mmc: fsl_esdhc_imx: use VENDORSPEC_FRC_SDCLK_ON to control card clock output
Subsystems (2)
PATCH 1/1 mmc: fsl_esdhc_imx: use VENDORSPEC_FRC_SDCLK_ON to control card clock output
From: Haibo Chen <[email protected]>
For FSL_USDHC, it do not implement VENDORSPEC_CKEN/PEREN/HCKEN/IPGEN, these are reserved bits. Instead, use VENDORSPEC_FRC_SDCLK_ON to gate on/off the card clock output.
After commit b5874b552ffa ("mmc: fsl_esdhc_imx: add wait_dat0() support"), we meet SD3.0 card can't work at UHS mode, mmc_switch_voltage() fail because the second mmc_wait_dat0 return -ETIMEDOUT. According to SD spec, during voltage switch, need to gate off/on the card clock. If not set the FRC_SDCLK_ON, after CMD11, hardware will gate off the card clock automatically, so card do not detect the clock off/on behavior, so will draw the data0 line low until next command.
--- drivers/mmc/fsl_esdhc_imx.c | 29 +++++++++++++++++++++-------- include/fsl_esdhc_imx.h | 2 ++ 2 files changed, 23 insertions(+), 8 deletions(-)
drivers/mmc/fsl_esdhc_imx.c +21 -8
@@ -653,7 +653,10 @@ static void set_sysctl(struct fsl_esdhc_priv *priv, struct mmc *mmc, uint clock) clk = (pre_div <<8 ) | (div <<4 );#ifdef CONFIG_FSL_USDHC -esdhc_clrbits32 (®s->vendorspec,VENDORSPEC_CKEN ); +esdhc_clrbits32 (®s->vendorspec,VENDORSPEC_FRC_SDCLK_ON ); + ret =readx_poll_timeout (esdhc_read32, ®s->prsstat, tmp, tmp &PRSSTAT_SDOFF ,100 ); +if (ret) +pr_warn ("fsl_esdhc_imx: Internal clock never gate off.\n" );#else esdhc_clrbits32 (®s->sysctl,SYSCTL_CKEN );#endif
@@ -665,7 +668,7 @@ static void set_sysctl(struct fsl_esdhc_priv *priv, struct mmc *mmc, uint clock)pr_warn ("fsl_esdhc_imx: Internal clock never stabilised.\n" );#ifdef CONFIG_FSL_USDHC -esdhc_setbits32 (®s->vendorspec,VENDORSPEC_PEREN |VENDORSPEC_CKEN ); +esdhc_setbits32 (®s->vendorspec,VENDORSPEC_FRC_SDCLK_ON );#else esdhc_setbits32 (®s->sysctl,SYSCTL_PEREN |SYSCTL_CKEN );#endif
@@ -720,8 +723,14 @@ static void esdhc_set_strobe_dll(struct mmc *mmc)struct fsl_esdhc_priv *priv =dev_get_priv (mmc->dev);struct fsl_esdhc *regs = priv->esdhc_regs;u32 val; +u32 tmp; +int ret;if (priv->clock >ESDHC_STROBE_DLL_CLK_FREQ ) { +esdhc_clrbits32 (®s->vendorspec,VENDORSPEC_FRC_SDCLK_ON ); + ret =readx_poll_timeout (esdhc_read32, ®s->prsstat, tmp, tmp &PRSSTAT_SDOFF ,100 ); +if (ret) +pr_warn ("fsl_esdhc_imx: Internal clock never gate off.\n" );esdhc_write32 (®s->strobe_dllctrl,ESDHC_STROBE_DLL_CTRL_RESET );/*
@@ -739,6 +748,7 @@ static void esdhc_set_strobe_dll(struct mmc *mmc)pr_warn ("HS400 strobe DLL status REF not lock!\n" );if (!(val &ESDHC_STROBE_DLL_STS_SLV_LOCK ))pr_warn ("HS400 strobe DLL status SLV not lock!\n" ); +esdhc_setbits32 (®s->vendorspec,VENDORSPEC_FRC_SDCLK_ON ); } }
@@ -962,14 +972,18 @@ static int esdhc_set_ios_common(struct fsl_esdhc_priv *priv, struct mmc *mmc)#ifdef MMC_SUPPORTS_TUNING if (mmc->clk_disable) {#ifdef CONFIG_FSL_USDHC -esdhc_clrbits32 (®s->vendorspec,VENDORSPEC_CKEN ); +u32 tmp; + +esdhc_clrbits32 (®s->vendorspec,VENDORSPEC_FRC_SDCLK_ON ); + ret =readx_poll_timeout (esdhc_read32, ®s->prsstat, tmp, tmp &PRSSTAT_SDOFF ,100 ); +if (ret) +pr_warn ("fsl_esdhc_imx: Internal clock never gate off.\n" );#else esdhc_clrbits32 (®s->sysctl,SYSCTL_CKEN );#endif }else {#ifdef CONFIG_FSL_USDHC -esdhc_setbits32 (®s->vendorspec,VENDORSPEC_PEREN | -VENDORSPEC_CKEN ); +esdhc_setbits32 (®s->vendorspec,VENDORSPEC_FRC_SDCLK_ON );#else esdhc_setbits32 (®s->sysctl,SYSCTL_PEREN |SYSCTL_CKEN );#endif
@@ -1045,7 +1059,7 @@ static int esdhc_init_common(struct fsl_esdhc_priv *priv, struct mmc *mmc)#ifndef CONFIG_FSL_USDHC esdhc_setbits32 (®s->sysctl,SYSCTL_HCKEN |SYSCTL_IPGEN );#else -esdhc_setbits32 (®s->vendorspec,VENDORSPEC_HCKEN |VENDORSPEC_IPGEN ); +esdhc_setbits32 (®s->vendorspec,VENDORSPEC_FRC_SDCLK_ON );#endif /* Set the initial clock speed */
@@ -1183,8 +1197,7 @@ static int fsl_esdhc_init(struct fsl_esdhc_priv *priv,esdhc_write32 (®s->autoc12err,0 );esdhc_write32 (®s->clktunectrlstatus,0 );#else -esdhc_setbits32 (®s->vendorspec,VENDORSPEC_PEREN | -VENDORSPEC_HCKEN |VENDORSPEC_IPGEN |VENDORSPEC_CKEN ); +esdhc_setbits32 (®s->vendorspec,VENDORSPEC_FRC_SDCLK_ON );#endif if (priv->vs18_enable)
include/fsl_esdhc_imx.h +2 -0
@@ -39,6 +39,7 @@#define VENDORSPEC_HCKEN 0x00001000#define VENDORSPEC_IPGEN 0x00000800#define VENDORSPEC_INIT 0x20007809 +#define VENDORSPEC_FRC_SDCLK_ON 0x00000100#define IRQSTAT 0x0002e030#define IRQSTAT_DMAE (0x10000000)
@@ -96,6 +97,7 @@#define PRSSTAT_CINS (0x00010000)#define PRSSTAT_BREN (0x00000800)#define PRSSTAT_BWEN (0x00000400) +#define PRSSTAT_SDOFF (0x00000080)#define PRSSTAT_SDSTB (0X00000008)#define PRSSTAT_DLA (0x00000004)#define PRSSTAT_CICHB (0x00000002)
Signature · 1 line
2.17.1
PATCH 1/1 mmc: fsl_esdhc_imx: use VENDORSPEC_FRC_SDCLK_ON to control card clock output
Hello Haibo,
Quoted 27 lines
-----Original Message-----
From: haibo.chen at nxp.com <[email protected]>
Sent: Wednesday, January 27, 2021 11:47 AM
To: peng.fan at nxp.com; u-boot at lists.denx.de
Cc: haibo.chen at nxp.com; uboot-imx at nxp.com; tharvey at gateworks.com; ZHIZHIKIN Andrey <[email protected]>
Subject: [PATCH] mmc:fsl_esdhc_imx: useVENDORSPEC_FRC_SDCLK_ONto control card clock outputThis email is not from Hexagon?s Office 365 instance. Please be careful while clicking links, opening attachments, or replying to this email.
From: Haibo Chen <[email protected]>
For
FSL_USDHC, it do not implement VENDORSPEC_CKEN/PEREN/HCKEN/IPGEN, these are reserved bits. Instead, useVENDORSPEC_FRC_SDCLK_ONto gate on/off the card clock output.After commit
b5874b552ffa("mmc:fsl_esdhc_imx: addwait_dat0()support"), we meet SD3.0 card can't work at UHS mode,mmc_switch_voltage()fail because the secondmmc_wait_dat0return-ETIMEDOUT. According to SD spec, during voltage switch, need to gate off/on the card clock. If not set theFRC_SDCLK_ON, after CMD11, hardware will gate off the card clock automatically, so card do not detect the clock off/on behavior, so will draw the data0 line low until next command.
Quoted 1 line
Fixes b5874b552ffa ("mmc: fsl_esdhc_imx: add wait_dat0() support")
This patch does not fix the switch of uSDHC to 1v8...
I've applied it locally on the imx8mmevk, and had following log during the boot when tried to query SD Card info:
-----------------------------------------------
U-Boot SPL 2021.01-01004-gb852007333 (Feb 01 2021 - 09:45:42 +0100)
Normal Boot
WDT: Started with servicing (60s timeout)
Trying to boot from MMC1
NOTICE: BL31: v2.2(release):rel_imx_5.4.70_2.3.0-0-gf1d7187f2 NOTICE: BL31: Built : 22:29:05, Jan 17 2021
U-Boot 2021.01-01004-gb852007333 (Feb 01 2021 - 09:45:42 +0100)
CPU: Freescale i.MX8MMQ rev1.0 at 1200 MHz Reset cause: POR Model: FSL i.MX8MM EVK board DRAM: 2 GiB WDT: Started with servicing (60s timeout) MMC: FSL_SDHC: 1, FSL_SDHC: 2
Loading Environment from MMC... Run CMD11 1.8V switch Card did not respond to voltage select! : -110 *** Warning - No block device, using default environment
In: serial Out: serial Err: serial Net: eth0: ethernet at 30be0000 Hit any key to stop autoboot: 0
u-boot=> mmc dev 1
Card did not respond to voltage select! : -110 u-boot=> mmc info
MMC Device 0 not found
no mmc device at slot 0
u-boot=> mmc dev 2
switch to partitions #0, OK
mmc2(part 0) is current device
u-boot=> mmc info
Device: FSL_SDHC
Manufacturer ID: 45
OEM: 100
Name: DG401
Bus Speed: 200000000
Mode: HS400ES (200MHz)
Rd Block Len: 512
MMC version 5.1
High Capacity: Yes
Capacity: 14.7 GiB
Bus Width: 8-bit DDR
Erase Group Size: 512 KiB
HC WP Group Size: 8 MiB
User Capacity: 14.7 GiB WRREL
Boot Capacity: 4 MiB ENH
RPMB Capacity: 4 MiB ENH
Boot area 0 is not write protected
Boot area 1 is not write protected
-----------------------------------------------
Note, that eMMC is not affected and is operating in HS400ES mode without any issues.
Reverting patch b5874b552ffa ("mmc: fsl_esdhc_imx: add wait_dat0() support") with this patch applied in tree does switch the SD Card to high-speed mode, following log is produced:
-----------------------------------------------
U-Boot SPL 2021.01-01005-gb8aeb689a2 (Feb 01 2021 - 10:38:01 +0100)
Normal Boot
WDT: Started with servicing (60s timeout)
Trying to boot from MMC1
NOTICE: BL31: v2.2(release):rel_imx_5.4.70_2.3.0-0-gf1d7187f2 NOTICE: BL31: Built : 22:29:05, Jan 17 2021
U-Boot 2021.01-01005-gb8aeb689a2 (Feb 01 2021 - 10:38:01 +0100)
CPU: Freescale i.MX8MMQ rev1.0 at 1200 MHz Reset cause: POR Model: FSL i.MX8MM EVK board DRAM: 2 GiB WDT: Started with servicing (60s timeout) MMC: FSL_SDHC: 1, FSL_SDHC: 2
Loading Environment from MMC... Run CMD11 1.8V switch *** Warning - bad CRC, using default environment
In: serial Out: serial Err: serial Net: eth0: ethernet at 30be0000 Hit any key to stop autoboot: 0
u-boot=> mmc dev 1
Run CMD11 1.8V switch
switch to partitions #0, OK mmc1 is current device u-boot=> mmc info
Device: FSL_SDHC
Manufacturer ID: 41
OEM: 3432
Name: SD32G
Bus Speed: 200000000
Mode: UHS SDR104 (208MHz)
Rd Block Len: 512
SD version 3.0
High Capacity: Yes
Capacity: 29.3 GiB
Bus Width: 4-bit
Erase Group Size: 512 Bytes u-boot=> mmc dev 2
switch to partitions #0, OK mmc2(part 0) is current device u-boot=> mmc info
Device: FSL_SDHC
Manufacturer ID: 45
OEM: 100
Name: DG401
Bus Speed: 200000000
Mode: HS400ES (200MHz)
Rd Block Len: 512
MMC version 5.1
High Capacity: Yes
Capacity: 14.7 GiB
Bus Width: 8-bit DDR
Erase Group Size: 512 KiB
HC WP Group Size: 8 MiB
User Capacity: 14.7 GiB WRREL Boot Capacity: 4 MiB ENH
RPMB Capacity: 4 MiB ENH
Boot area 0 is not write protected Boot area 1 is not write protected
-----------------------------------------------
I guess high-level modifications are also required, since the uSDHC does not follow the SD Specification in regards to voltage switching, which is currently implemented in U-Boot, this patch alone does not solve the mode switch.
Quoted 122 lines
Signed-off-by Haibo Chen <[email protected]>--- drivers/mmc/fsl_esdhc_imx.c | 29 +++++++++++++++++++++-------- include/fsl_esdhc_imx.h | 2 ++ 2 files changed, 23 insertions(+), 8 deletions(-)drivers/mmc/fsl_esdhc_imx.c index renamed from drivers/mmc/fsl_esdhc_imx.c +0
-0
8ac859797f..da33ee8253100644drivers/mmc/fsl_esdhc_imx.c +5
-1@@ -653,7 +653,10 @@ static void set_sysctl(struct fsl_esdhc_priv *priv, struct mmc *mmc, uint clock) clk = (pre_div <<8 ) | (div <<4 );#ifdef CONFIG_FSL_USDHC -esdhc_clrbits32 (®s->vendorspec,VENDORSPEC_CKEN ); +esdhc_clrbits32 (®s->vendorspec,VENDORSPEC_FRC_SDCLK_ON ); + ret =readx_poll_timeout (esdhc_read32, ®s->prsstat, tmp, tmp &PRSSTAT_SDOFF ,100 ); +if (ret) + pr_warn("fsl_esdhc_imx: Internal clock never gate + off.\n");#else esdhc_clrbits32(®s->sysctl, SYSCTL_CKEN); #endif @@ -665,7 +668,7 @@static void
set_sysctl(struct fsl_esdhc_priv *priv, struct mmc *mmc, uint clock)pr_warn("fsl_esdhc_imx: Internal clock never stabilised.\n");#ifdef
CONFIG_FSL_USDHC
- esdhc_setbits32(®s->vendorspec,VENDORSPEC_PEREN|VENDORSPEC_CKEN);
+esdhc_setbits32(®s->vendorspec, VENDORSPEC_FRC_SDCLK_ON);#else esdhc_setbits32(®s->sysctl, SYSCTL_PEREN | SYSCTL_CKEN); #endif @@ - 720,8 +723,14 @@ static void esdhc_set_strobe_dll(struct mmc *mmc) struct fsl_esdhc_priv *priv = dev_get_priv(mmc->dev); struct fsl_esdhc *regs = priv->esdhc_regs; u32 val;drivers/mmc/fsl_esdhc_imx.c +4
-0+u32 tmp; +int ret;if (priv->clock >ESDHC_STROBE_DLL_CLK_FREQ ) { +esdhc_clrbits32 (®s->vendorspec,VENDORSPEC_FRC_SDCLK_ON ); + ret = readx_poll_timeout(esdhc_read32, ®s->prsstat, tmp, tmp &
PRSSTAT_SDOFF, 100);drivers/mmc/fsl_esdhc_imx.c +15
-4+if (ret) + pr_warn("fsl_esdhc_imx: Internal clock never + gate off.\n");esdhc_write32 (®s->strobe_dllctrl,ESDHC_STROBE_DLL_CTRL_RESET );/* @@ -739,6 +748,7 @@ static void esdhc_set_strobe_dll(struct mmc *mmc)pr_warn ("HS400 strobe DLL status REF not lock!\n" );if (!(val &ESDHC_STROBE_DLL_STS_SLV_LOCK ))pr_warn ("HS400 strobe DLL status SLV not lock!\n" ); +esdhc_setbits32 (®s->vendorspec, +VENDORSPEC_FRC_SDCLK_ON ); } }@@ -962,14 +972,18 @@ static int esdhc_set_ios_common(struct fsl_esdhc_priv *priv, struct mmc *mmc) #ifdef MMC_SUPPORTS_TUNINGif (mmc->clk_disable) {#ifdef CONFIG_FSL_USDHC -esdhc_clrbits32 (®s->vendorspec,VENDORSPEC_CKEN ); +u32 tmp; + +esdhc_clrbits32 (®s->vendorspec,VENDORSPEC_FRC_SDCLK_ON ); + ret =readx_poll_timeout (esdhc_read32, ®s->prsstat, tmp, tmp &PRSSTAT_SDOFF ,100 ); +if (ret) + pr_warn("fsl_esdhc_imx: Internal clock never + gate off.\n");#else esdhc_clrbits32 (®s->sysctl,SYSCTL_CKEN );#endif }else {#ifdef CONFIG_FSL_USDHC -esdhc_setbits32 (®s->vendorspec,VENDORSPEC_PEREN | -VENDORSPEC_CKEN ); +esdhc_setbits32 (®s->vendorspec, +VENDORSPEC_FRC_SDCLK_ON );#else esdhc_setbits32 (®s->sysctl,SYSCTL_PEREN |SYSCTL_CKEN );#endif @@ -1045,7 +1059,7 @@ static int esdhc_init_common(struct fsl_esdhc_priv *priv, struct mmc *mmc) #ifndef CONFIG_FSL_USDHCesdhc_setbits32 (®s->sysctl,SYSCTL_HCKEN |SYSCTL_IPGEN );#else -esdhc_setbits32 (®s->vendorspec,VENDORSPEC_HCKEN | VENDORSPEC_IPGEN); +esdhc_setbits32 (®s->vendorspec,VENDORSPEC_FRC_SDCLK_ON );#endif /* Set the initial clock speed */ @@ -1183,8 +1197,7 @@ static intfsl_esdhc_init(struct
fsl_esdhc_priv*priv,esdhc_write32(®s->autoc12err, 0); esdhc_write32(®s->clktunectrlstatus, 0); #elsedrivers/mmc/fsl_esdhc_imx.c +1
-2-esdhc_setbits32 (®s->vendorspec,VENDORSPEC_PEREN | -VENDORSPEC_HCKEN |VENDORSPEC_IPGEN |VENDORSPEC_CKEN ); +esdhc_setbits32 (®s->vendorspec,VENDORSPEC_FRC_SDCLK_ON );#endif if (priv->vs18_enable)include/fsl_esdhc_imx.h index renamed from include/fsl_esdhc_imx.h +0
-0
45ed635a77..b092034464100644include/fsl_esdhc_imx.h +2
-1@@ -39,6 +39,7 @@#define VENDORSPEC_HCKEN 0x00001000#define VENDORSPEC_IPGEN 0x00000800#define VENDORSPEC_INIT 0x20007809 +#define VENDORSPEC_FRC_SDCLK_ON 0x00000100#define IRQSTAT 0x0002e030#define IRQSTAT_DMAE (0x10000000)@@ -96,6 +97,7 @@#define PRSSTAT_CINS (0x00010000)#define PRSSTAT_BREN (0x00000800)#define PRSSTAT_BWEN (0x00000400) +#define PRSSTAT_SDOFF (0x00000080)#define PRSSTAT_SDSTB (0X00000008)#define PRSSTAT_DLA (0x00000004)#define PRSSTAT_CICHB (0x00000002) --2.17.1
PATCH 1/1 mmc: fsl_esdhc_imx: use VENDORSPEC_FRC_SDCLK_ON to control card clock output
Quoted 58 lines
-----Original Message-----
From: ZHIZHIKIN Andrey [mailto:andrey.zhizhikin at leica-geosystems.com]
Sent: 2021?2?1? 17:52
To: Bough Chen <[email protected]>; Peng Fan <[email protected]>; u-boot at lists.denx.de
Cc: dl-uboot-imx <[email protected]>; tharvey at gateworks.com
Subject: RE: [PATCH] mmc:fsl_esdhc_imx: useVENDORSPEC_FRC_SDCLK_ONto control card clock outputHello Haibo,
Quoted 9 lines
-----Original Message-----
From: haibo.chen at nxp.com <[email protected]>
Sent: Wednesday, January 27, 2021 11:47 AM
To: peng.fan at nxp.com; u-boot at lists.denx.de
Cc: haibo.chen at nxp.com; uboot-imx at nxp.com; tharvey at gateworks.com; ZHIZHIKIN Andrey <[email protected]>
Subject: [PATCH] mmc:fsl_esdhc_imx: useVENDORSPEC_FRC_SDCLK_ONto control card clock outputQuoted 7 lines
This email is not from Hexagon?s Office 365 instance. Please be careful while clicking links, opening attachments, or replying to this email.
From: Haibo Chen <[email protected]>
For
FSL_USDHC, it do not implementVENDORSPEC_CKEN/PEREN/HCKEN/IPGEN,
Quoted 10 lines
these are reserved bits. Instead, use
VENDORSPEC_FRC_SDCLK_ONto gate on/off the card clock output.After commit
b5874b552ffa("mmc:fsl_esdhc_imx: addwait_dat0()support"), we meet SD3.0 card can't work at UHS mode,mmc_switch_voltage()fail because the secondmmc_wait_dat0return-ETIMEDOUT. According to SD spec, during voltage switch, need to gate off/on the card clock. If not set theFRC_SDCLK_ON, after CMD11, hardware will gate off the card clock automatically, so card do not detect the clock off/on behavior, so will draw the data0 line low until nextcommand.
Quoted 2 lines
Fixes b5874b552ffa ("mmc: fsl_esdhc_imx: add wait_dat0() support")This patch does not fix the switch of uSDHC to 1v8...
I've applied it locally on the imx8mmevk, and had following log during the boot when tried to query SD Card info:
-----------------------------------------------
U-Boot SPL 2021.01-01004-gb852007333 (Feb 01 2021 - 09:45:42 +0100) Normal Boot
WDT: Started with servicing (60s timeout)
Trying to boot from MMC1NOTICE: BL31: v2.2(release):rel_imx_5.4.70_2.3.0-0-gf1d7187f2 NOTICE: BL31: Built : 22:29:05, Jan 17 2021U-Boot 2021.01-01004-gb852007333 (Feb 01 2021 - 09:45:42 +0100)
Quoted 8 lines
CPU: Freescale i.MX8MMQ rev1.0 at 1200 MHz Reset cause: POR Model: FSL i.MX8MM EVK board DRAM: 2 GiB WDT: Started with servicing (60s timeout) MMC: FSL_SDHC: 1, FSL_SDHC: 2Loading Environment from MMC... Run CMD11 1.8V switch Card did not respond to voltage select! : -110
This do not align with my test result. Can you help identify which function first return the timeout on your side? Or can you try a different SD card?
Best Regards
Haibo
Quoted 251 lines
*** Warning - No block device, using default environment
In: serial Out: serial Err: serial Net: eth0: ethernet at 30be0000 Hit any key to stop autoboot: 0u-boot=> mmc dev 1
Card did not respond to voltage select! : -110 u-boot=> mmc info MMC Device 0 not found no mmc device at slot 0 u-boot=> mmc dev 2 switch to partitions #0, OK mmc2(part 0) is current device u-boot=> mmc info
Device:FSL_SDHC
Manufacturer ID: 45
OEM: 100
Name: DG401
Bus Speed: 200000000
Mode: HS400ES (200MHz)
Rd Block Len: 512
MMC version 5.1
High Capacity: Yes
Capacity: 14.7 GiB
Bus Width: 8-bit DDR
Erase Group Size: 512 KiB
HC WP Group Size: 8 MiB
User Capacity: 14.7 GiB WRREL
Boot Capacity: 4 MiB ENH
RPMB Capacity: 4 MiB ENH
Boot area 0 is not write protected
Boot area 1 is not write protected
-----------------------------------------------Note, that eMMC is not affected and is operating in HS400ES mode without any issues.
Reverting patch
b5874b552ffa("mmc:fsl_esdhc_imx: addwait_dat0()support") with this patch applied in tree does switch the SD Card to high-speed mode, following log is produced:
-----------------------------------------------
U-Boot SPL 2021.01-01005-gb8aeb689a2 (Feb 01 2021 - 10:38:01 +0100) Normal Boot
WDT: Started with servicing (60s timeout)
Trying to boot from MMC1NOTICE: BL31: v2.2(release):rel_imx_5.4.70_2.3.0-0-gf1d7187f2 NOTICE: BL31: Built : 22:29:05, Jan 17 2021U-Boot 2021.01-01005-gb8aeb689a2 (Feb 01 2021 - 10:38:01 +0100)
CPU: Freescale i.MX8MMQ rev1.0 at 1200 MHz Reset cause: POR Model: FSL i.MX8MM EVK board DRAM: 2 GiB WDT: Started with servicing (60s timeout) MMC: FSL_SDHC: 1, FSL_SDHC: 2Loading Environment from MMC... Run CMD11 1.8V switch *** Warning - bad CRC, using default environment
In: serial Out: serial Err: serial Net: eth0: ethernet at 30be0000 Hit any key to stop autoboot: 0u-boot=> mmc dev 1
Run CMD11 1.8V switch
switch to partitions #0, OK
mmc1 is current device
u-boot=> mmc info
Device:FSL_SDHC
Manufacturer ID: 41
OEM: 3432
Name: SD32G
Bus Speed: 200000000
Mode: UHS SDR104 (208MHz)
Rd Block Len: 512
SD version 3.0
High Capacity: Yes
Capacity: 29.3 GiB
Bus Width: 4-bit
Erase Group Size: 512 Bytes
u-boot=> mmc dev 2
switch to partitions #0, OK
mmc2(part 0) is current device
u-boot=> mmc info
Device:FSL_SDHC
Manufacturer ID: 45
OEM: 100
Name: DG401
Bus Speed: 200000000
Mode: HS400ES (200MHz)
Rd Block Len: 512
MMC version 5.1
High Capacity: Yes
Capacity: 14.7 GiB
Bus Width: 8-bit DDR
Erase Group Size: 512 KiB
HC WP Group Size: 8 MiB
User Capacity: 14.7 GiB WRREL
Boot Capacity: 4 MiB ENH
RPMB Capacity: 4 MiB ENH
Boot area 0 is not write protected
Boot area 1 is not write protected
-----------------------------------------------I guess high-level modifications are also required, since the uSDHC does not follow the SD Specification in regards to voltage switching, which is currently implemented in U-Boot, this patch alone does not solve the mode switch.
Quoted 12 lines
Signed-off-by Haibo Chen <[email protected]>--- drivers/mmc/fsl_esdhc_imx.c | 29 +++++++++++++++++++++-------- include/fsl_esdhc_imx.h | 2 ++ 2 files changed, 23 insertions(+), 8 deletions(-)drivers/mmc/fsl_esdhc_imx.c +0
-0index
8ac859797f..da33ee8253100644drivers/mmc/fsl_esdhc_imx.c static void set_sysctl(struct fsl_esdhc_priv · line 657
Quoted 6 lines
drivers/mmc/fsl_esdhc_imx.c +1
-1@@ -653,7 +653,10 @@ static void set_sysctl(struct fsl_esdhc_priv *priv, struct mmc *mmc, uint clock) clk = (pre_div <<8 ) | (div <<4 );#ifdef CONFIG_FSL_USDHC -esdhc_clrbits32 (®s->vendorspec,VENDORSPEC_CKEN ); + esdhc_clrbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);drivers/mmc/fsl_esdhc_imx.c static void set_sysctl(struct fsl_esdhc_priv
Quoted 8 lines
drivers/mmc/fsl_esdhc_imx.c +5
-0+ ret =readx_poll_timeout (esdhc_read32, ®s->prsstat, tmp, + tmp &PRSSTAT_SDOFF ,100 ); +if (ret) + pr_warn("fsl_esdhc_imx: Internal clock never gate + off.\n");#else esdhc_clrbits32(®s->sysctl, SYSCTL_CKEN); #endif @@-665,7
Quoted 1 line
+668,7 @@ static void set_sysctl(struct
fsl_esdhc_priv*priv, struct mmc*mmc, uint clock)
Quoted 7 lines
pr_warn("fsl_esdhc_imx: Internal clock never
stabilised.\n");#ifdef
CONFIG_FSL_USDHC
- esdhc_setbits32(®s->vendorspec,VENDORSPEC_PEREN|VENDORSPEC_CKEN);
+ esdhc_setbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);drivers/mmc/fsl_esdhc_imx.c static void esdhc_set_strobe_dll(struct mmc *mmc)
Quoted 12 lines
#else esdhc_setbits32(®s->sysctl, SYSCTL_PEREN | SYSCTL_CKEN); #endif @@ - 720,8 +723,14 @@ static void esdhc_set_strobe_dll(struct mmc *mmc) struct fsl_esdhc_priv *priv = dev_get_priv(mmc->dev); struct fsl_esdhc *regs = priv->esdhc_regs; u32 val;drivers/mmc/fsl_esdhc_imx.c +3
-0+u32 tmp; +int ret;if (priv->clock >ESDHC_STROBE_DLL_CLK_FREQ ) { + esdhc_clrbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);drivers/mmc/fsl_esdhc_imx.c static void esdhc_set_strobe_dll(struct mmc *mmc) · line 748
Quoted 12 lines
+ ret = readx_poll_timeout(esdhc_read32,
®s->prsstat,
+ tmp, tmp &PRSSTAT_SDOFF, 100);drivers/mmc/fsl_esdhc_imx.c +3
-0+if (ret) + pr_warn("fsl_esdhc_imx: Internal clock never + gate off.\n"); esdhc_write32(®s->strobe_dllctrl,
ESDHC_STROBE_DLL_CTRL_RESET);/*drivers/mmc/fsl_esdhc_imx.c +0
-0@@ -739,6 +748,7 @@ static void esdhc_set_strobe_dll(struct mmc *mmc) pr_warn("HS400 strobe DLL status REF notlock!\n");
Quoted 10 lines
drivers/mmc/fsl_esdhc_imx.c +2
-0if (!(val &ESDHC_STROBE_DLL_STS_SLV_LOCK ))pr_warn ("HS400 strobe DLL status SLV not lock!\n"); +esdhc_setbits32 (®s->vendorspec, +VENDORSPEC_FRC_SDCLK_ON ); } }@@ -962,14 +972,18 @@ static int esdhc_set_ios_common(struct
fsl_esdhc_priv*priv, struct mmc *mmc) #ifdefMMC_SUPPORTS_TUNINGQuoted 3 lines
if (mmc->clk_disable) {#ifdef
CONFIG_FSL_USDHC
- esdhc_clrbits32(®s->vendorspec,
VENDORSPEC_CKEN);drivers/mmc/fsl_esdhc_imx.c static int esdhc_set_ios_common(struct fsl_esdhc_priv *priv, struct mmc *mmc)
Quoted 3 lines
drivers/mmc/fsl_esdhc_imx.c +3
-0+u32 tmp; + + esdhc_clrbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);drivers/mmc/fsl_esdhc_imx.c static int esdhc_set_ios_common(struct fsl_esdhc_priv *priv, struct mmc *mmc)
Quoted 11 lines
+ ret = readx_poll_timeout(esdhc_read32,
®s->prsstat,
+ tmp, tmp &PRSSTAT_SDOFF, 100);drivers/mmc/fsl_esdhc_imx.c +3
-1+if (ret) + pr_warn("fsl_esdhc_imx: Internal clock never + gate off.\n");#else esdhc_clrbits32 (®s->sysctl,SYSCTL_CKEN );#endif }else {#ifdef CONFIG_FSL_USDHC - esdhc_setbits32(®s->vendorspec,
VENDORSPEC_PEREN|Quoted 7 lines
drivers/mmc/fsl_esdhc_imx.c +2
-1- VENDORSPEC_CKEN); +esdhc_setbits32 (®s->vendorspec, +VENDORSPEC_FRC_SDCLK_ON );#else esdhc_setbits32(®s->sysctl, SYSCTL_PEREN |
SYSCTL_CKEN); #endif @@ -1045,7 +1059,7 @@ static intesdhc_init_common(struct fsl_esdhc_priv *priv, struct mmc *mmc)#ifndef
CONFIG_FSL_USDHCQuoted 1 line
esdhc_setbits32(®s->sysctl, SYSCTL_HCKEN | SYSCTL_IPGEN);#else
Quoted 3 lines
- esdhc_setbits32(®s->vendorspec,
VENDORSPEC_HCKEN|VENDORSPEC_IPGEN);
+ esdhc_setbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);drivers/mmc/fsl_esdhc_imx.c
Quoted 8 lines
#endif
/* Set the initial clock speed */ @@ -1183,8 +1197,7 @@ static int fsl_esdhc_init(struct fsl_esdhc_priv *priv, esdhc_write32(®s->autoc12err, 0); esdhc_write32(®s->clktunectrlstatus, 0); #elsedrivers/mmc/fsl_esdhc_imx.c +0
-2- esdhc_setbits32(®s->vendorspec, VENDORSPEC_PEREN | - VENDORSPEC_HCKEN | VENDORSPEC_IPGEN |
VENDORSPEC_CKEN);Quoted 1 line
+ esdhc_setbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);Quoted 25 lines
#endif
if (priv->vs18_enable)include/fsl_esdhc_imx.h index renamed from include/fsl_esdhc_imx.h +0
-0
45ed635a77..b092034464100644include/fsl_esdhc_imx.h +2
-1@@ -39,6 +39,7 @@#define VENDORSPEC_HCKEN 0x00001000#define VENDORSPEC_IPGEN 0x00000800#define VENDORSPEC_INIT 0x20007809 +#define VENDORSPEC_FRC_SDCLK_ON 0x00000100#define IRQSTAT 0x0002e030#define IRQSTAT_DMAE (0x10000000)@@ -96,6 +97,7 @@#define PRSSTAT_CINS (0x00010000)#define PRSSTAT_BREN (0x00000800)#define PRSSTAT_BWEN (0x00000400) +#define PRSSTAT_SDOFF (0x00000080)#define PRSSTAT_SDSTB (0X00000008)#define PRSSTAT_DLA (0x00000004)#define PRSSTAT_CICHB (0x00000002) --2.17.1
PATCH 1/1 mmc: fsl_esdhc_imx: use VENDORSPEC_FRC_SDCLK_ON to control card clock output
Hello Haibo,
Quoted 72 lines
-----Original Message-----
From: Bough Chen <[email protected]>
Sent: Monday, February 1, 2021 11:10 AM
To: ZHIZHIKIN Andrey <[email protected]>; Peng Fan <[email protected]>; u-boot at lists.denx.de
Cc: dl-uboot-imx <[email protected]>; tharvey at gateworks.com
Subject: RE: [PATCH] mmc:fsl_esdhc_imx: useVENDORSPEC_FRC_SDCLK_ONto control card clock outputQuoted 54 lines
-----Original Message-----
From: ZHIZHIKIN Andrey [mailto:andrey.zhizhikin at leica-geosystems.com]
Sent: 2021?2?1? 17:52
To: Bough Chen <[email protected]>; Peng Fan <[email protected]>; u-boot at lists.denx.de
Cc: dl-uboot-imx <[email protected]>; tharvey at gateworks.com
Subject: RE: [PATCH] mmc:fsl_esdhc_imx: useVENDORSPEC_FRC_SDCLK_ONto control card clock outputHello Haibo,
Quoted 9 lines
-----Original Message-----
From: haibo.chen at nxp.com <[email protected]>
Sent: Wednesday, January 27, 2021 11:47 AM
To: peng.fan at nxp.com; u-boot at lists.denx.de
Cc: haibo.chen at nxp.com; uboot-imx at nxp.com; tharvey at gateworks.com; ZHIZHIKIN Andrey <[email protected]>
Subject: [PATCH] mmc:fsl_esdhc_imx: useVENDORSPEC_FRC_SDCLK_ONto control card clock outputQuoted 3 lines
From: Haibo Chen <[email protected]>
For
FSL_USDHC, it do not implementVENDORSPEC_CKEN/PEREN/HCKEN/IPGEN,
Quoted 11 lines
these are reserved bits. Instead, use
VENDORSPEC_FRC_SDCLK_ONto gate on/off the card clock output.After commit
b5874b552ffa("mmc:fsl_esdhc_imx: addwait_dat0()support"), we meet SD3.0 card can't work at UHS mode,mmc_switch_voltage()fail because the secondmmc_wait_dat0return-ETIMEDOUT. According to SD spec, during voltage switch, need to gate off/on the card clock. If not set theFRC_SDCLK_ON, after CMD11, hardware will gate off the card clock automatically, so card do not detect the clock off/on behavior, so will draw the data0 line low until nextcommand.
Quoted 2 lines
Fixes b5874b552ffa ("mmc: fsl_esdhc_imx: add wait_dat0() support")This patch does not fix the switch of uSDHC to 1v8...
I've applied it locally on the imx8mmevk, and had following log during the boot when tried to query SD Card info:
-----------------------------------------------
U-Boot SPL 2021.01-01004-gb852007333 (Feb 01 2021 - 09:45:42 +0100) Normal Boot
WDT: Started with servicing (60s timeout)
Trying to boot from MMC1NOTICE: BL31: v2.2(release):rel_imx_5.4.70_2.3.0-0-gf1d7187f2 NOTICE: BL31: Built : 22:29:05, Jan 17 2021U-Boot 2021.01-01004-gb852007333 (Feb 01 2021 - 09:45:42 +0100)
Quoted 8 lines
CPU: Freescale i.MX8MMQ rev1.0 at 1200 MHz Reset cause: POR Model: FSL i.MX8MM EVK board DRAM: 2 GiB WDT: Started with servicing (60s timeout) MMC: FSL_SDHC: 1, FSL_SDHC: 2Loading Environment from MMC... Run CMD11 1.8V switch Card did not respond to voltage select! : -110
Quoted 2 lines
This do not align with my test result. Can you help identify which function first return the timeout on your side?
I would have a look at the exact function, but it seems to me that it would be wait_dat0() since if I revert the patch introducing it - high speed mode switch is not timing out.
Quoted 1 line
Or can you try a different SD card?
It is rather strange, it seems like it is dependent on the SD Card used.
So far, I've tried 3 SD Cards I had on hands, one of which being operable:
== Working:
"Transcend 32GB"
Manufacturer ID: 74
OEM: 4a60
Name: USDU1
== Failed:
1. (Kingston 32GB)
Manufacturer ID: 41
OEM: 3432
Name: SD32G
2. (Intenso 32 GB)
Manufacturer ID: 9f
OEM: 5449
Name: 00000
This raises a concern that a regular off-the-shelf SD Card might exhibit this issue or might not - it is not tight to a one specific SD Card. Hence I believe it is quite important to look at it further, as there might be a lot of reports from various users.
Please note, that for all SD Cards I've tested NVCC_SD2 is indeed physically switched to 1v8.
Quoted 259 lines
Best Regards
HaiboQuoted 104 lines
*** Warning - No block device, using default environment
In: serial Out: serial Err: serial Net: eth0: ethernet at 30be0000 Hit any key to stop autoboot: 0u-boot=> mmc dev 1
Card did not respond to voltage select! : -110 u-boot=> mmc info MMC Device
0 not found no mmc device at slot 0 u-boot=> mmc dev 2 switch to partitions #0, OK mmc2(part 0) is current device u-boot=> mmc info
Device:FSL_SDHC
Manufacturer ID: 45
OEM: 100
Name: DG401
Bus Speed: 200000000
Mode: HS400ES (200MHz)
Rd Block Len: 512
MMC version 5.1
High Capacity: Yes
Capacity: 14.7 GiB
Bus Width: 8-bit DDR
Erase Group Size: 512 KiB
HC WP Group Size: 8 MiB
User Capacity: 14.7 GiB WRREL
Boot Capacity: 4 MiB ENH
RPMB Capacity: 4 MiB ENH
Boot area 0 is not write protected
Boot area 1 is not write protected
-----------------------------------------------Note, that eMMC is not affected and is operating in HS400ES mode without any issues.
Reverting patch
b5874b552ffa("mmc:fsl_esdhc_imx: addwait_dat0()support") with this patch applied in tree does switch the SD Card to high-speed mode, following log is produced:
-----------------------------------------------
U-Boot SPL 2021.01-01005-gb8aeb689a2 (Feb 01 2021 - 10:38:01 +0100) Normal Boot
WDT: Started with servicing (60s timeout)
Trying to boot from MMC1NOTICE: BL31: v2.2(release):rel_imx_5.4.70_2.3.0-0-gf1d7187f2 NOTICE: BL31: Built : 22:29:05, Jan 17 2021U-Boot 2021.01-01005-gb8aeb689a2 (Feb 01 2021 - 10:38:01 +0100)
CPU: Freescale i.MX8MMQ rev1.0 at 1200 MHz Reset cause: POR Model: FSL i.MX8MM EVK board DRAM: 2 GiB WDT: Started with servicing (60s timeout) MMC: FSL_SDHC: 1, FSL_SDHC: 2Loading Environment from MMC... Run CMD11 1.8V switch *** Warning - bad CRC, using default environment
In: serial Out: serial Err: serial Net: eth0: ethernet at 30be0000 Hit any key to stop autoboot: 0u-boot=> mmc dev 1
Run CMD11 1.8V switch
switch to partitions #0, OK
mmc1 is current device
u-boot=> mmc info
Device:FSL_SDHC
Manufacturer ID: 41
OEM: 3432
Name: SD32G
Bus Speed: 200000000
Mode: UHS SDR104 (208MHz)
Rd Block Len: 512
SD version 3.0
High Capacity: Yes
Capacity: 29.3 GiB
Bus Width: 4-bit
Erase Group Size: 512 Bytes
u-boot=> mmc dev 2
switch to partitions #0, OK
mmc2(part 0) is current device
u-boot=> mmc info
Device:FSL_SDHC
Manufacturer ID: 45
OEM: 100
Name: DG401
Bus Speed: 200000000
Mode: HS400ES (200MHz)
Rd Block Len: 512
MMC version 5.1
High Capacity: Yes
Capacity: 14.7 GiB
Bus Width: 8-bit DDR
Erase Group Size: 512 KiB
HC WP Group Size: 8 MiB
User Capacity: 14.7 GiB WRREL
Boot Capacity: 4 MiB ENH
RPMB Capacity: 4 MiB ENH
Boot area 0 is not write protected
Boot area 1 is not write protected
-----------------------------------------------Quoted 3 lines
I guess high-level modifications are also required, since the uSDHC does not follow the SD Specification in regards to voltage switching, which is currently implemented in U-Boot, this patch alone does not solve the
mode switch.
Quoted 147 lines
Quoted 13 lines
Signed-off-by Haibo Chen <[email protected]>--- drivers/mmc/fsl_esdhc_imx.c | 29 +++++++++++++++++++++-------- include/fsl_esdhc_imx.h | 2 ++ 2 files changed, 23 insertions(+), 8 deletions(-)(quoted diff) +0
-0
b/drivers/mmc/fsl_esdhc_imx.cindex8ac859797f..da33ee8253100644drivers/mmc/fsl_esdhc_imx.c static void set_sysctl(struct fsl_esdhc_priv · line 657
Quoted 6 lines
drivers/mmc/fsl_esdhc_imx.c +1
-1@@ -653,7 +653,10 @@ static void set_sysctl(struct fsl_esdhc_priv *priv, struct mmc *mmc, uint clock) clk = (pre_div <<8 ) | (div <<4 );#ifdef CONFIG_FSL_USDHC -esdhc_clrbits32 (®s->vendorspec,VENDORSPEC_CKEN ); + esdhc_clrbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);drivers/mmc/fsl_esdhc_imx.c static void set_sysctl(struct fsl_esdhc_priv
Quoted 8 lines
drivers/mmc/fsl_esdhc_imx.c +5
-0+ ret =readx_poll_timeout (esdhc_read32, ®s->prsstat, tmp, + tmp &PRSSTAT_SDOFF ,100 ); +if (ret) + pr_warn("fsl_esdhc_imx: Internal clock never gate + off.\n");#else esdhc_clrbits32(®s->sysctl, SYSCTL_CKEN); #endif @@-665,7
Quoted 2 lines
+668,7 @@ static void set_sysctl(struct
fsl_esdhc_priv*priv, struct +mmc*mmc, uint clock)
Quoted 7 lines
pr_warn("fsl_esdhc_imx: Internal clock never
stabilised.\n");#ifdef
CONFIG_FSL_USDHC
- esdhc_setbits32(®s->vendorspec,VENDORSPEC_PEREN|VENDORSPEC_CKEN);
+ esdhc_setbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);drivers/mmc/fsl_esdhc_imx.c static void esdhc_set_strobe_dll(struct mmc *mmc)
Quoted 12 lines
#else esdhc_setbits32(®s->sysctl, SYSCTL_PEREN | SYSCTL_CKEN); #endif @@ - 720,8 +723,14 @@ static void esdhc_set_strobe_dll(struct mmc *mmc) struct fsl_esdhc_priv *priv = dev_get_priv(mmc->dev); struct fsl_esdhc *regs = priv->esdhc_regs; u32 val;drivers/mmc/fsl_esdhc_imx.c +3
-0+u32 tmp; +int ret;if (priv->clock >ESDHC_STROBE_DLL_CLK_FREQ ) { + esdhc_clrbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);drivers/mmc/fsl_esdhc_imx.c static void esdhc_set_strobe_dll(struct mmc *mmc) · line 748
Quoted 12 lines
+ ret = readx_poll_timeout(esdhc_read32,
+®s->prsstat, tmp, tmp &PRSSTAT_SDOFF, 100);drivers/mmc/fsl_esdhc_imx.c +3
-0+if (ret) + pr_warn("fsl_esdhc_imx: Internal clock never + gate off.\n"); esdhc_write32(®s->strobe_dllctrl,
ESDHC_STROBE_DLL_CTRL_RESET);/*drivers/mmc/fsl_esdhc_imx.c +0
-0@@ -739,6 +748,7 @@ static void esdhc_set_strobe_dll(struct mmc *mmc) pr_warn("HS400 strobe DLL status REF notlock!\n");
Quoted 10 lines
drivers/mmc/fsl_esdhc_imx.c +2
-0if (!(val &ESDHC_STROBE_DLL_STS_SLV_LOCK ))pr_warn ("HS400 strobe DLL status SLV not lock!\n"); +esdhc_setbits32 (®s->vendorspec, +VENDORSPEC_FRC_SDCLK_ON ); } }@@ -962,14 +972,18 @@ static int esdhc_set_ios_common(struct
fsl_esdhc_priv*priv, struct mmc *mmc) #ifdefMMC_SUPPORTS_TUNINGQuoted 3 lines
if (mmc->clk_disable) {#ifdef
CONFIG_FSL_USDHC
- esdhc_clrbits32(®s->vendorspec,
VENDORSPEC_CKEN);drivers/mmc/fsl_esdhc_imx.c static int esdhc_set_ios_common(struct fsl_esdhc_priv *priv, struct mmc *mmc)
Quoted 3 lines
drivers/mmc/fsl_esdhc_imx.c +3
-0+u32 tmp; + + esdhc_clrbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);drivers/mmc/fsl_esdhc_imx.c static int esdhc_set_ios_common(struct fsl_esdhc_priv *priv, struct mmc *mmc)
Quoted 11 lines
+ ret = readx_poll_timeout(esdhc_read32,
+®s->prsstat, tmp, tmp &PRSSTAT_SDOFF, 100);drivers/mmc/fsl_esdhc_imx.c +3
-1+if (ret) + pr_warn("fsl_esdhc_imx: Internal clock never + gate off.\n");#else esdhc_clrbits32 (®s->sysctl,SYSCTL_CKEN );#endif }else {#ifdef CONFIG_FSL_USDHC - esdhc_setbits32(®s->vendorspec,
VENDORSPEC_PEREN|Quoted 8 lines
drivers/mmc/fsl_esdhc_imx.c +2
-1- VENDORSPEC_CKEN); +esdhc_setbits32 (®s->vendorspec, +VENDORSPEC_FRC_SDCLK_ON );#else esdhc_setbits32(®s->sysctl, SYSCTL_PEREN |
SYSCTL_CKEN); #endif @@ -1045,7 +1059,7 @@ static intesdhc_init_common(struct fsl_esdhc_priv *priv, struct mmc *mmc)#ifndef
CONFIG_FSL_USDHCQuoted 1 line
esdhc_setbits32(®s->sysctl, SYSCTL_HCKEN | SYSCTL_IPGEN);#else
Quoted 3 lines
- esdhc_setbits32(®s->vendorspec,
VENDORSPEC_HCKEN|VENDORSPEC_IPGEN);
+ esdhc_setbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);drivers/mmc/fsl_esdhc_imx.c
Quoted 8 lines
#endif
/* Set the initial clock speed */ @@ -1183,8 +1197,7 @@ static int fsl_esdhc_init(struct fsl_esdhc_priv *priv, esdhc_write32(®s->autoc12err, 0); esdhc_write32(®s->clktunectrlstatus, 0); #elsedrivers/mmc/fsl_esdhc_imx.c +0
-2- esdhc_setbits32(®s->vendorspec, VENDORSPEC_PEREN | - VENDORSPEC_HCKEN | VENDORSPEC_IPGEN |
VENDORSPEC_CKEN);Quoted 1 line
+ esdhc_setbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);Quoted 25 lines
#endif
if (priv->vs18_enable)include/fsl_esdhc_imx.h index renamed from include/fsl_esdhc_imx.h +0
-0
45ed635a77..b092034464100644include/fsl_esdhc_imx.h +2
-1@@ -39,6 +39,7 @@#define VENDORSPEC_HCKEN 0x00001000#define VENDORSPEC_IPGEN 0x00000800#define VENDORSPEC_INIT 0x20007809 +#define VENDORSPEC_FRC_SDCLK_ON 0x00000100#define IRQSTAT 0x0002e030#define IRQSTAT_DMAE (0x10000000)@@ -96,6 +97,7 @@#define PRSSTAT_CINS (0x00010000)#define PRSSTAT_BREN (0x00000800)#define PRSSTAT_BWEN (0x00000400) +#define PRSSTAT_SDOFF (0x00000080)#define PRSSTAT_SDSTB (0X00000008)#define PRSSTAT_DLA (0x00000004)#define PRSSTAT_CICHB (0x00000002) --2.17.1
PATCH 1/1 mmc: fsl_esdhc_imx: use VENDORSPEC_FRC_SDCLK_ON to control card clock output
Quoted 112 lines
-----Original Message-----
From: ZHIZHIKIN Andrey [mailto:andrey.zhizhikin at leica-geosystems.com]
Sent: 2021?2?1? 19:41
To: Bough Chen <[email protected]>; Peng Fan <[email protected]>; u-boot at lists.denx.de
Cc: dl-uboot-imx <[email protected]>; tharvey at gateworks.com
Subject: RE: [PATCH] mmc:fsl_esdhc_imx: useVENDORSPEC_FRC_SDCLK_ONto control card clock outputHello Haibo,
Quoted 7 lines
-----Original Message-----
From: Bough Chen <[email protected]>
Sent: Monday, February 1, 2021 11:10 AM
To: ZHIZHIKIN Andrey <[email protected]>; Peng Fan <[email protected]>; u-boot at lists.denx.de
Cc: dl-uboot-imx <[email protected]>; tharvey at gateworks.com
Subject: RE: [PATCH] mmc:fsl_esdhc_imx: use
VENDORSPEC_FRC_SDCLK_ONQuoted 10 lines
to control card clock output
Quoted 8 lines
-----Original Message-----
From: ZHIZHIKIN Andrey
[mailto:andrey.zhizhikin at leica-geosystems.com]
Sent: 2021?2?1? 17:52
To: Bough Chen <[email protected]>; Peng Fan <[email protected]>; u-boot at lists.denx.de
Cc: dl-uboot-imx <[email protected]>; tharvey at gateworks.com
Subject: RE: [PATCH] mmc:fsl_esdhc_imx: use
VENDORSPEC_FRC_SDCLK_ONQuoted 9 lines
Quoted 9 lines
to control card clock output
Hello Haibo,
Quoted 5 lines
-----Original Message-----
From: haibo.chen at nxp.com <[email protected]>
Sent: Wednesday, January 27, 2021 11:47 AM
To: peng.fan at nxp.com; u-boot at lists.denx.de
Cc: haibo.chen at nxp.com; uboot-imx at nxp.com;tharvey at gateworks.com;
Quoted 48 lines
Quoted 39 lines
Quoted 7 lines
ZHIZHIKIN Andrey <[email protected]>
Subject: [PATCH] mmc:fsl_esdhc_imx: useVENDORSPEC_FRC_SDCLK_ONto control card clock outputFrom: Haibo Chen <[email protected]>
For
FSL_USDHC, it do not implementVENDORSPEC_CKEN/PEREN/HCKEN/IPGEN,
Quoted 11 lines
these are reserved bits. Instead, use
VENDORSPEC_FRC_SDCLK_ONto gate on/off the card clock output.After commit
b5874b552ffa("mmc:fsl_esdhc_imx: addwait_dat0()support"), we meet SD3.0 card can't work at UHS mode,mmc_switch_voltage()fail because the secondmmc_wait_dat0return-ETIMEDOUT. According to SD spec, during voltage switch, need to gate off/on the card clock. If not set theFRC_SDCLK_ON, after CMD11, hardware will gate off the card clock automatically, so card do not detect the clock off/on behavior, so will draw the data0 line low until nextcommand.
Quoted 3 lines
Fixes b5874b552ffa ("mmc: fsl_esdhc_imx: add wait_dat0()support")
This patch does not fix the switch of uSDHC to 1v8...
I've applied it locally on the imx8mmevk, and had following log during the boot when tried to query SD Card info:
-----------------------------------------------
U-Boot SPL 2021.01-01004-gb852007333 (Feb 01 2021 - 09:45:42 +0100) Normal Boot
WDT: Started with servicing (60s timeout)
Trying to boot from MMC1NOTICE: BL31: v2.2(release):rel_imx_5.4.70_2.3.0-0-gf1d7187f2 NOTICE: BL31: Built : 22:29:05, Jan 17 2021U-Boot 2021.01-01004-gb852007333 (Feb 01 2021 - 09:45:42 +0100)
Quoted 8 lines
CPU: Freescale i.MX8MMQ rev1.0 at 1200 MHz Reset cause: POR Model: FSL i.MX8MM EVK board DRAM: 2 GiB WDT: Started with servicing (60s timeout) MMC: FSL_SDHC: 1, FSL_SDHC: 2Loading Environment from MMC... Run CMD11 1.8V switch Card did not respond to voltage select! : -110
Quoted 2 lines
This do not align with my test result. Can you help identify which function first return the timeout on your side?
I would have a look at the exact function, but it seems to me that it would be
wait_dat0()since if I revert the patch introducing it - high speed mode switch is not timing out.Quoted 1 line
Or can you try a different SD card?
It is rather strange, it seems like it is dependent on the SD Card used.
So far, I've tried 3 SD Cards I had on hands, one of which being operable:
== Working:
"Transcend 32GB"
Manufacturer ID: 74
OEM: 4a60
Name: USDU1== Failed:
1. (Kingston 32GB)
Manufacturer ID: 41
OEM: 3432
Name: SD32G
Quoted 5 lines
2. (Intenso 32 GB)
Manufacturer ID: 9f
OEM: 5449
Name: 00000
Hi Andrey,
With this patch, can you also add the following change to test again, I double check the code logic, only the following code do not follow the SD spec for voltage switch. Seems the two failed SD cards follow the sd voltage spec strictly.
drivers/mmc/fsl_esdhc_imx.c +2 -1
@@ -513,7 +513,7 @@ static int esdhc_send_cmd_common(struct fsl_esdhc_priv *priv, struct mmc *mmc, err = -ETIMEDOUT ; goto out; } - +#if 0 /* Switch voltage to 1.8V if CMD11 succeeded */ if (cmd->cmdidx ==SD_CMD_SWITCH_UHS18V ) {esdhc_setbits32 (®s->vendorspec,ESDHC_VENDORSPEC_VSELECT );
@@ -522,6 +522,7 @@ static int esdhc_send_cmd_common(struct fsl_esdhc_priv *priv, struct mmc *mmc,/* Sleep for 5 ms - max time for card to switch to 1.8V */ udelay(5000); } +#endif
Quoted 4 lines
This raises a concern that a regular off-the-shelf SD Card might exhibit this issue or might not - it is not tight to a one specific SD Card. Hence I believe it is quite important to look at it further, as there might be a lot of reports from various users.
Yes, I will take care of that, but I think current issue is not caused by this, because uboot already print out the log " Run CMD11 1.8V", which means the S18A is set in the response of ACMD41. Currently, common code only delay 2ms for power cycle, but some boards may need more time, few i.mx boards need 3.8ms ~4.7ms. for my new imx8mm board, need 2.7ms to let VDD change to 0.5v from 3.3v. And spec need to keep VDD under 0.5v for at least 1ms, so totally need about 3.7ms for the power cycle. According to the log, seems you use an old version board, not sure about the actual delay time on your board.
In Linux, regulator driver support the compatible "off-on-delay-us", let me double check whether uboot also support this feature.
Best Regards
Haibo
Quoted 270 lines
Please note, that for all SD Cards I've tested
NVCC_SD2is indeed physically switched to 1v8.Quoted 175 lines
Best Regards
HaiboQuoted 104 lines
*** Warning - No block device, using default environment
In: serial Out: serial Err: serial Net: eth0: ethernet at 30be0000 Hit any key to stop autoboot: 0u-boot=> mmc dev 1
Card did not respond to voltage select! : -110 u-boot=> mmc info MMC Device
0 not found no mmc device at slot 0 u-boot=> mmc dev 2 switch to partitions #0, OK mmc2(part 0) is current device u-boot=> mmc info
Device:FSL_SDHC
Manufacturer ID: 45
OEM: 100
Name: DG401
Bus Speed: 200000000
Mode: HS400ES (200MHz)
Rd Block Len: 512
MMC version 5.1
High Capacity: Yes
Capacity: 14.7 GiB
Bus Width: 8-bit DDR
Erase Group Size: 512 KiB
HC WP Group Size: 8 MiB
User Capacity: 14.7 GiB WRREL
Boot Capacity: 4 MiB ENH
RPMB Capacity: 4 MiB ENH
Boot area 0 is not write protected
Boot area 1 is not write protected
-----------------------------------------------Note, that eMMC is not affected and is operating in HS400ES mode without any issues.
Reverting patch
b5874b552ffa("mmc:fsl_esdhc_imx: addwait_dat0()support") with this patch applied in tree does switch the SD Card to high-speed mode, following log is produced:
-----------------------------------------------
U-Boot SPL 2021.01-01005-gb8aeb689a2 (Feb 01 2021 - 10:38:01 +0100) Normal Boot
WDT: Started with servicing (60s timeout)
Trying to boot from MMC1NOTICE: BL31: v2.2(release):rel_imx_5.4.70_2.3.0-0-gf1d7187f2 NOTICE: BL31: Built : 22:29:05, Jan 17 2021U-Boot 2021.01-01005-gb8aeb689a2 (Feb 01 2021 - 10:38:01 +0100)
CPU: Freescale i.MX8MMQ rev1.0 at 1200 MHz Reset cause: POR Model: FSL i.MX8MM EVK board DRAM: 2 GiB WDT: Started with servicing (60s timeout) MMC: FSL_SDHC: 1, FSL_SDHC: 2Loading Environment from MMC... Run CMD11 1.8V switch *** Warning - bad CRC, using default environment
In: serial Out: serial Err: serial Net: eth0: ethernet at 30be0000 Hit any key to stop autoboot: 0u-boot=> mmc dev 1
Run CMD11 1.8V switch
switch to partitions #0, OK
mmc1 is current device
u-boot=> mmc info
Device:FSL_SDHC
Manufacturer ID: 41
OEM: 3432
Name: SD32G
Bus Speed: 200000000
Mode: UHS SDR104 (208MHz)
Rd Block Len: 512
SD version 3.0
High Capacity: Yes
Capacity: 29.3 GiB
Bus Width: 4-bit
Erase Group Size: 512 Bytes
u-boot=> mmc dev 2
switch to partitions #0, OK
mmc2(part 0) is current device
u-boot=> mmc info
Device:FSL_SDHC
Manufacturer ID: 45
OEM: 100
Name: DG401
Bus Speed: 200000000
Mode: HS400ES (200MHz)
Rd Block Len: 512
MMC version 5.1
High Capacity: Yes
Capacity: 14.7 GiB
Bus Width: 8-bit DDR
Erase Group Size: 512 KiB
HC WP Group Size: 8 MiB
User Capacity: 14.7 GiB WRREL
Boot Capacity: 4 MiB ENH
RPMB Capacity: 4 MiB ENH
Boot area 0 is not write protected
Boot area 1 is not write protected
-----------------------------------------------Quoted 4 lines
I guess high-level modifications are also required, since the uSDHC does not follow the SD Specification in regards to voltage switching, which is currently implemented in U-Boot, this patch alone does not solve the
mode switch.
Quoted 62 lines
Quoted 13 lines
Signed-off-by Haibo Chen <[email protected]>--- drivers/mmc/fsl_esdhc_imx.c | 29 +++++++++++++++++++++-------- include/fsl_esdhc_imx.h | 2 ++ 2 files changed, 23 insertions(+), 8 deletions(-)(quoted diff) +0
-0
b/drivers/mmc/fsl_esdhc_imx.cindex8ac859797f..da33ee8253100644drivers/mmc/fsl_esdhc_imx.c static void set_sysctl(struct fsl_esdhc_priv · line 657
Quoted 6 lines
drivers/mmc/fsl_esdhc_imx.c +1
-1@@ -653,7 +653,10 @@ static void set_sysctl(struct fsl_esdhc_priv *priv, struct mmc *mmc, uint clock) clk = (pre_div <<8 ) | (div <<4 );#ifdef CONFIG_FSL_USDHC -esdhc_clrbits32 (®s->vendorspec,VENDORSPEC_CKEN ); + esdhc_clrbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);drivers/mmc/fsl_esdhc_imx.c static void set_sysctl(struct fsl_esdhc_priv
Quoted 8 lines
drivers/mmc/fsl_esdhc_imx.c +5
-0+ ret =readx_poll_timeout (esdhc_read32, ®s->prsstat, + tmp, tmp &PRSSTAT_SDOFF ,100 ); +if (ret) + pr_warn("fsl_esdhc_imx: Internal clock never gate + off.\n");#else esdhc_clrbits32(®s->sysctl, SYSCTL_CKEN); #endif @@-665,7
Quoted 2 lines
+668,7 @@ static void set_sysctl(struct
fsl_esdhc_priv*priv, +struct mmc*mmc, uint clock)
Quoted 7 lines
pr_warn("fsl_esdhc_imx: Internal clock never
stabilised.\n");#ifdef
CONFIG_FSL_USDHC
- esdhc_setbits32(®s->vendorspec,VENDORSPEC_PEREN|VENDORSPEC_CKEN);
+ esdhc_setbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);drivers/mmc/fsl_esdhc_imx.c
Quoted 12 lines
#else esdhc_setbits32(®s->sysctl, SYSCTL_PEREN | SYSCTL_CKEN); #endif @@ - 720,8 +723,14 @@ static void esdhc_set_strobe_dll(struct mmc *mmc) struct fsl_esdhc_priv *priv = dev_get_priv(mmc->dev); struct fsl_esdhc *regs = priv->esdhc_regs; u32 val;drivers/mmc/fsl_esdhc_imx.c +3
-0+u32 tmp; +int ret;if (priv->clock >ESDHC_STROBE_DLL_CLK_FREQ ) { + esdhc_clrbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);Quoted 9 lines
+ ret = readx_poll_timeout(esdhc_read32,
+®s->prsstat, tmp, tmp &PRSSTAT_SDOFF, 100);drivers/mmc/fsl_esdhc_imx.c +3
-0+if (ret) + pr_warn("fsl_esdhc_imx: Internal clock + never gate off.\n"); esdhc_write32(®s->strobe_dllctrl,
ESDHC_STROBE_DLL_CTRL_RESET);Quoted 2 lines
Quoted 2 lines
drivers/mmc/fsl_esdhc_imx.c static void esdhc_set_strobe_dll(struct mmc
Quoted 2 lines
/*drivers/mmc/fsl_esdhc_imx.c +0
-0@@ -739,6 +748,7 @@ static void esdhc_set_strobe_dll(struct mmc*mmc)
Quoted 10 lines
Quoted 10 lines
drivers/mmc/fsl_esdhc_imx.c static void esdhc_set_strobe_dll(struct mmc · line 748
Quoted 1 line
drivers/mmc/fsl_esdhc_imx.c +0
-0pr_warn("HS400 strobe DLL status REF notlock!\n");
Quoted 8 lines
drivers/mmc/fsl_esdhc_imx.c +2
-0if (!(val &ESDHC_STROBE_DLL_STS_SLV_LOCK ))pr_warn ("HS400 strobe DLL status SLV not lock!\n"); +esdhc_setbits32 (®s->vendorspec, +VENDORSPEC_FRC_SDCLK_ON ); } }Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
drivers/mmc/fsl_esdhc_imx.c +0
-0@@ -962,14 +972,18 @@ static int esdhc_set_ios_common(struct
fsl_esdhc_priv*priv, struct mmc *mmc) #ifdef
MMC_SUPPORTS_TUNINGQuoted 8 lines
Quoted 8 lines
Quoted 3 lines
if (mmc->clk_disable) {#ifdef
CONFIG_FSL_USDHC
- esdhc_clrbits32(®s->vendorspec,
VENDORSPEC_CKEN);drivers/mmc/fsl_esdhc_imx.c
Quoted 3 lines
drivers/mmc/fsl_esdhc_imx.c +3
-0+u32 tmp; + + esdhc_clrbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);Quoted 8 lines
Quoted 8 lines
drivers/mmc/fsl_esdhc_imx.c
Quoted 8 lines
+ ret = readx_poll_timeout(esdhc_read32,
+®s->prsstat, tmp, tmp &PRSSTAT_SDOFF, 100);drivers/mmc/fsl_esdhc_imx.c +3
-0+if (ret) + pr_warn("fsl_esdhc_imx: Internal clock + never gate off.\n");#else esdhc_clrbits32 (®s->sysctl,SYSCTL_CKEN );#endif
Quoted 20 lines
Quoted 20 lines
Quoted 3 lines
} else {#ifdef
CONFIG_FSL_USDHC
- esdhc_setbits32(®s->vendorspec,
VENDORSPEC_PEREN|Quoted 8 lines
drivers/mmc/fsl_esdhc_imx.c +2
-1- VENDORSPEC_CKEN); +esdhc_setbits32 (®s->vendorspec, +VENDORSPEC_FRC_SDCLK_ON );#else esdhc_setbits32(®s->sysctl, SYSCTL_PEREN |
SYSCTL_CKEN); #endif @@ -1045,7 +1059,7 @@ static intesdhc_init_common(struct fsl_esdhc_priv *priv, struct mmc *mmc)#ifndef
CONFIG_FSL_USDHCQuoted 2 lines
esdhc_setbits32(®s->sysctl, SYSCTL_HCKEN |
SYSCTL_IPGEN);#else
Quoted 3 lines
- esdhc_setbits32(®s->vendorspec,
VENDORSPEC_HCKEN|VENDORSPEC_IPGEN);
+ esdhc_setbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);Quoted 8 lines
Quoted 8 lines
drivers/mmc/fsl_esdhc_imx.c
Quoted 8 lines
#endif
/* Set the initial clock speed */ @@ -1183,8 +1197,7 @@ static int fsl_esdhc_init(struct fsl_esdhc_priv *priv, esdhc_write32(®s->autoc12err, 0); esdhc_write32(®s->clktunectrlstatus, 0); #elsedrivers/mmc/fsl_esdhc_imx.c +0
-2- esdhc_setbits32(®s->vendorspec, VENDORSPEC_PEREN | - VENDORSPEC_HCKEN |
VENDORSPEC_IPGEN|Quoted 29 lines
Quoted 29 lines
VENDORSPEC_CKEN);Quoted 1 line
+ esdhc_setbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);Quoted 26 lines
#endif
if (priv->vs18_enable)include/fsl_esdhc_imx.h +0
-0index
45ed635a77..b092034464100644include/fsl_esdhc_imx.h +2
-1@@ -39,6 +39,7 @@#define VENDORSPEC_HCKEN 0x00001000#define VENDORSPEC_IPGEN 0x00000800#define VENDORSPEC_INIT 0x20007809 +#define VENDORSPEC_FRC_SDCLK_ON 0x00000100#define IRQSTAT 0x0002e030#define IRQSTAT_DMAE (0x10000000)@@ -96,6 +97,7 @@#define PRSSTAT_CINS (0x00010000)#define PRSSTAT_BREN (0x00000800)#define PRSSTAT_BWEN (0x00000400) +#define PRSSTAT_SDOFF (0x00000080)#define PRSSTAT_SDSTB (0X00000008)#define PRSSTAT_DLA (0x00000004)#define PRSSTAT_CICHB (0x00000002) --2.17.1
PATCH 1/1 mmc: fsl_esdhc_imx: use VENDORSPEC_FRC_SDCLK_ON to control card clock output
Hi all,
Quoted 148 lines · Bough Chen
Quoted 112 lines
-----Original Message-----
From: ZHIZHIKIN Andrey [mailto:andrey.zhizhikin at leica-geosystems.com]
Sent: 2021?2?1? 19:41
To: Bough Chen <[email protected]>; Peng Fan <[email protected]>; u-boot at lists.denx.de
Cc: dl-uboot-imx <[email protected]>; tharvey at gateworks.com
Subject: RE: [PATCH] mmc:fsl_esdhc_imx: useVENDORSPEC_FRC_SDCLK_ONto control card clock outputHello Haibo,
Quoted 7 lines
-----Original Message-----
From: Bough Chen <[email protected]>
Sent: Monday, February 1, 2021 11:10 AM
To: ZHIZHIKIN Andrey <[email protected]>; Peng Fan <[email protected]>; u-boot at lists.denx.de
Cc: dl-uboot-imx <[email protected]>; tharvey at gateworks.com
Subject: RE: [PATCH] mmc:fsl_esdhc_imx: use
VENDORSPEC_FRC_SDCLK_ONQuoted 10 lines
to control card clock output
Quoted 8 lines
-----Original Message-----
From: ZHIZHIKIN Andrey
[mailto:andrey.zhizhikin at leica-geosystems.com]
Sent: 2021?2?1? 17:52
To: Bough Chen <[email protected]>; Peng Fan <[email protected]>; u-boot at lists.denx.de
Cc: dl-uboot-imx <[email protected]>; tharvey at gateworks.com
Subject: RE: [PATCH] mmc:fsl_esdhc_imx: use
VENDORSPEC_FRC_SDCLK_ONQuoted 9 lines
Quoted 9 lines
to control card clock output
Hello Haibo,
Quoted 5 lines
-----Original Message-----
From: haibo.chen at nxp.com <[email protected]>
Sent: Wednesday, January 27, 2021 11:47 AM
To: peng.fan at nxp.com; u-boot at lists.denx.de
Cc: haibo.chen at nxp.com; uboot-imx at nxp.com;tharvey at gateworks.com;
Quoted 48 lines
Quoted 39 lines
Quoted 7 lines
ZHIZHIKIN Andrey <[email protected]>
Subject: [PATCH] mmc:fsl_esdhc_imx: useVENDORSPEC_FRC_SDCLK_ONto control card clock outputFrom: Haibo Chen <[email protected]>
For
FSL_USDHC, it do not implementVENDORSPEC_CKEN/PEREN/HCKEN/IPGEN,
Quoted 11 lines
these are reserved bits. Instead, use
VENDORSPEC_FRC_SDCLK_ONto gate on/off the card clock output.After commit
b5874b552ffa("mmc:fsl_esdhc_imx: addwait_dat0()support"), we meet SD3.0 card can't work at UHS mode,mmc_switch_voltage()fail because the secondmmc_wait_dat0return-ETIMEDOUT. According to SD spec, during voltage switch, need to gate off/on the card clock. If not set theFRC_SDCLK_ON, after CMD11, hardware will gate off the card clock automatically, so card do not detect the clock off/on behavior, so will draw the data0 line low until nextcommand.
Quoted 3 lines
Fixes b5874b552ffa ("mmc: fsl_esdhc_imx: add wait_dat0()support")
This patch does not fix the switch of uSDHC to 1v8...
I've applied it locally on the imx8mmevk, and had following log during the boot when tried to query SD Card info:
-----------------------------------------------
U-Boot SPL 2021.01-01004-gb852007333 (Feb 01 2021 - 09:45:42 +0100) Normal Boot
WDT: Started with servicing (60s timeout)
Trying to boot from MMC1NOTICE: BL31: v2.2(release):rel_imx_5.4.70_2.3.0-0-gf1d7187f2 NOTICE: BL31: Built : 22:29:05, Jan 17 2021U-Boot 2021.01-01004-gb852007333 (Feb 01 2021 - 09:45:42 +0100)
Quoted 8 lines
CPU: Freescale i.MX8MMQ rev1.0 at 1200 MHz Reset cause: POR Model: FSL i.MX8MM EVK board DRAM: 2 GiB WDT: Started with servicing (60s timeout) MMC: FSL_SDHC: 1, FSL_SDHC: 2Loading Environment from MMC... Run CMD11 1.8V switch Card did not respond to voltage select! : -110
Quoted 2 lines
This do not align with my test result. Can you help identify which function first return the timeout on your side?
I would have a look at the exact function, but it seems to me that it would be
wait_dat0()since if I revert the patch introducing it - high speed mode switch is not timing out.Quoted 1 line
Or can you try a different SD card?
It is rather strange, it seems like it is dependent on the SD Card used.
So far, I've tried 3 SD Cards I had on hands, one of which being operable:
== Working:
"Transcend 32GB"
Manufacturer ID: 74
OEM: 4a60
Name: USDU1== Failed:
1. (Kingston 32GB)
Manufacturer ID: 41
OEM: 3432
Name: SD32GQuoted 5 lines
2. (Intenso 32 GB)
Manufacturer ID: 9f
OEM: 5449
Name: 00000Hi Andrey,
With this patch, can you also add the following change to test again, I double check the code logic, only the following code do not follow the SD spec for voltage switch. Seems the two failed SD cards follow the sd voltage spec strictly.
drivers/mmc/fsl_esdhc_imx.c +2
-1@@ -513,7 +513,7 @@ static int esdhc_send_cmd_common(struct fsl_esdhc_priv *priv, struct mmc *mmc, err = -ETIMEDOUT ; goto out; } - +#if 0 /* Switch voltage to 1.8V if CMD11 succeeded */ if (cmd->cmdidx ==SD_CMD_SWITCH_UHS18V ) {esdhc_setbits32 (®s->vendorspec,ESDHC_VENDORSPEC_VSELECT );@@ -522,6 +522,7 @@ static int esdhc_send_cmd_common(struct fsl_esdhc_priv *priv, struct mmc *mmc,/* Sleep for 5 ms - max time for card to switch to 1.8V */ udelay(5000); } +#endif Quoted 4 lines
This raises a concern that a regular off-the-shelf SD Card might exhibit this issue or might not - it is not tight to a one specific SD Card. Hence I believe it is quite important to look at it further, as there might be a lot of reports from various users.
Quoted 5 lines
Yes, I will take care of that, but I think current issue is not caused by this, because uboot already print out the log " Run CMD11 1.8V", which means the S18A is set in the response of ACMD41. Currently, common code only delay 2ms for power cycle, but some boards may need more time, few i.mx boards need 3.8ms ~4.7ms. for my new imx8mm board, need 2.7ms to let VDD change to 0.5v from 3.3v. And spec need to keep VDD under 0.5v for at least 1ms, so totally need about 3.7ms for the power cycle. According to the log, seems you use an old version board, not sure about the actual delay time on your board.
In Linux, regulator driver support the compatible "off-on-delay-us", let me double check whether uboot also support this feature.
I didn't have target that fsl_esdhc_imx is used. But when i have checked fsl_esdhc_imx.c, google coral dev board is using it, right? Then is it possible to produce this issue with its board?
Best Regards,
Jaehoon Chung
Quoted 277 lines
Best Regards
HaiboQuoted 270 lines
Please note, that for all SD Cards I've tested
NVCC_SD2is indeed physically switched to 1v8.Quoted 175 lines
Best Regards
HaiboQuoted 104 lines
*** Warning - No block device, using default environment
In: serial Out: serial Err: serial Net: eth0: ethernet at 30be0000 Hit any key to stop autoboot: 0u-boot=> mmc dev 1
Card did not respond to voltage select! : -110 u-boot=> mmc info MMC Device
0 not found no mmc device at slot 0 u-boot=> mmc dev 2 switch to partitions #0, OK mmc2(part 0) is current device u-boot=> mmc info
Device:FSL_SDHC
Manufacturer ID: 45
OEM: 100
Name: DG401
Bus Speed: 200000000
Mode: HS400ES (200MHz)
Rd Block Len: 512
MMC version 5.1
High Capacity: Yes
Capacity: 14.7 GiB
Bus Width: 8-bit DDR
Erase Group Size: 512 KiB
HC WP Group Size: 8 MiB
User Capacity: 14.7 GiB WRREL
Boot Capacity: 4 MiB ENH
RPMB Capacity: 4 MiB ENH
Boot area 0 is not write protected
Boot area 1 is not write protected
-----------------------------------------------Note, that eMMC is not affected and is operating in HS400ES mode without any issues.
Reverting patch
b5874b552ffa("mmc:fsl_esdhc_imx: addwait_dat0()support") with this patch applied in tree does switch the SD Card to high-speed mode, following log is produced:
-----------------------------------------------
U-Boot SPL 2021.01-01005-gb8aeb689a2 (Feb 01 2021 - 10:38:01 +0100) Normal Boot
WDT: Started with servicing (60s timeout)
Trying to boot from MMC1NOTICE: BL31: v2.2(release):rel_imx_5.4.70_2.3.0-0-gf1d7187f2 NOTICE: BL31: Built : 22:29:05, Jan 17 2021U-Boot 2021.01-01005-gb8aeb689a2 (Feb 01 2021 - 10:38:01 +0100)
CPU: Freescale i.MX8MMQ rev1.0 at 1200 MHz Reset cause: POR Model: FSL i.MX8MM EVK board DRAM: 2 GiB WDT: Started with servicing (60s timeout) MMC: FSL_SDHC: 1, FSL_SDHC: 2Loading Environment from MMC... Run CMD11 1.8V switch *** Warning - bad CRC, using default environment
In: serial Out: serial Err: serial Net: eth0: ethernet at 30be0000 Hit any key to stop autoboot: 0u-boot=> mmc dev 1
Run CMD11 1.8V switch
switch to partitions #0, OK
mmc1 is current device
u-boot=> mmc info
Device:FSL_SDHC
Manufacturer ID: 41
OEM: 3432
Name: SD32G
Bus Speed: 200000000
Mode: UHS SDR104 (208MHz)
Rd Block Len: 512
SD version 3.0
High Capacity: Yes
Capacity: 29.3 GiB
Bus Width: 4-bit
Erase Group Size: 512 Bytes
u-boot=> mmc dev 2
switch to partitions #0, OK
mmc2(part 0) is current device
u-boot=> mmc info
Device:FSL_SDHC
Manufacturer ID: 45
OEM: 100
Name: DG401
Bus Speed: 200000000
Mode: HS400ES (200MHz)
Rd Block Len: 512
MMC version 5.1
High Capacity: Yes
Capacity: 14.7 GiB
Bus Width: 8-bit DDR
Erase Group Size: 512 KiB
HC WP Group Size: 8 MiB
User Capacity: 14.7 GiB WRREL
Boot Capacity: 4 MiB ENH
RPMB Capacity: 4 MiB ENH
Boot area 0 is not write protected
Boot area 1 is not write protected
-----------------------------------------------Quoted 4 lines
I guess high-level modifications are also required, since the uSDHC does not follow the SD Specification in regards to voltage switching, which is currently implemented in U-Boot, this patch alone does not solve the
mode switch.
Quoted 62 lines
Quoted 13 lines
Signed-off-by Haibo Chen <[email protected]>--- drivers/mmc/fsl_esdhc_imx.c | 29 +++++++++++++++++++++-------- include/fsl_esdhc_imx.h | 2 ++ 2 files changed, 23 insertions(+), 8 deletions(-)(quoted diff) +0
-0
b/drivers/mmc/fsl_esdhc_imx.cindex8ac859797f..da33ee8253100644drivers/mmc/fsl_esdhc_imx.c static void set_sysctl(struct fsl_esdhc_priv · line 657
Quoted 6 lines
drivers/mmc/fsl_esdhc_imx.c +1
-1@@ -653,7 +653,10 @@ static void set_sysctl(struct fsl_esdhc_priv *priv, struct mmc *mmc, uint clock) clk = (pre_div <<8 ) | (div <<4 );#ifdef CONFIG_FSL_USDHC -esdhc_clrbits32 (®s->vendorspec,VENDORSPEC_CKEN ); + esdhc_clrbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);drivers/mmc/fsl_esdhc_imx.c static void set_sysctl(struct fsl_esdhc_priv
Quoted 8 lines
drivers/mmc/fsl_esdhc_imx.c +5
-0+ ret =readx_poll_timeout (esdhc_read32, ®s->prsstat, + tmp, tmp &PRSSTAT_SDOFF ,100 ); +if (ret) + pr_warn("fsl_esdhc_imx: Internal clock never gate + off.\n");#else esdhc_clrbits32(®s->sysctl, SYSCTL_CKEN); #endif @@-665,7
Quoted 2 lines
+668,7 @@ static void set_sysctl(struct
fsl_esdhc_priv*priv, +struct mmc*mmc, uint clock)
Quoted 7 lines
pr_warn("fsl_esdhc_imx: Internal clock never
stabilised.\n");#ifdef
CONFIG_FSL_USDHC
- esdhc_setbits32(®s->vendorspec,VENDORSPEC_PEREN|VENDORSPEC_CKEN);
+ esdhc_setbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);drivers/mmc/fsl_esdhc_imx.c
Quoted 12 lines
#else esdhc_setbits32(®s->sysctl, SYSCTL_PEREN | SYSCTL_CKEN); #endif @@ - 720,8 +723,14 @@ static void esdhc_set_strobe_dll(struct mmc *mmc) struct fsl_esdhc_priv *priv = dev_get_priv(mmc->dev); struct fsl_esdhc *regs = priv->esdhc_regs; u32 val;drivers/mmc/fsl_esdhc_imx.c +3
-0+u32 tmp; +int ret;if (priv->clock >ESDHC_STROBE_DLL_CLK_FREQ ) { + esdhc_clrbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);Quoted 9 lines
+ ret = readx_poll_timeout(esdhc_read32,
+®s->prsstat, tmp, tmp &PRSSTAT_SDOFF, 100);drivers/mmc/fsl_esdhc_imx.c +3
-0+if (ret) + pr_warn("fsl_esdhc_imx: Internal clock + never gate off.\n"); esdhc_write32(®s->strobe_dllctrl,
ESDHC_STROBE_DLL_CTRL_RESET);Quoted 2 lines
Quoted 2 lines
drivers/mmc/fsl_esdhc_imx.c static void esdhc_set_strobe_dll(struct mmc
Quoted 2 lines
/*drivers/mmc/fsl_esdhc_imx.c +0
-0@@ -739,6 +748,7 @@ static void esdhc_set_strobe_dll(struct mmc*mmc)
Quoted 10 lines
Quoted 10 lines
drivers/mmc/fsl_esdhc_imx.c static void esdhc_set_strobe_dll(struct mmc · line 748
Quoted 1 line
drivers/mmc/fsl_esdhc_imx.c +0
-0pr_warn("HS400 strobe DLL status REF notlock!\n");
Quoted 8 lines
drivers/mmc/fsl_esdhc_imx.c +2
-0if (!(val &ESDHC_STROBE_DLL_STS_SLV_LOCK ))pr_warn ("HS400 strobe DLL status SLV not lock!\n"); +esdhc_setbits32 (®s->vendorspec, +VENDORSPEC_FRC_SDCLK_ON ); } }Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
drivers/mmc/fsl_esdhc_imx.c +0
-0@@ -962,14 +972,18 @@ static int esdhc_set_ios_common(struct
fsl_esdhc_priv*priv, struct mmc *mmc) #ifdef
MMC_SUPPORTS_TUNINGQuoted 8 lines
Quoted 8 lines
Quoted 3 lines
if (mmc->clk_disable) {#ifdef
CONFIG_FSL_USDHC
- esdhc_clrbits32(®s->vendorspec,
VENDORSPEC_CKEN);drivers/mmc/fsl_esdhc_imx.c
Quoted 3 lines
drivers/mmc/fsl_esdhc_imx.c +3
-0+u32 tmp; + + esdhc_clrbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);Quoted 8 lines
Quoted 8 lines
drivers/mmc/fsl_esdhc_imx.c
Quoted 8 lines
+ ret = readx_poll_timeout(esdhc_read32,
+®s->prsstat, tmp, tmp &PRSSTAT_SDOFF, 100);drivers/mmc/fsl_esdhc_imx.c +3
-0+if (ret) + pr_warn("fsl_esdhc_imx: Internal clock + never gate off.\n");#else esdhc_clrbits32 (®s->sysctl,SYSCTL_CKEN );#endif
Quoted 20 lines
Quoted 20 lines
Quoted 3 lines
} else {#ifdef
CONFIG_FSL_USDHC
- esdhc_setbits32(®s->vendorspec,
VENDORSPEC_PEREN|Quoted 8 lines
drivers/mmc/fsl_esdhc_imx.c +2
-1- VENDORSPEC_CKEN); +esdhc_setbits32 (®s->vendorspec, +VENDORSPEC_FRC_SDCLK_ON );#else esdhc_setbits32(®s->sysctl, SYSCTL_PEREN |
SYSCTL_CKEN); #endif @@ -1045,7 +1059,7 @@ static intesdhc_init_common(struct fsl_esdhc_priv *priv, struct mmc *mmc)#ifndef
CONFIG_FSL_USDHCQuoted 2 lines
esdhc_setbits32(®s->sysctl, SYSCTL_HCKEN |
SYSCTL_IPGEN);#else
Quoted 3 lines
- esdhc_setbits32(®s->vendorspec,
VENDORSPEC_HCKEN|VENDORSPEC_IPGEN);
+ esdhc_setbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);Quoted 8 lines
Quoted 8 lines
drivers/mmc/fsl_esdhc_imx.c
Quoted 8 lines
#endif
/* Set the initial clock speed */ @@ -1183,8 +1197,7 @@ static int fsl_esdhc_init(struct fsl_esdhc_priv *priv, esdhc_write32(®s->autoc12err, 0); esdhc_write32(®s->clktunectrlstatus, 0); #elsedrivers/mmc/fsl_esdhc_imx.c +0
-2- esdhc_setbits32(®s->vendorspec, VENDORSPEC_PEREN | - VENDORSPEC_HCKEN |
VENDORSPEC_IPGEN|Quoted 29 lines
Quoted 29 lines
VENDORSPEC_CKEN);Quoted 1 line
+ esdhc_setbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);Quoted 26 lines
#endif
if (priv->vs18_enable)include/fsl_esdhc_imx.h +0
-0index
45ed635a77..b092034464100644include/fsl_esdhc_imx.h +2
-1@@ -39,6 +39,7 @@#define VENDORSPEC_HCKEN 0x00001000#define VENDORSPEC_IPGEN 0x00000800#define VENDORSPEC_INIT 0x20007809 +#define VENDORSPEC_FRC_SDCLK_ON 0x00000100#define IRQSTAT 0x0002e030#define IRQSTAT_DMAE (0x10000000)@@ -96,6 +97,7 @@#define PRSSTAT_CINS (0x00010000)#define PRSSTAT_BREN (0x00000800)#define PRSSTAT_BWEN (0x00000400) +#define PRSSTAT_SDOFF (0x00000080)#define PRSSTAT_SDSTB (0X00000008)#define PRSSTAT_DLA (0x00000004)#define PRSSTAT_CICHB (0x00000002) --2.17.1
PATCH 1/1 mmc: fsl_esdhc_imx: use VENDORSPEC_FRC_SDCLK_ON to control card clock output
In reply to Bough ChenHello Haibo,
Quoted 151 lines
-----Original Message-----
From: Bough Chen <[email protected]>
Sent: Friday, February 5, 2021 8:24 AM
To: ZHIZHIKIN Andrey <[email protected]>; Peng Fan <[email protected]>; u-boot at lists.denx.de
Cc: dl-uboot-imx <[email protected]>; tharvey at gateworks.com
Subject: RE: [PATCH] mmc:fsl_esdhc_imx: useVENDORSPEC_FRC_SDCLK_ONto control card clock outputQuoted 114 lines
-----Original Message-----
From: ZHIZHIKIN Andrey [mailto:andrey.zhizhikin at leica-geosystems.com]
Sent: 2021?2?1? 19:41
To: Bough Chen <[email protected]>; Peng Fan <[email protected]>; u-boot at lists.denx.de
Cc: dl-uboot-imx <[email protected]>; tharvey at gateworks.com
Subject: RE: [PATCH] mmc:fsl_esdhc_imx: useVENDORSPEC_FRC_SDCLK_ONto control card clock outputHello Haibo,
Quoted 7 lines
-----Original Message-----
From: Bough Chen <[email protected]>
Sent: Monday, February 1, 2021 11:10 AM
To: ZHIZHIKIN Andrey <[email protected]>; Peng Fan <[email protected]>; u-boot at lists.denx.de
Cc: dl-uboot-imx <[email protected]>; tharvey at gateworks.com
Subject: RE: [PATCH] mmc:fsl_esdhc_imx: use
VENDORSPEC_FRC_SDCLK_ONQuoted 10 lines
to control card clock output
Quoted 8 lines
-----Original Message-----
From: ZHIZHIKIN Andrey
[mailto:andrey.zhizhikin at leica-geosystems.com]
Sent: 2021?2?1? 17:52
To: Bough Chen <[email protected]>; Peng Fan <[email protected]>; u-boot at lists.denx.de
Cc: dl-uboot-imx <[email protected]>; tharvey at gateworks.com
Subject: RE: [PATCH] mmc:fsl_esdhc_imx: use
VENDORSPEC_FRC_SDCLK_ONQuoted 9 lines
Quoted 9 lines
to control card clock output
Hello Haibo,
Quoted 5 lines
-----Original Message-----
From: haibo.chen at nxp.com <[email protected]>
Sent: Wednesday, January 27, 2021 11:47 AM
To: peng.fan at nxp.com; u-boot at lists.denx.de
Cc: haibo.chen at nxp.com; uboot-imx at nxp.com;tharvey at gateworks.com;
Quoted 49 lines
Quoted 40 lines
Quoted 7 lines
ZHIZHIKIN Andrey <[email protected]>
Subject: [PATCH] mmc:fsl_esdhc_imx: useVENDORSPEC_FRC_SDCLK_ONto control card clock outputFrom: Haibo Chen <[email protected]>
For
FSL_USDHC, it do not implementVENDORSPEC_CKEN/PEREN/HCKEN/IPGEN,
Quoted 12 lines
these are reserved bits. Instead, use
VENDORSPEC_FRC_SDCLK_ONto gate on/off the card clock output.After commit
b5874b552ffa("mmc:fsl_esdhc_imx: addwait_dat0()support"), we meet SD3.0 card can't work at UHS mode,mmc_switch_voltage()fail because the secondmmc_wait_dat0return-ETIMEDOUT. According to SD spec, during voltage switch, need to gate off/on the card clock. If not set theFRC_SDCLK_ON, after CMD11, hardware will gate off the card clock automatically, so card do not detect the clock off/on behavior, so will draw the
data0 line low until nextcommand.
Quoted 3 lines
Fixes b5874b552ffa ("mmc: fsl_esdhc_imx: add wait_dat0()support")
This patch does not fix the switch of uSDHC to 1v8...
I've applied it locally on the imx8mmevk, and had following log during the boot when tried to query SD Card info:
-----------------------------------------------
U-Boot SPL 2021.01-01004-gb852007333 (Feb 01 2021 - 09:45:42 +0100) Normal Boot
WDT: Started with servicing (60s timeout)
Trying to boot from MMC1NOTICE: BL31: v2.2(release):rel_imx_5.4.70_2.3.0-0-gf1d7187f2 NOTICE: BL31: Built : 22:29:05, Jan 17 2021U-Boot 2021.01-01004-gb852007333 (Feb 01 2021 - 09:45:42 +0100)
Quoted 8 lines
CPU: Freescale i.MX8MMQ rev1.0 at 1200 MHz Reset cause: POR Model: FSL i.MX8MM EVK board DRAM: 2 GiB WDT: Started with servicing (60s timeout) MMC: FSL_SDHC: 1, FSL_SDHC: 2Loading Environment from MMC... Run CMD11 1.8V switch Card did not respond to voltage select! : -110
Quoted 2 lines
This do not align with my test result. Can you help identify which function first return the timeout on your side?
I would have a look at the exact function, but it seems to me that it would be
wait_dat0()since if I revert the patch introducing it - high speed mode switch is not timing out.Quoted 1 line
Or can you try a different SD card?
It is rather strange, it seems like it is dependent on the SD Card used.
So far, I've tried 3 SD Cards I had on hands, one of which being operable:
== Working:
"Transcend 32GB"
Manufacturer ID: 74
OEM: 4a60
Name: USDU1== Failed:
1. (Kingston 32GB)
Manufacturer ID: 41
OEM: 3432
Name: SD32GQuoted 5 lines
2. (Intenso 32 GB)
Manufacturer ID: 9f
OEM: 5449
Name: 00000Hi Andrey,
With this patch, can you also add the following change to test again, I double check the code logic, only the following code do not follow the SD spec for voltage switch.
Seems the two failed SD cards follow the sd voltage spec strictly.drivers/mmc/fsl_esdhc_imx.c index renamed from drivers/mmc/fsl_esdhc_imx.c +0
-0
da33ee8253..20337b0ed4100644drivers/mmc/fsl_esdhc_imx.c +1
-1@@ -513,7 +513,7 @@ static int esdhc_send_cmd_common(struct fsl_esdhc_priv *priv, struct mmc *mmc, err = -ETIMEDOUT ; goto out; } - +#if 0 /* Switch voltage to 1.8V if CMD11 succeeded */ if (cmd->cmdidx ==SD_CMD_SWITCH_UHS18V ) {esdhc_setbits32(®s->vendorspec, ESDHC_VENDORSPEC_VSELECT);
drivers/mmc/fsl_esdhc_imx.c static int esdhc_send_cmd_common(struct · line 527
Quoted 6 lines
drivers/mmc/fsl_esdhc_imx.c +1
-0@@ -522,6 +522,7 @@ static int esdhc_send_cmd_common(struct fsl_esdhc_priv *priv, struct mmc *mmc,/* Sleep for 5 ms - max time for card to switch to 1.8V */ udelay(5000); } +#endif
I've tried this change and can confirm: the failed Kingston 32GB card from the list above operates in high-speed mode.
u-boot=> mmc dev 1
switch to partitions #0, OK
mmc1 is current device
u-boot=> mmc info
Device: FSL_SDHC
Manufacturer ID: 41
OEM: 3432
Name: SD32G
Bus Speed: 200000000
Mode: UHS SDR104 (208MHz)
Rd Block Len: 512
SD version 3.0
High Capacity: Yes
Capacity: 29.3 GiB
Bus Width: 4-bit
Erase Group Size: 512 Bytes
Quoted 16 lines
Quoted 5 lines
This raises a concern that a regular off-the-shelf SD Card might exhibit this issue or might not - it is not tight to a one specific SD Card. Hence I believe it is quite important to look at it further, as there might be a lot of reports from various users.
Yes, I will take care of that, but I think current issue is not caused by this, because uboot already print out the log " Run CMD11 1.8V", which means the S18A is set in the response of ACMD41.
Currently, common code only delay 2ms for power cycle, but some boards may need more time, few i.mx boards need 3.8ms ~4.7ms. for my new imx8mm board, need 2.7ms to let VDD change to 0.5v from 3.3v.
And spec need to keep VDD under 0.5v for at least 1ms, so totally need about 3.7ms for the power cycle. According to the log, seems you use an old version board, not sure about the actual delay time on your board.
Quoted 2 lines
In Linux, regulator driver support the compatible "off-on-delay-us", let me double check whether uboot also support this feature.
This binding does exist in u-boot, and I've se it to 20 msec in commit 247bbeb74c ("ARM: dts: imx8m: increase off-on delay on the SD Vcc regulator").
Quoted 264 lines
Best Regards
HaiboQuoted 258 lines
Please note, that for all SD Cards I've tested
NVCC_SD2is indeed physically switched to 1v8.Quoted 175 lines
Best Regards
HaiboQuoted 104 lines
*** Warning - No block device, using default environment
In: serial Out: serial Err: serial Net: eth0: ethernet at 30be0000 Hit any key to stop autoboot: 0u-boot=> mmc dev 1
Card did not respond to voltage select! : -110 u-boot=> mmc info MMC Device
0 not found no mmc device at slot 0 u-boot=> mmc dev 2 switch to partitions #0, OK mmc2(part 0) is current device u-boot=> mmc info
Device:FSL_SDHC
Manufacturer ID: 45
OEM: 100
Name: DG401
Bus Speed: 200000000
Mode: HS400ES (200MHz)
Rd Block Len: 512
MMC version 5.1
High Capacity: Yes
Capacity: 14.7 GiB
Bus Width: 8-bit DDR
Erase Group Size: 512 KiB
HC WP Group Size: 8 MiB
User Capacity: 14.7 GiB WRREL
Boot Capacity: 4 MiB ENH
RPMB Capacity: 4 MiB ENH
Boot area 0 is not write protected Boot area 1 is not write protected
-----------------------------------------------Note, that eMMC is not affected and is operating in HS400ES mode without any issues.
Reverting patch
b5874b552ffa("mmc:fsl_esdhc_imx: addwait_dat0()support") with this patch applied in tree does switch the SD Card to high-speed mode, following log is produced:
-----------------------------------------------
U-Boot SPL 2021.01-01005-gb8aeb689a2 (Feb 01 2021 - 10:38:01 +0100) Normal Boot
WDT: Started with servicing (60s timeout)
Trying to boot from MMC1NOTICE: BL31: v2.2(release):rel_imx_5.4.70_2.3.0-0-gf1d7187f2 NOTICE: BL31: Built : 22:29:05, Jan 17 2021U-Boot 2021.01-01005-gb8aeb689a2 (Feb 01 2021 - 10:38:01 +0100)
CPU: Freescale i.MX8MMQ rev1.0 at 1200 MHz Reset cause: POR Model: FSL i.MX8MM EVK board DRAM: 2 GiB WDT: Started with servicing (60s timeout) MMC: FSL_SDHC: 1, FSL_SDHC: 2Loading Environment from MMC... Run CMD11 1.8V switch *** Warning - bad CRC, using default environment
In: serial Out: serial Err: serial Net: eth0: ethernet at 30be0000 Hit any key to stop autoboot: 0u-boot=> mmc dev 1
Run CMD11 1.8V switch
switch to partitions #0, OK
mmc1 is current device
u-boot=> mmc info
Device:FSL_SDHC
Manufacturer ID: 41
OEM: 3432
Name: SD32G
Bus Speed: 200000000
Mode: UHS SDR104 (208MHz)
Rd Block Len: 512
SD version 3.0
High Capacity: Yes
Capacity: 29.3 GiB
Bus Width: 4-bit
Erase Group Size: 512 Bytes
u-boot=> mmc dev 2
switch to partitions #0, OK
mmc2(part 0) is current device
u-boot=> mmc info
Device:FSL_SDHC
Manufacturer ID: 45
OEM: 100
Name: DG401
Bus Speed: 200000000
Mode: HS400ES (200MHz)
Rd Block Len: 512
MMC version 5.1
High Capacity: Yes
Capacity: 14.7 GiB
Bus Width: 8-bit DDR
Erase Group Size: 512 KiB
HC WP Group Size: 8 MiB
User Capacity: 14.7 GiB WRREL
Boot Capacity: 4 MiB ENH
RPMB Capacity: 4 MiB ENH
Boot area 0 is not write protected Boot area 1 is not write protected
-----------------------------------------------Quoted 4 lines
I guess high-level modifications are also required, since the uSDHC does not follow the SD Specification in regards to voltage switching, which is currently implemented in U-Boot, this patch alone does not solve the
mode switch.
Quoted 62 lines
Quoted 13 lines
Signed-off-by Haibo Chen <[email protected]>--- drivers/mmc/fsl_esdhc_imx.c | 29 +++++++++++++++++++++-------- include/fsl_esdhc_imx.h | 2 ++ 2 files changed, 23 insertions(+), 8 deletions(-)(quoted diff) +0
-0
b/drivers/mmc/fsl_esdhc_imx.cindex8ac859797f..da33ee8253100644drivers/mmc/fsl_esdhc_imx.c static void set_sysctl(struct · line 657
Quoted 6 lines
drivers/mmc/fsl_esdhc_imx.c +1
-1@@ -653,7 +653,10 @@ static void set_sysctl(struct fsl_esdhc_priv *priv, struct mmc *mmc, uint clock) clk = (pre_div <<8 ) | (div <<4 );#ifdef CONFIG_FSL_USDHC -esdhc_clrbits32 (®s->vendorspec,VENDORSPEC_CKEN ); + esdhc_clrbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);drivers/mmc/fsl_esdhc_imx.c static void set_sysctl(struct
Quoted 8 lines
drivers/mmc/fsl_esdhc_imx.c +5
-0+ ret =readx_poll_timeout (esdhc_read32, ®s->prsstat, + tmp, tmp &PRSSTAT_SDOFF ,100 ); +if (ret) + pr_warn("fsl_esdhc_imx: Internal clock never + gate off.\n");#else esdhc_clrbits32(®s->sysctl, SYSCTL_CKEN); #endif @@-665,7
Quoted 2 lines
+668,7 @@ static void set_sysctl(struct
fsl_esdhc_priv*priv, +struct mmc*mmc, uint clock)
Quoted 7 lines
pr_warn("fsl_esdhc_imx: Internal clock never
stabilised.\n");#ifdef
CONFIG_FSL_USDHC
- esdhc_setbits32(®s->vendorspec,VENDORSPEC_PEREN|VENDORSPEC_CKEN);
+ esdhc_setbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);drivers/mmc/fsl_esdhc_imx.c
Quoted 12 lines
#else esdhc_setbits32(®s->sysctl, SYSCTL_PEREN | SYSCTL_CKEN); #endif @@ - 720,8 +723,14 @@ static void esdhc_set_strobe_dll(struct mmc *mmc) struct fsl_esdhc_priv *priv = dev_get_priv(mmc->dev); struct fsl_esdhc *regs = priv->esdhc_regs; u32 val;drivers/mmc/fsl_esdhc_imx.c +3
-0+u32 tmp; +int ret;if (priv->clock >ESDHC_STROBE_DLL_CLK_FREQ ) { + esdhc_clrbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);Quoted 9 lines
+ ret = readx_poll_timeout(esdhc_read32,
+®s->prsstat, tmp, tmp &PRSSTAT_SDOFF, 100);drivers/mmc/fsl_esdhc_imx.c +3
-0+if (ret) + pr_warn("fsl_esdhc_imx: Internal clock + never gate off.\n"); esdhc_write32(®s->strobe_dllctrl,
ESDHC_STROBE_DLL_CTRL_RESET);Quoted 2 lines
Quoted 2 lines
drivers/mmc/fsl_esdhc_imx.c static void esdhc_set_strobe_dll(struct mmc
Quoted 2 lines
/*drivers/mmc/fsl_esdhc_imx.c +0
-0@@ -739,6 +748,7 @@ static void esdhc_set_strobe_dll(struct mmc*mmc)
Quoted 10 lines
Quoted 10 lines
drivers/mmc/fsl_esdhc_imx.c static void esdhc_set_strobe_dll(struct mmc · line 748
Quoted 1 line
drivers/mmc/fsl_esdhc_imx.c +0
-0pr_warn("HS400 strobe DLL status REF notlock!\n");
Quoted 8 lines
drivers/mmc/fsl_esdhc_imx.c +2
-0if (!(val &ESDHC_STROBE_DLL_STS_SLV_LOCK ))pr_warn ("HS400 strobe DLL status SLV not lock!\n"); +esdhc_setbits32 (®s->vendorspec, +VENDORSPEC_FRC_SDCLK_ON ); } }Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
drivers/mmc/fsl_esdhc_imx.c +0
-0@@ -962,14 +972,18 @@ static int esdhc_set_ios_common(struct
fsl_esdhc_priv*priv, struct mmc *mmc) #ifdef
MMC_SUPPORTS_TUNINGQuoted 8 lines
Quoted 8 lines
Quoted 3 lines
if (mmc->clk_disable) {#ifdef
CONFIG_FSL_USDHC
- esdhc_clrbits32(®s->vendorspec,
VENDORSPEC_CKEN);drivers/mmc/fsl_esdhc_imx.c
Quoted 3 lines
drivers/mmc/fsl_esdhc_imx.c +3
-0+u32 tmp; + + esdhc_clrbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);Quoted 8 lines
Quoted 8 lines
drivers/mmc/fsl_esdhc_imx.c
Quoted 8 lines
+ ret = readx_poll_timeout(esdhc_read32,
+®s->prsstat, tmp, tmp &PRSSTAT_SDOFF, 100);drivers/mmc/fsl_esdhc_imx.c +3
-0+if (ret) + pr_warn("fsl_esdhc_imx: Internal clock + never gate off.\n");#else esdhc_clrbits32 (®s->sysctl,SYSCTL_CKEN );#endif
Quoted 20 lines
Quoted 20 lines
Quoted 3 lines
} else {#ifdef
CONFIG_FSL_USDHC
- esdhc_setbits32(®s->vendorspec,
VENDORSPEC_PEREN|Quoted 8 lines
drivers/mmc/fsl_esdhc_imx.c +2
-1- VENDORSPEC_CKEN); +esdhc_setbits32 (®s->vendorspec, +VENDORSPEC_FRC_SDCLK_ON );#else esdhc_setbits32(®s->sysctl, SYSCTL_PEREN |
SYSCTL_CKEN); #endif @@ -1045,7 +1059,7 @@ static intesdhc_init_common(struct fsl_esdhc_priv *priv, struct mmc *mmc)#ifndef
CONFIG_FSL_USDHCQuoted 2 lines
esdhc_setbits32(®s->sysctl, SYSCTL_HCKEN |
SYSCTL_IPGEN);#else
Quoted 3 lines
- esdhc_setbits32(®s->vendorspec,
VENDORSPEC_HCKEN|VENDORSPEC_IPGEN);
+ esdhc_setbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);Quoted 8 lines
Quoted 8 lines
drivers/mmc/fsl_esdhc_imx.c
Quoted 8 lines
#endif
/* Set the initial clock speed */ @@ -1183,8 +1197,7 @@ static int fsl_esdhc_init(struct fsl_esdhc_priv *priv, esdhc_write32(®s->autoc12err, 0); esdhc_write32(®s->clktunectrlstatus, 0); #elsedrivers/mmc/fsl_esdhc_imx.c +0
-2- esdhc_setbits32(®s->vendorspec, VENDORSPEC_PEREN | - VENDORSPEC_HCKEN |
VENDORSPEC_IPGEN|Quoted 17 lines
Quoted 17 lines
VENDORSPEC_CKEN);Quoted 1 line
+ esdhc_setbits32(®s->vendorspec,
VENDORSPEC_FRC_SDCLK_ON);Quoted 14 lines
#endif
if (priv->vs18_enable)include/fsl_esdhc_imx.h +0
-0index
45ed635a77..b092034464100644include/fsl_esdhc_imx.h +1
-0@@ -39,6 +39,7 @@#define VENDORSPEC_HCKEN 0x00001000#define VENDORSPEC_IPGEN 0x00000800#define VENDORSPEC_INIT 0x20007809 +#define VENDORSPEC_FRC_SDCLK_ON 0x00000100
Quoted 12 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
include/fsl_esdhc_imx.h +0
-0#define IRQSTAT 0x0002e030#define IRQSTAT_DMAE (0x10000000)@@ -96,6 +97,7 @@#define PRSSTAT_CINS (0x00010000)Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
include/fsl_esdhc_imx.h +1
-1#define PRSSTAT_BREN (0x00000800)#define PRSSTAT_BWEN (0x00000400) +#define PRSSTAT_SDOFF (0x00000080)#define PRSSTAT_SDSTB (0X00000008)#define PRSSTAT_DLA (0x00000004)#define PRSSTAT_CICHB (0x00000002) --2.17.1
-- Andrey
PATCH 1/1 mmc: fsl_esdhc_imx: use VENDORSPEC_FRC_SDCLK_ON to control card clock output
In reply to Bough ChenQuoted 138 lines · Bough Chen
Quoted 113 lines
-----Original Message-----
From: ZHIZHIKIN Andrey [mailto:andrey.zhizhikin at leica-geosystems.com]
Sent: 2021?2?1? 19:41
To: Bough Chen <[email protected]>; Peng Fan <[email protected]>; u-boot at lists.denx.de
Cc: dl-uboot-imx <[email protected]>; tharvey at gateworks.com
Subject: RE: [PATCH] mmc:fsl_esdhc_imx: useVENDORSPEC_FRC_SDCLK_ONto control card clock outputHello Haibo,
Quoted 7 lines
-----Original Message-----
From: Bough Chen <[email protected]>
Sent: Monday, February 1, 2021 11:10 AM
To: ZHIZHIKIN Andrey <[email protected]>; Peng Fan <[email protected]>; u-boot at lists.denx.de
Cc: dl-uboot-imx <[email protected]>; tharvey at gateworks.com
Subject: RE: [PATCH] mmc:fsl_esdhc_imx: use
VENDORSPEC_FRC_SDCLK_ONQuoted 10 lines
to control card clock output
Quoted 8 lines
-----Original Message-----
From: ZHIZHIKIN Andrey
[mailto:andrey.zhizhikin at leica-geosystems.com]
Sent: 2021?2?1? 17:52
To: Bough Chen <[email protected]>; Peng Fan <[email protected]>; u-boot at lists.denx.de
Cc: dl-uboot-imx <[email protected]>; tharvey at gateworks.com
Subject: RE: [PATCH] mmc:fsl_esdhc_imx: use
VENDORSPEC_FRC_SDCLK_ONQuoted 9 lines
Quoted 9 lines
to control card clock output
Hello Haibo,
Quoted 5 lines
-----Original Message-----
From: haibo.chen at nxp.com <[email protected]>
Sent: Wednesday, January 27, 2021 11:47 AM
To: peng.fan at nxp.com; u-boot at lists.denx.de
Cc: haibo.chen at nxp.com; uboot-imx at nxp.com;tharvey at gateworks.com;
Quoted 48 lines
Quoted 39 lines
Quoted 7 lines
ZHIZHIKIN Andrey <[email protected]>
Subject: [PATCH] mmc:fsl_esdhc_imx: useVENDORSPEC_FRC_SDCLK_ONto control card clock outputFrom: Haibo Chen <[email protected]>
For
FSL_USDHC, it do not implementVENDORSPEC_CKEN/PEREN/HCKEN/IPGEN,
Quoted 11 lines
these are reserved bits. Instead, use
VENDORSPEC_FRC_SDCLK_ONto gate on/off the card clock output.After commit
b5874b552ffa("mmc:fsl_esdhc_imx: addwait_dat0()support"), we meet SD3.0 card can't work at UHS mode,mmc_switch_voltage()fail because the secondmmc_wait_dat0return-ETIMEDOUT. According to SD spec, during voltage switch, need to gate off/on the card clock. If not set theFRC_SDCLK_ON, after CMD11, hardware will gate off the card clock automatically, so card do not detect the clock off/on behavior, so will draw the data0 line low until nextcommand.
Quoted 3 lines
Fixes b5874b552ffa ("mmc: fsl_esdhc_imx: add wait_dat0()support")
This patch does not fix the switch of uSDHC to 1v8...
I've applied it locally on the imx8mmevk, and had following log during the boot when tried to query SD Card info:
-----------------------------------------------
U-Boot SPL 2021.01-01004-gb852007333 (Feb 01 2021 - 09:45:42 +0100) Normal Boot
WDT: Started with servicing (60s timeout)
Trying to boot from MMC1NOTICE: BL31: v2.2(release):rel_imx_5.4.70_2.3.0-0-gf1d7187f2 NOTICE: BL31: Built : 22:29:05, Jan 17 2021U-Boot 2021.01-01004-gb852007333 (Feb 01 2021 - 09:45:42 +0100)
Quoted 8 lines
CPU: Freescale i.MX8MMQ rev1.0 at 1200 MHz Reset cause: POR Model: FSL i.MX8MM EVK board DRAM: 2 GiB WDT: Started with servicing (60s timeout) MMC: FSL_SDHC: 1, FSL_SDHC: 2Loading Environment from MMC... Run CMD11 1.8V switch Card did not respond to voltage select! : -110
Quoted 2 lines
This do not align with my test result. Can you help identify which function first return the timeout on your side?
I would have a look at the exact function, but it seems to me that it would be
wait_dat0()since if I revert the patch introducing it - high speed mode switch is not timing out.Quoted 1 line
Or can you try a different SD card?
It is rather strange, it seems like it is dependent on the SD Card used.
So far, I've tried 3 SD Cards I had on hands, one of which being operable:
== Working:
"Transcend 32GB"
Manufacturer ID: 74
OEM: 4a60
Name: USDU1== Failed:
1. (Kingston 32GB)
Manufacturer ID: 41
OEM: 3432
Name: SD32GQuoted 5 lines
2. (Intenso 32 GB)
Manufacturer ID: 9f
OEM: 5449
Name: 00000Hi Andrey,
With this patch, can you also add the following change to test again, I double check the code logic, only the following code do not follow the SD spec for voltage switch. Seems the two failed SD cards follow the sd voltage spec strictly.
drivers/mmc/fsl_esdhc_imx.c +1
-1@@ -513,7 +513,7 @@ static int esdhc_send_cmd_common(struct fsl_esdhc_priv *priv, struct mmc *mmc, err = -ETIMEDOUT ; goto out; } - +#if 0 /* Switch voltage to 1.8V if CMD11 succeeded */ if (cmd->cmdidx ==SD_CMD_SWITCH_UHS18V ) {esdhc_setbits32 (®s->vendorspec,ESDHC_VENDORSPEC_VSELECT );
drivers/mmc/fsl_esdhc_imx.c static int esdhc_send_cmd_common(struct fsl_esdhc_priv *priv, struct mmc *mmc, · line 527
Quoted 6 lines
drivers/mmc/fsl_esdhc_imx.c +1
-0@@ -522,6 +522,7 @@ static int esdhc_send_cmd_common(struct fsl_esdhc_priv *priv, struct mmc *mmc,/* Sleep for 5 ms - max time for card to switch to 1.8V */ udelay(5000); } +#endif
Haibo,
I can confirm that this fixes UHS-I on the IMX8MM_EVK and I can now get SDR104 on the cards I have here. I would include this with your previous patch as without it your patch breaks MMC in my findings.
Thanks for keeping up on this!
Tim