Skip to contentSkip to conversation

staging: hv: Changed |typedef RING_BUFFER| to |struct hv_ring_buffer| throughout.

PATCH In review Steve Friedl 3 participants 3 messages Sep 22, 2009 – Oct 9, 2009 Staging

lore
Share & tools
lore
#0 lore raw

PATCH 2/8 staging: hv: Changed |typedef RING_BUFFER| to |struct hv_ring_buffer| throughout.

The style guide (and Greg's prior work) have suggested that the typedef UPPERCASENAME style is not appropriate in the Linux kernel (even though it's been de rigueur in Win32 development for 20 years), so I'm fixing a couple more of them that remain... one at a time.

~~~ Steve

---
 drivers/staging/hv/RingBuffer.c |    6 +++---
 drivers/staging/hv/RingBuffer.h |    6 +++---
 2 files changed, 6 insertions(+), 6 deletions(-)
drivers/staging/hv/RingBuffer.c +3 -3
index f69ae33..ffd64ac 100644
@@ -295,15 +295,15 @@ Description:
 --*/
 int RingBufferInit(RING_BUFFER_INFO *RingInfo, void *Buffer, u32 BufferLen)
 {
-	ASSERT(sizeof(RING_BUFFER) == PAGE_SIZE);
+	ASSERT(sizeof(struct hv_ring_buffer) == PAGE_SIZE);
 
 	memset(RingInfo, 0, sizeof(RING_BUFFER_INFO));
 
-	RingInfo->RingBuffer = (RING_BUFFER*)Buffer;
+	RingInfo->RingBuffer = (struct hv_ring_buffer*)Buffer;
 	RingInfo->RingBuffer->ReadIndex = RingInfo->RingBuffer->WriteIndex = 0;
 
 	RingInfo->RingSize = BufferLen;
-	RingInfo->RingDataSize = BufferLen - sizeof(RING_BUFFER);
+	RingInfo->RingDataSize = BufferLen - sizeof(struct hv_ring_buffer);
 
 	spin_lock_init(&RingInfo->ring_lock);
 
drivers/staging/hv/RingBuffer.h +3 -3
index 6202157..1ea2ced 100644
@@ -27,7 +27,7 @@
 
 #include <linux/scatterlist.h>
 
-typedef struct _RING_BUFFER {
+struct hv_ring_buffer {
 	/* Offset in bytes from the start of ring data below */
 	volatile u32 WriteIndex;
 
@@ -51,10 +51,10 @@ typedef struct _RING_BUFFER {
 	 * !!! DO NOT place any fields below this !!!
 	 */
 	u8 Buffer[0];
-} __attribute__((packed)) RING_BUFFER;
+} __attribute__((packed));
 
 typedef struct _RING_BUFFER_INFO {
-	RING_BUFFER *RingBuffer;
+	struct hv_ring_buffer *RingBuffer;
 	u32 RingSize;			/* Include the shared header */
 	spinlock_t ring_lock;
 
Signature · 1 line
1.6.5.rc1
#1 lore raw
Quoted 8 lines · Steve Friedl

The style guide (and Greg's prior work) have suggested that the typedef UPPERCASENAME style is not appropriate in the Linux kernel (even though it's been de rigueur in Win32 development for 20 years), so I'm fixing a couple more of them that remain... one at a time.

~~~ Steve

A few comments about your submission, not so much about this changes in this patch specifically.

First of all, welcome here and thanks for your first submission! Content-wise, all your patches look correct so far, nice work.

The patch comment above mixes both information about the contents of the patch and information about the submission. The latter (as well as your email signature) have no place in the changeset comment above the '---' line but should go below it. They are good to have in the mail but serve no purpose in the git history afterwards. As an example, you could have written the same introduction as:

|[PATCH 2/8] staging: hv: Change typedef RING_BUFFER to struct hv_ring_buffer
|
|The typedef UPPERCASENAME style is not appropriate in the Linux kernel
|(even though it's been de rigueur in Win32 development for 20 years),
|so change 'typedef RING_BUFFER' to 'struct hv_ring_buffer'.
|
|Signed-off-by: Steve Friedl <[email protected]>
|
|---
| The style guide (and Greg's prior work) have suggested doing that,
| and I'm fixing a couple more of them that remain... one at a time.
|
| ~~~ Steve
|
| drivers/staging/hv/RingBuffer.c |    6 +++---
| drivers/staging/hv/RingBuffer.h |    6 +++---
| 2 files changed, 6 insertions(+), 6 deletions(-)

It's also preferred to have the full series in a single email thread, with all patches as a reply to the introductory mail. git-send-email does that with the '--thread --no-chain-reply-to', git-format-patch does the same with 'thread=shallow' (great job on consistency ;-) ).

You should also Cc the right mailing lists and people for the subsystem you are patching. If you don't know them, use scripts/get_maintainer.pl. 'scripts/get_maintainer.pl -f drivers/staging/hv/*' currently returns Greg Kroah-Hartman <[email protected]>
Nicolas Palix <[email protected]>
Hank Janssen <[email protected]>
Haiyang Zhang <[email protected]>
Bill Pemberton <[email protected]>
Randy Dunlap <[email protected]>
Jan Beulich <[email protected]>
Moritz Muehlenhoff <[email protected]> [email protected]
[email protected]

By looking at 'git log drivers/staging/hv', you can also determine that Randy, Jan and Moritz each only contributed a single patch, so they are not likely to be interested in your changes, but it would not be harmful to Cc them.

Also, you should check if there is a maintainer tree that already addresses the problems you are fixing or if there are possibly conflicts with other changes in that tree, not sure if you did that but I've seen a number of similar patches for the hv drivers so it would be possible that someone else did the same already.

	Arnd <><
#2 lore raw
In reply to Steve Friedl
Quoted 11 lines · Steve Friedl

The style guide (and Greg's prior work) have suggested that the typedef UPPERCASENAME style is not appropriate in the Linux kernel (even though it's been de rigueur in Win32 development for 20 years), so I'm fixing a couple more of them that remain... one at a time.

~~~ Steve

---
 drivers/staging/hv/RingBuffer.c |    6 +++---

I'd rather not polish this turd anymore. Can someone just replace the ringbuffer.c code by using the in-kernel ringbuffer or kfifo code instead? I'd love to just delete this file completly, it does not belong in the tree at all.

thanks,

greg k-h