Re: [U-Boot] [PATCH 3/4] ARM: mmc: bcm283x: Remove get_timer_us() from mmc driver
On Thursday, June 18, 2015 at 02:51:18 PM, Jakub Kiciński wrote: On Thu, 18 Jun 2015 14:35:27 +0200, Marek Vasut wrote: On Wednesday, June 17, 2015 at 06:13:03 PM, Jakub Kiciński wrote: On Wed, 17 Jun 2015 12:44:15 +0200, Marek Vasut wrote: On Tuesday, June 16, 2015 at 05:44:06 AM, Stephen Warren wrote: On 05/04/2015 02:54 PM, Marek Vasut wrote: The get_timer_us() function is something which is no longer existing in case we use generic timer framework, so replace it with get_timer(). Marek, This patch causes saveenv to got from almost no time to nearly 50s on my RPi model A+. Can you take a look at that please? Can you try the attached diff ? ;-/ I think I mistakenly used get_timer(), which returns msecs instead of timer_get_us() which reports usecs, sorry. I can confirm this solves the regression. That's not a regression, that was a bug ;-) Ach OK, I thought Stephen said that MMC used to be fast before the offending patch ;) Anyway I don't really know what's the difference and wikipedia says regression is a bug too so we may both be right :P I always considered it to be [1] 2c , maybe I'm wrong though ;-) Just recently I thought I finally figured out the difference too, but now you surely made me unsure again :) 2 : a trend or shift toward a lower or less perfect state: as c : reversion to an earlier mental or behavioral level [1] http://www.merriam-webster.com/dictionary/regression Best regards, Marek Vasut ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] [PATCH 3/4] ARM: mmc: bcm283x: Remove get_timer_us() from mmc driver
On Thu, 18 Jun 2015 14:35:27 +0200, Marek Vasut wrote: On Wednesday, June 17, 2015 at 06:13:03 PM, Jakub Kiciński wrote: On Wed, 17 Jun 2015 12:44:15 +0200, Marek Vasut wrote: On Tuesday, June 16, 2015 at 05:44:06 AM, Stephen Warren wrote: On 05/04/2015 02:54 PM, Marek Vasut wrote: The get_timer_us() function is something which is no longer existing in case we use generic timer framework, so replace it with get_timer(). Marek, This patch causes saveenv to got from almost no time to nearly 50s on my RPi model A+. Can you take a look at that please? Can you try the attached diff ? ;-/ I think I mistakenly used get_timer(), which returns msecs instead of timer_get_us() which reports usecs, sorry. I can confirm this solves the regression. That's not a regression, that was a bug ;-) Ach OK, I thought Stephen said that MMC used to be fast before the offending patch ;) Anyway I don't really know what's the difference and wikipedia says regression is a bug too so we may both be right :P ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] [PATCH 3/4] ARM: mmc: bcm283x: Remove get_timer_us() from mmc driver
On Wednesday, June 17, 2015 at 06:13:03 PM, Jakub Kiciński wrote: On Wed, 17 Jun 2015 12:44:15 +0200, Marek Vasut wrote: On Tuesday, June 16, 2015 at 05:44:06 AM, Stephen Warren wrote: On 05/04/2015 02:54 PM, Marek Vasut wrote: The get_timer_us() function is something which is no longer existing in case we use generic timer framework, so replace it with get_timer(). Marek, This patch causes saveenv to got from almost no time to nearly 50s on my RPi model A+. Can you take a look at that please? Can you try the attached diff ? ;-/ I think I mistakenly used get_timer(), which returns msecs instead of timer_get_us() which reports usecs, sorry. I can confirm this solves the regression. That's not a regression, that was a bug ;-) Thanks for testing :) Best regards, Marek Vasut ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] [PATCH 3/4] ARM: mmc: bcm283x: Remove get_timer_us() from mmc driver
On 06/17/2015 04:44 AM, Marek Vasut wrote: On Tuesday, June 16, 2015 at 05:44:06 AM, Stephen Warren wrote: On 05/04/2015 02:54 PM, Marek Vasut wrote: The get_timer_us() function is something which is no longer existing in case we use generic timer framework, so replace it with get_timer(). Marek, This patch causes saveenv to got from almost no time to nearly 50s on my RPi model A+. Can you take a look at that please? Can you try the attached diff ? ;-/ I think I mistakenly used get_timer(), which returns msecs instead of timer_get_us() which reports usecs, sorry. Thanks, that fixes it. Tested-by: Stephen Warren swar...@wwwdotorg.org ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] [PATCH 3/4] ARM: mmc: bcm283x: Remove get_timer_us() from mmc driver
On Tuesday, June 16, 2015 at 05:44:06 AM, Stephen Warren wrote: On 05/04/2015 02:54 PM, Marek Vasut wrote: The get_timer_us() function is something which is no longer existing in case we use generic timer framework, so replace it with get_timer(). Marek, This patch causes saveenv to got from almost no time to nearly 50s on my RPi model A+. Can you take a look at that please? Can you try the attached diff ? ;-/ I think I mistakenly used get_timer(), which returns msecs instead of timer_get_us() which reports usecs, sorry. Best regards, Marek Vasut diff --git a/drivers/mmc/bcm2835_sdhci.c b/drivers/mmc/bcm2835_sdhci.c index 078ef05..74b497e 100644 --- a/drivers/mmc/bcm2835_sdhci.c +++ b/drivers/mmc/bcm2835_sdhci.c @@ -69,11 +69,11 @@ static inline void bcm2835_sdhci_raw_writel(struct sdhci_host *host, u32 val, * (Which is just as well - otherwise we'd have to nobble the DMA engine * too) */ - while (get_timer(bcm_host-last_write) bcm_host-twoticks_delay) + while (timer_get_us() bcm_host-last_write + bcm_host-twoticks_delay) ; writel(val, host-ioaddr + reg); - bcm_host-last_write = get_timer(0); + bcm_host-last_write = timer_get_us(); } static inline u32 bcm2835_sdhci_raw_readl(struct sdhci_host *host, int reg) ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] [PATCH 3/4] ARM: mmc: bcm283x: Remove get_timer_us() from mmc driver
On Wed, 17 Jun 2015 12:44:15 +0200, Marek Vasut wrote: On Tuesday, June 16, 2015 at 05:44:06 AM, Stephen Warren wrote: On 05/04/2015 02:54 PM, Marek Vasut wrote: The get_timer_us() function is something which is no longer existing in case we use generic timer framework, so replace it with get_timer(). Marek, This patch causes saveenv to got from almost no time to nearly 50s on my RPi model A+. Can you take a look at that please? Can you try the attached diff ? ;-/ I think I mistakenly used get_timer(), which returns msecs instead of timer_get_us() which reports usecs, sorry. I can confirm this solves the regression. ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] [PATCH 3/4] ARM: mmc: bcm283x: Remove get_timer_us() from mmc driver
On 05/04/2015 02:54 PM, Marek Vasut wrote: The get_timer_us() function is something which is no longer existing in case we use generic timer framework, so replace it with get_timer(). Marek, This patch causes saveenv to got from almost no time to nearly 50s on my RPi model A+. Can you take a look at that please? ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] [PATCH 3/4] ARM: mmc: bcm283x: Remove get_timer_us() from mmc driver
Hi Marek, On May 4, 2015, at 23:54 , Marek Vasut ma...@denx.de wrote: The get_timer_us() function is something which is no longer existing in case we use generic timer framework, so replace it with get_timer(). Signed-off-by: Marek Vasut ma...@denx.de Cc: Stephen Warren swar...@wwwdotorg.org Cc: Tyler Baker tyler.ba...@linaro.org --- drivers/mmc/bcm2835_sdhci.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/mmc/bcm2835_sdhci.c b/drivers/mmc/bcm2835_sdhci.c index 4ec2968..0f81add 100644 --- a/drivers/mmc/bcm2835_sdhci.c +++ b/drivers/mmc/bcm2835_sdhci.c @@ -69,11 +69,11 @@ static inline void bcm2835_sdhci_raw_writel(struct sdhci_host *host, u32 val, * (Which is just as well - otherwise we'd have to nobble the DMA engine * too) */ - while (get_timer_us(bcm_host-last_write) bcm_host-twoticks_delay) + while (get_timer(bcm_host-last_write) bcm_host-twoticks_delay) ; writel(val, host-ioaddr + reg); - bcm_host-last_write = get_timer_us(0); + bcm_host-last_write = get_timer(0); } static inline u32 bcm2835_sdhci_raw_readl(struct sdhci_host *host, int reg) -- 2.1.4 Applied, thanks. — Pantelis ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
[U-Boot] [PATCH 3/4] ARM: mmc: bcm283x: Remove get_timer_us() from mmc driver
The get_timer_us() function is something which is no longer existing in case we use generic timer framework, so replace it with get_timer(). Signed-off-by: Marek Vasut ma...@denx.de Cc: Stephen Warren swar...@wwwdotorg.org Cc: Tyler Baker tyler.ba...@linaro.org --- drivers/mmc/bcm2835_sdhci.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/mmc/bcm2835_sdhci.c b/drivers/mmc/bcm2835_sdhci.c index 4ec2968..0f81add 100644 --- a/drivers/mmc/bcm2835_sdhci.c +++ b/drivers/mmc/bcm2835_sdhci.c @@ -69,11 +69,11 @@ static inline void bcm2835_sdhci_raw_writel(struct sdhci_host *host, u32 val, * (Which is just as well - otherwise we'd have to nobble the DMA engine * too) */ - while (get_timer_us(bcm_host-last_write) bcm_host-twoticks_delay) + while (get_timer(bcm_host-last_write) bcm_host-twoticks_delay) ; writel(val, host-ioaddr + reg); - bcm_host-last_write = get_timer_us(0); + bcm_host-last_write = get_timer(0); } static inline u32 bcm2835_sdhci_raw_readl(struct sdhci_host *host, int reg) -- 2.1.4 ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot