linux-sh.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH v2] mmc: tmio: enable odd number size access
@ 2014-09-10  1:29 Kuninori Morimoto
  2014-09-10  7:11 ` Geert Uytterhoeven
  0 siblings, 1 reply; 3+ messages in thread
From: Kuninori Morimoto @ 2014-09-10  1:29 UTC (permalink / raw)
  To: Ulf Hansson, Chris Ball
  Cc: Simon, Kuninori Morimoto, Linux-SH, linux-mmc, Shinobu Uehara

From: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>

Current tmio is using sd_ctrl_read16/write16_rep()
for data transfer.
It works if transfer size was even number,
but, last 1 byte will be ignored if
transfer size was odd number.
This patch adds new tmio_mmc_transfer_data()
and solve this issue.

Tested-by: Shinobu Uehara <shinobu.uehara.xc@renesas.com>
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
---
v1 -> v2

 - use u8 pointer only
 - fix buf offset
 - remove unneeded -1

 drivers/mmc/host/tmio_mmc_pio.c |   46 +++++++++++++++++++++++++++++++++++----
 1 file changed, 42 insertions(+), 4 deletions(-)

diff --git a/drivers/mmc/host/tmio_mmc_pio.c b/drivers/mmc/host/tmio_mmc_pio.c
index ff5ff0f..2ee0e21 100644
--- a/drivers/mmc/host/tmio_mmc_pio.c
+++ b/drivers/mmc/host/tmio_mmc_pio.c
@@ -376,6 +376,47 @@ static int tmio_mmc_start_command(struct tmio_mmc_host *host, struct mmc_command
 	return 0;
 }
 
+static void tmio_mmc_transfer_data(struct tmio_mmc_host *host,
+				   unsigned short *buf,
+				   unsigned int count)
+{
+	int is_read = host->data->flags & MMC_DATA_READ;
+	u16 extra;
+	u8  *buf8;
+
+	/*
+	 * Transfer the data
+	 */
+	if (is_read)
+		sd_ctrl_read16_rep(host, CTL_SD_DATA_PORT, buf, count >> 1);
+	else
+		sd_ctrl_write16_rep(host, CTL_SD_DATA_PORT, buf, count >> 1);
+
+	/* count was even number */
+	if (!(count & 0x1))
+		return;
+
+	/* count was odd number */
+
+	buf8 = (u8 *)(buf + (count >> 1));
+
+	/*
+	 * FIXME
+	 *
+	 * driver and this function are assuming that
+	 * it is used as little endian
+	 */
+	if (is_read) {
+		extra = sd_ctrl_read16(host, CTL_SD_DATA_PORT);
+
+		*buf8 = (u8)(extra & 0xff);
+	} else {
+		extra = (u16)(*buf8);
+
+		sd_ctrl_write16(host, CTL_SD_DATA_PORT, extra);
+	}
+}
+
 /*
  * This chip always returns (at least?) as much data as you ask for.
  * I'm unsure what happens if you ask for less than a block. This should be
@@ -408,10 +449,7 @@ static void tmio_mmc_pio_irq(struct tmio_mmc_host *host)
 		 count, host->sg_off, data->flags);
 
 	/* Transfer the data */
-	if (data->flags & MMC_DATA_READ)
-		sd_ctrl_read16_rep(host, CTL_SD_DATA_PORT, buf, count >> 1);
-	else
-		sd_ctrl_write16_rep(host, CTL_SD_DATA_PORT, buf, count >> 1);
+	tmio_mmc_transfer_data(host, buf, count);
 
 	host->sg_off += count;
 
-- 
1.7.9.5


^ permalink raw reply related	[flat|nested] 3+ messages in thread

* Re: [PATCH v2] mmc: tmio: enable odd number size access
  2014-09-10  1:29 [PATCH v2] mmc: tmio: enable odd number size access Kuninori Morimoto
@ 2014-09-10  7:11 ` Geert Uytterhoeven
  2014-09-10  7:23   ` [PATCH v3] " Kuninori Morimoto
  0 siblings, 1 reply; 3+ messages in thread
From: Geert Uytterhoeven @ 2014-09-10  7:11 UTC (permalink / raw)
  To: Kuninori Morimoto
  Cc: Ulf Hansson, Chris Ball, Simon, Kuninori Morimoto, Linux-SH,
	linux-mmc, Shinobu Uehara

Hi Morimoto-san,

On Wed, Sep 10, 2014 at 3:29 AM, Kuninori Morimoto
<kuninori.morimoto.gx@gmail.com> wrote:
> +       if (is_read) {
> +               extra = sd_ctrl_read16(host, CTL_SD_DATA_PORT);
> +
> +               *buf8 = (u8)(extra & 0xff);
> +       } else {
> +               extra = (u16)(*buf8);
> +
> +               sd_ctrl_write16(host, CTL_SD_DATA_PORT, extra);

The casts to "u8" resp. "u16" are not needed.

You can even avoid introducing the "extra" variable:

if (is_read)
        *buf8 = sd_ctrl_read16(host, CTL_SD_DATA_PORT) & 0xff;
else
        sd_ctrl_write16(host, CTL_SD_DATA_PORT, *buf8);

Technically, the "& 0xff" is also not needed, but I'd keep it for clarity.

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

^ permalink raw reply	[flat|nested] 3+ messages in thread

* [PATCH v3] mmc: tmio: enable odd number size access
  2014-09-10  7:11 ` Geert Uytterhoeven
@ 2014-09-10  7:23   ` Kuninori Morimoto
  0 siblings, 0 replies; 3+ messages in thread
From: Kuninori Morimoto @ 2014-09-10  7:23 UTC (permalink / raw)
  To: Ulf Hansson
  Cc: Geert Uytterhoeven, Chris Ball, Simon, Kuninori Morimoto,
	Linux-SH, linux-mmc, Shinobu Uehara

From: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>

Current tmio is using sd_ctrl_read16/write16_rep()
for data transfer.
It works if transfer size was even number,
but, last 1 byte will be ignored if
transfer size was odd number.
This patch adds new tmio_mmc_transfer_data()
and solve this issue.

Tested-by: Shinobu Uehara <shinobu.uehara.xc@renesas.com>
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
---
v2 -> v3

 - remove cast
 - remove extra variable

 drivers/mmc/host/tmio_mmc_pio.c |   39 +++++++++++++++++++++++++++++++++++----
 1 file changed, 35 insertions(+), 4 deletions(-)

diff --git a/drivers/mmc/host/tmio_mmc_pio.c b/drivers/mmc/host/tmio_mmc_pio.c
index ff5ff0f..692e578 100644
--- a/drivers/mmc/host/tmio_mmc_pio.c
+++ b/drivers/mmc/host/tmio_mmc_pio.c
@@ -376,6 +376,40 @@ static int tmio_mmc_start_command(struct tmio_mmc_host *host, struct mmc_command
 	return 0;
 }
 
+static void tmio_mmc_transfer_data(struct tmio_mmc_host *host,
+				   unsigned short *buf,
+				   unsigned int count)
+{
+	int is_read = host->data->flags & MMC_DATA_READ;
+	u8  *buf8;
+
+	/*
+	 * Transfer the data
+	 */
+	if (is_read)
+		sd_ctrl_read16_rep(host, CTL_SD_DATA_PORT, buf, count >> 1);
+	else
+		sd_ctrl_write16_rep(host, CTL_SD_DATA_PORT, buf, count >> 1);
+
+	/* if count was even number */
+	if (!(count & 0x1))
+		return;
+
+	/* if count was odd number */
+	buf8 = (u8 *)(buf + (count >> 1));
+
+	/*
+	 * FIXME
+	 *
+	 * driver and this function are assuming that
+	 * it is used as little endian
+	 */
+	if (is_read)
+		*buf8 = sd_ctrl_read16(host, CTL_SD_DATA_PORT) & 0xff;
+	else
+		sd_ctrl_write16(host, CTL_SD_DATA_PORT, *buf8);
+}
+
 /*
  * This chip always returns (at least?) as much data as you ask for.
  * I'm unsure what happens if you ask for less than a block. This should be
@@ -408,10 +442,7 @@ static void tmio_mmc_pio_irq(struct tmio_mmc_host *host)
 		 count, host->sg_off, data->flags);
 
 	/* Transfer the data */
-	if (data->flags & MMC_DATA_READ)
-		sd_ctrl_read16_rep(host, CTL_SD_DATA_PORT, buf, count >> 1);
-	else
-		sd_ctrl_write16_rep(host, CTL_SD_DATA_PORT, buf, count >> 1);
+	tmio_mmc_transfer_data(host, buf, count);
 
 	host->sg_off += count;
 
-- 
1.7.9.5


^ permalink raw reply related	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2014-09-10  7:23 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2014-09-10  1:29 [PATCH v2] mmc: tmio: enable odd number size access Kuninori Morimoto
2014-09-10  7:11 ` Geert Uytterhoeven
2014-09-10  7:23   ` [PATCH v3] " Kuninori Morimoto

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).