2011/2/18 Andrei Warkentin <andr...@motorola.com>:
> On Fri, Feb 18, 2011 at 1:47 PM, Andrei Warkentin <andr...@motorola.com>
> wrote:
>> On Fri, Feb 18, 2011 at 7:44 AM, Arnd Bergmann <a...@arndb.de> wrote:
>>> I'm curious. Neither the manfid nor the oemid fields of either card
>>> match what I have seen on SD cards, I would expect them to be
>>>
>>> Sandisk: manfid 0x000003, oemid 0x5344
>>> Toshiba: manfid 0x000002, oemid 0x544d
>>>
>>> I have not actually seen any Toshiba SD cards, but I assume that they
>>> use the same controllers as Kingston.
>>>
>>> Does anyone know if the IDs have any correlation between MMC and SD
>>> controllers?
>>>
>>> Arnd
>>>
>>
>> I'm unsure about the older scheme (assigned by MMCA), but ever since
>> MMC is now JEDEC-controlled, the IDs have changed. Sandisk's new id
>> will be 0x45, and Toshiba I guess will be 0x11.
>>
>
> Flashbench timings for both Sandisk and Toshiba cards. Attaching due to size.
>
> Some interesting things that I don't understand. For the align test, I
> extended it to do a write align test (-A). I tried two partitions that
> I could write over, and both read and writes behaved differently for
> the two partitions on same device. Odd. They are both 4MB aligned.
>
> On the sandisk it was the write align that made the page size stand
> out. The read align had pretty constant results.
>
> On the toshiba the results varied wildly for the two partitions. For
> partition 6, there was a clear pattern in the diff values for read
> align. For 9, it was all over the place. For 9 with the write align,
> 8K and 16K the crossing writes took ~115ms!! Look in attached files
> for all the data.
>
> The AU tests were interesting too, especially how with several open
> AUs the throughput is higher for certain smaller sizes on sandisk, but
> if I interpret it correctly both cards have at least 4 AUs, as I
> didn't see yet a significant drop for small sizes. The larger ones I
> am running now on mmcblk0p9 which is sufficiently larger for these
> tests... (mmcblk0p6 is only 40mb, p9 is 314 mb)
>
> Thanks,
> A
>
I thought this was pretty interesting -
# echo 0 > /sys/block/mmcblk0/device/page_size
# ./flashbench -A -b 1024 /dev/block/mmcblk0p9
write align 8388608 pre 3.59ms on 6.54ms post 3.65ms diff
2.92ms
write align 4194304 pre 4.13ms on 7.37ms post 4.27ms diff
3.17ms
write align 2097152 pre 3.62ms on 6.81ms post 3.94ms diff
3.03ms
write align 1048576 pre 3.62ms on 6.53ms post 3.55ms diff
2.95ms
write align 524288 pre 3.62ms on 6.51ms post 3.63ms diff
2.88ms
write align 262144 pre 3.62ms on 6.51ms post 3.63ms diff
2.89ms
write align 131072 pre 3.62ms on 6.5ms post 3.63ms diff
2.88ms
write align 65536 pre 3.61ms on 6.49ms post 3.62ms diff
2.88ms
write align 32768 pre 3.61ms on 6.49ms post 3.61ms diff
2.88ms
write align 16384 pre 3.68ms on 107ms post 3.51ms diff
103ms
write align 8192 pre 3.74ms on 121ms post 3.91ms diff
117ms
write align 4096 pre 3.88ms on 3.87ms post 3.87ms diff
-2937ns
write align 2048 pre 3.89ms on 3.88ms post 3.88ms diff
-8734ns
# fjnh84@fjnh84-desktop:~/src/n/src/flash$ adb -s 17006185428011d7 shell
# echo 8192 > /sys/block/mmcblk0/device/page_size
# cd data
# ./flashbench -A -b 1024 /dev/block/mmcblk0p9
write align 8388608 pre 3.33ms on 6.8ms post 3.65ms diff
3.31ms
write align 4194304 pre 4.34ms on 8.14ms post 4.53ms diff
3.71ms
write align 2097152 pre 3.64ms on 7.31ms post 4.09ms diff
3.44ms
write align 1048576 pre 3.65ms on 7.52ms post 3.65ms diff
3.87ms
write align 524288 pre 3.62ms on 6.8ms post 3.63ms diff
3.17ms
write align 262144 pre 3.62ms on 6.84ms post 3.63ms diff
3.22ms
write align 131072 pre 3.62ms on 6.85ms post 3.44ms diff
3.32ms
write align 65536 pre 3.39ms on 6.8ms post 3.66ms diff
3.28ms
write align 32768 pre 3.64ms on 6.86ms post 3.66ms diff
3.21ms
write align 16384 pre 3.67ms on 6.86ms post 3.65ms diff
3.2ms
write align 8192 pre 3.66ms on 6.84ms post 3.64ms diff
3.19ms
write align 4096 pre 3.71ms on 3.71ms post 3.64ms diff
38.6µs
write align 2048 pre 3.71ms on 3.71ms post 3.72ms diff
-656ns
This was with the split unaligned accesses patch... Which I am
attaching for comments.
Thanks,
A
From b3e6a556a716e7cec86071342197e798b38c3cbf Mon Sep 17 00:00:00 2001
From: Andrei Warkentin <andr...@motorola.com>
Date: Fri, 18 Feb 2011 17:46:00 -0600
Subject: [PATCH] MMC: Split non-page-size aligned accesses.
If the card page size is known, splits the access into an unaligned
and an aligned portion, which helps with the performance.
Change-Id: I4ad7588d613d775212fac87436e418577909a22b
Signed-off-by: Andrei Warkentin <andr...@motorola.com>
---
drivers/mmc/card/block.c | 111 ++++++++++++++++++++++++++++++++++++++++++++++
include/linux/mmc/card.h | 1 +
2 files changed, 112 insertions(+), 0 deletions(-)
diff --git a/drivers/mmc/card/block.c b/drivers/mmc/card/block.c
index 7054fd5..be7d739 100644
--- a/drivers/mmc/card/block.c
+++ b/drivers/mmc/card/block.c
@@ -22,6 +22,7 @@
#include <linux/init.h>
#include <linux/kernel.h>
+#include <linux/ctype.h>
#include <linux/fs.h>
#include <linux/slab.h>
#include <linux/errno.h>
@@ -67,6 +68,74 @@ struct mmc_blk_data {
static DEFINE_MUTEX(open_lock);
+static ssize_t
+show_block_attr(struct device *dev, struct device_attribute *attr,
+ char *buf);
+
+static ssize_t
+set_block_attr(struct device *dev, struct device_attribute *attr,
+ const char *buf, size_t count);
+
+static DEVICE_ATTR(page_size, S_IRUGO | S_IWUSR, show_block_attr, set_block_attr);
+
+static ssize_t
+show_block_attr(struct device *dev, struct device_attribute *attr,
+ char *buf)
+{
+ unsigned int val;
+ ssize_t ret = 0;
+ struct mmc_card *card = container_of(dev, struct mmc_card, dev);
+ mmc_claim_host(card->host);
+ if (attr == &dev_attr_page_size)
+ val = card->page_size;
+ else
+ ret = -EINVAL;
+
+ mmc_release_host(card->host);
+ if (!ret)
+ ret = sprintf(buf, "%u\n", val);
+ return ret;
+}
+
+static ssize_t
+set_block_attr(struct device *dev, struct device_attribute *attr,
+ const char *buf, size_t count)
+{
+ ssize_t ret;
+ char *after;
+ unsigned int val, *dest = NULL;
+ struct mmc_card *card = container_of(dev, struct mmc_card, dev);
+ val = simple_strtoul(buf, &after, 10);
+ ret = after - buf;
+
+ while (isspace(*after++))
+ ret++;
+
+ if (ret != count)
+ return -EINVAL;
+
+ if (attr == &dev_attr_page_size)
+ dest = &card->page_size;
+ else
+ return -EINVAL;
+
+ if (dest) {
+ mmc_claim_host(card->host);
+ *dest = val;
+ mmc_release_host(card->host);
+ }
+ return ret;
+}
+
+static struct attribute *capability_attrs[] = {
+ &dev_attr_page_size.attr,
+ NULL,
+};
+
+static struct attribute_group attr_group = {
+ .attrs = capability_attrs,
+};
+
static struct mmc_blk_data *mmc_blk_get(struct gendisk *disk)
{
struct mmc_blk_data *md;
@@ -312,6 +381,38 @@ out:
return err ? 0 : 1;
}
+
+/*
+ * If the request is not aligned, split it into an unaligned
+ * and an aligned portion. Here we can adjust
+ * the size of the MMC request and let the block layer request handle
+ * deal with generating another MMC request.
+ */
+static bool mmc_adjust_write(struct mmc_card *card,
+ struct mmc_request *mrq)
+{
+ unsigned int left_in_page;
+ unsigned int page_size_blocks;
+
+ if (!card->page_size)
+ return false;
+
+ page_size_blocks = card->page_size / mrq->data->blksz;
+ left_in_page = page_size_blocks -
+ (mrq->cmd->arg % page_size_blocks);
+
+ /* Aligned access. */
+ if (left_in_page == page_size_blocks)
+ return false;
+
+ /* Not straddling page boundary. */
+ if (mrq->data->blocks <= left_in_page)
+ return false;
+
+ mrq->data->blocks = left_in_page;
+ return true;
+}
+
static int mmc_blk_issue_rw_rq(struct mmc_queue *mq, struct request *req)
{
struct mmc_blk_data *md = mq->data;
@@ -339,6 +440,10 @@ static int mmc_blk_issue_rw_rq(struct mmc_queue *mq, struct request *req)
brq.stop.flags = MMC_RSP_SPI_R1B | MMC_RSP_R1B | MMC_CMD_AC;
brq.data.blocks = blk_rq_sectors(req);
+ /* Check for unaligned accesses straddling pages. */
+ if (rq_data_dir(req) == WRITE)
+ mmc_adjust_write(card, &brq.mrq);
+
/*
* The block layer doesn't support all sector count
* restrictions, so we need to be prepared for too big
@@ -707,6 +812,10 @@ static int mmc_blk_probe(struct mmc_card *card)
if (err)
goto out;
+ err = sysfs_create_group(&card->dev.kobj, &attr_group);
+ if (err)
+ goto out;
+
string_get_size((u64)get_capacity(md->disk) << 9, STRING_UNITS_2,
cap_str, sizeof(cap_str));
printk(KERN_INFO "%s: %s %s %s %s\n",
@@ -735,6 +844,8 @@ static void mmc_blk_remove(struct mmc_card *card)
/* Stop new requests from getting into the queue */
del_gendisk(md->disk);
+ sysfs_remove_group(&card->dev.kobj, &attr_group);
+
/* Then flush out any already in there */
mmc_cleanup_queue(&md->queue);
diff --git a/include/linux/mmc/card.h b/include/linux/mmc/card.h
index 6b75250..d52768a 100644
--- a/include/linux/mmc/card.h
+++ b/include/linux/mmc/card.h
@@ -123,7 +123,7 @@ struct mmc_card {
unsigned int erase_size; /* erase size in sectors */
unsigned int erase_shift; /* if erase unit is power 2 */
unsigned int pref_erase; /* in sectors */
+ unsigned int page_size; /* page size in bytes */
u8 erased_byte; /* value of erased bytes */
u32 raw_cid[4]; /* raw card CID */
--
1.7.0.4
