[PATCH v7 1/3] mmc: core: Add packed command feature of eMMC4.5

[PATCH v7 1/3] mmc: core: Add packed command feature of eMMC4.5

[Seungwon Jeon]

This patch adds packed command feature of eMMC4.5.
The maximum number for packing read(or write) is offered
and exception event relevant to packed command which is
used for error handling is enabled. If host wants to use
this feature, MMC_CAP2_PACKED_CMD should be set.

Signed-off-by: Seungwon Jeon <tgih.jun@samsung.com>
---
 drivers/mmc/core/mmc.c   |   24 ++++++++++++++++++++++++
 include/linux/mmc/card.h |    3 +++
 include/linux/mmc/host.h |    4 ++++
 include/linux/mmc/mmc.h  |   15 +++++++++++++++
 4 files changed, 46 insertions(+), 0 deletions(-)

diff --git a/drivers/mmc/core/mmc.c b/drivers/mmc/core/mmc.c
index 258b203..f9d54b0 100644
--- a/drivers/mmc/core/mmc.c
+++ b/drivers/mmc/core/mmc.c
@@ -516,6 +516,11 @@ static int mmc_read_ext_csd(struct mmc_card *card, u8 *ext_csd)
 		} else {
 			card->ext_csd.data_tag_unit_size = 0;
 		}
+
+		card->ext_csd.max_packed_writes =
+			ext_csd[EXT_CSD_MAX_PACKED_WRITES];
+		card->ext_csd.max_packed_reads =
+			ext_csd[EXT_CSD_MAX_PACKED_READS];
 	} else {
 		card->ext_csd.data_sector_size = 512;
 	}
@@ -1246,6 +1251,25 @@ static int mmc_init_card(struct mmc_host *host, u32 ocr,
 		}
 	}
 
+	if ((host->caps2 & MMC_CAP2_PACKED_CMD) &&
+			((card->ext_csd.max_packed_writes > 0) ||
+			(card->ext_csd.max_packed_reads > 0))) {
+		err = mmc_switch(card, EXT_CSD_CMD_SET_NORMAL,
+				EXT_CSD_EXP_EVENTS_CTRL,
+				EXT_CSD_PACKED_EVENT_EN,
+				card->ext_csd.generic_cmd6_time);
+		if (err && err != -EBADMSG)
+			goto free_card;
+		if (err) {
+			pr_warning("%s: Enabling packed event failed\n",
+					mmc_hostname(card->host));
+			card->ext_csd.packed_event_en = 0;
+			err = 0;
+		} else {
+			card->ext_csd.packed_event_en = 1;
+		}
+	}
+
 	if (!oldcard)
 		host->card = card;
 
diff --git a/include/linux/mmc/card.h b/include/linux/mmc/card.h
index d76513b..4aeb4e9 100644
--- a/include/linux/mmc/card.h
+++ b/include/linux/mmc/card.h
@@ -53,6 +53,9 @@ struct mmc_ext_csd {
 	u8			part_config;
 	u8			cache_ctrl;
 	u8			rst_n_function;
+	u8			max_packed_writes;
+	u8			max_packed_reads;
+	u8			packed_event_en;
 	unsigned int		part_time;		/* Units: ms */
 	unsigned int		sa_timeout;		/* Units: 100ns */
 	unsigned int		generic_cmd6_time;	/* Units: 10ms */
diff --git a/include/linux/mmc/host.h b/include/linux/mmc/host.h
index 0707d22..9d0d946 100644
--- a/include/linux/mmc/host.h
+++ b/include/linux/mmc/host.h
@@ -238,6 +238,10 @@ struct mmc_host {
 #define MMC_CAP2_BROKEN_VOLTAGE	(1 << 7)	/* Use the broken voltage */
 #define MMC_CAP2_DETECT_ON_ERR	(1 << 8)	/* On I/O err check card removal */
 #define MMC_CAP2_HC_ERASE_SZ	(1 << 9)	/* High-capacity erase size */
+#define MMC_CAP2_PACKED_RD	(1 << 10)	/* Allow packed read */
+#define MMC_CAP2_PACKED_WR	(1 << 11)	/* Allow packed write */
+#define MMC_CAP2_PACKED_CMD	(MMC_CAP2_PACKED_RD | \
+				 MMC_CAP2_PACKED_WR) /* Allow packed commands */
 
 	mmc_pm_flag_t		pm_caps;	/* supported pm features */
 	unsigned int        power_notify_type;
diff --git a/include/linux/mmc/mmc.h b/include/linux/mmc/mmc.h
index d425cab..254901a 100644
--- a/include/linux/mmc/mmc.h
+++ b/include/linux/mmc/mmc.h
@@ -139,6 +139,7 @@ static inline bool mmc_op_multi(u32 opcode)
 #define R1_CURRENT_STATE(x)	((x & 0x00001E00) >> 9)	/* sx, b (4 bits) */
 #define R1_READY_FOR_DATA	(1 << 8)	/* sx, a */
 #define R1_SWITCH_ERROR		(1 << 7)	/* sx, c */
+#define R1_EXP_EVENT		(1 << 6)	/* sr, a */
 #define R1_APP_CMD		(1 << 5)	/* sr, c */
 
 #define R1_STATE_IDLE	0
@@ -274,6 +275,10 @@ struct _mmc_csd {
 #define EXT_CSD_FLUSH_CACHE		32      /* W */
 #define EXT_CSD_CACHE_CTRL		33      /* R/W */
 #define EXT_CSD_POWER_OFF_NOTIFICATION	34	/* R/W */
+#define EXT_CSD_PACKED_FAILURE_INDEX	35	/* RO */
+#define EXT_CSD_PACKED_CMD_STATUS	36	/* RO */
+#define EXT_CSD_EXP_EVENTS_STATUS	54	/* RO, 2 bytes */
+#define EXT_CSD_EXP_EVENTS_CTRL	56	/* R/W, 2 bytes */
 #define EXT_CSD_DATA_SECTOR_SIZE	61	/* R */
 #define EXT_CSD_GP_SIZE_MULT		143	/* R/W */
 #define EXT_CSD_PARTITION_ATTRIBUTE	156	/* R/W */
@@ -318,6 +323,8 @@ struct _mmc_csd {
 #define EXT_CSD_CACHE_SIZE		249	/* RO, 4 bytes */
 #define EXT_CSD_TAG_UNIT_SIZE		498	/* RO */
 #define EXT_CSD_DATA_TAG_SUPPORT	499	/* RO */
+#define EXT_CSD_MAX_PACKED_WRITES	500	/* RO */
+#define EXT_CSD_MAX_PACKED_READS	501	/* RO */
 #define EXT_CSD_HPI_FEATURES		503	/* RO */
 
 /*
@@ -377,6 +384,14 @@ struct _mmc_csd {
 #define EXT_CSD_PWR_CL_4BIT_MASK	0x0F	/* 8 bit PWR CLS */
 #define EXT_CSD_PWR_CL_8BIT_SHIFT	4
 #define EXT_CSD_PWR_CL_4BIT_SHIFT	0
+
+#define EXT_CSD_PACKED_EVENT_EN	(1 << 3)
+
+#define EXT_CSD_PACKED_FAILURE	(1 << 3)
+
+#define EXT_CSD_PACKED_GENERIC_ERROR	(1 << 0)
+#define EXT_CSD_PACKED_INDEXED_ERROR	(1 << 1)
+
 /*
  * MMC_SWITCH access modes
  */
-- 
1.7.0.4

Re: [PATCH v7 1/3] mmc: core: Add packed command feature of eMMC4.5

[merez]

> +	if ((host->caps2 & MMC_CAP2_PACKED_CMD) &&
> +			((card->ext_csd.max_packed_writes > 0) ||
> +			(card->ext_csd.max_packed_reads > 0))) {
> +		err = mmc_switch(card, EXT_CSD_CMD_SET_NORMAL,
> +				EXT_CSD_EXP_EVENTS_CTRL,
> +				EXT_CSD_PACKED_EVENT_EN,
> +				card->ext_csd.generic_cmd6_time);
> +		if (err && err != -EBADMSG)
> +			goto free_card;

Subhash suggestion to change to the following is missing:
     if (  (host->caps2 & MMC_CAP2_PACKED_WR &&
card->ext_csd.max_packed_writes > 0) ||
            (host->caps2 & MMC_CAP2_PACKED_RD &&
card->ext_csd.max_packed_reads > 0)

Thanks,
Maya Erez
Consultant for Qualcomm Innovation Center, Inc.
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum

RE: [PATCH v7 1/3] mmc: core: Add packed command feature of eMMC4.5

[Seungwon Jeon]

Maya Erez <merez@codeaurora.org> wrote:
> > +	if ((host->caps2 & MMC_CAP2_PACKED_CMD) &&
> > +			((card->ext_csd.max_packed_writes > 0) ||
> > +			(card->ext_csd.max_packed_reads > 0))) {
> > +		err = mmc_switch(card, EXT_CSD_CMD_SET_NORMAL,
> > +				EXT_CSD_EXP_EVENTS_CTRL,
> > +				EXT_CSD_PACKED_EVENT_EN,
> > +				card->ext_csd.generic_cmd6_time);
> > +		if (err && err != -EBADMSG)
> > +			goto free_card;
> 
> Subhash suggestion to change to the following is missing:
>      if (  (host->caps2 & MMC_CAP2_PACKED_WR &&
> card->ext_csd.max_packed_writes > 0) ||
>             (host->caps2 & MMC_CAP2_PACKED_RD &&
> card->ext_csd.max_packed_reads > 0)
Thanks to detect.
It'll be applied.

Best regards,
Seungwon Jeon

> 
> Thanks,
> Maya Erez
> Consultant for Qualcomm Innovation Center, Inc.
> Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-mmc" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH v7 1/3] mmc: core: Add packed command feature of eMMC4.5

[S, Venkatraman]

On Mon, Jun 11, 2012 at 1:53 PM, Seungwon Jeon <tgih.jun@samsung.com> wrote:
> This patch adds packed command feature of eMMC4.5.
> The maximum number for packing read(or write) is offered
> and exception event relevant to packed command which is
> used for error handling is enabled. If host wants to use
> this feature, MMC_CAP2_PACKED_CMD should be set.
>
> Signed-off-by: Seungwon Jeon <tgih.jun@samsung.com>

Can you please post some clear performance benchmarks with your patchset ?
Given that #merez claims to see a significant performance drop for
reads, it will be
good to compare notes.
If it's not too much trouble, both CFQ and deadline scheduler figures
would be useful, on a
set of read only, write only and parallel read write usecases.

I can also try to replicate your results if you can publish the exact
configuration you used
for testing (example: iozone parameters)

> ---
>  drivers/mmc/core/mmc.c   |   24 ++++++++++++++++++++++++
>  include/linux/mmc/card.h |    3 +++
>  include/linux/mmc/host.h |    4 ++++
>  include/linux/mmc/mmc.h  |   15 +++++++++++++++
>  4 files changed, 46 insertions(+), 0 deletions(-)
>
> diff --git a/drivers/mmc/core/mmc.c b/drivers/mmc/core/mmc.c
> index 258b203..f9d54b0 100644
> --- a/drivers/mmc/core/mmc.c
> +++ b/drivers/mmc/core/mmc.c
> @@ -516,6 +516,11 @@ static int mmc_read_ext_csd(struct mmc_card *card, u8 *ext_csd)
>                } else {
>                        card->ext_csd.data_tag_unit_size = 0;
>                }
> +
> +               card->ext_csd.max_packed_writes =
> +                       ext_csd[EXT_CSD_MAX_PACKED_WRITES];
> +               card->ext_csd.max_packed_reads =
> +                       ext_csd[EXT_CSD_MAX_PACKED_READS];
>        } else {
>                card->ext_csd.data_sector_size = 512;
>        }
> @@ -1246,6 +1251,25 @@ static int mmc_init_card(struct mmc_host *host, u32 ocr,
>                }
>        }
>
> +       if ((host->caps2 & MMC_CAP2_PACKED_CMD) &&
> +                       ((card->ext_csd.max_packed_writes > 0) ||
> +                       (card->ext_csd.max_packed_reads > 0))) {
> +               err = mmc_switch(card, EXT_CSD_CMD_SET_NORMAL,
> +                               EXT_CSD_EXP_EVENTS_CTRL,
> +                               EXT_CSD_PACKED_EVENT_EN,
> +                               card->ext_csd.generic_cmd6_time);
> +               if (err && err != -EBADMSG)
> +                       goto free_card;
> +               if (err) {
> +                       pr_warning("%s: Enabling packed event failed\n",
> +                                       mmc_hostname(card->host));
> +                       card->ext_csd.packed_event_en = 0;
> +                       err = 0;
> +               } else {
> +                       card->ext_csd.packed_event_en = 1;
> +               }
> +       }
> +
>        if (!oldcard)
>                host->card = card;
>
> diff --git a/include/linux/mmc/card.h b/include/linux/mmc/card.h
> index d76513b..4aeb4e9 100644
> --- a/include/linux/mmc/card.h
> +++ b/include/linux/mmc/card.h
> @@ -53,6 +53,9 @@ struct mmc_ext_csd {
>        u8                      part_config;
>        u8                      cache_ctrl;
>        u8                      rst_n_function;
> +       u8                      max_packed_writes;
> +       u8                      max_packed_reads;
> +       u8                      packed_event_en;
>        unsigned int            part_time;              /* Units: ms */
>        unsigned int            sa_timeout;             /* Units: 100ns */
>        unsigned int            generic_cmd6_time;      /* Units: 10ms */
> diff --git a/include/linux/mmc/host.h b/include/linux/mmc/host.h
> index 0707d22..9d0d946 100644
> --- a/include/linux/mmc/host.h
> +++ b/include/linux/mmc/host.h
> @@ -238,6 +238,10 @@ struct mmc_host {
>  #define MMC_CAP2_BROKEN_VOLTAGE        (1 << 7)        /* Use the broken voltage */
>  #define MMC_CAP2_DETECT_ON_ERR (1 << 8)        /* On I/O err check card removal */
>  #define MMC_CAP2_HC_ERASE_SZ   (1 << 9)        /* High-capacity erase size */
> +#define MMC_CAP2_PACKED_RD     (1 << 10)       /* Allow packed read */
> +#define MMC_CAP2_PACKED_WR     (1 << 11)       /* Allow packed write */
> +#define MMC_CAP2_PACKED_CMD    (MMC_CAP2_PACKED_RD | \
> +                                MMC_CAP2_PACKED_WR) /* Allow packed commands */
>
>        mmc_pm_flag_t           pm_caps;        /* supported pm features */
>        unsigned int        power_notify_type;
> diff --git a/include/linux/mmc/mmc.h b/include/linux/mmc/mmc.h
> index d425cab..254901a 100644
> --- a/include/linux/mmc/mmc.h
> +++ b/include/linux/mmc/mmc.h
> @@ -139,6 +139,7 @@ static inline bool mmc_op_multi(u32 opcode)
>  #define R1_CURRENT_STATE(x)    ((x & 0x00001E00) >> 9) /* sx, b (4 bits) */
>  #define R1_READY_FOR_DATA      (1 << 8)        /* sx, a */
>  #define R1_SWITCH_ERROR                (1 << 7)        /* sx, c */
> +#define R1_EXP_EVENT           (1 << 6)        /* sr, a */
>  #define R1_APP_CMD             (1 << 5)        /* sr, c */
>
>  #define R1_STATE_IDLE  0
> @@ -274,6 +275,10 @@ struct _mmc_csd {
>  #define EXT_CSD_FLUSH_CACHE            32      /* W */
>  #define EXT_CSD_CACHE_CTRL             33      /* R/W */
>  #define EXT_CSD_POWER_OFF_NOTIFICATION 34      /* R/W */
> +#define EXT_CSD_PACKED_FAILURE_INDEX   35      /* RO */
> +#define EXT_CSD_PACKED_CMD_STATUS      36      /* RO */
> +#define EXT_CSD_EXP_EVENTS_STATUS      54      /* RO, 2 bytes */
> +#define EXT_CSD_EXP_EVENTS_CTRL        56      /* R/W, 2 bytes */
>  #define EXT_CSD_DATA_SECTOR_SIZE       61      /* R */
>  #define EXT_CSD_GP_SIZE_MULT           143     /* R/W */
>  #define EXT_CSD_PARTITION_ATTRIBUTE    156     /* R/W */
> @@ -318,6 +323,8 @@ struct _mmc_csd {
>  #define EXT_CSD_CACHE_SIZE             249     /* RO, 4 bytes */
>  #define EXT_CSD_TAG_UNIT_SIZE          498     /* RO */
>  #define EXT_CSD_DATA_TAG_SUPPORT       499     /* RO */
> +#define EXT_CSD_MAX_PACKED_WRITES      500     /* RO */
> +#define EXT_CSD_MAX_PACKED_READS       501     /* RO */
>  #define EXT_CSD_HPI_FEATURES           503     /* RO */
>
>  /*
> @@ -377,6 +384,14 @@ struct _mmc_csd {
>  #define EXT_CSD_PWR_CL_4BIT_MASK       0x0F    /* 8 bit PWR CLS */
>  #define EXT_CSD_PWR_CL_8BIT_SHIFT      4
>  #define EXT_CSD_PWR_CL_4BIT_SHIFT      0
> +
> +#define EXT_CSD_PACKED_EVENT_EN        (1 << 3)
> +
> +#define EXT_CSD_PACKED_FAILURE (1 << 3)
> +
> +#define EXT_CSD_PACKED_GENERIC_ERROR   (1 << 0)
> +#define EXT_CSD_PACKED_INDEXED_ERROR   (1 << 1)
> +
>  /*
>  * MMC_SWITCH access modes
>  */
> --
> 1.7.0.4
>
>

RE: [PATCH v7 1/3] mmc: core: Add packed command feature of eMMC4.5

[Seungwon Jeon]

S, Venkatraman <svenkatr@ti.com> wrote:
> On Mon, Jun 11, 2012 at 1:53 PM, Seungwon Jeon <tgih.jun@samsung.com> wrote:
> > This patch adds packed command feature of eMMC4.5.
> > The maximum number for packing read(or write) is offered
> > and exception event relevant to packed command which is
> > used for error handling is enabled. If host wants to use
> > this feature, MMC_CAP2_PACKED_CMD should be set.
> >
> > Signed-off-by: Seungwon Jeon <tgih.jun@samsung.com>
> 
> Can you please post some clear performance benchmarks with your patchset ?
> Given that #merez claims to see a significant performance drop for
> reads, it will be
> good to compare notes.
> If it's not too much trouble, both CFQ and deadline scheduler figures
> would be useful, on a
> set of read only, write only and parallel read write usecases.
> 
> I can also try to replicate your results if you can publish the exact
> configuration you used
> for testing (example: iozone parameters)
I'm checking the merez's result.
Currently packed command is effective on write.
When running packed write with iozone, there is 25% performance gains.
(ex: iozone -az -i0 -I -s 10m -f /target/test -e)

> 
> > ---
> >  drivers/mmc/core/mmc.c   |   24 ++++++++++++++++++++++++
> >  include/linux/mmc/card.h |    3 +++
> >  include/linux/mmc/host.h |    4 ++++
> >  include/linux/mmc/mmc.h  |   15 +++++++++++++++
> >  4 files changed, 46 insertions(+), 0 deletions(-)
> >
> > diff --git a/drivers/mmc/core/mmc.c b/drivers/mmc/core/mmc.c
> > index 258b203..f9d54b0 100644
> > --- a/drivers/mmc/core/mmc.c
> > +++ b/drivers/mmc/core/mmc.c
> > @@ -516,6 +516,11 @@ static int mmc_read_ext_csd(struct mmc_card *card, u8 *ext_csd)
> >                } else {
> >                        card->ext_csd.data_tag_unit_size = 0;
> >                }
> > +
> > +               card->ext_csd.max_packed_writes =
> > +                       ext_csd[EXT_CSD_MAX_PACKED_WRITES];
> > +               card->ext_csd.max_packed_reads =
> > +                       ext_csd[EXT_CSD_MAX_PACKED_READS];
> >        } else {
> >                card->ext_csd.data_sector_size = 512;
> >        }
> > @@ -1246,6 +1251,25 @@ static int mmc_init_card(struct mmc_host *host, u32 ocr,
> >                }
> >        }
> >
> > +       if ((host->caps2 & MMC_CAP2_PACKED_CMD) &&
> > +                       ((card->ext_csd.max_packed_writes > 0) ||
> > +                       (card->ext_csd.max_packed_reads > 0))) {
> > +               err = mmc_switch(card, EXT_CSD_CMD_SET_NORMAL,
> > +                               EXT_CSD_EXP_EVENTS_CTRL,
> > +                               EXT_CSD_PACKED_EVENT_EN,
> > +                               card->ext_csd.generic_cmd6_time);
> > +               if (err && err != -EBADMSG)
> > +                       goto free_card;
> > +               if (err) {
> > +                       pr_warning("%s: Enabling packed event failed\n",
> > +                                       mmc_hostname(card->host));
> > +                       card->ext_csd.packed_event_en = 0;
> > +                       err = 0;
> > +               } else {
> > +                       card->ext_csd.packed_event_en = 1;
> > +               }
> > +       }
> > +
> >        if (!oldcard)
> >                host->card = card;
> >
> > diff --git a/include/linux/mmc/card.h b/include/linux/mmc/card.h
> > index d76513b..4aeb4e9 100644
> > --- a/include/linux/mmc/card.h
> > +++ b/include/linux/mmc/card.h
> > @@ -53,6 +53,9 @@ struct mmc_ext_csd {
> >        u8                      part_config;
> >        u8                      cache_ctrl;
> >        u8                      rst_n_function;
> > +       u8                      max_packed_writes;
> > +       u8                      max_packed_reads;
> > +       u8                      packed_event_en;
> >        unsigned int            part_time;              /* Units: ms */
> >        unsigned int            sa_timeout;             /* Units: 100ns */
> >        unsigned int            generic_cmd6_time;      /* Units: 10ms */
> > diff --git a/include/linux/mmc/host.h b/include/linux/mmc/host.h
> > index 0707d22..9d0d946 100644
> > --- a/include/linux/mmc/host.h
> > +++ b/include/linux/mmc/host.h
> > @@ -238,6 +238,10 @@ struct mmc_host {
> >  #define MMC_CAP2_BROKEN_VOLTAGE        (1 << 7)        /* Use the broken voltage */
> >  #define MMC_CAP2_DETECT_ON_ERR (1 << 8)        /* On I/O err check card removal */
> >  #define MMC_CAP2_HC_ERASE_SZ   (1 << 9)        /* High-capacity erase size */
> > +#define MMC_CAP2_PACKED_RD     (1 << 10)       /* Allow packed read */
> > +#define MMC_CAP2_PACKED_WR     (1 << 11)       /* Allow packed write */
> > +#define MMC_CAP2_PACKED_CMD    (MMC_CAP2_PACKED_RD | \
> > +                                MMC_CAP2_PACKED_WR) /* Allow packed commands */
> >
> >        mmc_pm_flag_t           pm_caps;        /* supported pm features */
> >        unsigned int        power_notify_type;
> > diff --git a/include/linux/mmc/mmc.h b/include/linux/mmc/mmc.h
> > index d425cab..254901a 100644
> > --- a/include/linux/mmc/mmc.h
> > +++ b/include/linux/mmc/mmc.h
> > @@ -139,6 +139,7 @@ static inline bool mmc_op_multi(u32 opcode)
> >  #define R1_CURRENT_STATE(x)    ((x & 0x00001E00) >> 9) /* sx, b (4 bits) */
> >  #define R1_READY_FOR_DATA      (1 << 8)        /* sx, a */
> >  #define R1_SWITCH_ERROR                (1 << 7)        /* sx, c */
> > +#define R1_EXP_EVENT           (1 << 6)        /* sr, a */
> >  #define R1_APP_CMD             (1 << 5)        /* sr, c */
> >
> >  #define R1_STATE_IDLE  0
> > @@ -274,6 +275,10 @@ struct _mmc_csd {
> >  #define EXT_CSD_FLUSH_CACHE            32      /* W */
> >  #define EXT_CSD_CACHE_CTRL             33      /* R/W */
> >  #define EXT_CSD_POWER_OFF_NOTIFICATION 34      /* R/W */
> > +#define EXT_CSD_PACKED_FAILURE_INDEX   35      /* RO */
> > +#define EXT_CSD_PACKED_CMD_STATUS      36      /* RO */
> > +#define EXT_CSD_EXP_EVENTS_STATUS      54      /* RO, 2 bytes */
> > +#define EXT_CSD_EXP_EVENTS_CTRL        56      /* R/W, 2 bytes */
> >  #define EXT_CSD_DATA_SECTOR_SIZE       61      /* R */
> >  #define EXT_CSD_GP_SIZE_MULT           143     /* R/W */
> >  #define EXT_CSD_PARTITION_ATTRIBUTE    156     /* R/W */
> > @@ -318,6 +323,8 @@ struct _mmc_csd {
> >  #define EXT_CSD_CACHE_SIZE             249     /* RO, 4 bytes */
> >  #define EXT_CSD_TAG_UNIT_SIZE          498     /* RO */
> >  #define EXT_CSD_DATA_TAG_SUPPORT       499     /* RO */
> > +#define EXT_CSD_MAX_PACKED_WRITES      500     /* RO */
> > +#define EXT_CSD_MAX_PACKED_READS       501     /* RO */
> >  #define EXT_CSD_HPI_FEATURES           503     /* RO */
> >
> >  /*
> > @@ -377,6 +384,14 @@ struct _mmc_csd {
> >  #define EXT_CSD_PWR_CL_4BIT_MASK       0x0F    /* 8 bit PWR CLS */
> >  #define EXT_CSD_PWR_CL_8BIT_SHIFT      4
> >  #define EXT_CSD_PWR_CL_4BIT_SHIFT      0
> > +
> > +#define EXT_CSD_PACKED_EVENT_EN        (1 << 3)
> > +
> > +#define EXT_CSD_PACKED_FAILURE (1 << 3)
> > +
> > +#define EXT_CSD_PACKED_GENERIC_ERROR   (1 << 0)
> > +#define EXT_CSD_PACKED_INDEXED_ERROR   (1 << 1)
> > +
> >  /*
> >  * MMC_SWITCH access modes
> >  */
> > --
> > 1.7.0.4
> >
> >
> --
> To unsubscribe from this list: send the line "unsubscribe linux-mmc" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

RE: [PATCH v7 1/3] mmc: core: Add packed command feature of eMMC4.5

[merez]

> S, Venkatraman <svenkatr@ti.com> wrote:
>> On Mon, Jun 11, 2012 at 1:53 PM, Seungwon Jeon <tgih.jun@samsung.com>
>> wrote:
>> > This patch adds packed command feature of eMMC4.5.
>> > The maximum number for packing read(or write) is offered
>> > and exception event relevant to packed command which is
>> > used for error handling is enabled. If host wants to use
>> > this feature, MMC_CAP2_PACKED_CMD should be set.
>> >
>> > Signed-off-by: Seungwon Jeon <tgih.jun@samsung.com>
>>
>> Can you please post some clear performance benchmarks with your patchset
>> ?
>> Given that #merez claims to see a significant performance drop for
>> reads, it will be
>> good to compare notes.
>> If it's not too much trouble, both CFQ and deadline scheduler figures
>> would be useful, on a
>> set of read only, write only and parallel read write usecases.
>>
>> I can also try to replicate your results if you can publish the exact
>> configuration you used
>> for testing (example: iozone parameters)
> I'm checking the merez's result.
> Currently packed command is effective on write.
> When running packed write with iozone, there is 25% performance gains.
> (ex: iozone -az -i0 -I -s 10m -f /target/test -e)
>
Our tests shows performance gain of 50-60% in scenarios of only write lmdd
operations.

As I mentioned in the write packing control thread the degradation of read
performance in case of mix read/write operations appears also without
write packing. Therefore I don't think it should stop us from approving
the write packing patch, that gives a significant improvement to the write
performance.
The read performance degradation should be resolved regardless of the
write packing patch.

Thanks,
Maya Erez
--
Sent by consultant of Qualcomm Innovation Center, Inc.
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum

Re: [PATCH v7 1/3] mmc: core: Add packed command feature of eMMC4.5

[S, Venkatraman]

On Wed, Jun 13, 2012 at 12:45 AM,  <merez@codeaurora.org> wrote:
>
>> S, Venkatraman <svenkatr@ti.com> wrote:
>>> On Mon, Jun 11, 2012 at 1:53 PM, Seungwon Jeon <tgih.jun@samsung.com>
>>> wrote:
>>> > This patch adds packed command feature of eMMC4.5.
>>> > The maximum number for packing read(or write) is offered
>>> > and exception event relevant to packed command which is
>>> > used for error handling is enabled. If host wants to use
>>> > this feature, MMC_CAP2_PACKED_CMD should be set.
>>> >
>>> > Signed-off-by: Seungwon Jeon <tgih.jun@samsung.com>
>>>
>>> Can you please post some clear performance benchmarks with your patchset
>>> ?
>>> Given that #merez claims to see a significant performance drop for
>>> reads, it will be
>>> good to compare notes.
>>> If it's not too much trouble, both CFQ and deadline scheduler figures
>>> would be useful, on a
>>> set of read only, write only and parallel read write usecases.
>>>
>>> I can also try to replicate your results if you can publish the exact
>>> configuration you used
>>> for testing (example: iozone parameters)
>> I'm checking the merez's result.
>> Currently packed command is effective on write.
>> When running packed write with iozone, there is 25% performance gains.
>> (ex: iozone -az -i0 -I -s 10m -f /target/test -e)
>>
> Our tests shows performance gain of 50-60% in scenarios of only write lmdd
> operations.
>
> As I mentioned in the write packing control thread the degradation of read
> performance in case of mix read/write operations appears also without
> write packing. Therefore I don't think it should stop us from approving
> the write packing patch, that gives a significant improvement to the write
> performance.
> The read performance degradation should be resolved regardless of the
> write packing patch.
>

One further question - when you say "degradation of read performance
in case of mix
read/write operations appears also without write packing", what
exactly does that mean?
Degradation w.r.to to read-only test ? Or any expected throughput ?

If the scenario you mention is accurate, I was actually thinking that
we should recommend to merge
read packing first, then merge write packing once the read performance
issue is well understood.

I am all for better performance with packing control etc, but the
overall code complexity is really
increasing more than necessary. I want to make sure that it is really
worth the effort.

Thanks,
Venkat.

Re: [PATCH v7 1/3] mmc: core: Add packed command feature of eMMC4.5

[merez]

On Wed, June 13, 2012 12:49 pm, S, Venkatraman wrote:
> On Wed, Jun 13, 2012 at 12:45 AM,  <merez@codeaurora.org> wrote:
>>
>>> S, Venkatraman <svenkatr@ti.com> wrote:
>>>> On Mon, Jun 11, 2012 at 1:53 PM, Seungwon Jeon <tgih.jun@samsung.com>
>>>> wrote:
>>>> > This patch adds packed command feature of eMMC4.5.
>>>> > The maximum number for packing read(or write) is offered
>>>> > and exception event relevant to packed command which is
>>>> > used for error handling is enabled. If host wants to use
>>>> > this feature, MMC_CAP2_PACKED_CMD should be set.
>>>> >
>>>> > Signed-off-by: Seungwon Jeon <tgih.jun@samsung.com>
>>>>
>>>> Can you please post some clear performance benchmarks with your
>>>> patchset
>>>> ?
>>>> Given that #merez claims to see a significant performance drop for
>>>> reads, it will be
>>>> good to compare notes.
>>>> If it's not too much trouble, both CFQ and deadline scheduler figures
>>>> would be useful, on a
>>>> set of read only, write only and parallel read write usecases.
>>>>
>>>> I can also try to replicate your results if you can publish the exact
>>>> configuration you used
>>>> for testing (example: iozone parameters)
>>> I'm checking the merez's result.
>>> Currently packed command is effective on write.
>>> When running packed write with iozone, there is 25% performance gains.
>>> (ex: iozone -az -i0 -I -s 10m -f /target/test -e)
>>>
>> Our tests shows performance gain of 50-60% in scenarios of only write
>> lmdd
>> operations.
>>
>> As I mentioned in the write packing control thread the degradation of
>> read
>> performance in case of mix read/write operations appears also without
>> write packing. Therefore I don't think it should stop us from approving
>> the write packing patch, that gives a significant improvement to the
>> write
>> performance.
>> The read performance degradation should be resolved regardless of the
>> write packing patch.
>>
>
> One further question - when you say "degradation of read performance
> in case of mix
> read/write operations appears also without write packing", what
> exactly does that mean?
> Degradation w.r.to to read-only test ? Or any expected throughput ?

I meant w.r. to read only test.

>
> If the scenario you mention is accurate, I was actually thinking that
> we should recommend to merge
> read packing first, then merge write packing once the read performance
> issue is well understood.

I don't know if you followed the early discussion of this patch but the
read throughput was not proved as efficient and in some of the cases it
also caused degradation of the read performance. Therefore, we don't
intend to merge it yet.
>
> I am all for better performance with packing control etc, but the
> overall code complexity is really
> increasing more than necessary. I want to make sure that it is really
> worth the effort.

In my opinion a gain of 50%-60% of the write performance worth the
complexity of the code and the effort to fix the issues it reveals.

>
> Thanks,
> Venkat.
>


-- 
Sent by consultant of Qualcomm Innovation Center, Inc.
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum

RE: [PATCH v7 1/3] mmc: core: Add packed command feature of eMMC4.5

[Seungwon Jeon]

Maya Erez <merez@codeaurora.org> wrote:
> On Wed, June 13, 2012 12:49 pm, S, Venkatraman wrote:
> > On Wed, Jun 13, 2012 at 12:45 AM,  <merez@codeaurora.org> wrote:
> >>
> >>> S, Venkatraman <svenkatr@ti.com> wrote:
> >>>> On Mon, Jun 11, 2012 at 1:53 PM, Seungwon Jeon <tgih.jun@samsung.com>
> >>>> wrote:
> >>>> > This patch adds packed command feature of eMMC4.5.
> >>>> > The maximum number for packing read(or write) is offered
> >>>> > and exception event relevant to packed command which is
> >>>> > used for error handling is enabled. If host wants to use
> >>>> > this feature, MMC_CAP2_PACKED_CMD should be set.
> >>>> >
> >>>> > Signed-off-by: Seungwon Jeon <tgih.jun@samsung.com>
> >>>>
> >>>> Can you please post some clear performance benchmarks with your
> >>>> patchset
> >>>> ?
> >>>> Given that #merez claims to see a significant performance drop for
> >>>> reads, it will be
> >>>> good to compare notes.
> >>>> If it's not too much trouble, both CFQ and deadline scheduler figures
> >>>> would be useful, on a
> >>>> set of read only, write only and parallel read write usecases.
> >>>>
> >>>> I can also try to replicate your results if you can publish the exact
> >>>> configuration you used
> >>>> for testing (example: iozone parameters)
> >>> I'm checking the merez's result.
> >>> Currently packed command is effective on write.
> >>> When running packed write with iozone, there is 25% performance gains.
> >>> (ex: iozone -az -i0 -I -s 10m -f /target/test -e)
> >>>
> >> Our tests shows performance gain of 50-60% in scenarios of only write
> >> lmdd
> >> operations.
> >>
> >> As I mentioned in the write packing control thread the degradation of
> >> read
> >> performance in case of mix read/write operations appears also without
> >> write packing. Therefore I don't think it should stop us from approving
> >> the write packing patch, that gives a significant improvement to the
> >> write
> >> performance.
> >> The read performance degradation should be resolved regardless of the
> >> write packing patch.
> >>
> >
> > One further question - when you say "degradation of read performance
> > in case of mix
> > read/write operations appears also without write packing", what
> > exactly does that mean?
> > Degradation w.r.to to read-only test ? Or any expected throughput ?
> 
> I meant w.r. to read only test.
> 
> >
> > If the scenario you mention is accurate, I was actually thinking that
> > we should recommend to merge
> > read packing first, then merge write packing once the read performance
> > issue is well understood.
> 
> I don't know if you followed the early discussion of this patch but the
> read throughput was not proved as efficient and in some of the cases it
> also caused degradation of the read performance. Therefore, we don't
> intend to merge it yet.
As I have mentioned in previous mailing, eMMC device which is tested with this patch
is not optimized for packed read. So currently it is difficult to ensure that
packed read is effective for performance. We need to test various vendor device
in regard to packed read.

Thanks,
Seungwon Jeon

> >
> > I am all for better performance with packing control etc, but the
> > overall code complexity is really
> > increasing more than necessary. I want to make sure that it is really
> > worth the effort.
> 
> In my opinion a gain of 50%-60% of the write performance worth the
> complexity of the code and the effort to fix the issues it reveals.
> 
> >
> > Thanks,
> > Venkat.
> >
> 
> 
> --
> Sent by consultant of Qualcomm Innovation Center, Inc.
> Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum