From: "Wolfgang Mües" <wolfgang.mues@auerswald.de>
To: Pierre Ossman <drzeus@drzeus.cx>
Cc: "Andrew Morton" <akpm@linux-foundation.org>,
"Matt Fleming" <matt@console-pimps.org>,
"David Brownell" <dbrownell@users.sourceforge.net>,
"Mike Frysinger" <vapier.adi@gmail.com>,
linux-kernel@vger.kernel.org
Subject: [PATCH 6/7 revised] mmc_spi: convert timeout handling to jiffies and avoid busy waiting
Date: Tue, 7 Apr 2009 15:26:30 +0100 [thread overview]
Message-ID: <200904071626.30917.wolfgang.mues@auerswald.de> (raw)
From: Wolfgang Muees <wolfgang.mues@auerswald.de>
o SD/MMC card timeouts can be very high. So avoid busy-waiting,
using the scheduler. Calculate all timeouts in jiffies units,
because this will give us the correct sign when to involve
the scheduler.
Signed-off-by: Wolfgang Muees <wolfgang.mues@auerswald.de>
---
This patch was discussed on lkml. Some where concerned about
performance loss, but no performance loss was visible in real-
life tests, and nobody came up with a better solution. And it
was clear that a busy waiting of several 100 of ms has to be
ended.
diff -uprN 2_6_29_rc7_patch12534/drivers/mmc/host/mmc_spi.c 2_6_29_rc7_patch125346/drivers/mmc/host/mmc_spi.c
--- 2_6_29_rc7_patch12534/drivers/mmc/host/mmc_spi.c 2009-03-16 12:15:36.000000000 +0100
+++ 2_6_29_rc7_patch125346/drivers/mmc/host/mmc_spi.c 2009-04-07 15:59:54.000000000 +0200
@@ -24,7 +24,7 @@
* along with this program; if not, write to the Free Software
* Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
*/
-#include <linux/hrtimer.h>
+#include <linux/sched.h>
#include <linux/delay.h>
#include <linux/bio.h>
#include <linux/dma-mapping.h>
@@ -95,7 +95,7 @@
* reads which takes nowhere near that long. Older cards may be able to use
* shorter timeouts ... but why bother?
*/
-#define r1b_timeout ktime_set(3, 0)
+#define r1b_timeout (HZ * 3)
/****************************************************************************/
@@ -183,12 +183,11 @@ mmc_spi_readbytes(struct mmc_spi_host *h
return status;
}
-static int
-mmc_spi_skip(struct mmc_spi_host *host, ktime_t timeout, unsigned n, u8 byte)
+static int mmc_spi_skip(struct mmc_spi_host *host, unsigned long timeout,
+ unsigned n, u8 byte)
{
u8 *cp = host->data->status;
-
- timeout = ktime_add(timeout, ktime_get());
+ unsigned long start = jiffies;
while (1) {
int status;
@@ -203,22 +202,26 @@ mmc_spi_skip(struct mmc_spi_host *host,
return cp[i];
}
- /* REVISIT investigate msleep() to avoid busy-wait I/O
- * in at least some cases.
- */
- if (ktime_to_ns(ktime_sub(ktime_get(), timeout)) > 0)
+ if (time_is_before_jiffies(start + timeout))
break;
+
+ /* If we need long timeouts, we may release the CPU.
+ * We use jiffies here because we want to have a relation
+ * between elapsed time and the blocking of the scheduler.
+ */
+ if (time_is_before_jiffies(start+1))
+ schedule();
}
return -ETIMEDOUT;
}
static inline int
-mmc_spi_wait_unbusy(struct mmc_spi_host *host, ktime_t timeout)
+mmc_spi_wait_unbusy(struct mmc_spi_host *host, unsigned long timeout)
{
return mmc_spi_skip(host, timeout, sizeof(host->data->status), 0);
}
-static int mmc_spi_readtoken(struct mmc_spi_host *host, ktime_t timeout)
+static int mmc_spi_readtoken(struct mmc_spi_host *host, unsigned long timeout)
{
return mmc_spi_skip(host, timeout, 1, 0xff);
}
@@ -607,7 +610,7 @@ mmc_spi_setup_data_message(
*/
static int
mmc_spi_writeblock(struct mmc_spi_host *host, struct spi_transfer *t,
- ktime_t timeout)
+ unsigned long timeout)
{
struct spi_device *spi = host->spi;
int status, i;
@@ -717,7 +720,7 @@ mmc_spi_writeblock(struct mmc_spi_host *
*/
static int
mmc_spi_readblock(struct mmc_spi_host *host, struct spi_transfer *t,
- ktime_t timeout)
+ unsigned long timeout)
{
struct spi_device *spi = host->spi;
int status;
@@ -803,7 +806,7 @@ mmc_spi_data_do(struct mmc_spi_host *hos
unsigned n_sg;
int multiple = (data->blocks > 1);
u32 clock_rate;
- ktime_t timeout;
+ unsigned long timeout;
if (data->flags & MMC_DATA_READ)
direction = DMA_FROM_DEVICE;
@@ -817,8 +820,9 @@ mmc_spi_data_do(struct mmc_spi_host *hos
else
clock_rate = spi->max_speed_hz;
- timeout = ktime_add_ns(ktime_set(0, 0), data->timeout_ns +
- data->timeout_clks * 1000000 / clock_rate);
+ timeout = data->timeout_ns +
+ data->timeout_clks * 1000000 / clock_rate;
+ timeout = usecs_to_jiffies((unsigned int)(timeout / 1000)) + 1;
/* Handle scatterlist segments one at a time, with synch for
* each 512-byte block
---
regards
i. A. Wolfgang Mües
--
Auerswald GmbH & Co. KG
Hardware Development
Telefon: +49 (0)5306 9219 0
Telefax: +49 (0)5306 9219 94
E-Mail: Wolfgang.Mues@Auerswald.de
Web: http://www.auerswald.de
--------------------------------------------------------------
Auerswald GmbH & Co. KG, Vor den Grashöfen 1, 38162 Cremlingen
Registriert beim AG Braunschweig HRA 13289
p.h.G Auerswald Geschäftsführungsges. mbH
Registriert beim AG Braunschweig HRB 7463
Geschäftsführer: Dipl-Ing. Gerhard Auerswald
next reply other threads:[~2009-04-07 14:27 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-04-07 14:26 Wolfgang Mües [this message]
2009-04-08 18:39 ` Pierre Ossman
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=200904071626.30917.wolfgang.mues@auerswald.de \
--to=wolfgang.mues@auerswald.de \
--cc=akpm@linux-foundation.org \
--cc=dbrownell@users.sourceforge.net \
--cc=drzeus@drzeus.cx \
--cc=linux-kernel@vger.kernel.org \
--cc=matt@console-pimps.org \
--cc=vapier.adi@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®