) per cache synchronization iteration, and that increasing the value would skip cache maintenance for intermediate cache lines.

This is a comment-only change with no functional impact.

---
  arch/powerpc/kernel/head_64.S | 20 ++++++++++++--------

  1 file changed, 12 insertions(+), 8 deletions(-)
arch/powerpc/kernel/head_64.S +0 -0
index 63432a33ec49..a54f6f979173 100644
@@ -713,14 +713,18 @@ p_end: .8byte _end - copy_to_here
   _GLOBAL(copy_and_flush)

      addi    r5,r5,-8
      addi    r6,r6,-8
-4:    li    r0,8            /* Use the smallest common    */ -                    /* denominator cache line    */
-                    /* size.  This results in    */
-                    /* extra cache line flushes    */
-                    /* but operation is correct.    */
-                    /* Can't get cache line size    */
-                    /* from NACA as it is being    */
-                    /* moved too.            */
+4:    li    r0,8            /* r0 is the number of 8-byte words    */ +                    /* to copy per cache sync iteration. */ +                    /* 8 words * 8 bytes = 64 bytes. 64B is    */ +                    /* the current default cache line size.    */

Quoted 8 lines

+                    /* This is a loop count, not a byte    */ +                    /* count. Increasing it may skip    */ +                    /* dcbst/icbi for lines in between and */ +                    /* leave stale instructions in icache. */ +                    /* This results in extra cache line    */ +                    /* flushes but operation is correct. */ +                    /* Can't get cache line size from NACA */ +                    /* as it is being moved too.        */

The comment looks fine. But the alignement of the comment is still different from earlier. Keep the alignment same as earlier, it looks a little nicer.

Regards,
Mukesh

Quoted 5 lines

        mtctr    r0            /* put # words/line in ctr    */ 3:    addi    r6,r6,8            /* copy a cache line */

Signature · 1 line
2.43.5
#9 lore raw
Quoted 13 lines · Nikhil Kumar Singh

Hi Mukesh,

Ideally, checkpatch should take care of this, I see no checkpatch errors.

```
./scripts/checkpatch.pl
v2-0001-powerpc-64s-Clarify-copy_and_flush-cache-sync-loo.patch
total: 0 errors, 0 warnings, 26 lines checked

v2-0001-powerpc-64s-Clarify-copy_and_flush-cache-sync-loo.patch has no
obvious style problems and is ready for submission.
```
Quoted 3 lines

Thanks
~ Nikhil

I am saying it looks nicer and easy to read for everyone if they are aligned. If you think it's not readable with the 80 char limit, put the whole comment before the function start. That will also makes it a little nicer. Right now it's getting wrapped around and it's difficult to read.

Regards,
Mukesh

Quoted 71 lines · Nikhil Kumar Singh
Quoted 70 lines

Hi Mukesh,

Thanks for reviewing the patch and for the suggestion. I'd prefer to keep the current alignment as I find it a bit more readable.

Thanks,
~ Nikhil

Quoted 60 lines · Mukesh Kumar Chaurasiya
Quoted 40 lines · Nikhil Kumar Singh

The value loaded into r0 in copy_and_flush() represents the number of 8-byte words processed between cache synchronization operations.

The existing comment refers to cache line size, which can make it appear
that the value is a cache line size in bytes rather than a loop count. Clarify the comment to explain that the loop processes 8 words (64 bytes)
per cache synchronization iteration, and that increasing the value would
skip cache maintenance for intermediate cache lines.

This is a comment-only change with no functional impact.

---
  arch/powerpc/kernel/head_64.S | 20 ++++++++++++--------

  1 file changed, 12 insertions(+), 8 deletions(-)
arch/powerpc/kernel/head_64.S +0 -0
index 63432a33ec49..a54f6f979173 100644
@@ -713,14 +713,18 @@ p_end: .8byte _end - copy_to_here
   _GLOBAL(copy_and_flush)

      addi    r5,r5,-8
      addi    r6,r6,-8
-4:    li    r0,8            /* Use the smallest common    */ -                    /* denominator cache line    */
-                    /* size.  This results in    */
-                    /* extra cache line flushes    */
-                    /* but operation is correct.    */
-                    /* Can't get cache line size    */
-                    /* from NACA as it is being    */
-                    /* moved too.            */
+4:    li    r0,8            /* r0 is the number of 8-byte words    */ +                    /* to copy per cache sync iteration. */ +                    /* 8 words * 8 bytes = 64 bytes. 64B is    */ +                    /* the current default cache line size.    */

Quoted 8 lines

+                    /* This is a loop count, not a byte    */ +                    /* count. Increasing it may skip    */ +                    /* dcbst/icbi for lines in between and */ +                    /* leave stale instructions in icache. */ +                    /* This results in extra cache line    */ +                    /* flushes but operation is correct. */ +                    /* Can't get cache line size from NACA */ +                    /* as it is being moved too.        */

The comment looks fine. But the alignement of the comment is still different from earlier. Keep the alignment same as earlier, it looks a little nicer.

Regards,
Mukesh

Quoted 5 lines

        mtctr    r0            /* put # words/line in ctr    */ 3:    addi    r6,r6,8            /* copy a cache line */

Signature · 1 line
2.43.5
#10 lore raw
In reply to Nikhil Kumar Singh
Quoted 3 lines · Nikhil Kumar Singh

Hi Mukesh,

Ideally, checkpatch should take care of this, I see no checkpatch errors.

That is because the default $max_line_length is set to 100 columns in checkpatch.pl scripts. The kernel coding style document recommends 80, but checkpatch enforces 100 unless you pass --max-line-length=80 explicitly.

  my $max_line_length = 100;

You will see the warnings being reported when you run the script as:

❯ scripts/checkpatch.pl ./v2_20260706_nikhilks_powerpc_64s_clarify_copy_and_flush_cache_sync_loop_comment.mbx --max-line-length=80
WARNING: line length of 82 exceeds 80 columns
#44: FILE: arch/powerpc/kernel/head_64.S:716:
+4: li r0,8 /* r0 is the number of 8-byte words */
[...]

total: 0 errors, 12 warnings, 26 lines checked

I agree with Mukesh here and I think the comment should be fixed to fit within 80 column width limit as staying within 80 columns is certainly still preferred.

Thanks,
Amit

Quoted 84 lines
```
./scripts/checkpatch.pl
v2-0001-powerpc-64s-Clarify-copy_and_flush-cache-sync-loo.patch
total: 0 errors, 0 warnings, 26 lines checked

v2-0001-powerpc-64s-Clarify-copy_and_flush-cache-sync-loo.patch has no
obvious style problems and is ready for submission.
```

Thanks
~ Nikhil

Quoted 70 lines · Nikhil Kumar Singh

Hi Mukesh,

Thanks for reviewing the patch and for the suggestion. I'd prefer to keep the current alignment as I find it a bit more readable.

Thanks,
~ Nikhil

Quoted 60 lines · Mukesh Kumar Chaurasiya
Quoted 40 lines · Nikhil Kumar Singh

The value loaded into r0 in copy_and_flush() represents the number of 8-byte words processed between cache synchronization operations.

The existing comment refers to cache line size, which can make it appear
that the value is a cache line size in bytes rather than a loop count. Clarify the comment to explain that the loop processes 8 words (64 bytes)
per cache synchronization iteration, and that increasing the value would
skip cache maintenance for intermediate cache lines.

This is a comment-only change with no functional impact.

---
  arch/powerpc/kernel/head_64.S | 20 ++++++++++++--------

  1 file changed, 12 insertions(+), 8 deletions(-)
arch/powerpc/kernel/head_64.S +0 -0
index 63432a33ec49..a54f6f979173 100644
@@ -713,14 +713,18 @@ p_end: .8byte _end - copy_to_here
   _GLOBAL(copy_and_flush)

      addi    r5,r5,-8
      addi    r6,r6,-8
-4:    li    r0,8            /* Use the smallest common    */ -                    /* denominator cache line    */
-                    /* size.  This results in    */
-                    /* extra cache line flushes    */
-                    /* but operation is correct.    */
-                    /* Can't get cache line size    */
-                    /* from NACA as it is being    */
-                    /* moved too.            */
+4:    li    r0,8            /* r0 is the number of 8-byte words    */ +                    /* to copy per cache sync iteration. */ +                    /* 8 words * 8 bytes = 64 bytes. 64B is    */ +                    /* the current default cache line size.    */

Quoted 8 lines

+                    /* This is a loop count, not a byte    */ +                    /* count. Increasing it may skip    */ +                    /* dcbst/icbi for lines in between and */ +                    /* leave stale instructions in icache. */ +                    /* This results in extra cache line    */ +                    /* flushes but operation is correct. */ +                    /* Can't get cache line size from NACA */ +                    /* as it is being moved too.        */

The comment looks fine. But the alignement of the comment is still different from earlier. Keep the alignment same as earlier, it looks a little nicer.

Regards,
Mukesh

Quoted 5 lines

        mtctr    r0            /* put # words/line in ctr    */ 3:    addi    r6,r6,8            /* copy a cache line */

Signature · 1 line
2.43.5
Reviewed-by on v2 1/1
#11 lore raw
In reply to Nikhil Kumar Singh
Quoted 37 lines · Nikhil Kumar Singh

The value loaded into r0 in copy_and_flush() represents the number of 8-byte words processed between cache synchronization operations.

The existing comment refers to cache line size, which can make it appear that the value is a cache line size in bytes rather than a loop count. Clarify the comment to explain that the loop processes 8 words (64 bytes) per cache synchronization iteration, and that increasing the value would skip cache maintenance for intermediate cache lines.

This is a comment-only change with no functional impact.

---
  arch/powerpc/kernel/head_64.S | 20 ++++++++++++--------
  1 file changed, 12 insertions(+), 8 deletions(-)
arch/powerpc/kernel/head_64.S +5 -8
index 63432a33ec49..a54f6f979173 100644
@@ -713,14 +713,18 @@ p_end: .8byte _end - copy_to_here
  _GLOBAL(copy_and_flush)
  	addi	r5,r5,-8
  	addi	r6,r6,-8
-4:	li	r0,8			/* Use the smallest common	*/
-					/* denominator cache line	*/
-					/* size.  This results in	*/
-					/* extra cache line flushes	*/
-					/* but operation is correct.	*/
-					/* Can't get cache line size	*/
-					/* from NACA as it is being	*/
-					/* moved too.			*/
+4:	li	r0,8			/* r0 is the number of 8-byte words	*/
+					/* to copy per cache sync iteration.	*/
+					/* 8 words * 8 bytes = 64 bytes. 64B is	*/
+					/* the current default cache line size.	*/
+					/* This is a loop count, not a byte	*/

arch/powerpc/kernel/head_64.S p_end: .8byte _end - copy_to_here · line 730

Quoted 10 lines
arch/powerpc/kernel/head_64.S +7 -0
+					/* count. Increasing it may skip	*/
+					/* dcbst/icbi for lines in between and	*/
+					/* leave stale instructions in icache.	*/
+					/* This results in extra cache line	*/
+					/* flushes but operation is correct.	*/
+					/* Can't get cache line size from NACA	*/
+					/* as it is being moved too.		*/
 
  	mtctr	r0			/* put # words/line in ctr	*/
  3:	addi	r6,r6,8			/* copy a cache line		*/

Makes sense. Feel free to keep the rb.

Thanks,

- Aditya G

Reviewed-by on v2 1/1
#12 lore raw
In reply to Nikhil Kumar Singh
Quoted 34 lines · Nikhil Kumar Singh

The value loaded into r0 in copy_and_flush() represents the number of 8-byte words processed between cache synchronization operations.

The existing comment refers to cache line size, which can make it appear that the value is a cache line size in bytes rather than a loop count. Clarify the comment to explain that the loop processes 8 words (64 bytes) per cache synchronization iteration, and that increasing the value would skip cache maintenance for intermediate cache lines.

This is a comment-only change with no functional impact.

---
 arch/powerpc/kernel/head_64.S | 20 ++++++++++++--------
 1 file changed, 12 insertions(+), 8 deletions(-)
arch/powerpc/kernel/head_64.S +2 -8
index 63432a33ec49..a54f6f979173 100644
@@ -713,14 +713,18 @@ p_end: .8byte _end - copy_to_here
 _GLOBAL(copy_and_flush)
 	addi	r5,r5,-8
 	addi	r6,r6,-8
-4:	li	r0,8			/* Use the smallest common	*/
-					/* denominator cache line	*/
-					/* size.  This results in	*/
-					/* extra cache line flushes	*/
-					/* but operation is correct.	*/
-					/* Can't get cache line size	*/
-					/* from NACA as it is being	*/
-					/* moved too.			*/
+4:	li	r0,8			/* r0 is the number of 8-byte words	*/
+					/* to copy per cache sync iteration.	*/

arch/powerpc/kernel/head_64.S p_end: .8byte _end - copy_to_here · line 727

Quoted 10 lines
arch/powerpc/kernel/head_64.S +10 -0
+					/* 8 words * 8 bytes = 64 bytes. 64B is	*/
+					/* the current default cache line size.	*/
+					/* This is a loop count, not a byte	*/
+					/* count. Increasing it may skip	*/
+					/* dcbst/icbi for lines in between and	*/
+					/* leave stale instructions in icache.	*/
+					/* This results in extra cache line	*/
+					/* flushes but operation is correct.	*/
+					/* Can't get cache line size from NACA	*/
+					/* as it is being moved too.		*/

Looks good to me. The previous comment was misleading which makes one to think r0 as cache line size which is not the case.

Thanks,
-Mahesh.

maintainer Applied
#13 lore raw
In reply to Nikhil Kumar Singh
Quoted 10 lines · Nikhil Kumar Singh

The value loaded into r0 in copy_and_flush() represents the number of 8-byte words processed between cache synchronization operations.

The existing comment refers to cache line size, which can make it appear that the value is a cache line size in bytes rather than a loop count. Clarify the comment to explain that the loop processes 8 words (64 bytes) per cache synchronization iteration, and that increasing the value would skip cache maintenance for intermediate cache lines.

[...]

Applied to powerpc/next.

[1/1] powerpc/64s: Clarify copy_and_flush() cache sync loop comment

      https://git.kernel.org/powerpc/c/76ea1257924f64521f708db91c99014ce7249bbf

cheers

Reviewed-by on v1 1/1
#14 lore raw
In reply to Nikhil Kumar Singh
Quoted 30 lines · Nikhil Kumar Singh

The value loaded into r0 in copy_and_flush() represents the number of 8-byte words processed between cache synchronization operations.

The existing comment refers to cache line size, which can make it appear that the value is a cache line size in bytes rather than a loop count. Clarify the comment to explain that the loop processes 8 words (64 bytes) per cache synchronization iteration, and that increasing the value would skip cache maintenance for intermediate cache lines.

This is a comment-only change with no functional impact.

---
 arch/powerpc/kernel/head_64.S | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)
arch/powerpc/kernel/head_64.S +0 -6
index 63432a33ec49..e21c2bce8f7e 100644
@@ -713,14 +713,14 @@ p_end: .8byte _end - copy_to_here
 _GLOBAL(copy_and_flush)
 	addi	r5,r5,-8
 	addi	r6,r6,-8
-4:	li	r0,8			/* Use the smallest common	*/
-					/* denominator cache line	*/
-					/* size.  This results in	*/
-					/* extra cache line flushes	*/
-					/* but operation is correct.	*/
-					/* Can't get cache line size	*/

arch/powerpc/kernel/head_64.S p_end: .8byte _end - copy_to_here · line 723

Quoted 10 lines
arch/powerpc/kernel/head_64.S +8 -2
-					/* from NACA as it is being	*/
-					/* moved too.			*/
+4:	li	r0,8			/* r0 is the number of 8-byte words       */
+					/* to copy per cache sync iteration.      */
+					/* 8 words * 8 bytes = 64 bytes. 64B is   */
+					/* the current default cache line size.   */
+					/* This is a loop count, not a byte       */
+					/* count. Increasing it will skip         */
+					/* dcbst/icbi for lines in between and    */
+					/* leave stale instructions in icache.    */

Looks good to me. The previous comment was misleading which makes one to think r0 as cache line size which is not the case.

Thanks,
-Mahesh.