[PATCH] [MMC] Use command class to determine read-only status.

[PATCH] [MMC] Use command class to determine read-only status.

[Pierre Ossman]

If a card doesn't support the "write block" command class then
any attempts to open the device should reflect this by denying
write access.

Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
---

 drivers/mmc/mmc_block.c |   11 +++++++----
 1 files changed, 7 insertions(+), 4 deletions(-)

diff --git a/drivers/mmc/mmc_block.c b/drivers/mmc/mmc_block.c
--- a/drivers/mmc/mmc_block.c
+++ b/drivers/mmc/mmc_block.c
@@ -97,7 +97,8 @@ static int mmc_blk_open(struct inode *in
 		ret = 0;
 
 		if ((filp->f_mode & FMODE_WRITE) &&
-			mmc_card_readonly(md->queue.card))
+			(!(md->queue.card->csd.cmdclass & CCC_BLOCK_WRITE) ||
+			mmc_card_readonly(md->queue.card)))
 			ret = -EROFS;
 	}
 
@@ -407,10 +408,12 @@ static int mmc_blk_probe(struct mmc_card
 	if (err)
 		goto out;
 
-	printk(KERN_INFO "%s: %s %s %dKiB %s\n",
+	printk(KERN_INFO "%s: %s %s %dKiB",
 		md->disk->disk_name, mmc_card_id(card), mmc_card_name(card),
-		(card->csd.capacity << card->csd.read_blkbits) / 1024,
-		mmc_card_readonly(card)?"(ro)":"");
+		(card->csd.capacity << card->csd.read_blkbits) / 1024);
+	if (mmc_card_readonly(card) || !(card->csd.cmdclass & CCC_BLOCK_WRITE))
+		printk("(ro)");
+	printk("\n");
 
 	mmc_set_drvdata(card, md);
 	add_disk(md->disk);

Re: [PATCH] [MMC] Use command class to determine read-only status.

[Russell King]

On Fri, Oct 28, 2005 at 09:36:05AM +0200, Pierre Ossman wrote:
> If a card doesn't support the "write block" command class then
> any attempts to open the device should reflect this by denying
> write access.

I'd rather we kept printk messages as one printk if at all possible.
How about encapsulating both of these conditions into an inline
function:

static inline int mmc_blk_readonly(struct mmc_card *card)
{
	return mmc_card_readonly(card) ||
	       !(card->csd.cmdclass & CCC_BLOCK_WRITE);
}

> diff --git a/drivers/mmc/mmc_block.c b/drivers/mmc/mmc_block.c
> --- a/drivers/mmc/mmc_block.c
> +++ b/drivers/mmc/mmc_block.c
> @@ -97,7 +97,8 @@ static int mmc_blk_open(struct inode *in
>  		ret = 0;
>  
>  		if ((filp->f_mode & FMODE_WRITE) &&
> -			mmc_card_readonly(md->queue.card))

+		     mmc_blk_readonly(md->queue.card))

>  	printk(KERN_INFO "%s: %s %s %dKiB %s\n",
>  		md->disk->disk_name, mmc_card_id(card), mmc_card_name(card),
> -		(card->csd.capacity << card->csd.read_blkbits) / 1024,
> -		mmc_card_readonly(card)?"(ro)":"");

+		mmc_blk_readonly(card) ? "(ro)" : "");

As a bonus, I think this makes the code a lot more readable... but
then I am biased. 8)

-- 
Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of:  2.6 Serial core

Re: [PATCH] [MMC] Use command class to determine read-only status.

[Pierre Ossman]

[-- Attachment #1: Type: text/plain, Size: 422 bytes --]

Russell King wrote:
> On Fri, Oct 28, 2005 at 09:36:05AM +0200, Pierre Ossman wrote:
>> If a card doesn't support the "write block" command class then
>> any attempts to open the device should reflect this by denying
>> write access.
> 
> I'd rather we kept printk messages as one printk if at all possible.
> How about encapsulating both of these conditions into an inline
> function:
> 

Ok with me. New patch included.

[-- Attachment #2: mmc-readonly-ccc.patch --]
[-- Type: text/x-patch, Size: 1368 bytes --]

[MMC] Use command class to determine read-only status.

If a card doesn't support the "write block" command class then
any attempts to open the device should reflect this by denying
write access.

Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
---

 drivers/mmc/mmc_block.c |   10 ++++++++--
 1 files changed, 8 insertions(+), 2 deletions(-)

diff --git a/drivers/mmc/mmc_block.c b/drivers/mmc/mmc_block.c
--- a/drivers/mmc/mmc_block.c
+++ b/drivers/mmc/mmc_block.c
@@ -85,6 +85,12 @@ static void mmc_blk_put(struct mmc_blk_d
 	up(&open_lock);
 }
 
+static inline int mmc_blk_readonly(struct mmc_card *card)
+{
+	return mmc_card_readonly(card) ||
+	       !(card->csd.cmdclass & CCC_BLOCK_WRITE);
+}
+
 static int mmc_blk_open(struct inode *inode, struct file *filp)
 {
 	struct mmc_blk_data *md;
@@ -97,7 +103,7 @@ static int mmc_blk_open(struct inode *in
 		ret = 0;
 
 		if ((filp->f_mode & FMODE_WRITE) &&
-			mmc_card_readonly(md->queue.card))
+			mmc_blk_readonly(md->queue.card))
 			ret = -EROFS;
 	}
 
@@ -410,7 +416,7 @@ static int mmc_blk_probe(struct mmc_card
 	printk(KERN_INFO "%s: %s %s %dKiB %s\n",
 		md->disk->disk_name, mmc_card_id(card), mmc_card_name(card),
 		(card->csd.capacity << card->csd.read_blkbits) / 1024,
-		mmc_card_readonly(card)?"(ro)":"");
+		mmc_blk_readonly(card)?"(ro)":"");
 
 	mmc_set_drvdata(card, md);
 	add_disk(md->disk);

Re: [PATCH] [MMC] Use command class to determine read-only status.

[Russell King]

On Sun, Oct 30, 2005 at 12:39:58AM +0200, Pierre Ossman wrote:
> Russell King wrote:
> >On Fri, Oct 28, 2005 at 09:36:05AM +0200, Pierre Ossman wrote:
> >>If a card doesn't support the "write block" command class then
> >>any attempts to open the device should reflect this by denying
> >>write access.
> >
> >I'd rather we kept printk messages as one printk if at all possible.
> >How about encapsulating both of these conditions into an inline
> >function:
> >
> 
> Ok with me. New patch included.

Applied, thanks.

-- 
Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of:  2.6 Serial core