Fix abnormal clock scaling behaviors
Subsystems (3)
v2 cover Fix abnormal clock scaling behaviors
From: Peter Wang <[email protected]>
This patch fix abnormal clock scaling behaviors.
Peter Wang (3):
ufs: core: only suspend clock scaling if scale down
ufs: core: fix abnormal scale up after last cmd finish
ufs: core: fix abnormal scale up after scale down
drivers/ufs/core/ufshcd.c | 26 ++++++++++++-------------- 1 file changed, 12 insertions(+), 14 deletions(-)
Signature · 1 line
2.18.0
PATCH v2 1/3 ufs: core: only suspend clock scaling if scale down
From: Peter Wang <[email protected]>
If clock scale up and suspend clock scaling, ufs will keep high performance/power mode but no read/write requests on going. It is logic wrong and have power concern.
--- drivers/ufs/core/ufshcd.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/ufs/core/ufshcd.c +1 -1
@@ -1458,7 +1458,7 @@ static int ufshcd_devfreq_target(struct device *dev, ktime_to_us(ktime_sub(ktime_get(), start)), ret); out: -if (sched_clk_scaling_suspend_work) +if (sched_clk_scaling_suspend_work && !scale_up)queue_work (hba->clk_scaling.workq, &hba->clk_scaling.suspend_work);
Signature · 1 line
2.18.0
PATCH v2 2/3 ufs: core: fix abnormal scale up after last cmd finish
In reply to [email protected]From: Peter Wang <[email protected]>
When ufshcd_clk_scaling_suspend_work(Thread A) running and new command coming, ufshcd_clk_scaling_start_busy(Thread B) may get host_lock after Thread A first time release host_lock. Then Thread A second time get host_lock will set clk_scaling.window_start_t = 0 which scale up clock abnormal next polling_ms time.
Also inlines another __ufshcd_suspend_clkscaling calls.
Below is racing step:
1 hba->clk_scaling.suspend_work (Thread A)
ufshcd_clk_scaling_suspend_work
2 spin_lock_irqsave(hba->host->host_lock, irq_flags);
3 hba->clk_scaling.is_suspended = true;
4 spin_unlock_irqrestore(hba->host->host_lock, irq_flags);
__ufshcd_suspend_clkscaling
7 spin_lock_irqsave(hba->host->host_lock, flags);
8 hba->clk_scaling.window_start_t = 0;
9 spin_unlock_irqrestore(hba->host->host_lock, flags);
ufshcd_send_command (Thread B)
ufshcd_clk_scaling_start_busy
5 spin_lock_irqsave(hba->host->host_lock, flags); ....
6 spin_unlock_irqrestore(hba->host->host_lock, flags);
--- drivers/ufs/core/ufshcd.c | 17 ++++------------- 1 file changed, 4 insertions(+), 13 deletions(-)
drivers/ufs/core/ufshcd.c +4 -13
@@ -275,7 +275,6 @@ static inline void ufshcd_add_delay_before_dme_cmd(struct ufs_hba *hba);static int ufshcd_host_reset_and_restore (struct ufs_hba *hba);static void ufshcd_resume_clkscaling (struct ufs_hba *hba);static void ufshcd_suspend_clkscaling (struct ufs_hba *hba); -static void __ufshcd_suspend_clkscaling (struct ufs_hba *hba);static int ufshcd_scale_clks (struct ufs_hba *hba,bool scale_up);static irqreturn_t ufshcd_intr (int irq,void *__hba); static int ufshcd_change_power_mode(struct ufs_hba *hba,
@@ -1385,9 +1384,10 @@ static void ufshcd_clk_scaling_suspend_work(struct work_struct *work)return ; } hba->clk_scaling.is_suspended = true; + hba->clk_scaling.window_start_t =0 ;spin_unlock_irqrestore (hba->host->host_lock, irq_flags); -__ufshcd_suspend_clkscaling (hba); +devfreq_suspend_device (hba->devfreq); }static void ufshcd_clk_scaling_resume_work (struct work_struct *work)
@@ -1564,16 +1564,6 @@ static void ufshcd_devfreq_remove(struct ufs_hba *hba)dev_pm_opp_remove (hba->dev, clki->max_freq); } -static void __ufshcd_suspend_clkscaling (struct ufs_hba *hba) -{ -unsigned long flags; - -devfreq_suspend_device (hba->devfreq); -spin_lock_irqsave (hba->host->host_lock, flags); - hba->clk_scaling.window_start_t =0 ; -spin_unlock_irqrestore (hba->host->host_lock, flags); -} -static void ufshcd_suspend_clkscaling (struct ufs_hba *hba) {unsigned long flags;
@@ -1586,11 +1576,12 @@ static void ufshcd_suspend_clkscaling(struct ufs_hba *hba)if (!hba->clk_scaling.is_suspended) { suspend = true; hba->clk_scaling.is_suspended = true; + hba->clk_scaling.window_start_t =0 ; }spin_unlock_irqrestore (hba->host->host_lock, flags);if (suspend) -__ufshcd_suspend_clkscaling (hba); +devfreq_suspend_device (hba->devfreq); }static void ufshcd_resume_clkscaling (struct ufs_hba *hba)
Signature · 1 line
2.18.0
Re: [PATCH v2 2/3] ufs: core: fix abnormal scale up after last cmd finish
Hi Bart,
This patch still need your review.
Thanks.
Peter
Quoted 95 lines · On Thu, 2023-08-31 at 21:08 +0800, [email protected] wrote:
From: Peter Wang <[email protected]>
When
ufshcd_clk_scaling_suspend_work(Thread A)running and new command
coming,ufshcd_clk_scaling_start_busy(Thread B)may gethost_lockafter Thread A first time releasehost_lock. Then Thread A second time
gethost_lockwill set clk_scaling.window_start_t = 0 which scale up clock abnormal nextpolling_mstime.
Also inlines another__ufshcd_suspend_clkscalingcalls.Below is racing step:
1hba->clk_scaling.suspend_work(Thread A)
ufshcd_clk_scaling_suspend_work
2spin_lock_irqsave(hba->host->host_lock, irq_flags);
3hba->clk_scaling.is_suspended= true;
4 spin_unlock_irqrestore(hba->host->host_lock,irq_flags);
__ufshcd_suspend_clkscaling
7spin_lock_irqsave(hba->host->host_lock, flags);
8hba->clk_scaling.window_start_t= 0;
9 spin_unlock_irqrestore(hba->host->host_lock,
flags);
ufshcd_send_command(Thread B)
ufshcd_clk_scaling_start_busy5 spin_lock_irqsave(hba->host->host_lock, flags); ....6 spin_unlock_irqrestore(hba->host->host_lock,
flags);Signed-off-by Peter Wang <[email protected]>--- drivers/ufs/core/ufshcd.c | 17 ++++------------- 1 file changed, 4 insertions(+), 13 deletions(-)drivers/ufs/core/ufshcd.c +0
-1@@ -275,7 +275,6 @@ static inline void ufshcd_add_delay_before_dme_cmd(struct ufs_hba *hba);static int ufshcd_host_reset_and_restore (struct ufs_hba *hba);static void ufshcd_resume_clkscaling (struct ufs_hba *hba);static void ufshcd_suspend_clkscaling (struct ufs_hba *hba); -static void __ufshcd_suspend_clkscaling (struct ufs_hba *hba);static int ufshcd_scale_clks (struct ufs_hba *hba,bool scale_up);static irqreturn_t ufshcd_intr (int irq,void *__hba);static int ufshcd_change_power_mode(struct
ufs_hba*hba,drivers/ufs/core/ufshcd.c +2
-1@@ -1385,9 +1384,10 @@ static void ufshcd_clk_scaling_suspend_work(struct work_struct *work) return; } hba->clk_scaling.is_suspended = true; + hba->clk_scaling.window_start_t =0 ;spin_unlock_irqrestore (hba->host->host_lock, irq_flags); -__ufshcd_suspend_clkscaling (hba); +devfreq_suspend_device (hba->devfreq); }static void
ufshcd_clk_scaling_resume_work(struct work_struct *work)drivers/ufs/core/ufshcd.c +0
-10@@ -1564,16 +1564,6 @@ static void ufshcd_devfreq_remove(struct ufs_hba *hba) dev_pm_opp_remove(hba->dev, clki->max_freq); } -static void __ufshcd_suspend_clkscaling (struct ufs_hba *hba) -{ -unsigned long flags; - -devfreq_suspend_device (hba->devfreq); -spin_lock_irqsave (hba->host->host_lock, flags); - hba->clk_scaling.window_start_t =0 ; -spin_unlock_irqrestore (hba->host->host_lock, flags); -} -static void ufshcd_suspend_clkscaling (struct ufs_hba *hba) {drivers/ufs/core/ufshcd.c +2
-1unsigned long flags;@@ -1586,11 +1576,12 @@ static void ufshcd_suspend_clkscaling(struct ufs_hba *hba)if (!hba->clk_scaling.is_suspended) { suspend = true; hba->clk_scaling.is_suspended = true; + hba->clk_scaling.window_start_t =0 ; }spin_unlock_irqrestore (hba->host->host_lock, flags);if (suspend) -__ufshcd_suspend_clkscaling (hba); +devfreq_suspend_device (hba->devfreq); }static void
ufshcd_resume_clkscaling(struct ufs_hba *hba)
Bart Van Assche
reviewerRe: [PATCH v2 2/3] ufs: core: fix abnormal scale up after last cmd finish
Quoted 1 line · Peter Wang (王信友)
This patch still need your review.
Hi Peter,
Although I can see that patch 2/3 made it to lore.kernel.org, I did not receive that patch in my mailbox. This probably was caused by an issue by the mailserver I'm using. I will download the emails I'm missing from lore.kernel.org and review these patches.
Thanks,
Bart.
Re: [PATCH v2 2/3] ufs: core: fix abnormal scale up after last cmd finish
In reply to [email protected]Quoted 6 lines · On 8/31/23 06:08, [email protected] wrote:
When
ufshcd_clk_scaling_suspend_work(Thread A)running and new command coming,ufshcd_clk_scaling_start_busy(Thread B)may gethost_lockafter Thread A first time releasehost_lock. Then Thread A second time gethost_lockwill set clk_scaling.window_start_t = 0 which scale up clock abnormal nextpolling_mstime.
Also inlines another__ufshcd_suspend_clkscalingcalls.
Re: [PATCH v2 2/3] ufs: core: fix abnormal scale up after last cmd finish
Quoted 10 lines · On 8/31/23 06:08, [email protected] wrote:
Quoted 6 lines
When
ufshcd_clk_scaling_suspend_work(Thread A)running and new command coming,ufshcd_clk_scaling_start_busy(Thread B)may gethost_lockafter Thread A first time releasehost_lock. Then Thread A second time gethost_lockwill set clk_scaling.window_start_t = 0 which scale up clock abnormal nextpolling_mstime.
Also inlines another__ufshcd_suspend_clkscalingcalls.Reviewed-by Bart Van Assche <[email protected]>
Applied 1-3 to 6.7/scsi-staging, thanks!
Signature · 1 line
Martin K. Petersen Oracle Linux Engineering
PATCH v2 3/3 ufs: core: fix abnormal scale up after scale down
In reply to [email protected]From: Peter Wang <[email protected]>
When no active_reqs, devfreq_monitor(Thread A) will suspend clock scaling. But it may have racing with clk_scaling.suspend_work(Thread B) and actually not suspend clock scaling(requue after suspend).
Next time after polling_ms, devfreq_monitor read clk_scaling.window_start_t = 0 then scale up clock abnormal.
Below is racing step:
devfreq->work (Thread A) devfreq_monitor update_devfreq ..... ufshcd_devfreq_target queue_work(hba->clk_scaling.workq, 1 &hba->clk_scaling.suspend_work) ..... 5 queue_delayed_work(devfreq_wq, &devfreq->work, msecs_to_jiffies(devfreq->profile->polling_ms));
2 hba->clk_scaling.suspend_work (Thread B)
ufshcd_clk_scaling_suspend_work __ufshcd_suspend_clkscaling devfreq_suspend_device(hba->devfreq); 3 cancel_delayed_work_sync(&devfreq->work); 4 hba->clk_scaling.window_start_t = 0; .....
--- drivers/ufs/core/ufshcd.c | 7 +++++++ 1 file changed, 7 insertions(+)
drivers/ufs/core/ufshcd.c +7 -0
@@ -1430,6 +1430,13 @@ static int ufshcd_devfreq_target(struct device *dev,return 0 ; } +/* Skip scaling clock when clock scaling is suspended */ +if (hba->clk_scaling.is_suspended) { +spin_unlock_irqrestore (hba->host->host_lock, irq_flags); +dev_warn (hba->dev,"clock scaling is suspended, skip" ); +return 0 ; + } +if (!hba->clk_scaling.active_reqs) sched_clk_scaling_suspend_work = true;
Signature · 1 line
2.18.0
Martin K. Petersen
AppliedQuoted 8 lines · On Thu, 31 Aug 2023 21:08:23 +0800, [email protected] wrote:
This patch fix abnormal clock scaling behaviors.
Peter Wang (3):
ufs: core: only suspend clock scaling if scale down
ufs: core: fix abnormal scale up after last cmd finish
ufs: core: fix abnormal scale up after scale down[...]
Applied to 6.7/scsi-queue, thanks!
[1/3] ufs: core: only suspend clock scaling if scale down
https://git.kernel.org/mkp/scsi/c/1d969731b87f [2/3] ufs: core: fix abnormal scale up after last cmd finish https://git.kernel.org/mkp/scsi/c/6fd53da45bbc [3/3] ufs: core: fix abnormal scale up after scale down https://git.kernel.org/mkp/scsi/c/b50d9c27a31e
Signature · 1 line
Martin K. Petersen Oracle Linux Engineering