mtd concat device driver
Subsystems (2)
v2 cover mtd concat device driver
Hi everybody,
when porting my router board from a mach-file based OpenWRT target to a device-tree based target, I found that there is no generic way to create a mtd_concat device from within the dts. The following patches attempt to provide that possibility.
This is a second roll of that patch series, the first one can be seen at [1]. Apologies for not including the correct recipients in the first roll.
In this first discussion, concerns were raised that a driver for a "virtual" device like this might have no place in the device tree system. However, I would argue that this very similar to specifying the partitions of a mtd device, which can also done in the device tree. In fact, I believe this is the only way to be able to specify the partitions of such a concat device in the dts file (but I'm happy to be corrected if I'm mistaken).
I have made the example in the dt-binding documentation a little bit more expressive in this detail.
In this second roll I have also addressed all issues that reviewers have brought up so far, hopefully to their satisfaction.
Best Regards
Bernhard
[1] http://lists.infradead.org/pipermail/linux-mtd/2018-September/083832.html
Bernhard Frauendienst (3):
mtd: core: add get_mtd_device_by_node
dt-bindings: add bindings for mtd-concat devices
mtd: mtdconcat: add dt driver for concat devices
.../devicetree/bindings/mtd/mtd-concat.txt | 36 +++++ drivers/mtd/Kconfig | 2 + drivers/mtd/Makefile | 3 + drivers/mtd/composite/Kconfig | 12 ++ drivers/mtd/composite/Makefile | 7 + drivers/mtd/composite/virt_concat.c | 128 ++++++++++++++++++ drivers/mtd/mtdcore.c | 38 ++++++ include/linux/mtd/mtd.h | 2 + 8 files changed, 228 insertions(+) create mode 100644 Documentation/devicetree/bindings/mtd/mtd-concat.txt create mode 100644 drivers/mtd/composite/Kconfig create mode 100644 drivers/mtd/composite/Makefile create mode 100644 drivers/mtd/composite/virt_concat.c
Signature · 1 line
2.17.1
PATCH v2 1/3 mtd: core: add get_mtd_device_by_node
Add function to retrieve a mtd device by its OF node. Since drivers can assign arbitrary names to mtd devices in the absence of a label property, there is no other reliable way to retrieve a mtd device for a given OF node.
--- drivers/mtd/mtdcore.c | 38 ++++++++++++++++++++++++++++++++++++++ include/linux/mtd/mtd.h | 2 ++ 2 files changed, 40 insertions(+)
drivers/mtd/mtdcore.c +38 -0
@@ -926,6 +926,44 @@ struct mtd_info *get_mtd_device_nm(const char *name) }EXPORT_SYMBOL_GPL (get_mtd_device_nm); +/** +* get_mtd_device_by_node - obtain a validated handle for an MTD device +* by of_node +* @of_node: OF node of MTD device to open +* +* This function returns MTD device description structure in case of +* success and an error code in case of failure. +*/ +struct mtd_info *get_mtd_device_by_node (const struct device_node *of_node) +{ +int err = -ENODEV ; +struct mtd_info *mtd =NULL , *other; + +mutex_lock (&mtd_table_mutex); + +mtd_for_each_device (other) { +if (of_node == other->dev.of_node) { + mtd = other; +break ; + } + } + +if (!mtd) + goto out_unlock; + + err =__get_mtd_device (mtd); +if (err) + goto out_unlock; + +mutex_unlock (&mtd_table_mutex); +return mtd; + +out_unlock: +mutex_unlock (&mtd_table_mutex); +return ERR_PTR (err); +} +EXPORT_SYMBOL_GPL (get_mtd_device_by_node); +void put_mtd_device (struct mtd_info *mtd) {mutex_lock (&mtd_table_mutex);
include/linux/mtd/mtd.h +2 -0
@@ -570,6 +570,8 @@ extern struct mtd_info *get_mtd_device(struct mtd_info *mtd, int num);extern int __get_mtd_device (struct mtd_info *mtd);extern void __put_mtd_device (struct mtd_info *mtd);extern struct mtd_info *get_mtd_device_nm (const char *name); +extern struct mtd_info *get_mtd_device_by_node ( +const struct device_node *of_node);extern void put_mtd_device (struct mtd_info *mtd);
Signature · 1 line
2.17.1
Apologies, again, I seem not to be able to handle git-send-mail correctly, the cover letter got lost in operation (using get_maintainers on a cover letter is not a good idea). Here it is again:
Hi everybody,
when porting my router board from a mach-file based OpenWRT target to a device-tree based target, I found that there is no generic way to create a mtd_concat device from within the dts. The following patches attempt to provide that possibility.
This is a second roll of that patch series, the first one can be seen at [1]. Apologies for not including the correct recipients in the first roll.
In this first discussion, concerns were raised that a driver for a "virtual" device like this might have no place in the device tree system. However, I would argue that this very similar to specifying the partitions of a mtd device, which can also done in the device tree. In fact, I believe this is the only way to be able to specify the partitions of such a concat device in the dts file (but I'm happy to be corrected if I'm mistaken).
I have made the example in the dt-binding documentation a little bit more expressive in this detail.
In this second roll I have also addressed all issues that reviewers have brought up so far, hopefully to their satisfaction.
Best Regards
Bernhard
[1] http://lists.infradead.org/pipermail/linux-mtd/2018-September/083832.html
Bernhard Frauendienst (3):
mtd: core: add get_mtd_device_by_node
dt-bindings: add bindings for mtd-concat devices
mtd: mtdconcat: add dt driver for concat devices
.../devicetree/bindings/mtd/mtd-concat.txt | 36 +++++ drivers/mtd/Kconfig | 2 + drivers/mtd/Makefile | 3 + drivers/mtd/composite/Kconfig | 12 ++ drivers/mtd/composite/Makefile | 7 + drivers/mtd/composite/virt_concat.c | 128 ++++++++++++++++++ drivers/mtd/mtdcore.c | 38 ++++++ include/linux/mtd/mtd.h | 2 +
8 files changed, 228 insertions(+)
create mode 100644 Documentation/devicetree/bindings/mtd/mtd-concat.txt create mode 100644 drivers/mtd/composite/Kconfig
create mode 100644 drivers/mtd/composite/Makefile
create mode 100644 drivers/mtd/composite/virt_concat.c
Signature · 1 line
2.17.1
Hello,
it looks like I didn't even manage to correctly CC the device tree maintainers and list, which was the main reason for sending out this second roll. This is getting embarrassing, but I guess I will simply prepare a 3rd roll and try even harder to send it to the right people this time (without relying on any script magic).
Of course I'm happy to include any suggestions you have in the meantime.
Sorry for the disturbance!
Best regards,
Bernhard
Am 07.09.2018 um 20:12 schrieb Bernhard Frauendienst:
Quoted 1 line
Apologies, again, I seem not to be able to handle git-send-mail > correctly, the cover letter got lost in operation (using
get_maintainers > on a cover letter is not a good idea). Here it is again: > > > Hi everybody, > > when porting my router board from a mach-file based OpenWRT target to a > device-tree based target, I found that there is no generic way to create > a mtd_concat device from within the dts. The following patches attempt > to provide that possibility. > > This is a second roll of that patch series, the first one can be seen at > [1]. Apologies for not including the correct recipients in the first > roll. > > In this first discussion, concerns were raised that a driver for a > "virtual" device like this might have no place in the device tree > system. However, I would argue that this very similar to specifying the > partitions of a mtd device, which can also done in the device tree. In > fact, I believe this is the only way to be able to specify the > partitions of such a concat device in the dts file (but I'm happy to be > corrected if I'm mistaken). > I have made the example in the dt-binding documentation a little bit > more expressive in this detail. > > In this second roll I have also addressed all issues that reviewers have > brought up so far, hopefully to their satisfaction. > > Best Regards > Bernhard > > [1] http://lists.infradead.org/pipermail/linux-mtd/2018-September/083832.html > > > Bernhard Frauendienst (3): > mtd: core: add get_mtd_device_by_node > dt-bindings: add bindings for mtd-concat devices > mtd: mtdconcat: add dt driver for concat devices > > .../devicetree/bindings/mtd/mtd-concat.txt | 36 +++++ > drivers/mtd/Kconfig | 2 + > drivers/mtd/Makefile | 3 + > drivers/mtd/composite/Kconfig | 12 ++ > drivers/mtd/composite/Makefile | 7 + > drivers/mtd/composite/virt_concat.c | 128 ++++++++++++++++++ > drivers/mtd/mtdcore.c | 38 ++++++ > include/linux/mtd/mtd.h | 2 + > 8 files changed, 228 insertions(+) > create mode 100644 Documentation/devicetree/bindings/mtd/mtd-concat.txt > create mode 100644 drivers/mtd/composite/Kconfig > create mode 100644 drivers/mtd/composite/Makefile > create mode 100644 drivers/mtd/composite/virt_concat.c >
PATCH v2 2/3 dt-bindings: add bindings for mtd-concat devices
In reply to Bernhard Frauendienst--- .../devicetree/bindings/mtd/mtd-concat.txt | 36 +++++++++++++++++++ 1 file changed, 36 insertions(+) create mode 100644 Documentation/devicetree/bindings/mtd/mtd-concat.txt
Documentation/devicetree/bindings/mtd/mtd-concat.txt new file +36 -0
@@ -0,0 +1,36 @@ +Virtual MTD concat device + +Requires properties: +- devices: list of phandles to mtd nodes that should be concatenated + +Example: + +&spi { + flash0: flash@0 { + ... + }; + flash1: flash@1 { + ... + }; +}; + +flash { + compatible = "mtd-concat"; + + devices = <&flash0 &flash1>; + + partitions { + compatible = "fixed-partitions"; + + partition@0 { + label = "boot"; + reg = <0x0000000 0x0040000>; + read-only; + }; + + partition@40000 { + label = "firmware"; + reg = <0x0040000 0x1fc0000>; + }; + } +}
Signature · 6 lines
2.17.1 ______________________________________________________ Linux MTD discussion mailing list http://lists.infradead.org/mailman/listinfo/linux-mtd/
PATCH v2 3/3 mtd: mtdconcat: add dt driver for concat devices
In reply to Bernhard FrauendienstSome mtd drivers like physmap variants have support for concatenating multiple mtd devices, but there is no generic way to define such a concat device from within the device tree.
This is useful for some SoC boards that use multiple flash chips as memory banks of a single mtd device, with partitions spanning chip borders.
This commit adds a driver for creating virtual mtd-concat devices. They must have a compatible = "mtd-concat" line, and define a list of devices to concat in the 'devices' property, for example:
flash {
compatible = "mtd-concat";
devices = <&flash0 &flash1>;
partitions {
... };
};
The driver is added to the very end of the mtd Makefile to increase the likelyhood of all child devices already being loaded at the time of probing, preventing unnecessary deferred probes.
--- drivers/mtd/Kconfig | 2 + drivers/mtd/Makefile | 3 + drivers/mtd/composite/Kconfig | 12 +++ drivers/mtd/composite/Makefile | 7 ++ drivers/mtd/composite/virt_concat.c | 128 ++++++++++++++++++++++++++++ 5 files changed, 152 insertions(+) create mode 100644 drivers/mtd/composite/Kconfig create mode 100644 drivers/mtd/composite/Makefile create mode 100644 drivers/mtd/composite/virt_concat.c
drivers/mtd/Kconfig +2 -0
@@ -339,4 +339,6 @@ source "drivers/mtd/spi-nor/Kconfig" source "drivers/mtd/ubi/Kconfig" +source "drivers/mtd/composite/Kconfig" + endif # MTD
drivers/mtd/Makefile +3 -0
@@ -36,3 +36,6 @@ obj-y += chips/ lpddr/ maps/ devices/ nand/ tests/ obj-$(CONFIG_MTD_SPI_NOR) += spi-nor/ obj-$(CONFIG_MTD_UBI) += ubi/ + +# Composite drivers must be loaded last +obj-y += composite/
drivers/mtd/composite/Kconfig new file +12 -0
@@ -0,0 +1,12 @@ +menu "Composite MTD device drivers" + depends on MTD!=n + +config MTD_VIRT_CONCAT + tristate "Virtual concat MTD device" + help + This driver allows creation of a virtual MTD concat device, which + concatenates multiple underlying MTD devices to a single device. + This is required by some SoC boards where multiple memory banks are + used as one device with partitions spanning across device boundaries. + +endmenu
drivers/mtd/composite/Makefile new file +7 -0
@@ -0,0 +1,7 @@ +# SPDX-License-Identifier: GPL-2.0 +# +# linux/drivers/mtd/composite/Makefile +# + +obj-$(CONFIG_MTD_VIRT_CONCAT) += virt_concat.o +
drivers/mtd/composite/virt_concat.c new file +128 -0
128 lines not on this page · raw message
Signature · 1 line
2.17.1