| Message ID | 20250201093126.7322-1-biju.das.jz@bp.renesas.com (mailing list archive) |
|---|---|
| State | New |
| Delegated to: | Geert Uytterhoeven |
| Headers | show |
| Series | [v2] of: base: Add of_get_available_child_by_name() | expand |
On Sat, Feb 1, 2025 at 3:31 AM Biju Das <biju.das.jz@bp.renesas.com> wrote: > > There are lot of drivers using of_get_child_by_name() followed by > of_device_is_available() to find the available child node by name for a > given parent. Provide a helper for these users to simplify the code. > > Suggested-by: Geert Uytterhoeven <geert+renesas@glider.be> > Signed-off-by: Biju Das <biju.das.jz@bp.renesas.com> Reviewed-by: Rob Herring <robh@kernel.org> > --- > v1->v2: > * Updated commit description. > * Updated kerneldoc comment block > * Avoided code duplication by using of_get_child_by_name(). > > Note: > grep showed the below files will be the users for this new API. > I will be updating these drivers once this patch is in mainline. No need to wait. Please convert all the net ones and send this patch with them. > > drivers/net/dsa/rzn1_a5psw.c > drivers/net/can/rcar/rcar_canfd.c > > drivers/net/ethernet/mediatek/mtk_star_emac.c > drivers/net/dsa/mt7530.c > drivers/net/dsa/sja1105/sja1105_mdio.c > drivers/net/dsa/qca/qca8k-8xxx.c > drivers/net/wireless/mediatek/mt76/mac80211.c > drivers/net/ethernet/ibm/emac/core.c > drivers/net/ethernet/ti/am65-cpsw-nuss.c > drivers/net/ethernet/actions/owl-emac.c > drivers/net/ethernet/mediatek/mtk_eth_soc.c > drivers/media/platform/samsung/exynos4-is/media-dev.h > drivers/gpu/drm/tegra/rgb.c > drivers/gpu/drm/msm/adreno/adreno_gpu.c > drivers/clk/davinci/pll.c
Hi Rob, +Cc relevant subsystems. > -----Original Message----- > From: Rob Herring <robh@kernel.org> > Sent: 03 February 2025 16:53 > Subject: Re: [PATCH v2] of: base: Add of_get_available_child_by_name() > > On Sat, Feb 1, 2025 at 3:31 AM Biju Das <biju.das.jz@bp.renesas.com> wrote: > > > > There are lot of drivers using of_get_child_by_name() followed by > > of_device_is_available() to find the available child node by name for > > a given parent. Provide a helper for these users to simplify the code. > > > > Suggested-by: Geert Uytterhoeven <geert+renesas@glider.be> > > Signed-off-by: Biju Das <biju.das.jz@bp.renesas.com> > > Reviewed-by: Rob Herring <robh@kernel.org> > > > --- > > v1->v2: > > * Updated commit description. > > * Updated kerneldoc comment block > > * Avoided code duplication by using of_get_child_by_name(). > > > > Note: > > grep showed the below files will be the users for this new API. > > I will be updating these drivers once this patch is in mainline. > > No need to wait. Please convert all the net ones and send this patch with them. Thanks for the feedback. Subsequently, I have send the patches. However, Andrew[1]/Krystoff[2] mentioned me to wait till this patch appear in -rc , Can it be fast tracked to 6.14-rcX?? Otherwise, it needs to wait till 6.15-rc1 and other patches will then appear on 6.16-rc1. [1] https://lore.kernel.org/all/96fbccd3-fd79-4b2f-8f41-bd0e3fdb2c69@lunn.ch/ [2] https://lore.kernel.org/all/7fe9dad9-85e2-4cf0-98bc-cca20ff62df5@kernel.org/ Cheers, Biju
On Mon, Feb 3, 2025 at 11:17 AM Biju Das <biju.das.jz@bp.renesas.com> wrote: > > Hi Rob, > > +Cc relevant subsystems. > > > -----Original Message----- > > From: Rob Herring <robh@kernel.org> > > Sent: 03 February 2025 16:53 > > Subject: Re: [PATCH v2] of: base: Add of_get_available_child_by_name() > > > > On Sat, Feb 1, 2025 at 3:31 AM Biju Das <biju.das.jz@bp.renesas.com> wrote: > > > > > > There are lot of drivers using of_get_child_by_name() followed by > > > of_device_is_available() to find the available child node by name for > > > a given parent. Provide a helper for these users to simplify the code. > > > > > > Suggested-by: Geert Uytterhoeven <geert+renesas@glider.be> > > > Signed-off-by: Biju Das <biju.das.jz@bp.renesas.com> > > > > Reviewed-by: Rob Herring <robh@kernel.org> > > > > > --- > > > v1->v2: > > > * Updated commit description. > > > * Updated kerneldoc comment block > > > * Avoided code duplication by using of_get_child_by_name(). > > > > > > Note: > > > grep showed the below files will be the users for this new API. > > > I will be updating these drivers once this patch is in mainline. > > > > No need to wait. Please convert all the net ones and send this patch with them. > > Thanks for the feedback. > > Subsequently, I have send the patches. However, Andrew[1]/Krystoff[2] > mentioned me to wait till this patch appear in -rc , > > Can it be fast tracked to 6.14-rcX?? Otherwise, it needs to wait till 6.15-rc1 > and other patches will then appear on 6.16-rc1. Most maintainer trees are based on rc1. So are you sure everyone is going to be fine with a rc2 dependency? Generally, new APIs don't go in without a user. That being said, if this was 10s of different trees I'd reconsider, but since most of the callers are in net, I'm less willing to apply "not a fix" to fixes. > [1] https://lore.kernel.org/all/96fbccd3-fd79-4b2f-8f41-bd0e3fdb2c69@lunn.ch/ > > [2] https://lore.kernel.org/all/7fe9dad9-85e2-4cf0-98bc-cca20ff62df5@kernel.org/ It's not like they are saying to do the opposite of what I said. If the dependency is not part of your series, then it needs to be in rc1. Rob
Hi Rob, > -----Original Message----- > From: Rob Herring <robh@kernel.org> > Sent: 03 February 2025 21:06 > Subject: Re: [PATCH v2] of: base: Add of_get_available_child_by_name() > > On Mon, Feb 3, 2025 at 11:17 AM Biju Das <biju.das.jz@bp.renesas.com> wrote: > > > > Hi Rob, > > > > +Cc relevant subsystems. > > > > > -----Original Message----- > > > From: Rob Herring <robh@kernel.org> > > > Sent: 03 February 2025 16:53 > > > Subject: Re: [PATCH v2] of: base: Add > > > of_get_available_child_by_name() > > > > > > On Sat, Feb 1, 2025 at 3:31 AM Biju Das <biju.das.jz@bp.renesas.com> wrote: > > > > > > > > There are lot of drivers using of_get_child_by_name() followed by > > > > of_device_is_available() to find the available child node by name > > > > for a given parent. Provide a helper for these users to simplify the code. > > > > > > > > Suggested-by: Geert Uytterhoeven <geert+renesas@glider.be> > > > > Signed-off-by: Biju Das <biju.das.jz@bp.renesas.com> > > > > > > Reviewed-by: Rob Herring <robh@kernel.org> > > > > > > > --- > > > > v1->v2: > > > > * Updated commit description. > > > > * Updated kerneldoc comment block > > > > * Avoided code duplication by using of_get_child_by_name(). > > > > > > > > Note: > > > > grep showed the below files will be the users for this new API. > > > > I will be updating these drivers once this patch is in mainline. > > > > > > No need to wait. Please convert all the net ones and send this patch with them. > > > > Thanks for the feedback. > > > > Subsequently, I have send the patches. However, Andrew[1]/Krystoff[2] > > mentioned me to wait till this patch appear in -rc , > > > > Can it be fast tracked to 6.14-rcX?? Otherwise, it needs to wait till > > 6.15-rc1 and other patches will then appear on 6.16-rc1. > > Most maintainer trees are based on rc1. So are you sure everyone is going to be fine with a rc2 > dependency? Generally, new APIs don't go in without a user. From [1], looks like 'net' rebase on rcX for fixes and 'net-next' rebase on 'net'? [1] https://www.kernel.org/doc/Documentation/networking/netdev-FAQ.txt > > That being said, if this was 10s of different trees I'd reconsider, but since most of the callers are > in net, I'm less willing to apply "not a fix" to fixes. > > > [1] > > https://lore.kernel.org/all/96fbccd3-fd79-4b2f-8f41-bd0e3fdb2c69@lunn. > > ch/ > > > > [2] > > https://lore.kernel.org/all/7fe9dad9-85e2-4cf0-98bc-cca20ff62df5@kerne > > l.org/ > > It's not like they are saying to do the opposite of what I said. If the dependency is not part of your > series, then it needs to be in rc1. OK. As you suggested, since 'net' is the main user of this API, I will send this patch along with other net patches, to net subsystem as net-next is open now. Cheers, Biju
diff --git a/drivers/of/base.c b/drivers/of/base.c index af6c68bbb427..e37b088f1fad 100644 --- a/drivers/of/base.c +++ b/drivers/of/base.c @@ -824,6 +824,33 @@ struct device_node *of_get_child_by_name(const struct device_node *node, } EXPORT_SYMBOL(of_get_child_by_name); +/** + * of_get_available_child_by_name - Find the available child node by name for a given parent + * @node: parent node + * @name: child name to look for. + * + * This function looks for child node for given matching name and checks the + * device's availability for use. + * + * Return: A node pointer if found, with refcount incremented, use + * of_node_put() on it when done. + * Returns NULL if node is not found. + */ +struct device_node *of_get_available_child_by_name(const struct device_node *node, + const char *name) +{ + struct device_node *child; + + child = of_get_child_by_name(node, name); + if (child && !of_device_is_available(child)) { + of_node_put(child); + return NULL; + } + + return child; +} +EXPORT_SYMBOL(of_get_available_child_by_name); + struct device_node *__of_find_node_by_path(const struct device_node *parent, const char *path) { diff --git a/include/linux/of.h b/include/linux/of.h index eaf0e2a2b75c..9d6b8a61607f 100644 --- a/include/linux/of.h +++ b/include/linux/of.h @@ -301,6 +301,8 @@ extern struct device_node *of_get_compatible_child(const struct device_node *par const char *compatible); extern struct device_node *of_get_child_by_name(const struct device_node *node, const char *name); +extern struct device_node *of_get_available_child_by_name(const struct device_node *node, + const char *name); /* cache lookup */ extern struct device_node *of_find_next_cache_node(const struct device_node *); @@ -578,6 +580,13 @@ static inline struct device_node *of_get_child_by_name( return NULL; } +static inline struct device_node *of_get_available_child_by_name( + const struct device_node *node, + const char *name) +{ + return NULL; +} + static inline int of_device_is_compatible(const struct device_node *device, const char *name) {
There are lot of drivers using of_get_child_by_name() followed by of_device_is_available() to find the available child node by name for a given parent. Provide a helper for these users to simplify the code. Suggested-by: Geert Uytterhoeven <geert+renesas@glider.be> Signed-off-by: Biju Das <biju.das.jz@bp.renesas.com> --- v1->v2: * Updated commit description. * Updated kerneldoc comment block * Avoided code duplication by using of_get_child_by_name(). Note: grep showed the below files will be the users for this new API. I will be updating these drivers once this patch is in mainline. drivers/net/dsa/rzn1_a5psw.c drivers/net/can/rcar/rcar_canfd.c drivers/net/ethernet/mediatek/mtk_star_emac.c drivers/net/dsa/mt7530.c drivers/net/dsa/sja1105/sja1105_mdio.c drivers/net/dsa/qca/qca8k-8xxx.c drivers/net/wireless/mediatek/mt76/mac80211.c drivers/net/ethernet/ibm/emac/core.c drivers/net/ethernet/ti/am65-cpsw-nuss.c drivers/net/ethernet/actions/owl-emac.c drivers/net/ethernet/mediatek/mtk_eth_soc.c drivers/media/platform/samsung/exynos4-is/media-dev.h drivers/gpu/drm/tegra/rgb.c drivers/gpu/drm/msm/adreno/adreno_gpu.c drivers/clk/davinci/pll.c --- drivers/of/base.c | 27 +++++++++++++++++++++++++++ include/linux/of.h | 9 +++++++++ 2 files changed, 36 insertions(+)