MediaTek UFS vendor implemenation part III and Auto-Hibern8 fix
Subsystems (3)
v3 cover MediaTek UFS vendor implemenation part III and Auto-Hibern8 fix
Hi,
This series provides MediaTek vendor implementations and some general fixes.
- General fixes
- Fix Auto-Hibern8 error detection
- MediaTek vendor implementations
- Ensure UniPro is powered on before every link startup
- Support linkoff state during suspend
- Gate reference clock for Auto-Hibern8 case
v3 (Resend)
- Fix "Fixes" tag in patch "scsi: ufs: fix Auto-Hibern8 error detection" (Greg KH)
v2 -> v3
- Squash below patches to a single patch (Bean Huo)
- scsi: ufs: add ufshcd_is_auto_hibern8_enabled facility
- scsi: ufs: fix auto-hibern8 error detection
- Add Fixes tag in patch "scsi: ufs: fix Auto-Hibern8 error detection" (Bean Huo)
- Rename VS_LINK_HIBER8 to VS_LINK_HIBERN8 in patch "scsi: ufs-mediatek: gate ref-clk during Auto-Hibern8"
v1 -> v2
- Fix and refine commit messages.
Stanley Chu (4):
scsi: ufs-mediatek: ensure UniPro is not powered down before linkup
scsi: ufs-mediatek: support linkoff state during suspend
scsi: ufs: fix Auto-Hibern8 error detection
scsi: ufs-mediatek: gate ref-clk during Auto-Hibern8
drivers/scsi/ufs/ufs-mediatek.c | 67 +++++++++++++++++++++------------ drivers/scsi/ufs/ufs-mediatek.h | 12 ++++++ drivers/scsi/ufs/ufshcd.c | 3 +- drivers/scsi/ufs/ufshcd.h | 6 +++ 4 files changed, 63 insertions(+), 25 deletions(-)
Signature · 5 lines
2.18.0 _______________________________________________ linux-arm-kernel mailing list [email protected] http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
PATCH RESEND v3 1/4 scsi: ufs-mediatek: ensure UniPro is not powered down before linkup
MediaTek Chipsets can enter proprietary UniPro low-power mode during suspend while link is in hibern8 state. Make sure leaving low-power mode before every link startup to prevent lockup in any possible error recovery path.
In the same time, re-factor related funcitons to improve code readability.
--- drivers/scsi/ufs/ufs-mediatek.c | 19 ++++++++++--------- 1 file changed, 10 insertions(+), 9 deletions(-)
drivers/scsi/ufs/ufs-mediatek.c +10 -9
@@ -30,6 +30,11 @@#define ufs_mtk_device_reset_ctrl (high, res) \ ufs_mtk_smc(UFS_MTK_SIP_DEVICE_RESET, high, res) +#define ufs_mtk_unipro_powerdown (hba, powerdown) \ + ufshcd_dme_set(hba, \ + UIC_ARG_MIB_SEL(VS_UNIPROPOWERDOWNCONTROL, 0), \ + powerdown) +static void ufs_mtk_cfg_unipro_cg (struct ufs_hba *hba,bool enable) {u32 tmp;
@@ -290,6 +295,8 @@ static int ufs_mtk_pre_link(struct ufs_hba *hba)int ret;u32 tmp; +ufs_mtk_unipro_powerdown (hba,0 ); +/* disable deep stall */ ret =ufshcd_dme_get (hba,UIC_ARG_MIB (VS_SAVEPOWERCONTROL ), &tmp);if (ret)
@@ -390,9 +397,7 @@ static int ufs_mtk_link_set_hpm(struct ufs_hba *hba)if (err)return err; - err =ufshcd_dme_set (hba, -UIC_ARG_MIB_SEL (VS_UNIPROPOWERDOWNCONTROL ,0 ), -0 ); + err =ufs_mtk_unipro_powerdown (hba,0 );if (err)return err;
@@ -413,14 +418,10 @@ static int ufs_mtk_link_set_lpm(struct ufs_hba *hba) {int err; - err =ufshcd_dme_set (hba, -UIC_ARG_MIB_SEL (VS_UNIPROPOWERDOWNCONTROL ,0 ), -1 ); + err =ufs_mtk_unipro_powerdown (hba,1 );if (err) {/* Resume UniPro state for following error recovery */ -ufshcd_dme_set (hba, -UIC_ARG_MIB_SEL (VS_UNIPROPOWERDOWNCONTROL ,0 ), -0 ); +ufs_mtk_unipro_powerdown (hba,0 );return err; }
Signature · 1 line
2.18.0
RE: [EXT] [PATCH RESEND v3 1/4] scsi: ufs-mediatek: ensure UniPro is not powered down before linkup
Quoted 2 lines
Signed-off-by Stanley Chu <[email protected]>
PATCH RESEND v3 2/4 scsi: ufs-mediatek: support linkoff state during suspend
In reply to Stanley ChuIf system suspend or runtime suspend mode is configured as linkoff state, phy can be powered off and reference clock can be gated in MediaTek Chipsets.
In the same time, remove redundant reference clock control in suspend and resume callbacks because such control can be well-handled in setup_clocks callback..
--- drivers/scsi/ufs/ufs-mediatek.c | 12 +++++++----- 1 file changed, 7 insertions(+), 5 deletions(-)
drivers/scsi/ufs/ufs-mediatek.c +7 -5
@@ -167,7 +167,7 @@ static int ufs_mtk_setup_clocks(struct ufs_hba *hba, bool on,switch (status) {case PRE_CHANGE : -if (!on) { +if (!on && !ufshcd_is_link_active (hba)) {ufs_mtk_setup_ref_clk (hba, on); ret =phy_power_off (host->mphy); }
@@ -437,10 +437,11 @@ static int ufs_mtk_suspend(struct ufs_hba *hba, enum ufs_pm_op pm_op) err =ufs_mtk_link_set_lpm (hba);if (err)return -EAGAIN ; -phy_power_off (host->mphy); -ufs_mtk_setup_ref_clk (hba, false); } +if (!ufshcd_is_link_active (hba)) +phy_power_off (host->mphy); +return 0 ; }
@@ -449,9 +450,10 @@ static int ufs_mtk_resume(struct ufs_hba *hba, enum ufs_pm_op pm_op)struct ufs_mtk_host *host =ufshcd_get_variant (hba);int err; -if (ufshcd_is_link_hibern8 (hba)) { -ufs_mtk_setup_ref_clk (hba, true); +if (!ufshcd_is_link_active (hba))phy_power_on (host->mphy); + +if (ufshcd_is_link_hibern8 (hba)) { err =ufs_mtk_link_set_hpm (hba);if (err)return err;
Signature · 1 line
2.18.0
RE: [EXT] [PATCH RESEND v3 2/4] scsi: ufs-mediatek: support linkoff state during suspend
Quoted 1 line
Signed-off-by Stanley Chu <[email protected]>
PATCH RESEND v3 3/4 scsi: ufs: fix Auto-Hibern8 error detection
In reply to Stanley ChuAuto-Hibern8 may be disabled by some vendors or sysfs in runtime even if Auto-Hibern8 capability is supported by host. If Auto-Hibern8 capability is supported by host but not actually enabled, Auto-Hibern8 error shall not happen.
To fix this, provide a way to detect if Auto-Hibern8 is actually enabled first, and bypass Auto-Hibern8 disabling case in ufshcd_is_auto_hibern8_error().
--- drivers/scsi/ufs/ufshcd.c | 3 ++- drivers/scsi/ufs/ufshcd.h | 6 ++++++ 2 files changed, 8 insertions(+), 1 deletion(-)
drivers/scsi/ufs/ufshcd.c +2 -1
@@ -5479,7 +5479,8 @@ static irqreturn_t ufshcd_update_uic_error(struct ufs_hba *hba)static bool ufshcd_is_auto_hibern8_error (struct ufs_hba *hba,u32 intr_mask) { -if (!ufshcd_is_auto_hibern8_supported (hba)) +if (!ufshcd_is_auto_hibern8_supported (hba) || + !ufshcd_is_auto_hibern8_enabled (hba))return false;if (!(intr_mask &UFSHCD_UIC_HIBERN8_MASK ))
drivers/scsi/ufs/ufshcd.h +6 -0
@@ -55,6 +55,7 @@#include <linux/clk.h> #include <linux/completion.h> #include <linux/regulator/consumer.h> +#include <linux/bitfield.h> #include "unipro.h" #include <asm/irq.h>
@@ -773,6 +774,11 @@ static inline bool ufshcd_is_auto_hibern8_supported(struct ufs_hba *hba)return (hba->capabilities &MASK_AUTO_HIBERN8_SUPPORT ); } +static inline bool ufshcd_is_auto_hibern8_enabled (struct ufs_hba *hba) +{ +return FIELD_GET (UFSHCI_AHIBERN8_TIMER_MASK , hba->ahit) ? true : false; +} +#define ufshcd_writel (hba, val, reg) \ writel((val), (hba)->mmio_base + (reg))#define ufshcd_readl (hba, reg) \
Signature · 1 line
2.18.0
Re: [PATCH RESEND v3 3/4] scsi: ufs: fix Auto-Hibern8 error detection
Quoted 9 lines · Stanley Chu
Auto-Hibern8 may be disabled by some vendors or sysfs in runtime even if Auto-Hibern8 capability is supported by host. If Auto-Hibern8 capability is supported by host but not actually enabled, Auto-Hibern8 error shall not happen.
To fix this, provide a way to detect if Auto-Hibern8 is actually enabled first, and bypass Auto-Hibern8 disabling case in
ufshcd_is_auto_hibern8_error().
Quoted 5 lines
Fixes 821744403913 ("scsi: ufs: Add error-handling of Auto-Hibernate")Signed-off-by Stanley Chu <[email protected]>Reviewed-by Bean Huo <[email protected]>---
Quoted 43 lines
drivers/scsi/ufs/ufshcd.c | 3 ++- drivers/scsi/ufs/ufshcd.h | 6 ++++++ 2 files changed, 8 insertions(+), 1 deletion(-)drivers/scsi/ufs/ufshcd.c +2
-1@@ -5479,7 +5479,8 @@ static irqreturn_t ufshcd_update_uic_error(struct ufs_hba *hba)static bool ufshcd_is_auto_hibern8_error (struct ufs_hba *hba,u32 intr_mask) { -if (!ufshcd_is_auto_hibern8_supported (hba)) +if (!ufshcd_is_auto_hibern8_supported (hba) || + !ufshcd_is_auto_hibern8_enabled (hba))return false;if (!(intr_mask &UFSHCD_UIC_HIBERN8_MASK ))drivers/scsi/ufs/ufshcd.h +6
-0@@ -55,6 +55,7 @@#include <linux/clk.h> #include <linux/completion.h> #include <linux/regulator/consumer.h> +#include <linux/bitfield.h> #include "unipro.h" #include <asm/irq.h> @@ -773,6 +774,11 @@ static inline bool ufshcd_is_auto_hibern8_supported(struct ufs_hba *hba)return (hba->capabilities &MASK_AUTO_HIBERN8_SUPPORT ); } +static inline bool ufshcd_is_auto_hibern8_enabled (struct ufs_hba *hba) +{ +return FIELD_GET (UFSHCI_AHIBERN8_TIMER_MASK , hba->ahit) ? true : false; +} +#define ufshcd_writel (hba, val, reg) \ writel((val), (hba)->mmio_base + (reg))#define ufshcd_readl (hba, reg) \
Signature · 2 lines
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, Linux Foundation Collaborative Project
Re: [PATCH RESEND v3 3/4] scsi: ufs: fix Auto-Hibern8 error detection
In reply to Stanley ChuQuoted 9 lines · Stanley Chu
Auto-Hibern8 may be disabled by some vendors or sysfs in runtime even if Auto-Hibern8 capability is supported by host. If Auto-Hibern8 capability is supported by host but not actually enabled, Auto-Hibern8 error shall not happen.
To fix this, provide a way to detect if Auto-Hibern8 is actually enabled first, and bypass Auto-Hibern8 disabling case in
ufshcd_is_auto_hibern8_error().
Quoted 5 lines
Fixes 821744403913 ("scsi: ufs: Add error-handling of Auto-Hibernate")Signed-off-by Stanley Chu <[email protected]>Reviewed-by Bean Huo <[email protected]>---
Quoted 45 lines
drivers/scsi/ufs/ufshcd.c | 3 ++- drivers/scsi/ufs/ufshcd.h | 6 ++++++ 2 files changed, 8 insertions(+), 1 deletion(-)drivers/scsi/ufs/ufshcd.c +2
-1@@ -5479,7 +5479,8 @@ static irqreturn_t ufshcd_update_uic_error(struct ufs_hba *hba) static bool ufshcd_is_auto_hibern8_error(struct ufs_hba *hba,u32 intr_mask) { -if (!ufshcd_is_auto_hibern8_supported (hba)) +if (!ufshcd_is_auto_hibern8_supported (hba) || + !ufshcd_is_auto_hibern8_enabled (hba))return false;drivers/scsi/ufs/ufshcd.c +0
-0if (!(intr_mask &UFSHCD_UIC_HIBERN8_MASK ))drivers/scsi/ufs/ufshcd.h +6
-0@@ -55,6 +55,7 @@#include <linux/clk.h> #include <linux/completion.h> #include <linux/regulator/consumer.h> +#include <linux/bitfield.h> #include "unipro.h" #include <asm/irq.h> @@ -773,6 +774,11 @@ static inline bool ufshcd_is_auto_hibern8_supported(struct ufs_hba *hba) return (hba->capabilities & MASK_AUTO_HIBERN8_SUPPORT); } +static inline bool ufshcd_is_auto_hibern8_enabled (struct ufs_hba *hba) +{ +return FIELD_GET (UFSHCI_AHIBERN8_TIMER_MASK , hba->ahit) ? true : false; +} +#define ufshcd_writel (hba, val, reg) \
writel((val), (hba)->mmio_base + (reg))
#defineufshcd_readl(hba, reg)\
PATCH RESEND v3 4/4 scsi: ufs-mediatek: gate ref-clk during Auto-Hibern8
In reply to Stanley ChuIn current UFS driver design, hba->uic_link_state will not be changed after link enters Hibern8 state by Auto-Hibern8 mechanism. In this case, reference clock gating will be skipped unless special handling is implemented in vendor's callbacks.
Support reference clock gating during Auto-Hibern8 period in MediaTek Chipsets: If link state is already in Hibern8 while Auto-Hibern8 feature is enabled, gate reference clock in setup_clocks callback.
--- drivers/scsi/ufs/ufs-mediatek.c | 38 +++++++++++++++++++++++---------- drivers/scsi/ufs/ufs-mediatek.h | 12 +++++++++++ 2 files changed, 39 insertions(+), 11 deletions(-)
drivers/scsi/ufs/ufs-mediatek.c +27 -11
@@ -143,6 +143,17 @@ static int ufs_mtk_setup_ref_clk(struct ufs_hba *hba, bool on)return 0 ; } +static u32 ufs_mtk_link_get_state (struct ufs_hba *hba) +{ +u32 val; + +ufshcd_writel (hba,0x20 ,REG_UFS_DEBUG_SEL ); + val =ufshcd_readl (hba,REG_UFS_PROBE ); + val = val >>28 ; + +return val; +} +/** * ufs_mtk_setup_clocks - enables/disable clocks * @hba: host controller instance
@@ -155,7 +166,7 @@ static int ufs_mtk_setup_clocks(struct ufs_hba *hba, bool on, enum ufs_notify_change_status status) {struct ufs_mtk_host *host =ufshcd_get_variant (hba); -int ret = -EINVAL ; +int ret =0 ;/* * In case ufs_mtk_init() is not yet done, simply ignore.
@@ -165,19 +176,24 @@ static int ufs_mtk_setup_clocks(struct ufs_hba *hba, bool on,if (!host)return 0 ; -switch (status) { -case PRE_CHANGE : -if (!on && !ufshcd_is_link_active (hba)) { +if (!on && status ==PRE_CHANGE ) { +if (!ufshcd_is_link_active (hba)) {ufs_mtk_setup_ref_clk (hba, on); ret =phy_power_off (host->mphy); + }else { +/* +* Gate ref-clk if link state is in Hibern8 +* triggered by Auto-Hibern8. +*/ +if (!ufshcd_can_hibern8_during_gating (hba) && +ufshcd_is_auto_hibern8_enabled (hba) && +ufs_mtk_link_get_state (hba) == +VS_LINK_HIBERN8 ) +ufs_mtk_setup_ref_clk (hba, on); } -break ; -case POST_CHANGE : -if (on) { - ret =phy_power_on (host->mphy); -ufs_mtk_setup_ref_clk (hba, on); - } -break ; + }else if (on && status ==POST_CHANGE ) { + ret =phy_power_on (host->mphy); +ufs_mtk_setup_ref_clk (hba, on); }return ret;
drivers/scsi/ufs/ufs-mediatek.h +12 -0
@@ -53,6 +53,18 @@#define VS_SAVEPOWERCONTROL 0xD0A6#define VS_UNIPROPOWERDOWNCONTROL 0xD0A8 +/* +* Vendor specific link state +*/ +enum { +VS_LINK_DISABLED =0 , +VS_LINK_DOWN =1 , +VS_LINK_UP =2 , +VS_LINK_HIBERN8 =3 , +VS_LINK_LOST =4 , +VS_LINK_CFG =5 , +}; +/* * SiP commands */
Signature · 1 line
2.18.0
RE: [EXT] [PATCH RESEND v3 4/4] scsi: ufs-mediatek: gate ref-clk during Auto-Hibern8
Quoted 1 line
Signed-off-by Stanley Chu <[email protected]>
RE: [PATCH RESEND v3 4/4] scsi: ufs-mediatek: gate ref-clk during Auto-Hibern8
In reply to Stanley Chudrivers/scsi/ufs/ufs-mediatek.c static int ufs_mtk_setup_ref_clk(struct ufs_hba *hba, bool on) · line 155
Quoted 11 lines
drivers/scsi/ufs/ufs-mediatek.c +10
-0+static u32 ufs_mtk_link_get_state (struct ufs_hba *hba) +{ +u32 val; + +ufshcd_writel (hba,0x20 ,REG_UFS_DEBUG_SEL ); + val =ufshcd_readl (hba,REG_UFS_PROBE ); + val = val >>28 ; + +return val; +}
A little bit strange that you are relying on debug registers to setup your ref-clock. Is this this debug info is always available?
Thanks,
Avri
RE: [PATCH RESEND v3 4/4] scsi: ufs-mediatek: gate ref-clk during Auto-Hibern8
Hi Avri,
Quoted 14 lines · Avri Altman
drivers/scsi/ufs/ufs-mediatek.c static int ufs_mtk_setup_ref_clk(struct ufs_hba *hba, bool on) · line 155
Quoted 11 lines
drivers/scsi/ufs/ufs-mediatek.c +10
-0+static u32 ufs_mtk_link_get_state (struct ufs_hba *hba) +{ +u32 val; + +ufshcd_writel (hba,0x20 ,REG_UFS_DEBUG_SEL ); + val =ufshcd_readl (hba,REG_UFS_PROBE ); + val = val >>28 ; + +return val; +}A little bit strange that you are relying on debug registers to setup your ref-clock. Is this this debug info is always available?
Yes, this register is only for this purpose now (query link state) and always existed in MediaTek UFS host.
Thanks,
Stanley
hello Stanley
Quoted 23 lines · Stanley Chu
Hi,
This series provides MediaTek vendor implementations and some general fixes.
- General fixes
- Fix Auto-Hibern8 error detection- MediaTek vendor implementations
- Ensure UniPro is powered on before every link startup
- Support linkoff state during suspend
- Gate reference clock for Auto-Hibern8 casev3 (Resend)
- Fix "Fixes" tag in patch "scsi: ufs: fix Auto-Hibern8 error detection" (Greg KH)v2 -> v3
- Squash below patches to a single patch (Bean Huo)
- scsi: ufs: addufshcd_is_auto_hibern8_enabledfacility
- scsi: ufs: fix auto-hibern8 error detection
- Add Fixes tag in patch "scsi: ufs: fix Auto-Hibern8 error detection" (Bean Huo)
- RenameVS_LINK_HIBER8toVS_LINK_HIBERN8in patch "scsi: ufs-mediatek: gate ref-clk during Auto-Hibern8"
Quoted 9 lines
v1 -> v2
- Fix and refine commit messages.Stanley Chu (4):
scsi: ufs-mediatek: ensure UniPro is not powered down before linkup
scsi: ufs-mediatek: support linkoff state during suspend
scsi: ufs: fix Auto-Hibern8 error detection
scsi: ufs-mediatek: gate ref-clk during Auto-Hibern8
For this series, feel free to add
Thanks
Quoted 8 lines
drivers/scsi/ufs/ufs-mediatek.c | 67 +++++++++++++++++++++------------ drivers/scsi/ufs/ufs-mediatek.h | 12 ++++++ drivers/scsi/ufs/ufshcd.c | 3 +- drivers/scsi/ufs/ufshcd.h | 6 +++ 4 files changed, 63 insertions(+), 25 deletions(-)--
2.18.0
Signature · 2 lines
Regards, Alim
Martin K. Petersen
AppliedStanley,
Quoted 2 lines
v3 (Resend) - Fix "Fixes" tag in patch "scsi: ufs: fix Auto-Hibern8 error detection" (Greg KH)
Next time, if you make changes, please bump the version number. Doesn't matter how small the change is.
Applied to 5.7/scsi-queue. Thanks!
Signature · 6 lines
Martin K. Petersen Oracle Linux Engineering _______________________________________________ linux-arm-kernel mailing list [email protected] http://lists.infradead.org/mailman/listinfo/linux-arm-kernel