[RFC 1/1] disk: PARTITIONS should depend on HAVE_BLOCK_DEVICE
Partitions are only relevant for block devices.
Signed-off-by: Heinrich Schuchardt xypron.glpk@gmx.de --- I still need to check if the part command is needed on boards w/o DISTRO_DEFAULTS. --- cmd/Kconfig | 4 ++-- disk/Kconfig | 1 + 2 files changed, 3 insertions(+), 2 deletions(-)
diff --git a/cmd/Kconfig b/cmd/Kconfig index 0625ee4050..e253fb7a68 100644 --- a/cmd/Kconfig +++ b/cmd/Kconfig @@ -1171,8 +1171,8 @@ config CMD_OSD
config CMD_PART bool "part" - select HAVE_BLOCK_DEVICE - select PARTITION_UUIDS + depends on PARTITION_UUIDS + default y if DISTRO_DEFAULTS help Read and display information about the partition table on various media. diff --git a/disk/Kconfig b/disk/Kconfig index cee16a80bc..1dfe3a0667 100644 --- a/disk/Kconfig +++ b/disk/Kconfig @@ -3,6 +3,7 @@ menu "Partition Types"
config PARTITIONS bool "Enable Partition Labels (disklabels) support" + depends on HAVE_BLOCK_DEVICE default y select SPL_SPRINTF if SPL select TPL_SPRINTF if TPL -- 2.29.2
On Mon, Jan 25, 2021 at 05:28:39PM +0100, Heinrich Schuchardt wrote:
Partitions are only relevant for block devices.
Signed-off-by: Heinrich Schuchardt xypron.glpk@gmx.de
I still need to check if the part command is needed on boards w/o DISTRO_DEFAULTS.
cmd/Kconfig | 4 ++-- disk/Kconfig | 1 + 2 files changed, 3 insertions(+), 2 deletions(-)
So, https://gist.github.com/trini/806edc6d37b0e53931a3b8195cbf9da2 shows all of the boards that had been building with this now invalid combination and are discarding a bunch of code. The next step is to look at these and see which of them likely have a use case not been described by these Kconfig choices.
On 1/25/21 10:02 PM, Tom Rini wrote:
On Mon, Jan 25, 2021 at 05:28:39PM +0100, Heinrich Schuchardt wrote:
Partitions are only relevant for block devices.
Signed-off-by: Heinrich Schuchardt xypron.glpk@gmx.de
I still need to check if the part command is needed on boards w/o DISTRO_DEFAULTS.
cmd/Kconfig | 4 ++-- disk/Kconfig | 1 + 2 files changed, 3 insertions(+), 2 deletions(-)
So, https://gist.github.com/trini/806edc6d37b0e53931a3b8195cbf9da2 shows all of the boards that had been building with this now invalid combination and are discarding a bunch of code. The next step is to look at these and see which of them likely have a use case not been described by these Kconfig choices.
With the patch for boards not using distro_defaults the following configuration items are dropped:
CONFIG_CMD_PART=y CONFIG_PARTITION_UUIDS=y
I will try to minimize the configuration changes produced by the next version of the patch.
Best regards
Heinrich
On Mon, Jan 25, 2021 at 11:18:04PM +0100, Heinrich Schuchardt wrote:
On 1/25/21 10:02 PM, Tom Rini wrote:
On Mon, Jan 25, 2021 at 05:28:39PM +0100, Heinrich Schuchardt wrote:
Partitions are only relevant for block devices.
Signed-off-by: Heinrich Schuchardt xypron.glpk@gmx.de
I still need to check if the part command is needed on boards w/o DISTRO_DEFAULTS.
cmd/Kconfig | 4 ++-- disk/Kconfig | 1 + 2 files changed, 3 insertions(+), 2 deletions(-)
So, https://gist.github.com/trini/806edc6d37b0e53931a3b8195cbf9da2 shows all of the boards that had been building with this now invalid combination and are discarding a bunch of code. The next step is to look at these and see which of them likely have a use case not been described by these Kconfig choices.
With the patch for boards not using distro_defaults the following configuration items are dropped:
CONFIG_CMD_PART=y CONFIG_PARTITION_UUIDS=y
I will try to minimize the configuration changes produced by the next version of the patch.
OK, but my expectation is that there's either zero change in sizes OR the board maintainer for those that do drop code agree it was functionally useless code today. Thanks!
participants (2)
-
Heinrich Schuchardt -
Tom Rini