[PATCH] block: remove unused REQ_TYPE_LINUX_BLOCK

[PATCH] block: remove unused REQ_TYPE_LINUX_BLOCK

[FUJITA Tomonori]

You prefer to keep this for future possible users?

This can be applied to block's for-2.6.36.

=
From: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Subject: [PATCH] block: remove unused REQ_TYPE_LINUX_BLOCK

Nobody uses REQ_TYPE_LINUX_BLOCK (and its REQ_LB_OP_*).

Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
---
 include/linux/blkdev.h |   15 ---------------
 1 files changed, 0 insertions(+), 15 deletions(-)

diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index 3a2c5d9..baf5258 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -60,7 +60,6 @@ enum rq_cmd_type_bits {
 	REQ_TYPE_PM_RESUME,		/* resume request */
 	REQ_TYPE_PM_SHUTDOWN,		/* shutdown request */
 	REQ_TYPE_SPECIAL,		/* driver defined type */
-	REQ_TYPE_LINUX_BLOCK,		/* generic block layer message */
 	/*
 	 * for ATA/ATAPI devices. this really doesn't belong here, ide should
 	 * use REQ_TYPE_SPECIAL and use rq->cmd[0] with the range of driver
@@ -70,20 +69,6 @@ enum rq_cmd_type_bits {
 	REQ_TYPE_ATA_PC,
 };
 
-/*
- * For request of type REQ_TYPE_LINUX_BLOCK, rq->cmd[0] is the opcode being
- * sent down (similar to how REQ_TYPE_BLOCK_PC means that ->cmd[] holds a
- * SCSI cdb.
- *
- * 0x00 -> 0x3f are driver private, to be used for whatever purpose they need,
- * typically to differentiate REQ_TYPE_SPECIAL requests.
- *
- */
-enum {
-	REQ_LB_OP_EJECT	= 0x40,		/* eject request */
-	REQ_LB_OP_FLUSH = 0x41,		/* flush request */
-};
-
 #define BLK_MAX_CDB	16
 
 /*
-- 
1.6.5

Re: [PATCH] block: remove unused REQ_TYPE_LINUX_BLOCK

[Jens Axboe]

On 2010-07-05 11:11, FUJITA Tomonori wrote:
> You prefer to keep this for future possible users?

No, lets just kill it if there are no users left. We can
always resurrect something like this later, if we want to.

-- 
Jens Axboe

Re: [PATCH] block: remove unused REQ_TYPE_LINUX_BLOCK

[FUJITA Tomonori]

On Mon, 05 Jul 2010 13:17:32 +0200
Jens Axboe <axboe@kernel.dk> wrote:

> On 2010-07-05 11:11, FUJITA Tomonori wrote:
> > You prefer to keep this for future possible users?
> 
> No, lets just kill it if there are no users left. We can
> always resurrect something like this later, if we want to.

Cool, please apply the patch then.

Re: [PATCH] block: remove unused REQ_TYPE_LINUX_BLOCK

[Jeff Garzik]

On 07/05/2010 05:11 AM, FUJITA Tomonori wrote:
> You prefer to keep this for future possible users?
>
> This can be applied to block's for-2.6.36.
>
> =
> From: FUJITA Tomonori<fujita.tomonori@lab.ntt.co.jp>
> Subject: [PATCH] block: remove unused REQ_TYPE_LINUX_BLOCK
>
> Nobody uses REQ_TYPE_LINUX_BLOCK (and its REQ_LB_OP_*).
>
> Signed-off-by: FUJITA Tomonori<fujita.tomonori@lab.ntt.co.jp>
> ---
>   include/linux/blkdev.h |   15 ---------------
>   1 files changed, 0 insertions(+), 15 deletions(-)
>
> diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
> index 3a2c5d9..baf5258 100644
> --- a/include/linux/blkdev.h
> +++ b/include/linux/blkdev.h
> @@ -60,7 +60,6 @@ enum rq_cmd_type_bits {
>   	REQ_TYPE_PM_RESUME,		/* resume request */
>   	REQ_TYPE_PM_SHUTDOWN,		/* shutdown request */
>   	REQ_TYPE_SPECIAL,		/* driver defined type */
> -	REQ_TYPE_LINUX_BLOCK,		/* generic block layer message */
>   	/*
>   	 * for ATA/ATAPI devices. this really doesn't belong here, ide should
>   	 * use REQ_TYPE_SPECIAL and use rq->cmd[0] with the range of driver
> @@ -70,20 +69,6 @@ enum rq_cmd_type_bits {
>   	REQ_TYPE_ATA_PC,
>   };
>
> -/*
> - * For request of type REQ_TYPE_LINUX_BLOCK, rq->cmd[0] is the opcode being
> - * sent down (similar to how REQ_TYPE_BLOCK_PC means that ->cmd[] holds a
> - * SCSI cdb.
> - *
> - * 0x00 ->  0x3f are driver private, to be used for whatever purpose they need,
> - * typically to differentiate REQ_TYPE_SPECIAL requests.
> - *
> - */
> -enum {
> -	REQ_LB_OP_EJECT	= 0x40,		/* eject request */
> -	REQ_LB_OP_FLUSH = 0x41,		/* flush request */
> -};

Acked-by: Jeff Garzik <jgarzik@redhat.com>

Having a second level of opcodes, under REQ_TYPE_LINUX_BLOCK, always 
seems less desirable than the current course of action you are now 
pursuing (REQ_FLUSH, etc.)...   nice cleanup.

	Jeff

Re: [PATCH] block: remove unused REQ_TYPE_LINUX_BLOCK

[FUJITA Tomonori]

On Mon, 05 Jul 2010 14:43:22 -0400
Jeff Garzik <jeff@garzik.org> wrote:

> > -/*
> > - * For request of type REQ_TYPE_LINUX_BLOCK, rq->cmd[0] is the opcode being
> > - * sent down (similar to how REQ_TYPE_BLOCK_PC means that ->cmd[] holds a
> > - * SCSI cdb.
> > - *
> > - * 0x00 ->  0x3f are driver private, to be used for whatever purpose they need,
> > - * typically to differentiate REQ_TYPE_SPECIAL requests.
> > - *
> > - */
> > -enum {
> > -	REQ_LB_OP_EJECT	= 0x40,		/* eject request */
> > -	REQ_LB_OP_FLUSH = 0x41,		/* flush request */
> > -};
> 
> Acked-by: Jeff Garzik <jgarzik@redhat.com>
> 
> Having a second level of opcodes, under REQ_TYPE_LINUX_BLOCK, always 
> seems less desirable than the current course of action you are now 
> pursuing (REQ_FLUSH, etc.)...   nice cleanup.

Agreed. We are running out of rq_flag_bits though. We can rethink when
we actually run out of it.