* [PATCH] gpio: ljca: reduce struct allocation @ 2026-03-08 2:12 Rosen Penev 2026-03-08 12:19 ` Sakari Ailus 2026-03-09 9:22 ` Bartosz Golaszewski 0 siblings, 2 replies; 8+ messages in thread From: Rosen Penev @ 2026-03-08 2:12 UTC (permalink / raw) To: linux-gpio Cc: Lixu Zhang, linux-hardening, gustavoars, Sakari Ailus, Linus Walleij, Bartosz Golaszewski, open list Convert connect_mode to a flexible array member to avoid calling kcalloc and to combine the allocations. Signed-off-by: Rosen Penev <rosenp@gmail.com> --- drivers/gpio/gpio-ljca.c | 16 +++++++--------- 1 file changed, 7 insertions(+), 9 deletions(-) diff --git a/drivers/gpio/gpio-ljca.c b/drivers/gpio/gpio-ljca.c index f32d1d237795..a531aaa0425f 100644 --- a/drivers/gpio/gpio-ljca.c +++ b/drivers/gpio/gpio-ljca.c @@ -63,7 +63,6 @@ struct ljca_gpio_dev { DECLARE_BITMAP(enabled_irqs, LJCA_MAX_GPIO_NUM); DECLARE_BITMAP(reenable_irqs, LJCA_MAX_GPIO_NUM); DECLARE_BITMAP(output_enabled, LJCA_MAX_GPIO_NUM); - u8 *connect_mode; /* protect irq bus */ struct mutex irq_lock; struct work_struct work; @@ -72,6 +71,8 @@ struct ljca_gpio_dev { u8 obuf[LJCA_GPIO_BUF_SIZE]; u8 ibuf[LJCA_GPIO_BUF_SIZE]; + + u8 connect_mode[]; }; static int ljca_gpio_config(struct ljca_gpio_dev *ljca_gpio, u8 gpio_id, @@ -400,22 +401,19 @@ static int ljca_gpio_probe(struct auxiliary_device *auxdev, const struct auxiliary_device_id *aux_dev_id) { struct ljca_client *ljca = auxiliary_dev_to_ljca_client(auxdev); + struct ljca_gpio_info *gpio_info; struct ljca_gpio_dev *ljca_gpio; struct gpio_irq_chip *girq; int ret; - ljca_gpio = devm_kzalloc(&auxdev->dev, sizeof(*ljca_gpio), GFP_KERNEL); + gpio_info = dev_get_platdata(&auxdev->dev); + ljca_gpio = devm_kzalloc(&auxdev->dev, struct_size(ljca_gpio, connect_mode, gpio_info->num), + GFP_KERNEL); if (!ljca_gpio) return -ENOMEM; ljca_gpio->ljca = ljca; - ljca_gpio->gpio_info = dev_get_platdata(&auxdev->dev); - ljca_gpio->connect_mode = devm_kcalloc(&auxdev->dev, - ljca_gpio->gpio_info->num, - sizeof(*ljca_gpio->connect_mode), - GFP_KERNEL); - if (!ljca_gpio->connect_mode) - return -ENOMEM; + ljca_gpio->gpio_info = gpio_info; ret = devm_mutex_init(&auxdev->dev, &ljca_gpio->irq_lock); if (ret) -- 2.53.0 ^ permalink raw reply related [flat|nested] 8+ messages in thread
* Re: [PATCH] gpio: ljca: reduce struct allocation 2026-03-08 2:12 [PATCH] gpio: ljca: reduce struct allocation Rosen Penev @ 2026-03-08 12:19 ` Sakari Ailus 2026-03-08 18:54 ` Rosen Penev 2026-03-09 9:22 ` Bartosz Golaszewski 1 sibling, 1 reply; 8+ messages in thread From: Sakari Ailus @ 2026-03-08 12:19 UTC (permalink / raw) To: Rosen Penev Cc: linux-gpio, Lixu Zhang, linux-hardening, gustavoars, Linus Walleij, Bartosz Golaszewski, open list Hi Rosen, Thanks or the patch. On Sat, Mar 07, 2026 at 06:12:01PM -0800, Rosen Penev wrote: > Convert connect_mode to a flexible array member to avoid calling > kcalloc and to combine the allocations. > > Signed-off-by: Rosen Penev <rosenp@gmail.com> > --- > drivers/gpio/gpio-ljca.c | 16 +++++++--------- > 1 file changed, 7 insertions(+), 9 deletions(-) > > diff --git a/drivers/gpio/gpio-ljca.c b/drivers/gpio/gpio-ljca.c > index f32d1d237795..a531aaa0425f 100644 > --- a/drivers/gpio/gpio-ljca.c > +++ b/drivers/gpio/gpio-ljca.c > @@ -63,7 +63,6 @@ struct ljca_gpio_dev { > DECLARE_BITMAP(enabled_irqs, LJCA_MAX_GPIO_NUM); > DECLARE_BITMAP(reenable_irqs, LJCA_MAX_GPIO_NUM); > DECLARE_BITMAP(output_enabled, LJCA_MAX_GPIO_NUM); > - u8 *connect_mode; > /* protect irq bus */ > struct mutex irq_lock; > struct work_struct work; > @@ -72,6 +71,8 @@ struct ljca_gpio_dev { > > u8 obuf[LJCA_GPIO_BUF_SIZE]; > u8 ibuf[LJCA_GPIO_BUF_SIZE]; > + > + u8 connect_mode[]; > }; > > static int ljca_gpio_config(struct ljca_gpio_dev *ljca_gpio, u8 gpio_id, > @@ -400,22 +401,19 @@ static int ljca_gpio_probe(struct auxiliary_device *auxdev, > const struct auxiliary_device_id *aux_dev_id) > { > struct ljca_client *ljca = auxiliary_dev_to_ljca_client(auxdev); > + struct ljca_gpio_info *gpio_info; > struct ljca_gpio_dev *ljca_gpio; > struct gpio_irq_chip *girq; > int ret; > > - ljca_gpio = devm_kzalloc(&auxdev->dev, sizeof(*ljca_gpio), GFP_KERNEL); > + gpio_info = dev_get_platdata(&auxdev->dev); You can do this assignment in variable declaration. > + ljca_gpio = devm_kzalloc(&auxdev->dev, struct_size(ljca_gpio, connect_mode, gpio_info->num), This line is a bit long, I'd wrap it. > + GFP_KERNEL); > if (!ljca_gpio) > return -ENOMEM; > > ljca_gpio->ljca = ljca; > - ljca_gpio->gpio_info = dev_get_platdata(&auxdev->dev); > - ljca_gpio->connect_mode = devm_kcalloc(&auxdev->dev, > - ljca_gpio->gpio_info->num, > - sizeof(*ljca_gpio->connect_mode), > - GFP_KERNEL); > - if (!ljca_gpio->connect_mode) > - return -ENOMEM; > + ljca_gpio->gpio_info = gpio_info; > > ret = devm_mutex_init(&auxdev->dev, &ljca_gpio->irq_lock); > if (ret) -- Kind regards, Sakari Ailus ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] gpio: ljca: reduce struct allocation 2026-03-08 12:19 ` Sakari Ailus @ 2026-03-08 18:54 ` Rosen Penev 2026-03-09 7:19 ` Sakari Ailus 0 siblings, 1 reply; 8+ messages in thread From: Rosen Penev @ 2026-03-08 18:54 UTC (permalink / raw) To: Sakari Ailus Cc: linux-gpio, Lixu Zhang, linux-hardening, gustavoars, Linus Walleij, Bartosz Golaszewski, open list On Sun, Mar 8, 2026 at 5:19 AM Sakari Ailus <sakari.ailus@linux.intel.com> wrote: > > Hi Rosen, > > Thanks or the patch. > > On Sat, Mar 07, 2026 at 06:12:01PM -0800, Rosen Penev wrote: > > Convert connect_mode to a flexible array member to avoid calling > > kcalloc and to combine the allocations. > > > > Signed-off-by: Rosen Penev <rosenp@gmail.com> > > --- > > drivers/gpio/gpio-ljca.c | 16 +++++++--------- > > 1 file changed, 7 insertions(+), 9 deletions(-) > > > > diff --git a/drivers/gpio/gpio-ljca.c b/drivers/gpio/gpio-ljca.c > > index f32d1d237795..a531aaa0425f 100644 > > --- a/drivers/gpio/gpio-ljca.c > > +++ b/drivers/gpio/gpio-ljca.c > > @@ -63,7 +63,6 @@ struct ljca_gpio_dev { > > DECLARE_BITMAP(enabled_irqs, LJCA_MAX_GPIO_NUM); > > DECLARE_BITMAP(reenable_irqs, LJCA_MAX_GPIO_NUM); > > DECLARE_BITMAP(output_enabled, LJCA_MAX_GPIO_NUM); > > - u8 *connect_mode; > > /* protect irq bus */ > > struct mutex irq_lock; > > struct work_struct work; > > @@ -72,6 +71,8 @@ struct ljca_gpio_dev { > > > > u8 obuf[LJCA_GPIO_BUF_SIZE]; > > u8 ibuf[LJCA_GPIO_BUF_SIZE]; > > + > > + u8 connect_mode[]; > > }; > > > > static int ljca_gpio_config(struct ljca_gpio_dev *ljca_gpio, u8 gpio_id, > > @@ -400,22 +401,19 @@ static int ljca_gpio_probe(struct auxiliary_device *auxdev, > > const struct auxiliary_device_id *aux_dev_id) > > { > > struct ljca_client *ljca = auxiliary_dev_to_ljca_client(auxdev); > > + struct ljca_gpio_info *gpio_info; > > struct ljca_gpio_dev *ljca_gpio; > > struct gpio_irq_chip *girq; > > int ret; > > > > - ljca_gpio = devm_kzalloc(&auxdev->dev, sizeof(*ljca_gpio), GFP_KERNEL); > > + gpio_info = dev_get_platdata(&auxdev->dev); > > You can do this assignment in variable declaration. I avoided doing so to keep reverse christmas tree order. I can still do so if desired. > > > + ljca_gpio = devm_kzalloc(&auxdev->dev, struct_size(ljca_gpio, connect_mode, gpio_info->num), > > This line is a bit long, I'd wrap it. Will fix. > > > + GFP_KERNEL); > > if (!ljca_gpio) > > return -ENOMEM; > > > > ljca_gpio->ljca = ljca; > > - ljca_gpio->gpio_info = dev_get_platdata(&auxdev->dev); > > - ljca_gpio->connect_mode = devm_kcalloc(&auxdev->dev, > > - ljca_gpio->gpio_info->num, > > - sizeof(*ljca_gpio->connect_mode), > > - GFP_KERNEL); > > - if (!ljca_gpio->connect_mode) > > - return -ENOMEM; > > + ljca_gpio->gpio_info = gpio_info; > > > > ret = devm_mutex_init(&auxdev->dev, &ljca_gpio->irq_lock); > > if (ret) > > -- > Kind regards, > > Sakari Ailus ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] gpio: ljca: reduce struct allocation 2026-03-08 18:54 ` Rosen Penev @ 2026-03-09 7:19 ` Sakari Ailus 2026-03-09 7:32 ` Rosen Penev 0 siblings, 1 reply; 8+ messages in thread From: Sakari Ailus @ 2026-03-09 7:19 UTC (permalink / raw) To: Rosen Penev Cc: linux-gpio, Lixu Zhang, linux-hardening, gustavoars, Linus Walleij, Bartosz Golaszewski, open list Hi Rosen, On Sun, Mar 08, 2026 at 11:54:07AM -0700, Rosen Penev wrote: > On Sun, Mar 8, 2026 at 5:19 AM Sakari Ailus > <sakari.ailus@linux.intel.com> wrote: > > > > Hi Rosen, > > > > Thanks or the patch. > > > > On Sat, Mar 07, 2026 at 06:12:01PM -0800, Rosen Penev wrote: > > > Convert connect_mode to a flexible array member to avoid calling > > > kcalloc and to combine the allocations. > > > > > > Signed-off-by: Rosen Penev <rosenp@gmail.com> > > > --- > > > drivers/gpio/gpio-ljca.c | 16 +++++++--------- > > > 1 file changed, 7 insertions(+), 9 deletions(-) > > > > > > diff --git a/drivers/gpio/gpio-ljca.c b/drivers/gpio/gpio-ljca.c > > > index f32d1d237795..a531aaa0425f 100644 > > > --- a/drivers/gpio/gpio-ljca.c > > > +++ b/drivers/gpio/gpio-ljca.c > > > @@ -63,7 +63,6 @@ struct ljca_gpio_dev { > > > DECLARE_BITMAP(enabled_irqs, LJCA_MAX_GPIO_NUM); > > > DECLARE_BITMAP(reenable_irqs, LJCA_MAX_GPIO_NUM); > > > DECLARE_BITMAP(output_enabled, LJCA_MAX_GPIO_NUM); > > > - u8 *connect_mode; > > > /* protect irq bus */ > > > struct mutex irq_lock; > > > struct work_struct work; > > > @@ -72,6 +71,8 @@ struct ljca_gpio_dev { > > > > > > u8 obuf[LJCA_GPIO_BUF_SIZE]; > > > u8 ibuf[LJCA_GPIO_BUF_SIZE]; > > > + > > > + u8 connect_mode[]; > > > }; > > > > > > static int ljca_gpio_config(struct ljca_gpio_dev *ljca_gpio, u8 gpio_id, > > > @@ -400,22 +401,19 @@ static int ljca_gpio_probe(struct auxiliary_device *auxdev, > > > const struct auxiliary_device_id *aux_dev_id) > > > { > > > struct ljca_client *ljca = auxiliary_dev_to_ljca_client(auxdev); > > > + struct ljca_gpio_info *gpio_info; > > > struct ljca_gpio_dev *ljca_gpio; > > > struct gpio_irq_chip *girq; > > > int ret; > > > > > > - ljca_gpio = devm_kzalloc(&auxdev->dev, sizeof(*ljca_gpio), GFP_KERNEL); > > > + gpio_info = dev_get_platdata(&auxdev->dev); > > > > You can do this assignment in variable declaration. > I avoided doing so to keep reverse christmas tree order. I can still > do so if desired. How does that affect it? There doesn't seem to be a dependency here -- and even if there was, the dependency obviously has priority. > > > > > + ljca_gpio = devm_kzalloc(&auxdev->dev, struct_size(ljca_gpio, connect_mode, gpio_info->num), > > > > This line is a bit long, I'd wrap it. > Will fix. > > > > > + GFP_KERNEL); > > > if (!ljca_gpio) > > > return -ENOMEM; > > > > > > ljca_gpio->ljca = ljca; > > > - ljca_gpio->gpio_info = dev_get_platdata(&auxdev->dev); > > > - ljca_gpio->connect_mode = devm_kcalloc(&auxdev->dev, > > > - ljca_gpio->gpio_info->num, > > > - sizeof(*ljca_gpio->connect_mode), > > > - GFP_KERNEL); > > > - if (!ljca_gpio->connect_mode) > > > - return -ENOMEM; > > > + ljca_gpio->gpio_info = gpio_info; > > > > > > ret = devm_mutex_init(&auxdev->dev, &ljca_gpio->irq_lock); > > > if (ret) > > -- Kind regards, Sakari Ailus ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] gpio: ljca: reduce struct allocation 2026-03-09 7:19 ` Sakari Ailus @ 2026-03-09 7:32 ` Rosen Penev 0 siblings, 0 replies; 8+ messages in thread From: Rosen Penev @ 2026-03-09 7:32 UTC (permalink / raw) To: Sakari Ailus Cc: linux-gpio, Lixu Zhang, linux-hardening, gustavoars, Linus Walleij, Bartosz Golaszewski, open list On Mon, Mar 9, 2026 at 12:20 AM Sakari Ailus <sakari.ailus@linux.intel.com> wrote: > > Hi Rosen, > > On Sun, Mar 08, 2026 at 11:54:07AM -0700, Rosen Penev wrote: > > On Sun, Mar 8, 2026 at 5:19 AM Sakari Ailus > > <sakari.ailus@linux.intel.com> wrote: > > > > > > Hi Rosen, > > > > > > Thanks or the patch. > > > > > > On Sat, Mar 07, 2026 at 06:12:01PM -0800, Rosen Penev wrote: > > > > Convert connect_mode to a flexible array member to avoid calling > > > > kcalloc and to combine the allocations. > > > > > > > > Signed-off-by: Rosen Penev <rosenp@gmail.com> > > > > --- > > > > drivers/gpio/gpio-ljca.c | 16 +++++++--------- > > > > 1 file changed, 7 insertions(+), 9 deletions(-) > > > > > > > > diff --git a/drivers/gpio/gpio-ljca.c b/drivers/gpio/gpio-ljca.c > > > > index f32d1d237795..a531aaa0425f 100644 > > > > --- a/drivers/gpio/gpio-ljca.c > > > > +++ b/drivers/gpio/gpio-ljca.c > > > > @@ -63,7 +63,6 @@ struct ljca_gpio_dev { > > > > DECLARE_BITMAP(enabled_irqs, LJCA_MAX_GPIO_NUM); > > > > DECLARE_BITMAP(reenable_irqs, LJCA_MAX_GPIO_NUM); > > > > DECLARE_BITMAP(output_enabled, LJCA_MAX_GPIO_NUM); > > > > - u8 *connect_mode; > > > > /* protect irq bus */ > > > > struct mutex irq_lock; > > > > struct work_struct work; > > > > @@ -72,6 +71,8 @@ struct ljca_gpio_dev { > > > > > > > > u8 obuf[LJCA_GPIO_BUF_SIZE]; > > > > u8 ibuf[LJCA_GPIO_BUF_SIZE]; > > > > + > > > > + u8 connect_mode[]; > > > > }; > > > > > > > > static int ljca_gpio_config(struct ljca_gpio_dev *ljca_gpio, u8 gpio_id, > > > > @@ -400,22 +401,19 @@ static int ljca_gpio_probe(struct auxiliary_device *auxdev, > > > > const struct auxiliary_device_id *aux_dev_id) > > > > { > > > > struct ljca_client *ljca = auxiliary_dev_to_ljca_client(auxdev); > > > > + struct ljca_gpio_info *gpio_info; > > > > struct ljca_gpio_dev *ljca_gpio; > > > > struct gpio_irq_chip *girq; > > > > int ret; > > > > > > > > - ljca_gpio = devm_kzalloc(&auxdev->dev, sizeof(*ljca_gpio), GFP_KERNEL); > > > > + gpio_info = dev_get_platdata(&auxdev->dev); > > > > > > You can do this assignment in variable declaration. > > I avoided doing so to keep reverse christmas tree order. I can still > > do so if desired. > > How does that affect it? There doesn't seem to be a dependency here -- and > even if there was, the dependency obviously has priority. It's just cosmetic. > > > > > > > > + ljca_gpio = devm_kzalloc(&auxdev->dev, struct_size(ljca_gpio, connect_mode, gpio_info->num), > > > > > > This line is a bit long, I'd wrap it. > > Will fix. > > > > > > > + GFP_KERNEL); > > > > if (!ljca_gpio) > > > > return -ENOMEM; > > > > > > > > ljca_gpio->ljca = ljca; > > > > - ljca_gpio->gpio_info = dev_get_platdata(&auxdev->dev); > > > > - ljca_gpio->connect_mode = devm_kcalloc(&auxdev->dev, > > > > - ljca_gpio->gpio_info->num, > > > > - sizeof(*ljca_gpio->connect_mode), > > > > - GFP_KERNEL); > > > > - if (!ljca_gpio->connect_mode) > > > > - return -ENOMEM; > > > > + ljca_gpio->gpio_info = gpio_info; > > > > > > > > ret = devm_mutex_init(&auxdev->dev, &ljca_gpio->irq_lock); > > > > if (ret) > > > > > -- > Kind regards, > > Sakari Ailus ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] gpio: ljca: reduce struct allocation 2026-03-08 2:12 [PATCH] gpio: ljca: reduce struct allocation Rosen Penev 2026-03-08 12:19 ` Sakari Ailus @ 2026-03-09 9:22 ` Bartosz Golaszewski 2026-03-09 21:48 ` Rosen Penev 1 sibling, 1 reply; 8+ messages in thread From: Bartosz Golaszewski @ 2026-03-09 9:22 UTC (permalink / raw) To: Rosen Penev Cc: linux-gpio, Lixu Zhang, linux-hardening, gustavoars, Sakari Ailus, Linus Walleij, open list On Sun, Mar 8, 2026 at 3:12 AM Rosen Penev <rosenp@gmail.com> wrote: > > Convert connect_mode to a flexible array member to avoid calling > kcalloc and to combine the allocations. > > Signed-off-by: Rosen Penev <rosenp@gmail.com> > --- > drivers/gpio/gpio-ljca.c | 16 +++++++--------- > 1 file changed, 7 insertions(+), 9 deletions(-) > > diff --git a/drivers/gpio/gpio-ljca.c b/drivers/gpio/gpio-ljca.c > index f32d1d237795..a531aaa0425f 100644 > --- a/drivers/gpio/gpio-ljca.c > +++ b/drivers/gpio/gpio-ljca.c > @@ -63,7 +63,6 @@ struct ljca_gpio_dev { > DECLARE_BITMAP(enabled_irqs, LJCA_MAX_GPIO_NUM); > DECLARE_BITMAP(reenable_irqs, LJCA_MAX_GPIO_NUM); > DECLARE_BITMAP(output_enabled, LJCA_MAX_GPIO_NUM); > - u8 *connect_mode; > /* protect irq bus */ > struct mutex irq_lock; > struct work_struct work; > @@ -72,6 +71,8 @@ struct ljca_gpio_dev { > > u8 obuf[LJCA_GPIO_BUF_SIZE]; > u8 ibuf[LJCA_GPIO_BUF_SIZE]; > + > + u8 connect_mode[]; Please use __counted_by(). > }; > > static int ljca_gpio_config(struct ljca_gpio_dev *ljca_gpio, u8 gpio_id, > @@ -400,22 +401,19 @@ static int ljca_gpio_probe(struct auxiliary_device *auxdev, > const struct auxiliary_device_id *aux_dev_id) > { > struct ljca_client *ljca = auxiliary_dev_to_ljca_client(auxdev); > + struct ljca_gpio_info *gpio_info; > struct ljca_gpio_dev *ljca_gpio; > struct gpio_irq_chip *girq; > int ret; > > - ljca_gpio = devm_kzalloc(&auxdev->dev, sizeof(*ljca_gpio), GFP_KERNEL); > + gpio_info = dev_get_platdata(&auxdev->dev); > + ljca_gpio = devm_kzalloc(&auxdev->dev, struct_size(ljca_gpio, connect_mode, gpio_info->num), I'm surprised to learn that there's no devm_kzalloc_flex() yet. I wanted to ask you to use it but it doesn't exist. Bart > + GFP_KERNEL); > if (!ljca_gpio) > return -ENOMEM; > > ljca_gpio->ljca = ljca; > - ljca_gpio->gpio_info = dev_get_platdata(&auxdev->dev); > - ljca_gpio->connect_mode = devm_kcalloc(&auxdev->dev, > - ljca_gpio->gpio_info->num, > - sizeof(*ljca_gpio->connect_mode), > - GFP_KERNEL); > - if (!ljca_gpio->connect_mode) > - return -ENOMEM; > + ljca_gpio->gpio_info = gpio_info; > > ret = devm_mutex_init(&auxdev->dev, &ljca_gpio->irq_lock); > if (ret) > -- > 2.53.0 > ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] gpio: ljca: reduce struct allocation 2026-03-09 9:22 ` Bartosz Golaszewski @ 2026-03-09 21:48 ` Rosen Penev 2026-03-10 11:52 ` Linus Walleij 0 siblings, 1 reply; 8+ messages in thread From: Rosen Penev @ 2026-03-09 21:48 UTC (permalink / raw) To: Bartosz Golaszewski Cc: linux-gpio, Lixu Zhang, linux-hardening, gustavoars, Sakari Ailus, Linus Walleij, open list On Mon, Mar 9, 2026 at 2:22 AM Bartosz Golaszewski <brgl@kernel.org> wrote: > > On Sun, Mar 8, 2026 at 3:12 AM Rosen Penev <rosenp@gmail.com> wrote: > > > > Convert connect_mode to a flexible array member to avoid calling > > kcalloc and to combine the allocations. > > > > Signed-off-by: Rosen Penev <rosenp@gmail.com> > > --- > > drivers/gpio/gpio-ljca.c | 16 +++++++--------- > > 1 file changed, 7 insertions(+), 9 deletions(-) > > > > diff --git a/drivers/gpio/gpio-ljca.c b/drivers/gpio/gpio-ljca.c > > index f32d1d237795..a531aaa0425f 100644 > > --- a/drivers/gpio/gpio-ljca.c > > +++ b/drivers/gpio/gpio-ljca.c > > @@ -63,7 +63,6 @@ struct ljca_gpio_dev { > > DECLARE_BITMAP(enabled_irqs, LJCA_MAX_GPIO_NUM); > > DECLARE_BITMAP(reenable_irqs, LJCA_MAX_GPIO_NUM); > > DECLARE_BITMAP(output_enabled, LJCA_MAX_GPIO_NUM); > > - u8 *connect_mode; > > /* protect irq bus */ > > struct mutex irq_lock; > > struct work_struct work; > > @@ -72,6 +71,8 @@ struct ljca_gpio_dev { > > > > u8 obuf[LJCA_GPIO_BUF_SIZE]; > > u8 ibuf[LJCA_GPIO_BUF_SIZE]; > > + > > + u8 connect_mode[]; > > Please use __counted_by(). Counted by what? __counted_by(gpio_info->num); does not work. > > }; > > > > static int ljca_gpio_config(struct ljca_gpio_dev *ljca_gpio, u8 gpio_id, > > @@ -400,22 +401,19 @@ static int ljca_gpio_probe(struct auxiliary_device *auxdev, > > const struct auxiliary_device_id *aux_dev_id) > > { > > struct ljca_client *ljca = auxiliary_dev_to_ljca_client(auxdev); > > + struct ljca_gpio_info *gpio_info; > > struct ljca_gpio_dev *ljca_gpio; > > struct gpio_irq_chip *girq; > > int ret; > > > > - ljca_gpio = devm_kzalloc(&auxdev->dev, sizeof(*ljca_gpio), GFP_KERNEL); > > + gpio_info = dev_get_platdata(&auxdev->dev); > > + ljca_gpio = devm_kzalloc(&auxdev->dev, struct_size(ljca_gpio, connect_mode, gpio_info->num), > > I'm surprised to learn that there's no devm_kzalloc_flex() yet. I > wanted to ask you to use it but it doesn't exist. I'm sure a treewide commit will be applied when it gets introduced. > > Bart > > > + GFP_KERNEL); > > if (!ljca_gpio) > > return -ENOMEM; > > > > ljca_gpio->ljca = ljca; > > - ljca_gpio->gpio_info = dev_get_platdata(&auxdev->dev); > > - ljca_gpio->connect_mode = devm_kcalloc(&auxdev->dev, > > - ljca_gpio->gpio_info->num, > > - sizeof(*ljca_gpio->connect_mode), > > - GFP_KERNEL); > > - if (!ljca_gpio->connect_mode) > > - return -ENOMEM; > > + ljca_gpio->gpio_info = gpio_info; > > > > ret = devm_mutex_init(&auxdev->dev, &ljca_gpio->irq_lock); > > if (ret) > > -- > > 2.53.0 > > ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] gpio: ljca: reduce struct allocation 2026-03-09 21:48 ` Rosen Penev @ 2026-03-10 11:52 ` Linus Walleij 0 siblings, 0 replies; 8+ messages in thread From: Linus Walleij @ 2026-03-10 11:52 UTC (permalink / raw) To: Rosen Penev Cc: Bartosz Golaszewski, linux-gpio, Lixu Zhang, linux-hardening, gustavoars, Sakari Ailus, open list On Mon, Mar 9, 2026 at 10:48 PM Rosen Penev <rosenp@gmail.com> wrote: > > > + u8 connect_mode[]; > > > > Please use __counted_by(). > Counted by what? > > __counted_by(gpio_info->num); does not work. As with the other patch (IIRC) just add a new member to the struct to hold the count. Yours, Linus Walleij ^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~2026-03-10 11:52 UTC | newest] Thread overview: 8+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2026-03-08 2:12 [PATCH] gpio: ljca: reduce struct allocation Rosen Penev 2026-03-08 12:19 ` Sakari Ailus 2026-03-08 18:54 ` Rosen Penev 2026-03-09 7:19 ` Sakari Ailus 2026-03-09 7:32 ` Rosen Penev 2026-03-09 9:22 ` Bartosz Golaszewski 2026-03-09 21:48 ` Rosen Penev 2026-03-10 11:52 ` Linus Walleij
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox