[PATCH 0/3] Even moar rpmh cleanups

[PATCH 0/3] Even moar rpmh cleanups

[Stephen Boyd]

Patches based on Doug's latest series[1] on top of linux-next. We remove
the tcs_is_free() API and then do super micro optimizations on the irq
handler. I haven't tested anything here so most likely there's a bug!

Stephen Boyd (3):
  soc: qcom: rpmh-rsc: Remove tcs_is_free() API
  soc: qcom: rpmh-rsc: Loop over less bits in irq handler
  soc: qcom: rpmh-rsc: Fold WARN_ON() into if condition

 drivers/soc/qcom/rpmh-rsc.c | 64 ++++++++++++-------------------------
 1 file changed, 20 insertions(+), 44 deletions(-)

Cc: Maulik Shah <mkshah@codeaurora.org>
Cc: Douglas Anderson <dianders@chromium.org>

[1] https://lore.kernel.org/r/20200422145408.v4.1.Ic7096b3b9b7828cdd41cd5469a6dee5eb6abf549@changeid

base-commit: 02d8ecc18b8f392389ac9e7b785b0230ecb80833
prerequisite-patch-id: 0d383ea46ef52ab4044886a7d88d85c3c506f4ed
prerequisite-patch-id: a02b0b018404d1a0c79270ab567051656f123b23
prerequisite-patch-id: e59d990462b004a9f8335e87c2d0d747afec49ea
prerequisite-patch-id: 00d26aeb99c48521f74b32d3a6c57919d82ac1b1
prerequisite-patch-id: bb479b9adbe28c58b3ac8f363a306de80b6dcb74
-- 
Sent by a computer, using git, on the internet

[PATCH 1/3] soc: qcom: rpmh-rsc: Remove tcs_is_free() API

[Stephen Boyd]

This API does very little. Let's replace all the callsites with the
normal operations that would be done on top of the bitmap that
tcs_in_use is. This simplifies and reduces the code size.

Cc: Maulik Shah <mkshah@codeaurora.org>
Cc: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Stephen Boyd <swboyd@chromium.org>
---
 drivers/soc/qcom/rpmh-rsc.c | 56 +++++++++++--------------------------
 1 file changed, 17 insertions(+), 39 deletions(-)

diff --git a/drivers/soc/qcom/rpmh-rsc.c b/drivers/soc/qcom/rpmh-rsc.c
index 060be10bc491..76e0294a672c 100644
--- a/drivers/soc/qcom/rpmh-rsc.c
+++ b/drivers/soc/qcom/rpmh-rsc.c
@@ -172,22 +172,6 @@ static void write_tcs_reg_sync(struct rsc_drv *drv, int reg, int tcs_id,
 	}
 }
 
-/**
- * tcs_is_free() - Return if a TCS is totally free.
- * @drv:    The RSC controller.
- * @tcs_id: The global ID of this TCS.
- *
- * Returns true if nobody has claimed this TCS (by setting tcs_in_use).
- *
- * Context: Must be called with the drv->lock held.
- *
- * Return: true if the given TCS is free.
- */
-static bool tcs_is_free(struct rsc_drv *drv, int tcs_id)
-{
-	return !test_bit(tcs_id, drv->tcs_in_use);
-}
-
 /**
  * tcs_invalidate() - Invalidate all TCSes of the given type (sleep or wake).
  * @drv:  The RSC controller.
@@ -500,7 +484,7 @@ static void __tcs_buffer_write(struct rsc_drv *drv, int tcs_id, int cmd_id,
  *
  * Return: 0 if nothing in flight or -EBUSY if we should try again later.
  *         The caller must re-enable interrupts between tries since that's
- *         the only way tcs_is_free() will ever return true and the only way
+ *         the only way tcs_in_use will ever be updated and the only way
  *         RSC_DRV_CMD_ENABLE will ever be cleared.
  */
 static int check_for_req_inflight(struct rsc_drv *drv, struct tcs_group *tcs,
@@ -508,17 +492,14 @@ static int check_for_req_inflight(struct rsc_drv *drv, struct tcs_group *tcs,
 {
 	unsigned long curr_enabled;
 	u32 addr;
-	int i, j, k;
-	int tcs_id = tcs->offset;
+	int j, k;
+	int i = tcs->offset;
 
-	for (i = 0; i < tcs->num_tcs; i++, tcs_id++) {
-		if (tcs_is_free(drv, tcs_id))
-			continue;
-
-		curr_enabled = read_tcs_reg(drv, RSC_DRV_CMD_ENABLE, tcs_id);
+	for_each_set_bit_from(i, drv->tcs_in_use, tcs->offset + tcs->num_tcs) {
+		curr_enabled = read_tcs_reg(drv, RSC_DRV_CMD_ENABLE, i);
 
 		for_each_set_bit(j, &curr_enabled, MAX_CMDS_PER_TCS) {
-			addr = read_tcs_cmd(drv, RSC_DRV_CMD_ADDR, tcs_id, j);
+			addr = read_tcs_cmd(drv, RSC_DRV_CMD_ADDR, i, j);
 			for (k = 0; k < msg->num_cmds; k++) {
 				if (addr == msg->cmds[k].addr)
 					return -EBUSY;
@@ -536,18 +517,18 @@ static int check_for_req_inflight(struct rsc_drv *drv, struct tcs_group *tcs,
  *
  * Must be called with the drv->lock held since that protects tcs_in_use.
  *
- * Return: The first tcs that's free.
+ * Return: The first tcs that's free or -EBUSY if all in use.
  */
 static int find_free_tcs(struct tcs_group *tcs)
 {
-	int i;
+	const struct rsc_drv *drv = tcs->drv;
+	unsigned long i;
 
-	for (i = 0; i < tcs->num_tcs; i++) {
-		if (tcs_is_free(tcs->drv, tcs->offset + i))
-			return tcs->offset + i;
-	}
+	i = find_next_zero_bit(drv->tcs_in_use, MAX_TCS_NR, tcs->offset);
+	if (i > tcs->offset + tcs->num_tcs)
+		return -EBUSY;
 
-	return -EBUSY;
+	return i;
 }
 
 /**
@@ -744,8 +725,8 @@ int rpmh_rsc_write_ctrl_data(struct rsc_drv *drv, const struct tcs_request *msg)
  */
 static bool rpmh_rsc_ctrlr_is_busy(struct rsc_drv *drv)
 {
-	int m;
-	struct tcs_group *tcs = &drv->tcs[ACTIVE_TCS];
+	unsigned long set;
+	const struct tcs_group *tcs = &drv->tcs[ACTIVE_TCS];
 
 	/*
 	 * If we made an active request on a RSC that does not have a
@@ -756,12 +737,9 @@ static bool rpmh_rsc_ctrlr_is_busy(struct rsc_drv *drv)
 	if (!tcs->num_tcs)
 		tcs = &drv->tcs[WAKE_TCS];
 
-	for (m = tcs->offset; m < tcs->offset + tcs->num_tcs; m++) {
-		if (!tcs_is_free(drv, m))
-			return true;
-	}
+	set = find_next_bit(drv->tcs_in_use, MAX_TCS_NR, tcs->offset);
 
-	return false;
+	return set < tcs->offset + tcs->num_tcs;
 }
 
 /**
-- 
Sent by a computer, using git, on the internet

[PATCH 2/3] soc: qcom: rpmh-rsc: Loop over less bits in irq handler

[Stephen Boyd]

readl returns a u32, and BITS_PER_LONG is different on 32-bit vs. 64-bit
architectures. Let's make the type we stash the readl into a u32 and
then loop over the bits set in that type instead of potentially looping
over more bits than we will ever need to.

Cc: Maulik Shah <mkshah@codeaurora.org>
Cc: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Stephen Boyd <swboyd@chromium.org>
---
 drivers/soc/qcom/rpmh-rsc.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/soc/qcom/rpmh-rsc.c b/drivers/soc/qcom/rpmh-rsc.c
index 76e0294a672c..462dd267afef 100644
--- a/drivers/soc/qcom/rpmh-rsc.c
+++ b/drivers/soc/qcom/rpmh-rsc.c
@@ -365,13 +365,13 @@ static irqreturn_t tcs_tx_done(int irq, void *p)
 {
 	struct rsc_drv *drv = p;
 	int i, j, err = 0;
-	unsigned long irq_status;
+	u32 irq_status;
 	const struct tcs_request *req;
 	struct tcs_cmd *cmd;
 
 	irq_status = readl_relaxed(drv->tcs_base + RSC_DRV_IRQ_STATUS);
 
-	for_each_set_bit(i, &irq_status, BITS_PER_LONG) {
+	for_each_set_bit(i, &irq_status, BITS_PER_TYPE(u32)) {
 		req = get_req_from_tcs(drv, i);
 		if (!req) {
 			WARN_ON(1);
-- 
Sent by a computer, using git, on the internet

[PATCH 3/3] soc: qcom: rpmh-rsc: Fold WARN_ON() into if condition

[Stephen Boyd]

Move the WARN_ON() into the if condition so the compiler can see that
the branch is unlikely() and possibly optimize it better.

Cc: Maulik Shah <mkshah@codeaurora.org>
Cc: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Stephen Boyd <swboyd@chromium.org>
---
 drivers/soc/qcom/rpmh-rsc.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/drivers/soc/qcom/rpmh-rsc.c b/drivers/soc/qcom/rpmh-rsc.c
index 462dd267afef..f7763f008e03 100644
--- a/drivers/soc/qcom/rpmh-rsc.c
+++ b/drivers/soc/qcom/rpmh-rsc.c
@@ -373,10 +373,8 @@ static irqreturn_t tcs_tx_done(int irq, void *p)
 
 	for_each_set_bit(i, &irq_status, BITS_PER_TYPE(u32)) {
 		req = get_req_from_tcs(drv, i);
-		if (!req) {
-			WARN_ON(1);
+		if (WARN_ON(!req))
 			goto skip;
-		}
 
 		err = 0;
 		for (j = 0; j < req->num_cmds; j++) {
-- 
Sent by a computer, using git, on the internet

Re: [PATCH 1/3] soc: qcom: rpmh-rsc: Remove tcs_is_free() API

[Bjorn Andersson]

On Thu 23 Apr 21:54 PDT 2020, Stephen Boyd wrote:
> diff --git a/drivers/soc/qcom/rpmh-rsc.c b/drivers/soc/qcom/rpmh-rsc.c
[..]
>  static int find_free_tcs(struct tcs_group *tcs)
>  {
> -	int i;
> +	const struct rsc_drv *drv = tcs->drv;
> +	unsigned long i;
>  
> -	for (i = 0; i < tcs->num_tcs; i++) {
> -		if (tcs_is_free(tcs->drv, tcs->offset + i))
> -			return tcs->offset + i;
> -	}
> +	i = find_next_zero_bit(drv->tcs_in_use, MAX_TCS_NR, tcs->offset);
> +	if (i > tcs->offset + tcs->num_tcs)

Afaict this should be >=

Regards,
Bjorn

Re: [PATCH 2/3] soc: qcom: rpmh-rsc: Loop over less bits in irq handler

[Bjorn Andersson]

On Thu 23 Apr 21:54 PDT 2020, Stephen Boyd wrote:

> readl returns a u32, and BITS_PER_LONG is different on 32-bit vs. 64-bit
> architectures. Let's make the type we stash the readl into a u32 and
> then loop over the bits set in that type instead of potentially looping
> over more bits than we will ever need to.
> 
> Cc: Maulik Shah <mkshah@codeaurora.org>
> Cc: Douglas Anderson <dianders@chromium.org>
> Signed-off-by: Stephen Boyd <swboyd@chromium.org>

Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>

> ---
>  drivers/soc/qcom/rpmh-rsc.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/soc/qcom/rpmh-rsc.c b/drivers/soc/qcom/rpmh-rsc.c
> index 76e0294a672c..462dd267afef 100644
> --- a/drivers/soc/qcom/rpmh-rsc.c
> +++ b/drivers/soc/qcom/rpmh-rsc.c
> @@ -365,13 +365,13 @@ static irqreturn_t tcs_tx_done(int irq, void *p)
>  {
>  	struct rsc_drv *drv = p;
>  	int i, j, err = 0;
> -	unsigned long irq_status;
> +	u32 irq_status;
>  	const struct tcs_request *req;
>  	struct tcs_cmd *cmd;
>  
>  	irq_status = readl_relaxed(drv->tcs_base + RSC_DRV_IRQ_STATUS);
>  
> -	for_each_set_bit(i, &irq_status, BITS_PER_LONG) {
> +	for_each_set_bit(i, &irq_status, BITS_PER_TYPE(u32)) {
>  		req = get_req_from_tcs(drv, i);
>  		if (!req) {
>  			WARN_ON(1);
> -- 
> Sent by a computer, using git, on the internet
> 

Re: [PATCH 3/3] soc: qcom: rpmh-rsc: Fold WARN_ON() into if condition

[Bjorn Andersson]

On Thu 23 Apr 21:54 PDT 2020, Stephen Boyd wrote:

> Move the WARN_ON() into the if condition so the compiler can see that
> the branch is unlikely() and possibly optimize it better.
> 
> Cc: Maulik Shah <mkshah@codeaurora.org>
> Cc: Douglas Anderson <dianders@chromium.org>
> Signed-off-by: Stephen Boyd <swboyd@chromium.org>

Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>

> ---
>  drivers/soc/qcom/rpmh-rsc.c | 4 +---
>  1 file changed, 1 insertion(+), 3 deletions(-)
> 
> diff --git a/drivers/soc/qcom/rpmh-rsc.c b/drivers/soc/qcom/rpmh-rsc.c
> index 462dd267afef..f7763f008e03 100644
> --- a/drivers/soc/qcom/rpmh-rsc.c
> +++ b/drivers/soc/qcom/rpmh-rsc.c
> @@ -373,10 +373,8 @@ static irqreturn_t tcs_tx_done(int irq, void *p)
>  
>  	for_each_set_bit(i, &irq_status, BITS_PER_TYPE(u32)) {
>  		req = get_req_from_tcs(drv, i);
> -		if (!req) {
> -			WARN_ON(1);
> +		if (WARN_ON(!req))
>  			goto skip;
> -		}
>  
>  		err = 0;
>  		for (j = 0; j < req->num_cmds; j++) {
> -- 
> Sent by a computer, using git, on the internet
> 

Re: [PATCH 3/3] soc: qcom: rpmh-rsc: Fold WARN_ON() into if condition

[Doug Anderson]

Hi,

On Thu, Apr 23, 2020 at 9:54 PM Stephen Boyd <swboyd@chromium.org> wrote:
>
> Move the WARN_ON() into the if condition so the compiler can see that
> the branch is unlikely() and possibly optimize it better.
>
> Cc: Maulik Shah <mkshah@codeaurora.org>
> Cc: Douglas Anderson <dianders@chromium.org>
> Signed-off-by: Stephen Boyd <swboyd@chromium.org>
> ---
>  drivers/soc/qcom/rpmh-rsc.c | 4 +---
>  1 file changed, 1 insertion(+), 3 deletions(-)

Reviewed-by: Douglas Anderson <dianders@chromium.org>

Re: [PATCH 2/3] soc: qcom: rpmh-rsc: Loop over less bits in irq handler

[Doug Anderson]

Hi,

On Thu, Apr 23, 2020 at 9:54 PM Stephen Boyd <swboyd@chromium.org> wrote:
>
> readl returns a u32, and BITS_PER_LONG is different on 32-bit vs. 64-bit
> architectures. Let's make the type we stash the readl into a u32 and
> then loop over the bits set in that type instead of potentially looping
> over more bits than we will ever need to.
>
> Cc: Maulik Shah <mkshah@codeaurora.org>
> Cc: Douglas Anderson <dianders@chromium.org>
> Signed-off-by: Stephen Boyd <swboyd@chromium.org>
> ---
>  drivers/soc/qcom/rpmh-rsc.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)

Reviewed-by: Douglas Anderson <dianders@chromium.org>

Re: [PATCH 1/3] soc: qcom: rpmh-rsc: Remove tcs_is_free() API

[Doug Anderson]

Hi,

On Thu, Apr 23, 2020 at 9:54 PM Stephen Boyd <swboyd@chromium.org> wrote:
>
> This API does very little. Let's replace all the callsites with the
> normal operations that would be done on top of the bitmap that
> tcs_in_use is. This simplifies and reduces the code size.
>
> Cc: Maulik Shah <mkshah@codeaurora.org>
> Cc: Douglas Anderson <dianders@chromium.org>
> Signed-off-by: Stephen Boyd <swboyd@chromium.org>
> ---
>  drivers/soc/qcom/rpmh-rsc.c | 56 +++++++++++--------------------------
>  1 file changed, 17 insertions(+), 39 deletions(-)
>
> diff --git a/drivers/soc/qcom/rpmh-rsc.c b/drivers/soc/qcom/rpmh-rsc.c
> index 060be10bc491..76e0294a672c 100644
> --- a/drivers/soc/qcom/rpmh-rsc.c
> +++ b/drivers/soc/qcom/rpmh-rsc.c
> @@ -172,22 +172,6 @@ static void write_tcs_reg_sync(struct rsc_drv *drv, int reg, int tcs_id,
>         }
>  }
>
> -/**
> - * tcs_is_free() - Return if a TCS is totally free.
> - * @drv:    The RSC controller.
> - * @tcs_id: The global ID of this TCS.
> - *
> - * Returns true if nobody has claimed this TCS (by setting tcs_in_use).
> - *
> - * Context: Must be called with the drv->lock held.
> - *
> - * Return: true if the given TCS is free.
> - */
> -static bool tcs_is_free(struct rsc_drv *drv, int tcs_id)
> -{
> -       return !test_bit(tcs_id, drv->tcs_in_use);
> -}
> -
>  /**
>   * tcs_invalidate() - Invalidate all TCSes of the given type (sleep or wake).
>   * @drv:  The RSC controller.
> @@ -500,7 +484,7 @@ static void __tcs_buffer_write(struct rsc_drv *drv, int tcs_id, int cmd_id,
>   *
>   * Return: 0 if nothing in flight or -EBUSY if we should try again later.
>   *         The caller must re-enable interrupts between tries since that's
> - *         the only way tcs_is_free() will ever return true and the only way
> + *         the only way tcs_in_use will ever be updated and the only way
>   *         RSC_DRV_CMD_ENABLE will ever be cleared.
>   */
>  static int check_for_req_inflight(struct rsc_drv *drv, struct tcs_group *tcs,
> @@ -508,17 +492,14 @@ static int check_for_req_inflight(struct rsc_drv *drv, struct tcs_group *tcs,
>  {
>         unsigned long curr_enabled;
>         u32 addr;
> -       int i, j, k;
> -       int tcs_id = tcs->offset;
> +       int j, k;
> +       int i = tcs->offset;
>
> -       for (i = 0; i < tcs->num_tcs; i++, tcs_id++) {
> -               if (tcs_is_free(drv, tcs_id))
> -                       continue;
> -
> -               curr_enabled = read_tcs_reg(drv, RSC_DRV_CMD_ENABLE, tcs_id);
> +       for_each_set_bit_from(i, drv->tcs_in_use, tcs->offset + tcs->num_tcs) {
> +               curr_enabled = read_tcs_reg(drv, RSC_DRV_CMD_ENABLE, i);
>
>                 for_each_set_bit(j, &curr_enabled, MAX_CMDS_PER_TCS) {
> -                       addr = read_tcs_cmd(drv, RSC_DRV_CMD_ADDR, tcs_id, j);
> +                       addr = read_tcs_cmd(drv, RSC_DRV_CMD_ADDR, i, j);
>                         for (k = 0; k < msg->num_cmds; k++) {
>                                 if (addr == msg->cmds[k].addr)
>                                         return -EBUSY;
> @@ -536,18 +517,18 @@ static int check_for_req_inflight(struct rsc_drv *drv, struct tcs_group *tcs,
>   *
>   * Must be called with the drv->lock held since that protects tcs_in_use.
>   *
> - * Return: The first tcs that's free.
> + * Return: The first tcs that's free or -EBUSY if all in use.
>   */
>  static int find_free_tcs(struct tcs_group *tcs)
>  {
> -       int i;
> +       const struct rsc_drv *drv = tcs->drv;
> +       unsigned long i;
>
> -       for (i = 0; i < tcs->num_tcs; i++) {
> -               if (tcs_is_free(tcs->drv, tcs->offset + i))
> -                       return tcs->offset + i;
> -       }
> +       i = find_next_zero_bit(drv->tcs_in_use, MAX_TCS_NR, tcs->offset);

Rather than passing MAX_TCS_NR would it be legit to pass "tcs->offset
+ tcs->num_tcs"?  You are passing that as "size" above in
check_for_req_inflight().


> +       if (i > tcs->offset + tcs->num_tcs)

Should the above be ">=" ?  Oh, I guess Bjorn found the same thing.  ;-)


> +               return -EBUSY;
>
> -       return -EBUSY;
> +       return i;
>  }
>
>  /**
> @@ -744,8 +725,8 @@ int rpmh_rsc_write_ctrl_data(struct rsc_drv *drv, const struct tcs_request *msg)
>   */
>  static bool rpmh_rsc_ctrlr_is_busy(struct rsc_drv *drv)
>  {
> -       int m;
> -       struct tcs_group *tcs = &drv->tcs[ACTIVE_TCS];
> +       unsigned long set;
> +       const struct tcs_group *tcs = &drv->tcs[ACTIVE_TCS];
>
>         /*
>          * If we made an active request on a RSC that does not have a
> @@ -756,12 +737,9 @@ static bool rpmh_rsc_ctrlr_is_busy(struct rsc_drv *drv)
>         if (!tcs->num_tcs)
>                 tcs = &drv->tcs[WAKE_TCS];
>
> -       for (m = tcs->offset; m < tcs->offset + tcs->num_tcs; m++) {
> -               if (!tcs_is_free(drv, m))
> -                       return true;
> -       }
> +       set = find_next_bit(drv->tcs_in_use, MAX_TCS_NR, tcs->offset);

Similar to above, can you pass "tcs->offset + tcs->num_tcs" instead of
MAX_TCS_NR?

Re: [PATCH 2/3] soc: qcom: rpmh-rsc: Loop over less bits in irq handler

[Doug Anderson]

Hi,

On Fri, Apr 24, 2020 at 10:11 AM Doug Anderson <dianders@chromium.org> wrote:
>
> Hi,
>
> On Thu, Apr 23, 2020 at 9:54 PM Stephen Boyd <swboyd@chromium.org> wrote:
> >
> > readl returns a u32, and BITS_PER_LONG is different on 32-bit vs. 64-bit
> > architectures. Let's make the type we stash the readl into a u32 and
> > then loop over the bits set in that type instead of potentially looping
> > over more bits than we will ever need to.
> >
> > Cc: Maulik Shah <mkshah@codeaurora.org>
> > Cc: Douglas Anderson <dianders@chromium.org>
> > Signed-off-by: Stephen Boyd <swboyd@chromium.org>
> > ---
> >  drivers/soc/qcom/rpmh-rsc.c | 4 ++--
> >  1 file changed, 2 insertions(+), 2 deletions(-)
>
> Reviewed-by: Douglas Anderson <dianders@chromium.org>

Oh, I suppose one nit is s/less/fewer/ in the subject.  "bits" are
discrete / countable, not continuous / uncountable.

-Doug

Re: [PATCH 1/3] soc: qcom: rpmh-rsc: Remove tcs_is_free() API

[Stephen Boyd]

Quoting Bjorn Andersson (2020-04-23 22:36:26)
> On Thu 23 Apr 21:54 PDT 2020, Stephen Boyd wrote:
> > diff --git a/drivers/soc/qcom/rpmh-rsc.c b/drivers/soc/qcom/rpmh-rsc.c
> [..]
> >  static int find_free_tcs(struct tcs_group *tcs)
> >  {
> > -     int i;
> > +     const struct rsc_drv *drv = tcs->drv;
> > +     unsigned long i;
> >  
> > -     for (i = 0; i < tcs->num_tcs; i++) {
> > -             if (tcs_is_free(tcs->drv, tcs->offset + i))
> > -                     return tcs->offset + i;
> > -     }
> > +     i = find_next_zero_bit(drv->tcs_in_use, MAX_TCS_NR, tcs->offset);
> > +     if (i > tcs->offset + tcs->num_tcs)
> 
> Afaict this should be >=
> 

Thanks!

Re: [PATCH 2/3] soc: qcom: rpmh-rsc: Loop over less bits in irq handler

[Stephen Boyd]

Quoting Doug Anderson (2020-04-24 10:13:43)
> Hi,
> 
> On Fri, Apr 24, 2020 at 10:11 AM Doug Anderson <dianders@chromium.org> wrote:
> >
> > Hi,
> >
> > On Thu, Apr 23, 2020 at 9:54 PM Stephen Boyd <swboyd@chromium.org> wrote:
> > >
> > > readl returns a u32, and BITS_PER_LONG is different on 32-bit vs. 64-bit
> > > architectures. Let's make the type we stash the readl into a u32 and
> > > then loop over the bits set in that type instead of potentially looping
> > > over more bits than we will ever need to.
> > >
> > > Cc: Maulik Shah <mkshah@codeaurora.org>
> > > Cc: Douglas Anderson <dianders@chromium.org>
> > > Signed-off-by: Stephen Boyd <swboyd@chromium.org>
> > > ---
> > >  drivers/soc/qcom/rpmh-rsc.c | 4 ++--
> > >  1 file changed, 2 insertions(+), 2 deletions(-)
> >
> > Reviewed-by: Douglas Anderson <dianders@chromium.org>
> 
> Oh, I suppose one nit is s/less/fewer/ in the subject.  "bits" are
> discrete / countable, not continuous / uncountable.
> 

Ok that's good.

Re: [PATCH 1/3] soc: qcom: rpmh-rsc: Remove tcs_is_free() API

[Stephen Boyd]

Quoting Doug Anderson (2020-04-24 10:12:04)
> On Thu, Apr 23, 2020 at 9:54 PM Stephen Boyd <swboyd@chromium.org> wrote:
> > +       unsigned long i;
> >
> > -       for (i = 0; i < tcs->num_tcs; i++) {
> > -               if (tcs_is_free(tcs->drv, tcs->offset + i))
> > -                       return tcs->offset + i;
> > -       }
> > +       i = find_next_zero_bit(drv->tcs_in_use, MAX_TCS_NR, tcs->offset);
> 
> Rather than passing MAX_TCS_NR would it be legit to pass "tcs->offset
> + tcs->num_tcs"?  You are passing that as "size" above in
> check_for_req_inflight().

Yeah I think that should work. I'll roll it in.