2022-06-23 12:45:27

by Paul Cercueil

[permalink] [raw]
Subject: [PATCH v2] brcmfmac: Remove #ifdef guards for PM related functions

Use the new DEFINE_SIMPLE_DEV_PM_OPS() and pm_sleep_ptr() macros to
handle the .suspend/.resume callbacks.

These macros allow the suspend and resume functions to be automatically
dropped by the compiler when CONFIG_SUSPEND is disabled, without having
to use #ifdef guards.

Some other functions not directly called by the .suspend/.resume
callbacks, but still related to PM were also taken outside #ifdef
guards.

The advantage is then that these functions are now always compiled
independently of any Kconfig option, and thanks to that bugs and
regressions are easier to catch.

Signed-off-by: Paul Cercueil <[email protected]>
---

Notes:
v2:
- Move checks for IS_ENABLED(CONFIG_PM_SLEEP) inside functions to keep
the calling functions intact.
- Reword the commit message to explain why this patch is useful.

.../broadcom/brcm80211/brcmfmac/bcmsdh.c | 38 +++++++------------
.../broadcom/brcm80211/brcmfmac/sdio.c | 5 +--
.../broadcom/brcm80211/brcmfmac/sdio.h | 16 --------
3 files changed, 16 insertions(+), 43 deletions(-)

diff --git a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/bcmsdh.c b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/bcmsdh.c
index 9c598ea97499..11ad878a906b 100644
--- a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/bcmsdh.c
+++ b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/bcmsdh.c
@@ -784,9 +784,11 @@ void brcmf_sdiod_sgtable_alloc(struct brcmf_sdio_dev *sdiodev)
sdiodev->txglomsz = sdiodev->settings->bus.sdio.txglomsz;
}

-#ifdef CONFIG_PM_SLEEP
static int brcmf_sdiod_freezer_attach(struct brcmf_sdio_dev *sdiodev)
{
+ if (!IS_ENABLED(CONFIG_PM_SLEEP))
+ return 0;
+
sdiodev->freezer = kzalloc(sizeof(*sdiodev->freezer), GFP_KERNEL);
if (!sdiodev->freezer)
return -ENOMEM;
@@ -799,7 +801,7 @@ static int brcmf_sdiod_freezer_attach(struct brcmf_sdio_dev *sdiodev)

static void brcmf_sdiod_freezer_detach(struct brcmf_sdio_dev *sdiodev)
{
- if (sdiodev->freezer) {
+ if (IS_ENABLED(CONFIG_PM_SLEEP) && sdiodev->freezer) {
WARN_ON(atomic_read(&sdiodev->freezer->freezing));
kfree(sdiodev->freezer);
}
@@ -833,7 +835,8 @@ static void brcmf_sdiod_freezer_off(struct brcmf_sdio_dev *sdiodev)

bool brcmf_sdiod_freezing(struct brcmf_sdio_dev *sdiodev)
{
- return atomic_read(&sdiodev->freezer->freezing);
+ return IS_ENABLED(CONFIG_PM_SLEEP) &&
+ atomic_read(&sdiodev->freezer->freezing);
}

void brcmf_sdiod_try_freeze(struct brcmf_sdio_dev *sdiodev)
@@ -847,24 +850,16 @@ void brcmf_sdiod_try_freeze(struct brcmf_sdio_dev *sdiodev)

void brcmf_sdiod_freezer_count(struct brcmf_sdio_dev *sdiodev)
{
- atomic_inc(&sdiodev->freezer->thread_count);
+ if (IS_ENABLED(CONFIG_PM_SLEEP))
+ atomic_inc(&sdiodev->freezer->thread_count);
}

void brcmf_sdiod_freezer_uncount(struct brcmf_sdio_dev *sdiodev)
{
- atomic_dec(&sdiodev->freezer->thread_count);
-}
-#else
-static int brcmf_sdiod_freezer_attach(struct brcmf_sdio_dev *sdiodev)
-{
- return 0;
+ if (IS_ENABLED(CONFIG_PM_SLEEP))
+ atomic_dec(&sdiodev->freezer->thread_count);
}

-static void brcmf_sdiod_freezer_detach(struct brcmf_sdio_dev *sdiodev)
-{
-}
-#endif /* CONFIG_PM_SLEEP */
-
int brcmf_sdiod_remove(struct brcmf_sdio_dev *sdiodev)
{
sdiodev->state = BRCMF_SDIOD_DOWN;
@@ -1136,7 +1131,6 @@ void brcmf_sdio_wowl_config(struct device *dev, bool enabled)
brcmf_dbg(SDIO, "WOWL not supported\n");
}

-#ifdef CONFIG_PM_SLEEP
static int brcmf_ops_sdio_suspend(struct device *dev)
{
struct sdio_func *func;
@@ -1204,11 +1198,9 @@ static int brcmf_ops_sdio_resume(struct device *dev)
return ret;
}

-static const struct dev_pm_ops brcmf_sdio_pm_ops = {
- .suspend = brcmf_ops_sdio_suspend,
- .resume = brcmf_ops_sdio_resume,
-};
-#endif /* CONFIG_PM_SLEEP */
+static DEFINE_SIMPLE_DEV_PM_OPS(brcmf_sdio_pm_ops,
+ brcmf_ops_sdio_suspend,
+ brcmf_ops_sdio_resume);

static struct sdio_driver brcmf_sdmmc_driver = {
.probe = brcmf_ops_sdio_probe,
@@ -1217,9 +1209,7 @@ static struct sdio_driver brcmf_sdmmc_driver = {
.id_table = brcmf_sdmmc_ids,
.drv = {
.owner = THIS_MODULE,
-#ifdef CONFIG_PM_SLEEP
- .pm = &brcmf_sdio_pm_ops,
-#endif /* CONFIG_PM_SLEEP */
+ .pm = pm_sleep_ptr(&brcmf_sdio_pm_ops),
.coredump = brcmf_dev_coredump,
},
};
diff --git a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
index 212fbbe1cd7e..58ed6b70f8c5 100644
--- a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
+++ b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
@@ -4020,15 +4020,14 @@ brcmf_sdio_probe_attach(struct brcmf_sdio *bus)
*/
brcmf_sdiod_sgtable_alloc(sdiodev);

-#ifdef CONFIG_PM_SLEEP
/* wowl can be supported when KEEP_POWER is true and (WAKE_SDIO_IRQ
* is true or when platform data OOB irq is true).
*/
- if ((sdio_get_host_pm_caps(sdiodev->func1) & MMC_PM_KEEP_POWER) &&
+ if (IS_ENABLED(CONFIG_PM_SLEEP) &&
+ (sdio_get_host_pm_caps(sdiodev->func1) & MMC_PM_KEEP_POWER) &&
((sdio_get_host_pm_caps(sdiodev->func1) & MMC_PM_WAKE_SDIO_IRQ) ||
(sdiodev->settings->bus.sdio.oob_irq_supported)))
sdiodev->bus_if->wowl_supported = true;
-#endif

if (brcmf_sdio_kso_init(bus)) {
brcmf_err("error enabling KSO\n");
diff --git a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.h b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.h
index 15d2c02fa3ec..47351ff458ca 100644
--- a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.h
+++ b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.h
@@ -346,26 +346,10 @@ int brcmf_sdiod_abort(struct brcmf_sdio_dev *sdiodev, struct sdio_func *func);
void brcmf_sdiod_sgtable_alloc(struct brcmf_sdio_dev *sdiodev);
void brcmf_sdiod_change_state(struct brcmf_sdio_dev *sdiodev,
enum brcmf_sdiod_state state);
-#ifdef CONFIG_PM_SLEEP
bool brcmf_sdiod_freezing(struct brcmf_sdio_dev *sdiodev);
void brcmf_sdiod_try_freeze(struct brcmf_sdio_dev *sdiodev);
void brcmf_sdiod_freezer_count(struct brcmf_sdio_dev *sdiodev);
void brcmf_sdiod_freezer_uncount(struct brcmf_sdio_dev *sdiodev);
-#else
-static inline bool brcmf_sdiod_freezing(struct brcmf_sdio_dev *sdiodev)
-{
- return false;
-}
-static inline void brcmf_sdiod_try_freeze(struct brcmf_sdio_dev *sdiodev)
-{
-}
-static inline void brcmf_sdiod_freezer_count(struct brcmf_sdio_dev *sdiodev)
-{
-}
-static inline void brcmf_sdiod_freezer_uncount(struct brcmf_sdio_dev *sdiodev)
-{
-}
-#endif /* CONFIG_PM_SLEEP */

int brcmf_sdiod_probe(struct brcmf_sdio_dev *sdiodev);
int brcmf_sdiod_remove(struct brcmf_sdio_dev *sdiodev);
--
2.35.1


2022-06-24 09:24:40

by Arend Van Spriel

[permalink] [raw]
Subject: Re: [PATCH v2] brcmfmac: Remove #ifdef guards for PM related functions

On 6/23/2022 2:42 PM, Paul Cercueil wrote:
> Use the new DEFINE_SIMPLE_DEV_PM_OPS() and pm_sleep_ptr() macros to
> handle the .suspend/.resume callbacks.
>
> These macros allow the suspend and resume functions to be automatically
> dropped by the compiler when CONFIG_SUSPEND is disabled, without having
> to use #ifdef guards.
>
> Some other functions not directly called by the .suspend/.resume
> callbacks, but still related to PM were also taken outside #ifdef
> guards.
>
> The advantage is then that these functions are now always compiled
> independently of any Kconfig option, and thanks to that bugs and
> regressions are easier to catch.

Reviewed-by: Arend van Spriel <[email protected]>
> Signed-off-by: Paul Cercueil <[email protected]>
> ---
>
> Notes:
> v2:
> - Move checks for IS_ENABLED(CONFIG_PM_SLEEP) inside functions to keep
> the calling functions intact.
> - Reword the commit message to explain why this patch is useful.
>
> .../broadcom/brcm80211/brcmfmac/bcmsdh.c | 38 +++++++------------
> .../broadcom/brcm80211/brcmfmac/sdio.c | 5 +--
> .../broadcom/brcm80211/brcmfmac/sdio.h | 16 --------
> 3 files changed, 16 insertions(+), 43 deletions(-)
>
> diff --git a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/bcmsdh.c b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/bcmsdh.c
> index 9c598ea97499..11ad878a906b 100644
> --- a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/bcmsdh.c
> +++ b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/bcmsdh.c
> @@ -784,9 +784,11 @@ void brcmf_sdiod_sgtable_alloc(struct brcmf_sdio_dev *sdiodev)
> sdiodev->txglomsz = sdiodev->settings->bus.sdio.txglomsz;
> }
>
> -#ifdef CONFIG_PM_SLEEP
> static int brcmf_sdiod_freezer_attach(struct brcmf_sdio_dev *sdiodev)
> {
> + if (!IS_ENABLED(CONFIG_PM_SLEEP))
> + return 0;
> +
> sdiodev->freezer = kzalloc(sizeof(*sdiodev->freezer), GFP_KERNEL);
> if (!sdiodev->freezer)
> return -ENOMEM;
> @@ -799,7 +801,7 @@ static int brcmf_sdiod_freezer_attach(struct brcmf_sdio_dev *sdiodev)
>
> static void brcmf_sdiod_freezer_detach(struct brcmf_sdio_dev *sdiodev)
> {
> - if (sdiodev->freezer) {
> + if (IS_ENABLED(CONFIG_PM_SLEEP) && sdiodev->freezer) {

This change is not necessary. sdiodev->freezer will be NULL when
CONFIG_PM_SLEEP is not enabled.

> WARN_ON(atomic_read(&sdiodev->freezer->freezing));
> kfree(sdiodev->freezer);
> }

2022-06-24 16:32:55

by Jeff Johnson

[permalink] [raw]
Subject: Re: [PATCH v2] brcmfmac: Remove #ifdef guards for PM related functions

On 6/24/2022 2:24 AM, Arend Van Spriel wrote:
> On 6/23/2022 2:42 PM, Paul Cercueil wrote:

[snip]

>> -    if (sdiodev->freezer) {
>> +    if (IS_ENABLED(CONFIG_PM_SLEEP) && sdiodev->freezer) {
>
> This change is not necessary. sdiodev->freezer will be NULL when
> CONFIG_PM_SLEEP is not enabled.

but won't the compiler be able to completely optimize the code away if
the change is present?

2022-06-24 18:34:16

by Paul Cercueil

[permalink] [raw]
Subject: Re: [PATCH v2] brcmfmac: Remove #ifdef guards for PM related functions

Hi,

Le ven., juin 24 2022 at 09:31:22 -0700, Jeff Johnson
<[email protected]> a ?crit :
> On 6/24/2022 2:24 AM, Arend Van Spriel wrote:
>> On 6/23/2022 2:42 PM, Paul Cercueil wrote:
>
> [snip]
>
>>> - if (sdiodev->freezer) {
>>> + if (IS_ENABLED(CONFIG_PM_SLEEP) && sdiodev->freezer) {
>>
>> This change is not necessary. sdiodev->freezer will be NULL when
>> CONFIG_PM_SLEEP is not enabled.
>
> but won't the compiler be able to completely optimize the code away
> if the change is present?

That's correct. But do we want to complexify a bit the code for the
sake of saving a few bytes? I leave that as an open question to the
maintainer, I'm really fine with both options.

Cheers,
-Paul


2022-06-24 19:02:51

by Jeff Johnson

[permalink] [raw]
Subject: Re: [PATCH v2] brcmfmac: Remove #ifdef guards for PM related functions

On 6/24/2022 11:32 AM, Paul Cercueil wrote:
> Hi,
>
> Le ven., juin 24 2022 at 09:31:22 -0700, Jeff Johnson
> <[email protected]> a écrit :
>> On 6/24/2022 2:24 AM, Arend Van Spriel wrote:
>>> On 6/23/2022 2:42 PM, Paul Cercueil wrote:
>>
>> [snip]
>>
>>>> -    if (sdiodev->freezer) {
>>>> +    if (IS_ENABLED(CONFIG_PM_SLEEP) && sdiodev->freezer) {
>>>
>>> This change is not necessary. sdiodev->freezer will be NULL when
>>> CONFIG_PM_SLEEP is not enabled.
>>
>> but won't the compiler be able to completely optimize the code away if
>> the change is present?
>
> That's correct. But do we want to complexify a bit the code for the sake
> of saving a few bytes? I leave that as an open question to the
> maintainer, I'm really fine with both options.

I'm fine either way as well. I'd have stronger opinions if ath drivers
were involved ;)