[PATCH] scsi: ufs: core: Ensure clk_gating.lock is used only after initialization

[PATCH] scsi: ufs: core: Ensure clk_gating.lock is used only after initialization

[Avri Altman]

This commit addresses a lockdep warning triggered by the use of the
clk_gating.lock before it is properly initialized. The warning is as
follows:

[    4.388838] INFO: trying to register non-static key.
[    4.395673] The code is fine but needs lockdep annotation, or maybe
[    4.402118] you didn't initialize this object before use?
[    4.407673] turning off the locking correctness validator.
[    4.413334] CPU: 5 UID: 0 PID: 58 Comm: kworker/u32:1 Not tainted 6.12-rc1 #185
[    4.413343] Hardware name: Qualcomm Technologies, Inc. Robotics RB5 (DT)
[    4.413362] Call trace:
[    4.413364]  show_stack+0x18/0x24 (C)
[    4.413374]  dump_stack_lvl+0x90/0xd0
[    4.413384]  dump_stack+0x18/0x24
[    4.413392]  register_lock_class+0x498/0x4a8
[    4.413400]  __lock_acquire+0xb4/0x1b90
[    4.413406]  lock_acquire+0x114/0x310
[    4.413413]  _raw_spin_lock_irqsave+0x60/0x88
[    4.413423]  ufshcd_setup_clocks+0x2c0/0x490
[    4.413433]  ufshcd_init+0x198/0x10ec
[    4.413437]  ufshcd_pltfrm_init+0x600/0x7c0
[    4.413444]  ufs_qcom_probe+0x20/0x58
[    4.413449]  platform_probe+0x68/0xd8
[    4.413459]  really_probe+0xbc/0x268
[    4.413466]  __driver_probe_device+0x78/0x12c
[    4.413473]  driver_probe_device+0x40/0x11c
[    4.413481]  __device_attach_driver+0xb8/0xf8
[    4.413489]  bus_for_each_drv+0x84/0xe4
[    4.413495]  __device_attach+0xfc/0x18c
[    4.413502]  device_initial_probe+0x14/0x20
[    4.413510]  bus_probe_device+0xb0/0xb4
[    4.413517]  deferred_probe_work_func+0x8c/0xc8
[    4.413524]  process_scheduled_works+0x250/0x658
[    4.413534]  worker_thread+0x15c/0x2c8
[    4.413542]  kthread+0x134/0x200
[    4.413550]  ret_from_fork+0x10/0x20

To fix this issue, we use the existing `is_initialized` flag in the
`clk_gating` structure to ensure that the spinlock is only used after it
has been properly initialized. We check this flag before using the
spinlock in the `ufshcd_setup_clocks` function.

It was incorrect in the first place to call `setup_clocks()` before
`ufshcd_init_clk_gating()`, and the introduction of the new lock
unmasked this bug.

Fixes: 209f4e43b806 ("scsi: ufs: core: Introduce a new clock_gating lock")
Reported-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Signed-off-by: Avri Altman <avri.altman@wdc.com>
---
 drivers/ufs/core/ufshcd.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
index f6c38cf10382..a778fc51ca2a 100644
--- a/drivers/ufs/core/ufshcd.c
+++ b/drivers/ufs/core/ufshcd.c
@@ -9142,7 +9142,7 @@ static int ufshcd_setup_clocks(struct ufs_hba *hba, bool on)
 			if (!IS_ERR_OR_NULL(clki->clk) && clki->enabled)
 				clk_disable_unprepare(clki->clk);
 		}
-	} else if (!ret && on) {
+	} else if (!ret && on && hba->clk_gating.is_initialized) {
 		scoped_guard(spinlock_irqsave, &hba->clk_gating.lock)
 			hba->clk_gating.state = CLKS_ON;
 		trace_ufshcd_clk_gating(dev_name(hba->dev),
-- 
2.25.1

Re: [PATCH] scsi: ufs: core: Ensure clk_gating.lock is used only after initialization

[Manivannan Sadhasivam]

On Wed, Jan 22, 2025 at 08:27:18AM +0200, Avri Altman wrote:
> This commit addresses a lockdep warning triggered by the use of the
> clk_gating.lock before it is properly initialized. The warning is as
> follows:
> 
> [    4.388838] INFO: trying to register non-static key.
> [    4.395673] The code is fine but needs lockdep annotation, or maybe
> [    4.402118] you didn't initialize this object before use?
> [    4.407673] turning off the locking correctness validator.
> [    4.413334] CPU: 5 UID: 0 PID: 58 Comm: kworker/u32:1 Not tainted 6.12-rc1 #185
> [    4.413343] Hardware name: Qualcomm Technologies, Inc. Robotics RB5 (DT)
> [    4.413362] Call trace:
> [    4.413364]  show_stack+0x18/0x24 (C)
> [    4.413374]  dump_stack_lvl+0x90/0xd0
> [    4.413384]  dump_stack+0x18/0x24
> [    4.413392]  register_lock_class+0x498/0x4a8
> [    4.413400]  __lock_acquire+0xb4/0x1b90
> [    4.413406]  lock_acquire+0x114/0x310
> [    4.413413]  _raw_spin_lock_irqsave+0x60/0x88
> [    4.413423]  ufshcd_setup_clocks+0x2c0/0x490
> [    4.413433]  ufshcd_init+0x198/0x10ec
> [    4.413437]  ufshcd_pltfrm_init+0x600/0x7c0
> [    4.413444]  ufs_qcom_probe+0x20/0x58
> [    4.413449]  platform_probe+0x68/0xd8
> [    4.413459]  really_probe+0xbc/0x268
> [    4.413466]  __driver_probe_device+0x78/0x12c
> [    4.413473]  driver_probe_device+0x40/0x11c
> [    4.413481]  __device_attach_driver+0xb8/0xf8
> [    4.413489]  bus_for_each_drv+0x84/0xe4
> [    4.413495]  __device_attach+0xfc/0x18c
> [    4.413502]  device_initial_probe+0x14/0x20
> [    4.413510]  bus_probe_device+0xb0/0xb4
> [    4.413517]  deferred_probe_work_func+0x8c/0xc8
> [    4.413524]  process_scheduled_works+0x250/0x658
> [    4.413534]  worker_thread+0x15c/0x2c8
> [    4.413542]  kthread+0x134/0x200
> [    4.413550]  ret_from_fork+0x10/0x20
> 
> To fix this issue, we use the existing `is_initialized` flag in the
> `clk_gating` structure to ensure that the spinlock is only used after it
> has been properly initialized. We check this flag before using the
> spinlock in the `ufshcd_setup_clocks` function.
> 
> It was incorrect in the first place to call `setup_clocks()` before
> `ufshcd_init_clk_gating()`, and the introduction of the new lock
> unmasked this bug.

If calling setup_clocks() before ufshcd_init_clk_gating() is incorrect, why are
you not reordering it?

Checking for 'clk_gating.is_initialized' looks like a hack.

- Mani

> 
> Fixes: 209f4e43b806 ("scsi: ufs: core: Introduce a new clock_gating lock")
> Reported-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
> Signed-off-by: Avri Altman <avri.altman@wdc.com>
> ---
>  drivers/ufs/core/ufshcd.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
> index f6c38cf10382..a778fc51ca2a 100644
> --- a/drivers/ufs/core/ufshcd.c
> +++ b/drivers/ufs/core/ufshcd.c
> @@ -9142,7 +9142,7 @@ static int ufshcd_setup_clocks(struct ufs_hba *hba, bool on)
>  			if (!IS_ERR_OR_NULL(clki->clk) && clki->enabled)
>  				clk_disable_unprepare(clki->clk);
>  		}
> -	} else if (!ret && on) {
> +	} else if (!ret && on && hba->clk_gating.is_initialized) {
>  		scoped_guard(spinlock_irqsave, &hba->clk_gating.lock)
>  			hba->clk_gating.state = CLKS_ON;
>  		trace_ufshcd_clk_gating(dev_name(hba->dev),
> -- 
> 2.25.1
> 

-- 
மணிவண்ணன் சதாசிவம்

RE: [PATCH] scsi: ufs: core: Ensure clk_gating.lock is used only after initialization

[Avri Altman]

> > To fix this issue, we use the existing `is_initialized` flag in the
> > `clk_gating` structure to ensure that the spinlock is only used after
> > it has been properly initialized. We check this flag before using the
> > spinlock in the `ufshcd_setup_clocks` function.
> >
> > It was incorrect in the first place to call `setup_clocks()` before
> > `ufshcd_init_clk_gating()`, and the introduction of the new lock
> > unmasked this bug.
> 
> If calling setup_clocks() before ufshcd_init_clk_gating() is incorrect, why are
> you not reordering it?
> 
> Checking for 'clk_gating.is_initialized' looks like a hack.
Actually 'clk_gating.is_initialized' seems like the standard way to do this - see e.g. in hold and release.
As for moving setup_clocks() around, I have some concerns about moving it out of ufshcd_hba_init.
Having considered the alternatives, it seems that using 'clk_gating.is_initialized' ,
despite its limitations, is the most practical solution we have.

I am open though for other suggestions.

Thanks,
Avri

> 
> - Mani
> 
> >
> > Fixes: 209f4e43b806 ("scsi: ufs: core: Introduce a new clock_gating
> > lock")
> > Reported-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
> > Signed-off-by: Avri Altman <avri.altman@wdc.com>
> > ---
> >  drivers/ufs/core/ufshcd.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
> > index f6c38cf10382..a778fc51ca2a 100644
> > --- a/drivers/ufs/core/ufshcd.c
> > +++ b/drivers/ufs/core/ufshcd.c
> > @@ -9142,7 +9142,7 @@ static int ufshcd_setup_clocks(struct ufs_hba
> *hba, bool on)
> >                       if (!IS_ERR_OR_NULL(clki->clk) && clki->enabled)
> >                               clk_disable_unprepare(clki->clk);
> >               }
> > -     } else if (!ret && on) {
> > +     } else if (!ret && on && hba->clk_gating.is_initialized) {
> >               scoped_guard(spinlock_irqsave, &hba->clk_gating.lock)
> >                       hba->clk_gating.state = CLKS_ON;
> >               trace_ufshcd_clk_gating(dev_name(hba->dev),
> > --
> > 2.25.1
> >
> 
> --
> மணிவண்ணன் சதாசிவம்

Re: [PATCH] scsi: ufs: core: Ensure clk_gating.lock is used only after initialization

[Manivannan Sadhasivam]

On Wed, Jan 22, 2025 at 11:04:12AM +0000, Avri Altman wrote:
> > > To fix this issue, we use the existing `is_initialized` flag in the
> > > `clk_gating` structure to ensure that the spinlock is only used after
> > > it has been properly initialized. We check this flag before using the
> > > spinlock in the `ufshcd_setup_clocks` function.
> > >
> > > It was incorrect in the first place to call `setup_clocks()` before
> > > `ufshcd_init_clk_gating()`, and the introduction of the new lock
> > > unmasked this bug.
> > 
> > If calling setup_clocks() before ufshcd_init_clk_gating() is incorrect, why are
> > you not reordering it?
> > 
> > Checking for 'clk_gating.is_initialized' looks like a hack.
> Actually 'clk_gating.is_initialized' seems like the standard way to do this - see e.g. in hold and release.
> As for moving setup_clocks() around, I have some concerns about moving it out of ufshcd_hba_init.
> Having considered the alternatives, it seems that using 'clk_gating.is_initialized' ,
> despite its limitations, is the most practical solution we have.
> 
> I am open though for other suggestions.
> 

Looking at the code again, I think it is OK to have this fix for now. But
someone should spend some time to revisit the locking part in this driver.

Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>

- Mani

-- 
மணிவண்ணன் சதாசிவம்

Re: [PATCH] scsi: ufs: core: Ensure clk_gating.lock is used only after initialization

[Bart Van Assche]

On 1/21/25 10:27 PM, Avri Altman wrote:
> This commit addresses a lockdep warning triggered by the use of the
> clk_gating.lock before it is properly initialized. The warning is as
> follows:
> 
> [    4.388838] INFO: trying to register non-static key.
> [    4.395673] The code is fine but needs lockdep annotation, or maybe
> [    4.402118] you didn't initialize this object before use?
> [    4.407673] turning off the locking correctness validator.
> [    4.413334] CPU: 5 UID: 0 PID: 58 Comm: kworker/u32:1 Not tainted 6.12-rc1 #185
> [    4.413343] Hardware name: Qualcomm Technologies, Inc. Robotics RB5 (DT)
> [    4.413362] Call trace:
> [    4.413364]  show_stack+0x18/0x24 (C)
> [    4.413374]  dump_stack_lvl+0x90/0xd0
> [    4.413384]  dump_stack+0x18/0x24
> [    4.413392]  register_lock_class+0x498/0x4a8
> [    4.413400]  __lock_acquire+0xb4/0x1b90
> [    4.413406]  lock_acquire+0x114/0x310
> [    4.413413]  _raw_spin_lock_irqsave+0x60/0x88
> [    4.413423]  ufshcd_setup_clocks+0x2c0/0x490
> [    4.413433]  ufshcd_init+0x198/0x10ec
> [    4.413437]  ufshcd_pltfrm_init+0x600/0x7c0
> [    4.413444]  ufs_qcom_probe+0x20/0x58
> [    4.413449]  platform_probe+0x68/0xd8
> [    4.413459]  really_probe+0xbc/0x268
> [    4.413466]  __driver_probe_device+0x78/0x12c
> [    4.413473]  driver_probe_device+0x40/0x11c
> [    4.413481]  __device_attach_driver+0xb8/0xf8
> [    4.413489]  bus_for_each_drv+0x84/0xe4
> [    4.413495]  __device_attach+0xfc/0x18c
> [    4.413502]  device_initial_probe+0x14/0x20
> [    4.413510]  bus_probe_device+0xb0/0xb4
> [    4.413517]  deferred_probe_work_func+0x8c/0xc8
> [    4.413524]  process_scheduled_works+0x250/0x658
> [    4.413534]  worker_thread+0x15c/0x2c8
> [    4.413542]  kthread+0x134/0x200
> [    4.413550]  ret_from_fork+0x10/0x20
> 
> To fix this issue, we use the existing `is_initialized` flag in the
> `clk_gating` structure to ensure that the spinlock is only used after it
> has been properly initialized. We check this flag before using the
> spinlock in the `ufshcd_setup_clocks` function.
> 
> It was incorrect in the first place to call `setup_clocks()` before
> `ufshcd_init_clk_gating()`, and the introduction of the new lock
> unmasked this bug.
> 
> Fixes: 209f4e43b806 ("scsi: ufs: core: Introduce a new clock_gating lock")
> Reported-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
> Signed-off-by: Avri Altman <avri.altman@wdc.com>
> ---
>   drivers/ufs/core/ufshcd.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
> index f6c38cf10382..a778fc51ca2a 100644
> --- a/drivers/ufs/core/ufshcd.c
> +++ b/drivers/ufs/core/ufshcd.c
> @@ -9142,7 +9142,7 @@ static int ufshcd_setup_clocks(struct ufs_hba *hba, bool on)
>   			if (!IS_ERR_OR_NULL(clki->clk) && clki->enabled)
>   				clk_disable_unprepare(clki->clk);
>   		}
> -	} else if (!ret && on) {
> +	} else if (!ret && on && hba->clk_gating.is_initialized) {
>   		scoped_guard(spinlock_irqsave, &hba->clk_gating.lock)
>   			hba->clk_gating.state = CLKS_ON;
>   		trace_ufshcd_clk_gating(dev_name(hba->dev),

Has it been considered to move the spin_lock_init(&hba->clk_gating.lock)
call from ufshcd_init_clk_gating() such that it occurs before its first
use, e.g. just before the ufshcd_hba_init() call in ufshcd_init()?

Thanks,

Bart.

RE: [PATCH] scsi: ufs: core: Ensure clk_gating.lock is used only after initialization

[Avri Altman]

> On 1/21/25 10:27 PM, Avri Altman wrote:
> > This commit addresses a lockdep warning triggered by the use of the
> > clk_gating.lock before it is properly initialized. The warning is as
> > follows:
> >
> > [    4.388838] INFO: trying to register non-static key.
> > [    4.395673] The code is fine but needs lockdep annotation, or maybe
> > [    4.402118] you didn't initialize this object before use?
> > [    4.407673] turning off the locking correctness validator.
> > [    4.413334] CPU: 5 UID: 0 PID: 58 Comm: kworker/u32:1 Not tainted 6.12-
> rc1 #185
> > [    4.413343] Hardware name: Qualcomm Technologies, Inc. Robotics RB5
> (DT)
> > [    4.413362] Call trace:
> > [    4.413364]  show_stack+0x18/0x24 (C)
> > [    4.413374]  dump_stack_lvl+0x90/0xd0
> > [    4.413384]  dump_stack+0x18/0x24
> > [    4.413392]  register_lock_class+0x498/0x4a8
> > [    4.413400]  __lock_acquire+0xb4/0x1b90
> > [    4.413406]  lock_acquire+0x114/0x310
> > [    4.413413]  _raw_spin_lock_irqsave+0x60/0x88
> > [    4.413423]  ufshcd_setup_clocks+0x2c0/0x490
> > [    4.413433]  ufshcd_init+0x198/0x10ec
> > [    4.413437]  ufshcd_pltfrm_init+0x600/0x7c0
> > [    4.413444]  ufs_qcom_probe+0x20/0x58
> > [    4.413449]  platform_probe+0x68/0xd8
> > [    4.413459]  really_probe+0xbc/0x268
> > [    4.413466]  __driver_probe_device+0x78/0x12c
> > [    4.413473]  driver_probe_device+0x40/0x11c
> > [    4.413481]  __device_attach_driver+0xb8/0xf8
> > [    4.413489]  bus_for_each_drv+0x84/0xe4
> > [    4.413495]  __device_attach+0xfc/0x18c
> > [    4.413502]  device_initial_probe+0x14/0x20
> > [    4.413510]  bus_probe_device+0xb0/0xb4
> > [    4.413517]  deferred_probe_work_func+0x8c/0xc8
> > [    4.413524]  process_scheduled_works+0x250/0x658
> > [    4.413534]  worker_thread+0x15c/0x2c8
> > [    4.413542]  kthread+0x134/0x200
> > [    4.413550]  ret_from_fork+0x10/0x20
> >
> > To fix this issue, we use the existing `is_initialized` flag in the
> > `clk_gating` structure to ensure that the spinlock is only used after
> > it has been properly initialized. We check this flag before using the
> > spinlock in the `ufshcd_setup_clocks` function.
> >
> > It was incorrect in the first place to call `setup_clocks()` before
> > `ufshcd_init_clk_gating()`, and the introduction of the new lock
> > unmasked this bug.
> >
> > Fixes: 209f4e43b806 ("scsi: ufs: core: Introduce a new clock_gating
> > lock")
> > Reported-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
> > Signed-off-by: Avri Altman <avri.altman@wdc.com>
> > ---
> >   drivers/ufs/core/ufshcd.c | 2 +-
> >   1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
> > index f6c38cf10382..a778fc51ca2a 100644
> > --- a/drivers/ufs/core/ufshcd.c
> > +++ b/drivers/ufs/core/ufshcd.c
> > @@ -9142,7 +9142,7 @@ static int ufshcd_setup_clocks(struct ufs_hba
> *hba, bool on)
> >                       if (!IS_ERR_OR_NULL(clki->clk) && clki->enabled)
> >                               clk_disable_unprepare(clki->clk);
> >               }
> > -     } else if (!ret && on) {
> > +     } else if (!ret && on && hba->clk_gating.is_initialized) {
> >               scoped_guard(spinlock_irqsave, &hba->clk_gating.lock)
> >                       hba->clk_gating.state = CLKS_ON;
> >               trace_ufshcd_clk_gating(dev_name(hba->dev),
> 
> Has it been considered to move the spin_lock_init(&hba->clk_gating.lock)
> call from ufshcd_init_clk_gating() such that it occurs before its first use, e.g.
> just before the ufshcd_hba_init() call in ufshcd_init()?
While your suggestion has merit, it would unfortunately break the fundamental concept of concentrating the initialization logic in one place, which is essential for maintaining a clean and manageable codebase.
Will do that if you think it's better.

Thanks,
Avri

> 
> Thanks,
> 
> Bart.

Re: [PATCH] scsi: ufs: core: Ensure clk_gating.lock is used only after initialization

[Geert Uytterhoeven]

Hi Avri,

On Wed, Jan 22, 2025 at 7:30 AM Avri Altman <avri.altman@wdc.com> wrote:
> This commit addresses a lockdep warning triggered by the use of the
> clk_gating.lock before it is properly initialized. The warning is as
> follows:
>
> [    4.388838] INFO: trying to register non-static key.
> [    4.395673] The code is fine but needs lockdep annotation, or maybe
> [    4.402118] you didn't initialize this object before use?
> [    4.407673] turning off the locking correctness validator.
> [    4.413334] CPU: 5 UID: 0 PID: 58 Comm: kworker/u32:1 Not tainted 6.12-rc1 #185
> [    4.413343] Hardware name: Qualcomm Technologies, Inc. Robotics RB5 (DT)
> [    4.413362] Call trace:
> [    4.413364]  show_stack+0x18/0x24 (C)
> [    4.413374]  dump_stack_lvl+0x90/0xd0
> [    4.413384]  dump_stack+0x18/0x24
> [    4.413392]  register_lock_class+0x498/0x4a8
> [    4.413400]  __lock_acquire+0xb4/0x1b90
> [    4.413406]  lock_acquire+0x114/0x310
> [    4.413413]  _raw_spin_lock_irqsave+0x60/0x88
> [    4.413423]  ufshcd_setup_clocks+0x2c0/0x490
> [    4.413433]  ufshcd_init+0x198/0x10ec
> [    4.413437]  ufshcd_pltfrm_init+0x600/0x7c0
> [    4.413444]  ufs_qcom_probe+0x20/0x58
> [    4.413449]  platform_probe+0x68/0xd8
> [    4.413459]  really_probe+0xbc/0x268
> [    4.413466]  __driver_probe_device+0x78/0x12c
> [    4.413473]  driver_probe_device+0x40/0x11c
> [    4.413481]  __device_attach_driver+0xb8/0xf8
> [    4.413489]  bus_for_each_drv+0x84/0xe4
> [    4.413495]  __device_attach+0xfc/0x18c
> [    4.413502]  device_initial_probe+0x14/0x20
> [    4.413510]  bus_probe_device+0xb0/0xb4
> [    4.413517]  deferred_probe_work_func+0x8c/0xc8
> [    4.413524]  process_scheduled_works+0x250/0x658
> [    4.413534]  worker_thread+0x15c/0x2c8
> [    4.413542]  kthread+0x134/0x200
> [    4.413550]  ret_from_fork+0x10/0x20
>
> To fix this issue, we use the existing `is_initialized` flag in the
> `clk_gating` structure to ensure that the spinlock is only used after it
> has been properly initialized. We check this flag before using the
> spinlock in the `ufshcd_setup_clocks` function.
>
> It was incorrect in the first place to call `setup_clocks()` before
> `ufshcd_init_clk_gating()`, and the introduction of the new lock
> unmasked this bug.
>
> Fixes: 209f4e43b806 ("scsi: ufs: core: Introduce a new clock_gating lock")
> Reported-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
> Signed-off-by: Avri Altman <avri.altman@wdc.com>

Thanks for your patch!

I just ran into the same issue on R-Car S4 (S4 Starter Kit).

> --- a/drivers/ufs/core/ufshcd.c
> +++ b/drivers/ufs/core/ufshcd.c
> @@ -9142,7 +9142,7 @@ static int ufshcd_setup_clocks(struct ufs_hba *hba, bool on)
>                         if (!IS_ERR_OR_NULL(clki->clk) && clki->enabled)
>                                 clk_disable_unprepare(clki->clk);
>                 }
> -       } else if (!ret && on) {
> +       } else if (!ret && on && hba->clk_gating.is_initialized) {
>                 scoped_guard(spinlock_irqsave, &hba->clk_gating.lock)
>                         hba->clk_gating.state = CLKS_ON;
>                 trace_ufshcd_clk_gating(dev_name(hba->dev),

This looks like a very fragile solution to me...

In addition, while this change does fix this particular spinlock
warning, it just BUGs in a different place later:

  do_raw_spin_lock+0x34/0xb4
  _raw_spin_lock_irqsave+0x1c/0x30
  class_spinlock_irqsave_constructor+0x18/0x30
- ufshcd_setup_clocks+0x98/0x23c
- ufshcd_init+0x268/0xd2c
+ ufshcd_release+0x30/0x74
+ ufshcd_send_uic_cmd+0x70/0x90
+ ufshcd_link_startup.constprop.0+0x70/0x258
+ ufshcd_init+0xa38/0xd2c
  ufshcd_pltfrm_init+0x618/0x738
  ufs_renesas_probe+0x18/0x24
  platform_probe+0x68/0xb8

I think you should initialize all your spinlocks (and mutexes) early
in ufshcd_init(), instead of sprinkled across various helper functions.

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

RE: [PATCH] scsi: ufs: core: Ensure clk_gating.lock is used only after initialization

[Avri Altman]

> I just ran into the same issue on R-Car S4 (S4 Starter Kit).
> 
> > --- a/drivers/ufs/core/ufshcd.c
> > +++ b/drivers/ufs/core/ufshcd.c
> > @@ -9142,7 +9142,7 @@ static int ufshcd_setup_clocks(struct ufs_hba *hba,
> bool on)
> >                         if (!IS_ERR_OR_NULL(clki->clk) && clki->enabled)
> >                                 clk_disable_unprepare(clki->clk);
> >                 }
> > -       } else if (!ret && on) {
> > +       } else if (!ret && on && hba->clk_gating.is_initialized) {
> >                 scoped_guard(spinlock_irqsave, &hba->clk_gating.lock)
> >                         hba->clk_gating.state = CLKS_ON;
> >                 trace_ufshcd_clk_gating(dev_name(hba->dev),
> 
> This looks like a very fragile solution to me...
> 
> In addition, while this change does fix this particular spinlock warning, it just
> BUGs in a different place later:
> 
>   do_raw_spin_lock+0x34/0xb4
>   _raw_spin_lock_irqsave+0x1c/0x30
>   class_spinlock_irqsave_constructor+0x18/0x30
> - ufshcd_setup_clocks+0x98/0x23c
> - ufshcd_init+0x268/0xd2c
> + ufshcd_release+0x30/0x74
> + ufshcd_send_uic_cmd+0x70/0x90
> + ufshcd_link_startup.constprop.0+0x70/0x258
> + ufshcd_init+0xa38/0xd2c
>   ufshcd_pltfrm_init+0x618/0x738
>   ufs_renesas_probe+0x18/0x24
>   platform_probe+0x68/0xb8
I don't understand how it is possible that `ufshcd_init_clk_gating(hba)` is called after `ufshcd_link_startup(hba)` in 'ufshcd_init'.
Nor how concurrency could take place in this init flow.
Evidently, this is happening.

> 
> I think you should initialize all your spinlocks (and mutexes) early in
> ufshcd_init(), instead of sprinkled across various helper functions.
This is the case today.  Let me suggest a different fix.

Thanks,
Avri

> 
> Gr{oetje,eeting}s,
> 
>                         Geert
> 
> --
> Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-
> m68k.org
> 
> In personal conversations with technical people, I call myself a hacker. But when
> I'm talking to journalists I just say "programmer" or something like that.
>                                 -- Linus Torvalds