Fix mmc_test build order

PATCH In review Kyungmin Park 3 participants 7 messages Jul 26, 2010 – Jul 27, 2010 MMC, SD and SDIO

lore

Share & tools

lore

Kyungmin Park

author

2010-07-26 10:10 UTC #0 lore raw

PATCH 1/1 Fix mmc_test build order

To use the mmc_test, it should be build before mmc_block.

Reported-by Adrian Hunter <[email protected]>

Signed-off-by Kyungmin Park <[email protected]>

---

drivers/mmc/card/Makefile +1 -1

index 0d40751..9d50437 100644
@@ -6,9 +6,9 @@ ifeq ($(CONFIG_MMC_DEBUG),y)
 	EXTRA_CFLAGS		+= -DDEBUG
 endif
 
+obj-$(CONFIG_MMC_TEST)		+= mmc_test.o
 obj-$(CONFIG_MMC_BLOCK)		+= mmc_block.o
 mmc_block-objs			:= block.o queue.o
-obj-$(CONFIG_MMC_TEST)		+= mmc_test.o
 
 obj-$(CONFIG_SDIO_UART)		+= sdio_uart.o

Adrian Hunter

2010-07-26 11:06 UTC #1 lore raw

Quoted 3 lines · Kyungmin Park

To use the mmc_test, it should be build before mmc_block.
Reported-by Adrian Hunter <[email protected]>

I did not report this. I just pointed out that mmc_test will not work if mmc_block has already been given all the MMC devices.

You can always switch between mmc_test and mmc_block by unbinding / binding e.g.

# ls -al /sys/bus/mmc/drivers/mmcblk
drwxr-xr-x    2 root     0               0 Jan  1 00:00 .
drwxr-xr-x    4 root     0               0 Jan  1 00:00 ..
--w-------    1 root     0            4096 Jan  1 00:01 bind
lrwxrwxrwx    1 root     0               0 Jan  1 00:01 mmc0:0001 -> ../../../../devices/platform/mmci-omap-hs.1/mmc_host/mmc0/mmc0:0001
--w-------    1 root     0            4096 Jan  1 00:01 uevent
--w-------    1 root     0            4096 Jan  1 00:01 unbind

# echo mmc0:0001 > /sys/bus/mmc/drivers/mmcblk/unbind

# echo mmc0:0001 > /sys/bus/mmc/drivers/mmc_test/bind
[  128.221313] mmc_test mmc0:0001: Card claimed for testing.

I suggest adding explanation to the Kconfig help text for MMC_TEST and also the module decription in drivers/mmc/card/mmc_test.c

Quoted 18 lines

Signed-off-by Kyungmin Park <[email protected]>

---

drivers/mmc/card/Makefile +1 -1

index 0d40751..9d50437 100644
@@ -6,9 +6,9 @@ ifeq ($(CONFIG_MMC_DEBUG),y)
 	EXTRA_CFLAGS		+= -DDEBUG
 endif
 
+obj-$(CONFIG_MMC_TEST)		+= mmc_test.o
 obj-$(CONFIG_MMC_BLOCK)		+= mmc_block.o
 mmc_block-objs			:= block.o queue.o
-obj-$(CONFIG_MMC_TEST)		+= mmc_test.o
 
 obj-$(CONFIG_SDIO_UART)		+= sdio_uart.o

Ghorai, Sukumar

2010-07-26 12:28 UTC #2 lore raw

PATCH 1/1 mmc: add device class attribute for ext_csd

All,
EXT_CSD is 512byte information (read using CMD8) for the MMC (eMMC) device. And for the latest and upcoming specification, ext_csd are much relevant then past.

So please let me know your idea to add and get the same ext_csd as MMC device attribute. This is the same way currently available for csd, cid, src, etc. (# cat /sys/class/mmc_host/mmc1/mmc1\:0001/ext_csd)

Regards,
Sukumar Ghorai

From: Sukumar Ghorai <[email protected]>
Date: Mon, 26 Jul 2010 10:47:12 +0530
Subject: [PATCH] mmc: add device class attribute for ext_csd

ext_csd is getting more informative form eMMC 4.4 onwards and better to have this in device information as currently available for csd, cid, src.

Signed-off-by Sukumar Ghorai <[email protected]>

---
 drivers/mmc/core/bus.c   |    1 +
 drivers/mmc/core/mmc.c   |   23 +++++++++++++++++------
 include/linux/mmc/card.h |    1 +
 3 files changed, 19 insertions(+), 6 deletions(-)

drivers/mmc/core/bus.c +1 -0

index 49d9dca..57d035e 100644
@@ -191,6 +191,7 @@ static void mmc_release_card(struct device *dev)
 	if (card->info)
 		kfree(card->info);
 
+	kfree(card->raw_ext_csd);
 	kfree(card);
 }

drivers/mmc/core/mmc.c +17 -6

index 89f7a25..23b1813 100644
@@ -167,10 +167,6 @@ static int mmc_read_ext_csd(struct mmc_card *card)
 	if (card->csd.mmca_vsn < CSD_SPEC_VER_4)
 		return 0;
 
-	/*
-	 * As the ext_csd is so large and mostly unused, we don't store the
-	 * raw block in mmc_card.
-	 */
 	ext_csd = kmalloc(512, GFP_KERNEL);
 	if (!ext_csd) {
 		printk(KERN_ERR "%s: could not allocate a buffer to "

@@ -250,11 +246,25 @@ static int mmc_read_ext_csd(struct mmc_card *card)
 	}
 
 out:
-	kfree(ext_csd);
-
+	card->raw_ext_csd = ext_csd;
 	return err;
 }
 
+static ssize_t mmc_ext_csd_show(struct device *dev,
+			struct device_attribute *attr, char *buf)
+{
+int i;
+struct mmc_card *card = container_of(dev, struct mmc_card, dev);
+
+	for (i = 0; i < 512; i++)
+		buf += sprintf(buf, "%02x", card->raw_ext_csd[i]);
+
+	buf -= 512;
+	return 512;
+}
+
+static DEVICE_ATTR(ext_csd, S_IRUGO, mmc_ext_csd_show, NULL);
+
 MMC_DEV_ATTR(cid, "%08x%08x%08x%08x\n", card->raw_cid[0], card->raw_cid[1],
 	card->raw_cid[2], card->raw_cid[3]);
 MMC_DEV_ATTR(csd, "%08x%08x%08x%08x\n", card->raw_csd[0], card->raw_csd[1],

@@ -270,6 +280,7 @@ MMC_DEV_ATTR(serial, "0x%08x\n", card->cid.serial);
 static struct attribute *mmc_std_attrs[] = {
 	&dev_attr_cid.attr,
 	&dev_attr_csd.attr,
+	&dev_attr_ext_csd.attr,
 	&dev_attr_date.attr,
 	&dev_attr_fwrev.attr,
 	&dev_attr_hwrev.attr,

include/linux/mmc/card.h +1 -0

index d02d2c6..00332ca 100644
@@ -105,6 +105,7 @@ struct mmc_card {
 	u32			raw_cid[4];	/* raw card CID */
 	u32			raw_csd[4];	/* raw card CSD */
 	u32			raw_scr[2];	/* raw card SCR */
+	u8			*raw_ext_csd;	/* raw card EXT CSD for MMC */
 	struct mmc_cid		cid;		/* card identification */
 	struct mmc_csd		csd;		/* card specific */
 	struct mmc_ext_csd	ext_csd;	/* mmc v4 extended card specific */

From: Sukumar Ghorai <[email protected]>
Date: Mon, 26 Jul 2010 10:47:12 +0530
Subject: [PATCH] mmc: add device class attribute for ext_csd

ext_csd is getting more informative form eMMC 4.4 onwards and better to have this in device information as currently available for csd, cid, src.

Signed-off-by Sukumar Ghorai <[email protected]>

---
 drivers/mmc/core/bus.c   |    1 +
 drivers/mmc/core/mmc.c   |   23 +++++++++++++++++------
 include/linux/mmc/card.h |    1 +
 3 files changed, 19 insertions(+), 6 deletions(-)

drivers/mmc/core/bus.c +1 -0

index 49d9dca..57d035e 100644
@@ -191,6 +191,7 @@ static void mmc_release_card(struct device *dev)
 	if (card->info)
 		kfree(card->info);
 
+	kfree(card->raw_ext_csd);
 	kfree(card);
 }

drivers/mmc/core/mmc.c +17 -6

index 89f7a25..23b1813 100644
@@ -167,10 +167,6 @@ static int mmc_read_ext_csd(struct mmc_card *card)
 	if (card->csd.mmca_vsn < CSD_SPEC_VER_4)
 		return 0;
 
-	/*
-	 * As the ext_csd is so large and mostly unused, we don't store the
-	 * raw block in mmc_card.
-	 */
 	ext_csd = kmalloc(512, GFP_KERNEL);
 	if (!ext_csd) {
 		printk(KERN_ERR "%s: could not allocate a buffer to "

@@ -250,11 +246,25 @@ static int mmc_read_ext_csd(struct mmc_card *card)
 	}
 
 out:
-	kfree(ext_csd);
-
+	card->raw_ext_csd = ext_csd;
 	return err;
 }
 
+static ssize_t mmc_ext_csd_show(struct device *dev,
+			struct device_attribute *attr, char *buf)
+{
+int i;
+struct mmc_card *card = container_of(dev, struct mmc_card, dev);
+
+	for (i = 0; i < 512; i++)
+		buf += sprintf(buf, "%02x", card->raw_ext_csd[i]);
+
+	buf -= 512;
+	return 512;
+}
+
+static DEVICE_ATTR(ext_csd, S_IRUGO, mmc_ext_csd_show, NULL);
+
 MMC_DEV_ATTR(cid, "%08x%08x%08x%08x\n", card->raw_cid[0], card->raw_cid[1],
 	card->raw_cid[2], card->raw_cid[3]);
 MMC_DEV_ATTR(csd, "%08x%08x%08x%08x\n", card->raw_csd[0], card->raw_csd[1],

@@ -270,6 +280,7 @@ MMC_DEV_ATTR(serial, "0x%08x\n", card->cid.serial);
 static struct attribute *mmc_std_attrs[] = {
 	&dev_attr_cid.attr,
 	&dev_attr_csd.attr,
+	&dev_attr_ext_csd.attr,
 	&dev_attr_date.attr,
 	&dev_attr_fwrev.attr,
 	&dev_attr_hwrev.attr,

include/linux/mmc/card.h +1 -0

index d02d2c6..00332ca 100644
@@ -105,6 +105,7 @@ struct mmc_card {
 	u32			raw_cid[4];	/* raw card CID */
 	u32			raw_csd[4];	/* raw card CSD */
 	u32			raw_scr[2];	/* raw card SCR */
+	u8			*raw_ext_csd;	/* raw card EXT CSD for MMC */
 	struct mmc_cid		cid;		/* card identification */
 	struct mmc_csd		csd;		/* card specific */
 	struct mmc_ext_csd	ext_csd;	/* mmc v4 extended card specific */

Adrian Hunter

2010-07-26 12:43 UTC #3 lore raw

Re: [PATCH] mmc: add device class attribute for ext_csd

Quoted 6 lines · Sukumar

All,
EXT_CSD is 512byte information (read using CMD8) for the MMC (eMMC) device. And for the latest and upcoming specification, ext_csd are much relevant then past.
So please let me know your idea to add and get the same ext_csd as MMC device attribute. This is the same way currently available for csd, cid, src, etc. (# cat /sys/class/mmc_host/mmc1/mmc1\:0001/ext_csd)

Some of the values in Extended CSD can change so storing it could cause incorrect values to be reported.

Note that Extended CSD is available via debugfs
e.g. for device mmc0:0001 with debugfs mounted on /sys/kernel/debug

cat /sys/kernel/debug/mmc0/mmc0\:0001/ext_csd

Quoted 93 lines

Regards,
Sukumar Ghorai

From: Sukumar Ghorai <[email protected]>
Date: Mon, 26 Jul 2010 10:47:12 +0530
Subject: [PATCH] mmc: add device class attribute for ext_csd

ext_csd is getting more informative form eMMC 4.4 onwards and better to have this in device information as currently available for csd, cid, src.

Signed-off-by Sukumar Ghorai <[email protected]>

---
 drivers/mmc/core/bus.c   |    1 +
 drivers/mmc/core/mmc.c   |   23 +++++++++++++++++------
 include/linux/mmc/card.h |    1 +
 3 files changed, 19 insertions(+), 6 deletions(-)

drivers/mmc/core/bus.c +1 -0

index 49d9dca..57d035e 100644
@@ -191,6 +191,7 @@ static void mmc_release_card(struct device *dev)
 	if (card->info)
 		kfree(card->info);
 
+	kfree(card->raw_ext_csd);
 	kfree(card);
 }

drivers/mmc/core/mmc.c +17 -6

index 89f7a25..23b1813 100644
@@ -167,10 +167,6 @@ static int mmc_read_ext_csd(struct mmc_card *card)
 	if (card->csd.mmca_vsn < CSD_SPEC_VER_4)
 		return 0;
 
-	/*
-	 * As the ext_csd is so large and mostly unused, we don't store the
-	 * raw block in mmc_card.
-	 */
 	ext_csd = kmalloc(512, GFP_KERNEL);
 	if (!ext_csd) {
 		printk(KERN_ERR "%s: could not allocate a buffer to "

@@ -250,11 +246,25 @@ static int mmc_read_ext_csd(struct mmc_card *card)
 	}
 
 out:
-	kfree(ext_csd);
-
+	card->raw_ext_csd = ext_csd;
 	return err;
 }
 
+static ssize_t mmc_ext_csd_show(struct device *dev,
+			struct device_attribute *attr, char *buf)
+{
+int i;
+struct mmc_card *card = container_of(dev, struct mmc_card, dev);
+
+	for (i = 0; i < 512; i++)
+		buf += sprintf(buf, "%02x", card->raw_ext_csd[i]);
+
+	buf -= 512;
+	return 512;
+}
+
+static DEVICE_ATTR(ext_csd, S_IRUGO, mmc_ext_csd_show, NULL);
+
 MMC_DEV_ATTR(cid, "%08x%08x%08x%08x\n", card->raw_cid[0], card->raw_cid[1],
 	card->raw_cid[2], card->raw_cid[3]);
 MMC_DEV_ATTR(csd, "%08x%08x%08x%08x\n", card->raw_csd[0], card->raw_csd[1],

@@ -270,6 +280,7 @@ MMC_DEV_ATTR(serial, "0x%08x\n", card->cid.serial);
 static struct attribute *mmc_std_attrs[] = {
 	&dev_attr_cid.attr,
 	&dev_attr_csd.attr,
+	&dev_attr_ext_csd.attr,
 	&dev_attr_date.attr,
 	&dev_attr_fwrev.attr,
 	&dev_attr_hwrev.attr,

include/linux/mmc/card.h +1 -0

index d02d2c6..00332ca 100644
@@ -105,6 +105,7 @@ struct mmc_card {
 	u32			raw_cid[4];	/* raw card CID */
 	u32			raw_csd[4];	/* raw card CSD */
 	u32			raw_scr[2];	/* raw card SCR */
+	u8			*raw_ext_csd;	/* raw card EXT CSD for MMC */
 	struct mmc_cid		cid;		/* card identification */
 	struct mmc_csd		csd;		/* card specific */
 	struct mmc_ext_csd	ext_csd;	/* mmc v4 extended card specific */

Ghorai, Sukumar

2010-07-26 15:00 UTC #4 lore raw

RE: [PATCH] mmc: add device class attribute for ext_csd

Quoted 16 lines

-----Original Message-----
From: Adrian Hunter [mailto:[email protected]]
Sent: Monday, July 26, 2010 6:13 PM
To: Ghorai, Sukumar
Cc: [email protected]; Andrew Morton
Subject: Re: [PATCH] mmc: add device class attribute for ext_csd
Quoted 2 lines · Sukumar
All,
EXT_CSD is 512byte information (read using CMD8) for the MMC (eMMC)
device.
Quoted 1 line
And for the latest and upcoming specification, ext_csd are much relevant
then past.
Quoted 1 line

Quoted 7 lines

Quoted 1 line
So please let me know your idea to add and get the same ext_csd as MMC
device attribute. This is the same way currently available for csd, cid, src, etc.
Quoted 1 line
(# cat /sys/class/mmc_host/mmc1/mmc1\:0001/ext_csd)
Some of the values in Extended CSD can change so storing it could cause incorrect values to be reported.

[Ghorai] I was thinking to get the original value as reference, such that we know that host is capable for maximum negotiation, like 8-bit, DDR mode, voltage, etc.
1. Is there any other way to get the original value supported by card?
2. If yes, is it worth full to do?

Quoted 4 lines

Note that Extended CSD is available via debugfs
e.g. for device mmc0:0001 with debugfs mounted on /sys/kernel/debug
    cat /sys/kernel/debug/mmc0/mmc0\:0001/ext_csd

[Ghorai] Thanks. I got it.

<<snip>>

Adrian Hunter

2010-07-27 07:51 UTC #5 lore raw

Re: [PATCH] mmc: add device class attribute for ext_csd

Quoted 15 lines · Sukumar

Quoted 15 lines
-----Original Message-----
From: Adrian Hunter [mailto:[email protected]]
Sent: Monday, July 26, 2010 6:13 PM
To: Ghorai, Sukumar
Cc: [email protected]; Andrew Morton
Subject: Re: [PATCH] mmc: add device class attribute for ext_csd
Quoted 2 lines · Sukumar
All,
EXT_CSD is 512byte information (read using CMD8) for the MMC (eMMC)
device.
Quoted 1 line
And for the latest and upcoming specification, ext_csd are much relevant
then past.
Quoted 1 line
So please let me know your idea to add and get the same ext_csd as MMC

Quoted 8 lines

Quoted 5 lines
device attribute. This is the same way currently available for csd, cid, src, etc.
Quoted 1 line
(# cat /sys/class/mmc_host/mmc1/mmc1\:0001/ext_csd)
Some of the values in Extended CSD can change so storing it could cause incorrect values to be reported.
[Ghorai] I was thinking to get the original value as reference, such that we know that host is capable for maximum negotiation, like 8-bit, DDR mode, voltage, etc.
1. Is there any other way to get the original value supported by card?
2. If yes, is it worth full to do?

If it is for debugging, would a debug prints do instead e.g.

drivers/mmc/core/mmc.c +3 -0

index 89f7a25..cba87b0 100644
@@ -249,6 +249,9 @@ static int mmc_read_ext_csd(struct mmc_card *card)
                                       1 << ext_csd[EXT_CSD_S_A_TIMEOUT];
       }
 
+       pr_debug("%s: initial ext_csd bus width code %02x",
+                mmc_hostname(host), ext_csd[EXT_CSD_BUS_WIDTH]);
+

out:

       kfree(ext_csd);

Quoted 13 lines

Quoted 5 lines
Note that Extended CSD is available via debugfs
e.g. for device mmc0:0001 with debugfs mounted on /sys/kernel/debug
    cat /sys/kernel/debug/mmc0/mmc0\:0001/ext_csd
[Ghorai] Thanks. I got it.
<<snip>>
--
To unsubscribe from this list: send the line "unsubscribe linux-mmc" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

Ghorai, Sukumar

2010-07-27 10:50 UTC #6 lore raw

RE: [PATCH] mmc: add device class attribute for ext_csd

Quoted 44 lines

-----Original Message-----
From: Adrian Hunter [mailto:[email protected]]
Sent: Tuesday, July 27, 2010 1:22 PM
To: Ghorai, Sukumar
Cc: [email protected]; Andrew Morton
Subject: Re: [PATCH] mmc: add device class attribute for ext_csd
Quoted 8 lines · Sukumar
Quoted 8 lines
-----Original Message-----
From: Adrian Hunter [mailto:[email protected]]
Sent: Monday, July 26, 2010 6:13 PM
To: Ghorai, Sukumar
Cc: [email protected]; Andrew Morton
Subject: Re: [PATCH] mmc: add device class attribute for ext_csd
Quoted 5 lines
Quoted 5 lines
Quoted 2 lines · Sukumar
All,
EXT_CSD is 512byte information (read using CMD8) for the MMC (eMMC)
device.
Quoted 1 line
And for the latest and upcoming specification, ext_csd are much
relevant
Quoted 8 lines
Quoted 7 lines
then past.
Quoted 1 line
So please let me know your idea to add and get the same ext_csd as MMC
device attribute. This is the same way currently available for csd, cid, src, etc.
Quoted 1 line
(# cat /sys/class/mmc_host/mmc1/mmc1\:0001/ext_csd)
Some of the values in Extended CSD can change so storing it could cause incorrect values to be reported.
[Ghorai] I was thinking to get the original value as reference, such
that we know that host is capable for maximum negotiation, like 8-bit, DDR mode, voltage, etc.
Quoted 2 lines
1. Is there any other way to get the original value supported by card?
2. If yes, is it worth full to do?
If it is for debugging, would a debug prints do instead e.g.
drivers/mmc/core/mmc.c +0 -0
index 89f7a25..cba87b0 100644

drivers/mmc/core/mmc.c static int mmc_read_ext_csd(struct mmc_card *card) · line 254

Quoted 6 lines

drivers/mmc/core/mmc.c +3 -0

@@ -249,6 +249,9 @@ static int mmc_read_ext_csd(struct mmc_card *card)
                                        1 << ext_csd[EXT_CSD_S_A_TIMEOUT];
        }
 
+       pr_debug("%s: initial ext_csd bus width code %02x",
+                mmc_hostname(host), ext_csd[EXT_CSD_BUS_WIDTH]);
+

[Ghorai] yes, we need this kind messages in some format.
So let me add some more important debug print message(s) from ext_csd and submit as a patch.

Quoted 18 lines

drivers/mmc/core/mmc.c +0 -0
 out:
        kfree(ext_csd);
 
Quoted 15 lines
Quoted 7 lines
Note that Extended CSD is available via debugfs
e.g. for device mmc0:0001 with debugfs mounted on /sys/kernel/debug
    cat /sys/kernel/debug/mmc0/mmc0\:0001/ext_csd
[Ghorai] Thanks. I got it.
<<snip>>
--
To unsubscribe from this list: send the line "unsubscribe linux-mmc" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html