1. Patchew
  2. linux
  3. View series

[PATCH] i3c: dw: reduce do_daa time if there's no client

Jisheng Zhang posted 1 patch 1 month, 4 weeks ago
There is a newer version of this series
drivers/i3c/master/dw-i3c-master.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
[PATCH] i3c: dw: reduce do_daa time if there's no client
Posted by Jisheng Zhang 1 month, 4 weeks ago 
If there's no client, wait for completion will timeout, then there's no
need to add i3c dev at all. This can save boot time by ~6s if there's
no client connected to the i3c interface.

Signed-off-by: Jisheng Zhang <jszhang@kernel.org>
---
 drivers/i3c/master/dw-i3c-master.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/i3c/master/dw-i3c-master.c b/drivers/i3c/master/dw-i3c-master.c
index 0625587f1527..18ebf304029a 100644
--- a/drivers/i3c/master/dw-i3c-master.c
+++ b/drivers/i3c/master/dw-i3c-master.c
@@ -885,8 +885,10 @@ static int dw_i3c_master_daa(struct i3c_master_controller *m)
 		      COMMAND_PORT_ROC;
 
 	dw_i3c_master_enqueue_xfer(master, xfer);
-	if (!wait_for_completion_timeout(&xfer->comp, XFER_TIMEOUT))
+	if (!wait_for_completion_timeout(&xfer->comp, XFER_TIMEOUT)) {
 		dw_i3c_master_dequeue_xfer(master, xfer);
+		goto rpm_out;
+	}
 
 	newdevs = GENMASK(master->maxdevs - cmd->rx_len - 1, 0);
 	newdevs &= ~olddevs;
-- 
2.53.0
Re: [PATCH] i3c: dw: reduce do_daa time if there's no client
Posted by Frank Li 1 month, 3 weeks ago 
On Mon, Aug 03, 2026 at 10:03:07PM +0800, Jisheng Zhang wrote:
> If there's no client, wait for completion will timeout, then there's no
> need to add i3c dev at all. This can save boot time by ~6s if there's
> no client connected to the i3c interface.
>
> Signed-off-by: Jisheng Zhang <jszhang@kernel.org>
> ---
>  drivers/i3c/master/dw-i3c-master.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/i3c/master/dw-i3c-master.c b/drivers/i3c/master/dw-i3c-master.c
> index 0625587f1527..18ebf304029a 100644
> --- a/drivers/i3c/master/dw-i3c-master.c
> +++ b/drivers/i3c/master/dw-i3c-master.c
> @@ -885,8 +885,10 @@ static int dw_i3c_master_daa(struct i3c_master_controller *m)
>  		      COMMAND_PORT_ROC;
>
>  	dw_i3c_master_enqueue_xfer(master, xfer);
> -	if (!wait_for_completion_timeout(&xfer->comp, XFER_TIMEOUT))
> +	if (!wait_for_completion_timeout(&xfer->comp, XFER_TIMEOUT)) {
>  		dw_i3c_master_dequeue_xfer(master, xfer);
> +		goto rpm_out;

not sure why save 6s. timeout already happen, does register i3c device
takes 6s?

And if do_daa find 3 devices, then timeout, you skip register found 3
devices.

Frank

> +	}
>
>  	newdevs = GENMASK(master->maxdevs - cmd->rx_len - 1, 0);
>  	newdevs &= ~olddevs;
> --
> 2.53.0
>
Re: [PATCH] i3c: dw: reduce do_daa time if there's no client
Posted by Jisheng Zhang 1 month, 2 weeks ago 
On Wed, Aug 05, 2026 at 04:05:35PM -0400, Frank Li wrote:
> On Mon, Aug 03, 2026 at 10:03:07PM +0800, Jisheng Zhang wrote:
> > If there's no client, wait for completion will timeout, then there's no
> > need to add i3c dev at all. This can save boot time by ~6s if there's
> > no client connected to the i3c interface.
> >
> > Signed-off-by: Jisheng Zhang <jszhang@kernel.org>
> > ---
> >  drivers/i3c/master/dw-i3c-master.c | 4 +++-
> >  1 file changed, 3 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/i3c/master/dw-i3c-master.c b/drivers/i3c/master/dw-i3c-master.c
> > index 0625587f1527..18ebf304029a 100644
> > --- a/drivers/i3c/master/dw-i3c-master.c
> > +++ b/drivers/i3c/master/dw-i3c-master.c
> > @@ -885,8 +885,10 @@ static int dw_i3c_master_daa(struct i3c_master_controller *m)
> >  		      COMMAND_PORT_ROC;
> >
> >  	dw_i3c_master_enqueue_xfer(master, xfer);
> > -	if (!wait_for_completion_timeout(&xfer->comp, XFER_TIMEOUT))
> > +	if (!wait_for_completion_timeout(&xfer->comp, XFER_TIMEOUT)) {
> >  		dw_i3c_master_dequeue_xfer(master, xfer);
> > +		goto rpm_out;
> 
> not sure why save 6s. timeout already happen, does register i3c device
> takes 6s?

sorry, typo, s/6/4
For example if the maxdev == 4, when there's no i3c device connected, the
max cmd->rx_len = 0 after timeout, then newdevs = 0xf, then go through
the i3c_master_add_i3c_dev_locked(), each take ~1s, so adds up to 4s

> 
> And if do_daa find 3 devices, then timeout, you skip register found 3
> devices.

It's the addr assign cmd timeout, so if 3 devices respond, the addr
assign cmd won't timeout.

> 
> Frank
> 
> > +	}
> >
> >  	newdevs = GENMASK(master->maxdevs - cmd->rx_len - 1, 0);
> >  	newdevs &= ~olddevs;
> > --
> > 2.53.0
> >
Re: [PATCH] i3c: dw: reduce do_daa time if there's no client
Posted by Frank Li 1 month, 2 weeks ago 
On Thu, Aug 13, 2026 at 12:56:09PM +0800, Jisheng Zhang wrote:
> On Wed, Aug 05, 2026 at 04:05:35PM -0400, Frank Li wrote:
> > On Mon, Aug 03, 2026 at 10:03:07PM +0800, Jisheng Zhang wrote:
> > > If there's no client, wait for completion will timeout, then there's no
> > > need to add i3c dev at all. This can save boot time by ~6s if there's
> > > no client connected to the i3c interface.
> > >
> > > Signed-off-by: Jisheng Zhang <jszhang@kernel.org>
> > > ---
> > >  drivers/i3c/master/dw-i3c-master.c | 4 +++-
> > >  1 file changed, 3 insertions(+), 1 deletion(-)
> > >
> > > diff --git a/drivers/i3c/master/dw-i3c-master.c b/drivers/i3c/master/dw-i3c-master.c
> > > index 0625587f1527..18ebf304029a 100644
> > > --- a/drivers/i3c/master/dw-i3c-master.c
> > > +++ b/drivers/i3c/master/dw-i3c-master.c
> > > @@ -885,8 +885,10 @@ static int dw_i3c_master_daa(struct i3c_master_controller *m)
> > >  		      COMMAND_PORT_ROC;
> > >
> > >  	dw_i3c_master_enqueue_xfer(master, xfer);
> > > -	if (!wait_for_completion_timeout(&xfer->comp, XFER_TIMEOUT))
> > > +	if (!wait_for_completion_timeout(&xfer->comp, XFER_TIMEOUT)) {
> > >  		dw_i3c_master_dequeue_xfer(master, xfer);
> > > +		goto rpm_out;
> >
> > not sure why save 6s. timeout already happen, does register i3c device
> > takes 6s?
>
> sorry, typo, s/6/4
> For example if the maxdev == 4, when there's no i3c device connected, the
> max cmd->rx_len = 0 after timeout, then newdevs = 0xf, then go through
> the i3c_master_add_i3c_dev_locked(), each take ~1s, so adds up to 4s


https://lore.kernel.org/linux-i3c/20260731070150.2519825-1-tommaso.merciai.xr@bp.renesas.com/

Need fix newdevs, it should be 0, not 0xf.

Frank

>
> >
> > And if do_daa find 3 devices, then timeout, you skip register found 3
> > devices.
>
> It's the addr assign cmd timeout, so if 3 devices respond, the addr
> assign cmd won't timeout.
>
> >
> > Frank
> >
> > > +	}
> > >
> > >  	newdevs = GENMASK(master->maxdevs - cmd->rx_len - 1, 0);
> > >  	newdevs &= ~olddevs;
> > > --
> > > 2.53.0
> > >
Re: [PATCH] i3c: dw: reduce do_daa time if there's no client
Posted by Jisheng Zhang 1 month, 2 weeks ago 
On Thu, Aug 13, 2026 at 04:45:00PM -0400, Frank Li wrote:
> On Thu, Aug 13, 2026 at 12:56:09PM +0800, Jisheng Zhang wrote:
> > On Wed, Aug 05, 2026 at 04:05:35PM -0400, Frank Li wrote:
> > > On Mon, Aug 03, 2026 at 10:03:07PM +0800, Jisheng Zhang wrote:
> > > > If there's no client, wait for completion will timeout, then there's no
> > > > need to add i3c dev at all. This can save boot time by ~6s if there's
> > > > no client connected to the i3c interface.
> > > >
> > > > Signed-off-by: Jisheng Zhang <jszhang@kernel.org>
> > > > ---
> > > >  drivers/i3c/master/dw-i3c-master.c | 4 +++-
> > > >  1 file changed, 3 insertions(+), 1 deletion(-)
> > > >
> > > > diff --git a/drivers/i3c/master/dw-i3c-master.c b/drivers/i3c/master/dw-i3c-master.c
> > > > index 0625587f1527..18ebf304029a 100644
> > > > --- a/drivers/i3c/master/dw-i3c-master.c
> > > > +++ b/drivers/i3c/master/dw-i3c-master.c
> > > > @@ -885,8 +885,10 @@ static int dw_i3c_master_daa(struct i3c_master_controller *m)
> > > >  		      COMMAND_PORT_ROC;
> > > >
> > > >  	dw_i3c_master_enqueue_xfer(master, xfer);
> > > > -	if (!wait_for_completion_timeout(&xfer->comp, XFER_TIMEOUT))
> > > > +	if (!wait_for_completion_timeout(&xfer->comp, XFER_TIMEOUT)) {
> > > >  		dw_i3c_master_dequeue_xfer(master, xfer);
> > > > +		goto rpm_out;
> > >
> > > not sure why save 6s. timeout already happen, does register i3c device
> > > takes 6s?
> >
> > sorry, typo, s/6/4
> > For example if the maxdev == 4, when there's no i3c device connected, the
> > max cmd->rx_len = 0 after timeout, then newdevs = 0xf, then go through
> > the i3c_master_add_i3c_dev_locked(), each take ~1s, so adds up to 4s
> 
> 
> https://lore.kernel.org/linux-i3c/20260731070150.2519825-1-tommaso.merciai.xr@bp.renesas.com/
> 
> Need fix newdevs, it should be 0, not 0xf.

Aha, this method also works. Let me cook a new version

Thanks a lot
> 
> Frank
> 
> >
> > >
> > > And if do_daa find 3 devices, then timeout, you skip register found 3
> > > devices.
> >
> > It's the addr assign cmd timeout, so if 3 devices respond, the addr
> > assign cmd won't timeout.
> >
> > >
> > > Frank
> > >
> > > > +	}
> > > >
> > > >  	newdevs = GENMASK(master->maxdevs - cmd->rx_len - 1, 0);
> > > >  	newdevs &= ~olddevs;
> > > > --
> > > > 2.53.0
> > > >

Patchew 2.1-devel-ea86937

© 2016 - 2026 Red Hat, Inc.