aboutsummaryrefslogtreecommitdiffstats
path: root/target/linux/brcm2708/patches-4.1/0092-bcm2835-sdhost-Further-improve-overclock-back-off.patch
diff options
context:
space:
mode:
authorJohn Crispin <john@openwrt.org>2015-07-17 12:48:39 +0000
committerJohn Crispin <john@openwrt.org>2015-07-17 12:48:39 +0000
commit208ab54e3e9516dd9fcf1ce197950044e0bebd90 (patch)
tree425580d694235f06661dd2d6763925a1f69279c6 /target/linux/brcm2708/patches-4.1/0092-bcm2835-sdhost-Further-improve-overclock-back-off.patch
parent03b56c9350057ee77cfe2b8bb8f8d7146917fb79 (diff)
downloadupstream-208ab54e3e9516dd9fcf1ce197950044e0bebd90.tar.gz
upstream-208ab54e3e9516dd9fcf1ce197950044e0bebd90.tar.bz2
upstream-208ab54e3e9516dd9fcf1ce197950044e0bebd90.zip
brcm2708: add linux 4.1 support
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com> SVN-Revision: 46399
Diffstat (limited to 'target/linux/brcm2708/patches-4.1/0092-bcm2835-sdhost-Further-improve-overclock-back-off.patch')
-rw-r--r--target/linux/brcm2708/patches-4.1/0092-bcm2835-sdhost-Further-improve-overclock-back-off.patch292
1 files changed, 292 insertions, 0 deletions
diff --git a/target/linux/brcm2708/patches-4.1/0092-bcm2835-sdhost-Further-improve-overclock-back-off.patch b/target/linux/brcm2708/patches-4.1/0092-bcm2835-sdhost-Further-improve-overclock-back-off.patch
new file mode 100644
index 0000000000..810a338073
--- /dev/null
+++ b/target/linux/brcm2708/patches-4.1/0092-bcm2835-sdhost-Further-improve-overclock-back-off.patch
@@ -0,0 +1,292 @@
+From d645d31525be338b1baa22916217f3ac0c3705f9 Mon Sep 17 00:00:00 2001
+From: Phil Elwell <phil@raspberrypi.org>
+Date: Thu, 25 Jun 2015 08:47:09 +0100
+Subject: [PATCH 092/121] bcm2835-sdhost: Further improve overclock back-off
+
+---
+ drivers/mmc/host/bcm2835-sdhost.c | 144 +++++++++++++++++++++-----------------
+ 1 file changed, 78 insertions(+), 66 deletions(-)
+
+--- a/drivers/mmc/host/bcm2835-sdhost.c
++++ b/drivers/mmc/host/bcm2835-sdhost.c
+@@ -161,8 +161,6 @@ struct bcm2835_host {
+
+ unsigned int use_busy:1; /* Wait for busy interrupt */
+
+- unsigned int reduce_overclock:1; /* ...at the next opportunity */
+-
+ unsigned int debug:1; /* Enable debug output */
+
+ u32 thread_isr;
+@@ -466,36 +464,25 @@ static void bcm2835_sdhost_dma_complete(
+ spin_unlock_irqrestore(&host->lock, flags);
+ }
+
+-static bool data_transfer_wait(struct bcm2835_host *host, const char *caller)
++static bool data_transfer_wait(struct bcm2835_host *host)
+ {
+ unsigned long timeout = 1000000;
+- u32 hsts;
+ while (timeout)
+ {
+- hsts = bcm2835_sdhost_read(host, SDHSTS);
+- if (hsts & (SDHSTS_TRANSFER_ERROR_MASK |
+- SDHSTS_DATA_FLAG)) {
+- bcm2835_sdhost_write(host, SDHSTS_TRANSFER_ERROR_MASK,
+- SDHSTS);
++ u32 sdhsts = bcm2835_sdhost_read(host, SDHSTS);
++ if (sdhsts & SDHSTS_DATA_FLAG) {
++ bcm2835_sdhost_write(host, SDHSTS_DATA_FLAG, SDHSTS);
+ break;
+ }
+ timeout--;
+ }
+-
+- if (hsts & (SDHSTS_CRC16_ERROR |
+- SDHSTS_CRC7_ERROR |
+- SDHSTS_FIFO_ERROR)) {
+- pr_err("%s: data error in %s - HSTS %x\n",
+- mmc_hostname(host->mmc), caller, hsts);
+- host->data->error = -EILSEQ;
+- return false;
+- } else if ((timeout == 0) ||
+- (hsts & (SDHSTS_CMD_TIME_OUT |
+- SDHSTS_REW_TIME_OUT))) {
+- pr_err("%s: timeout in %s - HSTS %x\n",
+- mmc_hostname(host->mmc), caller, hsts);
+- host->data->error = -ETIMEDOUT;
+- return false;
++ if (timeout == 0) {
++ pr_err("%s: Data %s timeout\n",
++ mmc_hostname(host->mmc),
++ (host->data->flags & MMC_DATA_READ) ? "read" : "write");
++ bcm2835_sdhost_dumpregs(host);
++ host->data->error = -ETIMEDOUT;
++ return false;
+ }
+ return true;
+ }
+@@ -523,7 +510,7 @@ static void bcm2835_sdhost_read_block_pi
+ buf = (u32 *)host->sg_miter.addr;
+
+ while (len) {
+- if (!data_transfer_wait(host, "read_block_pio"))
++ if (!data_transfer_wait(host))
+ break;
+
+ *(buf++) = bcm2835_sdhost_read(host, SDDATA);
+@@ -562,7 +549,7 @@ static void bcm2835_sdhost_write_block_p
+ buf = host->sg_miter.addr;
+
+ while (len) {
+- if (!data_transfer_wait(host, "write_block_pio"))
++ if (!data_transfer_wait(host))
+ break;
+
+ bcm2835_sdhost_write(host, *(buf++), SDDATA);
+@@ -581,13 +568,33 @@ static void bcm2835_sdhost_write_block_p
+
+ static void bcm2835_sdhost_transfer_pio(struct bcm2835_host *host)
+ {
++ u32 sdhsts;
++ bool is_read;
+ BUG_ON(!host->data);
+
+- if (host->data->flags & MMC_DATA_READ) {
++ is_read = (host->data->flags & MMC_DATA_READ) != 0;
++ if (is_read)
+ bcm2835_sdhost_read_block_pio(host);
+- } else {
++ else
+ bcm2835_sdhost_write_block_pio(host);
+
++ sdhsts = bcm2835_sdhost_read(host, SDHSTS);
++ if (sdhsts & (SDHSTS_CRC16_ERROR |
++ SDHSTS_CRC7_ERROR |
++ SDHSTS_FIFO_ERROR)) {
++ pr_err("%s: %s transfer error - HSTS %x\n",
++ mmc_hostname(host->mmc),
++ is_read ? "read" : "write",
++ sdhsts);
++ host->data->error = -EILSEQ;
++ } else if ((sdhsts & (SDHSTS_CMD_TIME_OUT |
++ SDHSTS_REW_TIME_OUT))) {
++ pr_err("%s: %s timeout error - HSTS %x\n",
++ mmc_hostname(host->mmc),
++ is_read ? "read" : "write",
++ sdhsts);
++ host->data->error = -ETIMEDOUT;
++ } else if (!is_read && !host->data->error) {
+ /* Start a timer in case a transfer error occurs because
+ there is no error interrupt */
+ mod_timer(&host->pio_timer, jiffies + host->pio_timeout);
+@@ -701,8 +708,9 @@ static void bcm2835_sdhost_prepare_data(
+
+ void bcm2835_sdhost_send_command(struct bcm2835_host *host, struct mmc_command *cmd)
+ {
+- u32 sdcmd;
++ u32 sdcmd, sdhsts;
+ unsigned long timeout;
++ int delay;
+
+ WARN_ON(host->cmd);
+
+@@ -719,8 +727,8 @@ void bcm2835_sdhost_send_command(struct
+ mmc_hostname(host->mmc),
+ cmd->opcode, cmd->arg, cmd->flags);
+
+- /* Wait max 10 ms */
+- timeout = 1000;
++ /* Wait max 100 ms */
++ timeout = 10000;
+
+ while (bcm2835_sdhost_read(host, SDCMD) & SDCMD_NEW_FLAG) {
+ if (timeout == 0) {
+@@ -735,8 +743,9 @@ void bcm2835_sdhost_send_command(struct
+ udelay(10);
+ }
+
+- if ((1000-timeout)/100 > 1 && (1000-timeout)/100 > host->max_delay) {
+- host->max_delay = (1000-timeout)/100;
++ delay = (10000 - timeout)/100;
++ if (delay > host->max_delay) {
++ host->max_delay = delay;
+ pr_warning("%s: controller hung for %d ms\n",
+ mmc_hostname(host->mmc),
+ host->max_delay);
+@@ -751,6 +760,11 @@ void bcm2835_sdhost_send_command(struct
+
+ host->cmd = cmd;
+
++ /* Clear any error flags */
++ sdhsts = bcm2835_sdhost_read(host, SDHSTS);
++ if (sdhsts & SDHSTS_ERROR_MASK)
++ bcm2835_sdhost_write(host, sdhsts, SDHSTS);
++
+ bcm2835_sdhost_prepare_data(host, cmd);
+
+ bcm2835_sdhost_write(host, cmd->arg, SDARG);
+@@ -876,7 +890,7 @@ static void bcm2835_sdhost_transfer_comp
+ static void bcm2835_sdhost_finish_command(struct bcm2835_host *host)
+ {
+ u32 sdcmd;
+- int timeout = 1000;
++ unsigned long timeout;
+ #ifdef DEBUG
+ struct timeval before, after;
+ int timediff = 0;
+@@ -889,6 +903,8 @@ static void bcm2835_sdhost_finish_comman
+ #ifdef DEBUG
+ do_gettimeofday(&before);
+ #endif
++ /* Wait max 100 ms */
++ timeout = 10000;
+ for (sdcmd = bcm2835_sdhost_read(host, SDCMD);
+ (sdcmd & SDCMD_NEW_FLAG) && timeout;
+ timeout--) {
+@@ -1049,9 +1065,9 @@ static void bcm2835_sdhost_pio_timeout(u
+ spin_lock_irqsave(&host->lock, flags);
+
+ if (host->data) {
+- u32 hsts = bcm2835_sdhost_read(host, SDHSTS);
++ u32 sdhsts = bcm2835_sdhost_read(host, SDHSTS);
+
+- if (hsts & SDHSTS_REW_TIME_OUT) {
++ if (sdhsts & SDHSTS_REW_TIME_OUT) {
+ pr_err("%s: transfer timeout\n",
+ mmc_hostname(host->mmc));
+ if (host->debug)
+@@ -1380,19 +1396,10 @@ void bcm2835_sdhost_set_clock(struct bcm
+ if (host->debug)
+ pr_info("%s: set_clock(%d)\n", mmc_hostname(host->mmc), clock);
+
+- if ((clock == 0) && host->reduce_overclock) {
+- /* This is a reset following data corruption - reduce any
+- overclock */
+- host->reduce_overclock = 0;
+- if (host->overclock_50 > 50) {
+- pr_warn("%s: reducing overclock due to errors\n",
+- mmc_hostname(host->mmc));
+- host->overclock_50--;
+- }
+- }
+-
+- if (host->overclock_50 && (clock == 50*MHZ))
++ if ((host->overclock_50 > 50) &&
++ (clock == 50*MHZ)) {
+ clock = host->overclock_50 * MHZ + (MHZ - 1);
++ }
+
+ /* The SDCDIV register has 11 bits, and holds (div - 2).
+ But in data mode the max is 50MHz wihout a minimum, and only the
+@@ -1450,11 +1457,12 @@ void bcm2835_sdhost_set_clock(struct bcm
+ host->overclock = clock;
+ }
+ }
+- else if ((clock == 50 * MHZ) && host->overclock)
++ else if (host->overclock)
+ {
+- pr_warn("%s: cancelling overclock\n",
+- mmc_hostname(host->mmc));
+ host->overclock = 0;
++ if (clock == 50 * MHZ)
++ pr_warn("%s: cancelling overclock\n",
++ mmc_hostname(host->mmc));
+ }
+
+ host->cdiv = div;
+@@ -1492,6 +1500,14 @@ static void bcm2835_sdhost_request(struc
+ cmd->opcode, cmd->arg, cmd->flags);
+ }
+
++ /* Reset the error statuses in case this is a retry */
++ if (mrq->cmd)
++ mrq->cmd->error = 0;
++ if (mrq->data)
++ mrq->data->error = 0;
++ if (mrq->stop)
++ mrq->stop->error = 0;
++
+ if (mrq->data && !is_power_of_2(mrq->data->blksz)) {
+ pr_err("%s: unsupported block size (%d bytes)\n",
+ mmc_hostname(mmc), mrq->data->blksz);
+@@ -1613,21 +1629,16 @@ static void bcm2835_sdhost_tasklet_finis
+
+ /* Drop the overclock after any data corruption, or after any
+ error overclocked */
+- if (mrq->data && (mrq->data->error == -EILSEQ))
+- host->reduce_overclock = 1;
+- else if (host->overclock) {
+- /* Convert timeout errors while overclocked to data errors,
+- because the system recovers better. */
+- if (mrq->cmd && mrq->cmd->error) {
+- host->reduce_overclock = 1;
+- if (mrq->cmd->error == -ETIMEDOUT)
+- mrq->cmd->error = -EILSEQ;
+- }
+-
+- if (mrq->data && mrq->data->error) {
+- host->reduce_overclock = 1;
+- if (mrq->data->error == -ETIMEDOUT)
+- mrq->data->error = -EILSEQ;
++ if (host->overclock) {
++ if ((mrq->cmd && mrq->cmd->error) ||
++ (mrq->data && mrq->data->error) ||
++ (mrq->stop && mrq->stop->error)) {
++ host->overclock_50--;
++ pr_warn("%s: reducing overclock due to errors\n",
++ mmc_hostname(host->mmc));
++ bcm2835_sdhost_set_clock(host,50*MHZ);
++ mrq->cmd->error = -EILSEQ;
++ mrq->cmd->retries = 1;
+ }
+ }
+
+@@ -1769,6 +1780,7 @@ static int bcm2835_sdhost_probe(struct p
+ host = mmc_priv(mmc);
+ host->mmc = mmc;
+ host->pio_timeout = msecs_to_jiffies(500);
++ host->max_delay = 1; /* Warn if over 1ms */
+ spin_lock_init(&host->lock);
+
+ iomem = platform_get_resource(pdev, IORESOURCE_MEM, 0);