* [PATCH v4 0/3] introduce board-driver support @ 2020-07-20 6:39 Oleksij Rempel 2020-07-20 6:39 ` [PATCH v4 1/3] devinfo: do not dump the device node for the root node Oleksij Rempel ` (2 more replies) 0 siblings, 3 replies; 6+ messages in thread From: Oleksij Rempel @ 2020-07-20 6:39 UTC (permalink / raw) To: barebox; +Cc: Oleksij Rempel, david changes v4: - rework devinfo to print DT node without child nodes changes v3: - move devinfo patch to the first place - devinfo: provide better commit message changes v2: - move root device registration to the of_probe() - port one board as example - fix issue with devinfo Oleksij Rempel (3): devinfo: do not dump the device node for the root node of: base: register DT root as device ARM: embest-riotboard: port board file to the driver model arch/arm/boards/embest-riotboard/board.c | 18 ++++++++++++----- commands/devinfo.c | 2 +- drivers/of/base.c | 25 ++++++++++++++++++------ include/of.h | 1 + 4 files changed, 34 insertions(+), 12 deletions(-) -- 2.27.0 _______________________________________________ barebox mailing list barebox@lists.infradead.org http://lists.infradead.org/mailman/listinfo/barebox ^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH v4 1/3] devinfo: do not dump the device node for the root node 2020-07-20 6:39 [PATCH v4 0/3] introduce board-driver support Oleksij Rempel @ 2020-07-20 6:39 ` Oleksij Rempel 2020-07-20 8:35 ` Lucas Stach 2020-07-20 6:39 ` [PATCH v4 2/3] of: base: register DT root as device Oleksij Rempel 2020-07-20 6:39 ` [PATCH v4 3/3] ARM: embest-riotboard: port board file to the driver model Oleksij Rempel 2 siblings, 1 reply; 6+ messages in thread From: Oleksij Rempel @ 2020-07-20 6:39 UTC (permalink / raw) To: barebox; +Cc: Oleksij Rempel, david Calling the devinfo against a device which is linked to some device tree node will result a device tree dump of this node. For example: barebox@Protonic PRTI6Q board:/ devinfo devinfo ldb.of Bus: platform Device node: /ldb ldb { #address-cells = <0x1>; #size-cells = <0x0>; compatible = "fsl,imx6q-ldb", "fsl,imx53-ldb"; gpr = <0x5>; status = "okay"; clocks = <0x4 0x21 0x4 0x22 0x4 0x27 0x4 0x28 0x4 0x29 0x4 0x2a 0x4 0x87 0x4 0x88>; clock-names = "di0_pll", "di1_pll", "di0_sel", "di1_sel", "di2_sel", "di3_sel", "di0", "di1"; lvds-channel@0 { #address-cells = <0x1>; #size-cells = <0x0>; reg = <0x0>; status = "okay"; port@0 { reg = <0x0>; endpoint { remote-endpoint = <0x6>; phandle = <0x56>; }; }; port@1 { reg = <0x1>; endpoint { remote-endpoint = <0x7>; phandle = <0x5a>; }; }; port@2 { reg = <0x2>; endpoint { remote-endpoint = <0x8>; phandle = <0x60>; }; }; ......... Calling same command on a device which is linked to the root node of device tree, for example "machine.of", will trigger a dump of complete device tree. Since the same functionality is provided by the "of_dump" command, it is better to limit devinfo to print only exactly requested node, without printing the child nodes. After this patch we would get following output: barebox@Protonic PRTI6Q board:/ devinfo ldb.of Bus: platform Device node: /ldb ldb { #address-cells = <0x1>; #size-cells = <0x0>; compatible = "fsl,imx6q-ldb", "fsl,imx53-ldb"; gpr = <0x5>; status = "okay"; clocks = <0x4 0x21 0x4 0x22 0x4 0x27 0x4 0x28 0x4 0x29 0x4 0x2a 0x4 0x87 0x4 0x88>; clock-names = "di0_pll", "di1_pll", "di0_sel", "di1_sel", "di2_sel", "di3_sel", "di0", "di1"; }; Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de> --- commands/devinfo.c | 2 +- drivers/of/base.c | 20 ++++++++++++++------ include/of.h | 1 + 3 files changed, 16 insertions(+), 7 deletions(-) diff --git a/commands/devinfo.c b/commands/devinfo.c index 81956b1cc0..7036545cdb 100644 --- a/commands/devinfo.c +++ b/commands/devinfo.c @@ -101,7 +101,7 @@ static int do_devinfo(int argc, char *argv[]) #ifdef CONFIG_OFDEVICE if (dev->device_node) { printf("Device node: %s\n", dev->device_node->full_name); - of_print_nodes(dev->device_node, 0); + of_print_node(dev->device_node, 0); } #endif } diff --git a/drivers/of/base.c b/drivers/of/base.c index 4c633bcd49..4754fcb98f 100644 --- a/drivers/of/base.c +++ b/drivers/of/base.c @@ -1793,7 +1793,8 @@ int of_property_read_string_helper(const struct device_node *np, return i <= 0 ? -ENODATA : i; } -static void __of_print_nodes(struct device_node *node, int indent, const char *prefix) +static void __of_print_nodes(struct device_node *node, int indent, + const char *prefix, bool single) { struct device_node *n; struct property *p; @@ -1815,8 +1816,10 @@ static void __of_print_nodes(struct device_node *node, int indent, const char *p printf(";\n"); } - list_for_each_entry(n, &node->children, parent_list) { - __of_print_nodes(n, indent + 1, prefix); + if (!single) { + list_for_each_entry(n, &node->children, parent_list) { + __of_print_nodes(n, indent + 1, prefix, false); + } } printf("%s%*s};\n", prefix, indent * 8, ""); @@ -1824,7 +1827,12 @@ static void __of_print_nodes(struct device_node *node, int indent, const char *p void of_print_nodes(struct device_node *node, int indent) { - __of_print_nodes(node, indent, NULL); + __of_print_nodes(node, indent, NULL, false); +} + +void of_print_node(struct device_node *node, int indent) +{ + __of_print_nodes(node, indent, NULL, true); } static void __of_print_property(struct property *p, int indent) @@ -1934,14 +1942,14 @@ void of_diff(struct device_node *a, struct device_node *b, int indent) of_diff(ca, cb, indent + 1); } else { of_print_parents(a, &printed); - __of_print_nodes(ca, indent, "-"); + __of_print_nodes(ca, indent, "-", false); } } for_each_child_of_node(b, cb) { if (!of_get_child_by_name(a, cb->name)) { of_print_parents(a, &printed); - __of_print_nodes(cb, indent, "+"); + __of_print_nodes(cb, indent, "+", false); } } diff --git a/include/of.h b/include/of.h index 08bbeaf4d2..820a8ea0ed 100644 --- a/include/of.h +++ b/include/of.h @@ -104,6 +104,7 @@ static inline const void *of_property_get_value(struct property *pp) void of_print_property(const void *data, int len); void of_print_cmdline(struct device_node *root); +void of_print_node(struct device_node *node, int indent); void of_print_nodes(struct device_node *node, int indent); void of_diff(struct device_node *a, struct device_node *b, int indent); int of_probe(void); -- 2.27.0 _______________________________________________ barebox mailing list barebox@lists.infradead.org http://lists.infradead.org/mailman/listinfo/barebox ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH v4 1/3] devinfo: do not dump the device node for the root node 2020-07-20 6:39 ` [PATCH v4 1/3] devinfo: do not dump the device node for the root node Oleksij Rempel @ 2020-07-20 8:35 ` Lucas Stach 0 siblings, 0 replies; 6+ messages in thread From: Lucas Stach @ 2020-07-20 8:35 UTC (permalink / raw) To: Oleksij Rempel, barebox; +Cc: david Am Montag, den 20.07.2020, 08:39 +0200 schrieb Oleksij Rempel: > Calling the devinfo against a device which is linked to some device tree > node will result a device tree dump of this node. For example: > > barebox@Protonic PRTI6Q board:/ devinfo devinfo ldb.of > Bus: platform > Device node: /ldb > ldb { > > #address-cells = <0x1>; > #size-cells = <0x0>; > compatible = "fsl,imx6q-ldb", "fsl,imx53-ldb"; > gpr = <0x5>; > status = "okay"; > clocks = <0x4 0x21 0x4 0x22 0x4 0x27 0x4 0x28 0x4 0x29 0x4 0x2a 0x4 0x87 0x4 0x88>; > clock-names = "di0_pll", "di1_pll", "di0_sel", "di1_sel", "di2_sel", "di3_sel", "di0", "di1"; > lvds-channel@0 { > #address-cells = <0x1>; > #size-cells = <0x0>; > reg = <0x0>; > status = "okay"; > port@0 { > reg = <0x0>; > endpoint { > remote-endpoint = <0x6>; > phandle = <0x56>; > }; > }; > port@1 { > reg = <0x1>; > endpoint { > remote-endpoint = <0x7>; > phandle = <0x5a>; > }; > }; > port@2 { > reg = <0x2>; > endpoint { > remote-endpoint = <0x8>; > phandle = <0x60>; > }; > }; > ......... > > Calling same command on a device which is linked to the root node of > device tree, for example "machine.of", will trigger a dump of complete > device tree. Since the same functionality is provided by the "of_dump" > command, it is better to limit devinfo to print only exactly requested > node, without printing the child nodes. After this patch we would get > following output: > > barebox@Protonic PRTI6Q board:/ devinfo ldb.of > Bus: platform > Device node: /ldb > ldb { > #address-cells = <0x1>; > #size-cells = <0x0>; > compatible = "fsl,imx6q-ldb", "fsl,imx53-ldb"; > gpr = <0x5>; > status = "okay"; > clocks = <0x4 0x21 0x4 0x22 0x4 0x27 0x4 0x28 0x4 0x29 0x4 0x2a 0x4 0x87 0x4 0x88>; > clock-names = "di0_pll", "di1_pll", "di0_sel", "di1_sel", "di2_sel", "di3_sel", "di0", "di1"; > }; Bad example, the port nodes are actually part of the LDB device and are useful information when dumping the info of the LDB. Maybe this wasn't clear enough, but in my last mail I said "don't dump subnodes with a compatible" on purpose. If we just stop iterating when the subnode has its own compatible, we still dump subnodes like the port nodes, which are logically part of the parent OF node, but still don't dump subnodes which likely have a device on their own. Regards, Lucas > Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de> > --- > commands/devinfo.c | 2 +- > drivers/of/base.c | 20 ++++++++++++++------ > include/of.h | 1 + > 3 files changed, 16 insertions(+), 7 deletions(-) > > diff --git a/commands/devinfo.c b/commands/devinfo.c > index 81956b1cc0..7036545cdb 100644 > --- a/commands/devinfo.c > +++ b/commands/devinfo.c > @@ -101,7 +101,7 @@ static int do_devinfo(int argc, char *argv[]) > #ifdef CONFIG_OFDEVICE > if (dev->device_node) { > printf("Device node: %s\n", dev->device_node->full_name); > - of_print_nodes(dev->device_node, 0); > + of_print_node(dev->device_node, 0); > } > #endif > } > diff --git a/drivers/of/base.c b/drivers/of/base.c > index 4c633bcd49..4754fcb98f 100644 > --- a/drivers/of/base.c > +++ b/drivers/of/base.c > @@ -1793,7 +1793,8 @@ int of_property_read_string_helper(const struct device_node *np, > return i <= 0 ? -ENODATA : i; > } > > -static void __of_print_nodes(struct device_node *node, int indent, const char *prefix) > +static void __of_print_nodes(struct device_node *node, int indent, > + const char *prefix, bool single) > { > struct device_node *n; > struct property *p; > @@ -1815,8 +1816,10 @@ static void __of_print_nodes(struct device_node *node, int indent, const char *p > printf(";\n"); > } > > - list_for_each_entry(n, &node->children, parent_list) { > - __of_print_nodes(n, indent + 1, prefix); > + if (!single) { > + list_for_each_entry(n, &node->children, parent_list) { > + __of_print_nodes(n, indent + 1, prefix, false); > + } > } > > printf("%s%*s};\n", prefix, indent * 8, ""); > @@ -1824,7 +1827,12 @@ static void __of_print_nodes(struct device_node *node, int indent, const char *p > > void of_print_nodes(struct device_node *node, int indent) > { > - __of_print_nodes(node, indent, NULL); > + __of_print_nodes(node, indent, NULL, false); > +} > + > +void of_print_node(struct device_node *node, int indent) > +{ > + __of_print_nodes(node, indent, NULL, true); > } > > static void __of_print_property(struct property *p, int indent) > @@ -1934,14 +1942,14 @@ void of_diff(struct device_node *a, struct device_node *b, int indent) > of_diff(ca, cb, indent + 1); > } else { > of_print_parents(a, &printed); > - __of_print_nodes(ca, indent, "-"); > + __of_print_nodes(ca, indent, "-", false); > } > } > > for_each_child_of_node(b, cb) { > if (!of_get_child_by_name(a, cb->name)) { > of_print_parents(a, &printed); > - __of_print_nodes(cb, indent, "+"); > + __of_print_nodes(cb, indent, "+", false); > } > } > > diff --git a/include/of.h b/include/of.h > index 08bbeaf4d2..820a8ea0ed 100644 > --- a/include/of.h > +++ b/include/of.h > @@ -104,6 +104,7 @@ static inline const void *of_property_get_value(struct property *pp) > void of_print_property(const void *data, int len); > void of_print_cmdline(struct device_node *root); > > +void of_print_node(struct device_node *node, int indent); > void of_print_nodes(struct device_node *node, int indent); > void of_diff(struct device_node *a, struct device_node *b, int indent); > int of_probe(void); _______________________________________________ barebox mailing list barebox@lists.infradead.org http://lists.infradead.org/mailman/listinfo/barebox ^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH v4 2/3] of: base: register DT root as device 2020-07-20 6:39 [PATCH v4 0/3] introduce board-driver support Oleksij Rempel 2020-07-20 6:39 ` [PATCH v4 1/3] devinfo: do not dump the device node for the root node Oleksij Rempel @ 2020-07-20 6:39 ` Oleksij Rempel 2020-08-11 14:24 ` Ahmad Fatoum 2020-07-20 6:39 ` [PATCH v4 3/3] ARM: embest-riotboard: port board file to the driver model Oleksij Rempel 2 siblings, 1 reply; 6+ messages in thread From: Oleksij Rempel @ 2020-07-20 6:39 UTC (permalink / raw) To: barebox; +Cc: Oleksij Rempel, david A usual board file contains at least one of_machine_is_compatible(). Some of the have a rather long list with complicated version logic. To avoid own implementation for driver management, register the root node of device tree as platform device. So, the main platform bus can attach proper board driver. After this patch a typical board.c file can reuse existing driver infrastructure. After this patch, you will be able to see all registered board drivers with drvinfo as fallow: ... board-embest-riot board-protonic-imx6 machine.of ... With devinfo, you'll be able to get some board specific information, if this is implemented: barebox@Protonic PRTI6Q board:/ devinfo machine.of Driver: board-protonic-imx6 Bus: platform Parameters: boardid: 0 (type: uint32) boardrev: 1 (type: uint32) Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de> --- drivers/of/base.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/drivers/of/base.c b/drivers/of/base.c index 4754fcb98f..b76e424da0 100644 --- a/drivers/of/base.c +++ b/drivers/of/base.c @@ -2141,6 +2141,7 @@ static void of_probe_memory(void) int of_probe(void) { struct device_node *firmware; + struct device_d *dev; if(!root_node) return -ENODEV; @@ -2157,6 +2158,10 @@ int of_probe(void) if (firmware) of_platform_populate(firmware, NULL, NULL); + dev = of_platform_device_create(root_node, NULL); + if (dev) + dev_set_name(dev, "%s.of", "machine"); + of_clk_init(root_node, NULL); of_platform_populate(root_node, of_default_bus_match_table, NULL); -- 2.27.0 _______________________________________________ barebox mailing list barebox@lists.infradead.org http://lists.infradead.org/mailman/listinfo/barebox ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH v4 2/3] of: base: register DT root as device 2020-07-20 6:39 ` [PATCH v4 2/3] of: base: register DT root as device Oleksij Rempel @ 2020-08-11 14:24 ` Ahmad Fatoum 0 siblings, 0 replies; 6+ messages in thread From: Ahmad Fatoum @ 2020-08-11 14:24 UTC (permalink / raw) To: Oleksij Rempel, barebox; +Cc: david Hello Oleksij, On 7/20/20 8:39 AM, Oleksij Rempel wrote: > A usual board file contains at least one of_machine_is_compatible(). > Some of the have a rather long list with complicated version logic. > > To avoid own implementation for driver management, register the root node > of device tree as platform device. So, the main platform bus can attach > proper board driver. After this patch a typical board.c file can reuse > existing driver infrastructure. > > After this patch, you will be able to see all registered board drivers > with drvinfo as fallow: this patch breaks boot on the Linux Automation MC-1. of_platform_device_create() uses of_device_make_bus_id() to generate a name, but that function is a no-op for the root node. This results in a platform device with dev->name == NULL, which inside register_device() -> get_device_by_name_id() -> strcmp leads to a NULL pointer dereference. You should probably not use of_platform_device_create, but use platform_device_register directly and set a suitable name upfront in the struct device_d you initialize. Cheers Ahmad > ... > board-embest-riot > board-protonic-imx6 > machine.of > ... > > With devinfo, you'll be able to get some board specific information, > if this is implemented: > barebox@Protonic PRTI6Q board:/ devinfo machine.of > Driver: board-protonic-imx6 > Bus: platform > Parameters: > boardid: 0 (type: uint32) > boardrev: 1 (type: uint32) > > Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de> > --- > drivers/of/base.c | 5 +++++ > 1 file changed, 5 insertions(+) > > diff --git a/drivers/of/base.c b/drivers/of/base.c > index 4754fcb98f..b76e424da0 100644 > --- a/drivers/of/base.c > +++ b/drivers/of/base.c > @@ -2141,6 +2141,7 @@ static void of_probe_memory(void) > int of_probe(void) > { > struct device_node *firmware; > + struct device_d *dev; > > if(!root_node) > return -ENODEV; > @@ -2157,6 +2158,10 @@ int of_probe(void) > if (firmware) > of_platform_populate(firmware, NULL, NULL); > > + dev = of_platform_device_create(root_node, NULL); > + if (dev) > + dev_set_name(dev, "%s.of", "machine"); > + > of_clk_init(root_node, NULL); > of_platform_populate(root_node, of_default_bus_match_table, NULL); > > -- Pengutronix e.K. | | Steuerwalder Str. 21 | http://www.pengutronix.de/ | 31137 Hildesheim, Germany | Phone: +49-5121-206917-0 | Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 | _______________________________________________ barebox mailing list barebox@lists.infradead.org http://lists.infradead.org/mailman/listinfo/barebox ^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH v4 3/3] ARM: embest-riotboard: port board file to the driver model 2020-07-20 6:39 [PATCH v4 0/3] introduce board-driver support Oleksij Rempel 2020-07-20 6:39 ` [PATCH v4 1/3] devinfo: do not dump the device node for the root node Oleksij Rempel 2020-07-20 6:39 ` [PATCH v4 2/3] of: base: register DT root as device Oleksij Rempel @ 2020-07-20 6:39 ` Oleksij Rempel 2 siblings, 0 replies; 6+ messages in thread From: Oleksij Rempel @ 2020-07-20 6:39 UTC (permalink / raw) To: barebox; +Cc: Oleksij Rempel, david This patch can be used as example for the new board-driver functionality. Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de> --- arch/arm/boards/embest-riotboard/board.c | 18 +++++++++++++----- 1 file changed, 13 insertions(+), 5 deletions(-) diff --git a/arch/arm/boards/embest-riotboard/board.c b/arch/arm/boards/embest-riotboard/board.c index 2e0cc9f0ab..746ecf0562 100644 --- a/arch/arm/boards/embest-riotboard/board.c +++ b/arch/arm/boards/embest-riotboard/board.c @@ -51,11 +51,8 @@ static int ar8035_phy_fixup(struct phy_device *dev) return 0; } -static int riotboard_device_init(void) +static int riotboard_probe(struct device_d *dev) { - if (!of_machine_is_compatible("riot,imx6s-riotboard")) - return 0; - phy_register_fixup_for_uid(0x004dd072, 0xffffffef, ar8035_phy_fixup); imx6_bbu_internal_mmc_register_handler("emmc", "/dev/mmc3.barebox", @@ -65,4 +62,15 @@ static int riotboard_device_init(void) return 0; } -device_initcall(riotboard_device_init); + +static const struct of_device_id riotboard_of_match[] = { + { .compatible = "riot,imx6s-riotboard" }, + {}, +}; + +static struct driver_d riotboard_driver = { + .name = "board-embest-riot", + .probe = riotboard_probe, + .of_compatible = DRV_OF_COMPAT(riotboard_of_match), +}; +device_platform_driver(riotboard_driver); -- 2.27.0 _______________________________________________ barebox mailing list barebox@lists.infradead.org http://lists.infradead.org/mailman/listinfo/barebox ^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2020-08-11 14:25 UTC | newest] Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2020-07-20 6:39 [PATCH v4 0/3] introduce board-driver support Oleksij Rempel 2020-07-20 6:39 ` [PATCH v4 1/3] devinfo: do not dump the device node for the root node Oleksij Rempel 2020-07-20 8:35 ` Lucas Stach 2020-07-20 6:39 ` [PATCH v4 2/3] of: base: register DT root as device Oleksij Rempel 2020-08-11 14:24 ` Ahmad Fatoum 2020-07-20 6:39 ` [PATCH v4 3/3] ARM: embest-riotboard: port board file to the driver model Oleksij Rempel
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox