[v4 0/7] Fix Rockchip RK3399 bootstd migration
Hey all,
I took a look at Simon's v3 series to fix the rk3399 bootstd migration, and it changed too much for everything else. I took about half of that series and then reworked a few things. Now only rk3399 platforms change at all and aside from bootcmd changes, the only thing is they now disable true/test/sysboot/showvar/false/exit commands as those were being pulled in from distro and now we don't set that flag. I think the way I changed how we enable BOOTSTD_DEFAULTS should make it easier to perform more SoC migrations.
From: Simon Glass sjg@chromium.org
Unfortunately the IRAM used to stash the bootstage records in TPL becomes inaccessible after SPL runs. Presumably this is because of ATF taking it over.
We could move the stash to another address in SPL, before passing it to U-Boot proper. But it seems easier to wait until we have support for standard passage[1] which should not be too far away.
For now, disable it in TPL and SPL.
[1] https://patchwork.ozlabs.org/project/uboot/cover/ 20220117150428.1580273-1-sjg@chromium.org/
Signed-off-by: Simon Glass sjg@chromium.org Tested-by: Vagrant Cascadian vagrant@debian.org --- Changes in v4: None. --- configs/rockpro64-rk3399_defconfig | 5 ----- 1 file changed, 5 deletions(-)
diff --git a/configs/rockpro64-rk3399_defconfig b/configs/rockpro64-rk3399_defconfig index dd67f9dff64b..496142368195 100644 --- a/configs/rockpro64-rk3399_defconfig +++ b/configs/rockpro64-rk3399_defconfig @@ -13,7 +13,6 @@ CONFIG_DM_RESET=y CONFIG_ROCKCHIP_RK3399=y CONFIG_TARGET_ROCKPRO64_RK3399=y CONFIG_SPL_STACK=0x400000 -CONFIG_BOOTSTAGE_STASH_ADDR=0xff8e0000 CONFIG_DEBUG_UART_BASE=0xFF1A0000 CONFIG_DEBUG_UART_CLOCK=24000000 CONFIG_SPL_SPI_FLASH_SUPPORT=y @@ -21,11 +20,7 @@ CONFIG_SPL_SPI=y CONFIG_SYS_LOAD_ADDR=0x800800 CONFIG_DEBUG_UART=y CONFIG_BOOTSTAGE=y -CONFIG_SPL_BOOTSTAGE=y -CONFIG_TPL_BOOTSTAGE=y CONFIG_BOOTSTAGE_REPORT=y -CONFIG_SPL_BOOTSTAGE_RECORD_COUNT=10 -CONFIG_BOOTSTAGE_STASH=y CONFIG_USE_PREBOOT=y CONFIG_DEFAULT_FDT_FILE="rockchip/rk3399-rockpro64.dtb" CONFIG_DISPLAY_BOARDINFO_LATE=y
From: Simon Glass sjg@chromium.org
This relates to booting so move it in to that Kconfig file, before changing it.
Signed-off-by: Simon Glass sjg@chromium.org --- Changes in v4: None. --- Kconfig | 27 --------------------------- boot/Kconfig | 27 +++++++++++++++++++++++++++ 2 files changed, 27 insertions(+), 27 deletions(-)
diff --git a/Kconfig b/Kconfig index a75cce7e28fb..b8f65589f457 100644 --- a/Kconfig +++ b/Kconfig @@ -191,33 +191,6 @@ config XEN
[1] - https://xenproject.org/
-config DISTRO_DEFAULTS - bool "Select defaults suitable for booting general purpose Linux distributions" - select AUTO_COMPLETE - select CMDLINE_EDITING - select CMD_BOOTI if ARM64 - select CMD_BOOTZ if ARM && !ARM64 - select CMD_DHCP if CMD_NET - select CMD_ENV_EXISTS - select CMD_EXT2 - select CMD_EXT4 - select CMD_FAT - select CMD_FS_GENERIC - select CMD_PART if PARTITIONS - select CMD_PING if CMD_NET - select CMD_PXE if NET - select CMD_SYSBOOT - select ENV_VARS_UBOOT_CONFIG - select HUSH_PARSER - select SUPPORT_RAW_INITRD - select SYS_LONGHELP - imply CMD_MII if NET - imply USB_STORAGE - imply USE_BOOTCOMMAND - help - Select this to enable various options and commands which are suitable - for building u-boot for booting general purpose Linux distributions. - config ENV_VARS_UBOOT_CONFIG bool "Add arch, board, vendor and soc variables to default environment" help diff --git a/boot/Kconfig b/boot/Kconfig index 5f491625c820..170ab63f27dc 100644 --- a/boot/Kconfig +++ b/boot/Kconfig @@ -820,6 +820,33 @@ config SYS_BOOT_RAMDISK_HIGH
endmenu # Boot images
+config DISTRO_DEFAULTS + bool "Select defaults suitable for booting general purpose Linux distributions" + select AUTO_COMPLETE + select CMDLINE_EDITING + select CMD_BOOTI if ARM64 + select CMD_BOOTZ if ARM && !ARM64 + select CMD_DHCP if CMD_NET + select CMD_ENV_EXISTS + select CMD_EXT2 + select CMD_EXT4 + select CMD_FAT + select CMD_FS_GENERIC + select CMD_PART if PARTITIONS + select CMD_PING if CMD_NET + select CMD_PXE if NET + select CMD_SYSBOOT + select ENV_VARS_UBOOT_CONFIG + select HUSH_PARSER + select SUPPORT_RAW_INITRD + select SYS_LONGHELP + imply CMD_MII if NET + imply USB_STORAGE + imply USE_BOOTCOMMAND + help + Select this to enable various options and commands which are suitable + for building u-boot for booting general purpose Linux distributions. + menu "Boot timing"
config BOOTSTAGE
From: Simon Glass sjg@chromium.org
Ramdisk relocation requires LMB, so enable it automatically to avoid build errors.
Signed-off-by: Simon Glass sjg@chromium.org --- Changes in v4: None. --- boot/Kconfig | 1 + 1 file changed, 1 insertion(+)
diff --git a/boot/Kconfig b/boot/Kconfig index 170ab63f27dc..8fa3181d7e23 100644 --- a/boot/Kconfig +++ b/boot/Kconfig @@ -814,6 +814,7 @@ config SYS_BOOT_RAMDISK_HIGH depends on CMD_BOOTM || CMD_BOOTI || CMD_BOOTZ depends on !(NIOS2 || SANDBOX || SH || XTENSA) def_bool y + select LMB help Enable initrd_high functionality. If defined then the initrd_high feature is enabled and the boot* ramdisk subcommand is enabled.
From: Simon Glass sjg@chromium.org
These two features use a lot of common options. Move them into a common CONFIG to reduce duplication.
Use 'select' for most options since these are things that boards aren't supposed to override. For now it is not possible to disable BOOT_DEFAULTS but we may take another look later.
Note that five options use 'imply' to match existing behaviour.
Signed-off-by: Simon Glass sjg@chromium.org [trini: Rework a bit so we don't grow so many platforms unintentionally] Signed-off-by: Tom Rini trini@konsulko.com --- Changes in v4: - Reword the commit message - Rework the logic so that build-wise, nothing changes. --- boot/Kconfig | 62 ++++++++++++++++++++++++---------------------------- 1 file changed, 28 insertions(+), 34 deletions(-)
diff --git a/boot/Kconfig b/boot/Kconfig index 8fa3181d7e23..ad035695a4a3 100644 --- a/boot/Kconfig +++ b/boot/Kconfig @@ -350,6 +350,32 @@ config PXE_UTILS help Utilities for parsing PXE file formats.
+config BOOT_DEFAULTS + bool # Common defaults for standard boot and distroboot + imply USE_BOOTCOMMAND + select CMD_ENV_EXISTS + select CMD_EXT2 + select CMD_EXT4 + select CMD_FAT + select CMD_FS_GENERIC + select CMD_PART if PARTITIONS + select CMD_DHCP if CMD_NET + select CMD_PING if CMD_NET + select CMD_PXE if CMD_NET + select SUPPORT_RAW_INITRD + select ENV_VARS_UBOOT_CONFIG + select CMD_BOOTI if ARM64 + select CMD_BOOTZ if ARM && !ARM64 + imply CMD_MII if NET + imply USB_STORAGE + imply EFI_PARTITION + imply ISO_PARTITION + help + These are not required but are commonly needed to support a good + selection of booting methods. Enable this to improve the capability + of U-Boot to boot various images. Currently much functionality is + tied to enabling the command that exercises it. + config BOOTSTD bool "Standard boot support" default y @@ -410,24 +436,7 @@ config BOOTSTD_DEFAULTS bool "Select some common defaults for standard boot" depends on BOOTSTD imply USE_BOOTCOMMAND - # Bring in some defaults which are generally needed. Boards can drop - # these as needed to save code space. Bootstd does not generally require - # the commands themselves to be enabled, but this is how some of the - # functionality is controlled at present - imply CMD_EXT2 - imply CMD_EXT4 - imply CMD_FAT - imply CMD_FS_GENERIC - imply CMD_PART - imply CMD_DHCP if NET - imply CMD_MII if NET - imply CMD_PING if NET - imply CMD_PXE if NET - imply USB_STORAGE - imply SUPPORT_RAW_INITRD - imply ENV_VARS_UBOOT_CONFIG - imply EFI_PARTITION - imply ISO_PARTITION + select BOOT_DEFAULTS help These are not required but are commonly needed to support a good selection of booting methods. Enable this to improve the capability @@ -823,27 +832,12 @@ endmenu # Boot images
config DISTRO_DEFAULTS bool "Select defaults suitable for booting general purpose Linux distributions" + select BOOT_DEFAULTS select AUTO_COMPLETE select CMDLINE_EDITING - select CMD_BOOTI if ARM64 - select CMD_BOOTZ if ARM && !ARM64 - select CMD_DHCP if CMD_NET - select CMD_ENV_EXISTS - select CMD_EXT2 - select CMD_EXT4 - select CMD_FAT - select CMD_FS_GENERIC - select CMD_PART if PARTITIONS - select CMD_PING if CMD_NET - select CMD_PXE if NET select CMD_SYSBOOT - select ENV_VARS_UBOOT_CONFIG select HUSH_PARSER - select SUPPORT_RAW_INITRD select SYS_LONGHELP - imply CMD_MII if NET - imply USB_STORAGE - imply USE_BOOTCOMMAND help Select this to enable various options and commands which are suitable for building u-boot for booting general purpose Linux distributions.
From: Simon Glass sjg@chromium.org
These board have moved to standard boot but the old 'distro_bootcmd' command is still active. Disable DISTRO_DEFAULTS to fix this.
Signed-off-by: Simon Glass sjg@chromium.org Tested-by: Vagrant Cascadian vagrant@debian.org --- Changes in v4: None. --- arch/arm/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig index bd7fffcce0ba..4e7ebeaee87d 100644 --- a/arch/arm/Kconfig +++ b/arch/arm/Kconfig @@ -1955,7 +1955,7 @@ config ARCH_ROCKCHIP imply ADC imply CMD_DM imply DEBUG_UART_BOARD_INIT - imply DISTRO_DEFAULTS + imply DISTRO_DEFAULTS if !ROCKCHIP_RK3399 imply FAT_WRITE imply SARADC_ROCKCHIP imply SPL_SYSRESET
On 2023-03-24, Tom Rini wrote:
These board have moved to standard boot but the old 'distro_bootcmd' command is still active. Disable DISTRO_DEFAULTS to fix this.
I successfully tested the v4 series against v2023.04-rc4 on rockpro64-rk3399 and pinebook-pro-rk3399, and for good measure double-checked that it did not introduce regressions on rock64-rk3328, because it readily available to test. :)
Looks good to me, thanks!
live well, vagrant
Signed-off-by: Simon Glass sjg@chromium.org Tested-by: Vagrant Cascadian vagrant@debian.org
Changes in v4: None.
arch/arm/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig index bd7fffcce0ba..4e7ebeaee87d 100644 --- a/arch/arm/Kconfig +++ b/arch/arm/Kconfig @@ -1955,7 +1955,7 @@ config ARCH_ROCKCHIP imply ADC imply CMD_DM imply DEBUG_UART_BOARD_INIT
- imply DISTRO_DEFAULTS
- imply DISTRO_DEFAULTS if !ROCKCHIP_RK3399 imply FAT_WRITE imply SARADC_ROCKCHIP imply SPL_SYSRESET
-- 2.34.1
When we do not enable DISTRO_DEFAULTS (generally, to get distro_bootcmd) we instea do want to imply BOOTSTD_DEFAULTS so that when using bootstd the general distro boot functionality will still work.
Signed-off-by: Tom Rini trini@konsulko.com --- Changes in v4: - New patch --- arch/arm/Kconfig | 1 + 1 file changed, 1 insertion(+)
diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig index 4e7ebeaee87d..8a1e22342248 100644 --- a/arch/arm/Kconfig +++ b/arch/arm/Kconfig @@ -1956,6 +1956,7 @@ config ARCH_ROCKCHIP imply CMD_DM imply DEBUG_UART_BOARD_INIT imply DISTRO_DEFAULTS if !ROCKCHIP_RK3399 + imply BOOTSTD_DEFAULTS if !DISTRO_DEFAULTS imply FAT_WRITE imply SARADC_ROCKCHIP imply SPL_SYSRESET
The defined altbootcmd was specific to distro_bootcmd which is not longer in use on these platforms, so drop it.
Signed-off-by: Tom Rini trini@konsulko.com --- Changes in v4: - New patch
Cc: Kever Yang kever.yang@rock-chips.com Cc: Simon Glass sjg@chromium.org Cc: Jonas Karlman jonas@kwiboo.se --- include/configs/rk3399_common.h | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-)
diff --git a/include/configs/rk3399_common.h b/include/configs/rk3399_common.h index 04dcbf20bcb5..96ba19c659bd 100644 --- a/include/configs/rk3399_common.h +++ b/include/configs/rk3399_common.h @@ -52,10 +52,7 @@ "fdtfile=" CONFIG_DEFAULT_FDT_FILE "\0" \ "partitions=" PARTS_DEFAULT \ ROCKCHIP_DEVICE_SETTINGS \ - "boot_targets=" BOOT_TARGETS "\0" \ - "altbootcmd=" \ - "setenv boot_syslinux_conf extlinux/extlinux-rollback.conf;" \ - "run distro_bootcmd\0" + "boot_targets=" BOOT_TARGETS "\0"
#endif
Hi Tom,
On 2023-03-24 21:58, Tom Rini wrote:
The defined altbootcmd was specific to distro_bootcmd which is not longer in use on these platforms, so drop it.
Signed-off-by: Tom Rini trini@konsulko.com
Changes in v4:
- New patch
Cc: Kever Yang kever.yang@rock-chips.com Cc: Simon Glass sjg@chromium.org Cc: Jonas Karlman jonas@kwiboo.se
include/configs/rk3399_common.h | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-)
diff --git a/include/configs/rk3399_common.h b/include/configs/rk3399_common.h index 04dcbf20bcb5..96ba19c659bd 100644 --- a/include/configs/rk3399_common.h +++ b/include/configs/rk3399_common.h @@ -52,10 +52,7 @@ "fdtfile=" CONFIG_DEFAULT_FDT_FILE "\0" \ "partitions=" PARTS_DEFAULT \ ROCKCHIP_DEVICE_SETTINGS \
- "boot_targets=" BOOT_TARGETS "\0" \
- "altbootcmd=" \
"setenv boot_syslinux_conf extlinux/extlinux-rollback.conf;" \"run distro_bootcmd\0"
- "boot_targets=" BOOT_TARGETS "\0"
#endif
With this series my rk3399 boards stop trying to use distro_bootcmd and will now successfully boot using default bootcmd=bootflow scan.
So for the entire series:
Tested-by: Jonas Karlman jonas@kwiboo.se Reviewed-by: Jonas Karlman jonas@kwiboo.se
Regards, Jonas
Hi Tom,
On Sat, 25 Mar 2023 at 09:58, Tom Rini trini@konsulko.com wrote:
Hey all,
I took a look at Simon's v3 series to fix the rk3399 bootstd migration, and it changed too much for everything else. I took about half of that series and then reworked a few things. Now only rk3399 platforms change at all and aside from bootcmd changes, the only thing is they now disable true/test/sysboot/showvar/false/exit commands as those were being pulled in from distro and now we don't set that flag. I think the way I changed how we enable BOOTSTD_DEFAULTS should make it easier to perform more SoC migrations.
Thanks for digging into this. I haven't seen any comments on the rpi conversion, so perhaps people could test that?
Regards, Simon
On Mon, Mar 27, 2023 at 05:00:31PM +1300, Simon Glass wrote:
Hi Tom,
On Sat, 25 Mar 2023 at 09:58, Tom Rini trini@konsulko.com wrote:
Hey all,
I took a look at Simon's v3 series to fix the rk3399 bootstd migration, and it changed too much for everything else. I took about half of that series and then reworked a few things. Now only rk3399 platforms change at all and aside from bootcmd changes, the only thing is they now disable true/test/sysboot/showvar/false/exit commands as those were being pulled in from distro and now we don't set that flag. I think the way I changed how we enable BOOTSTD_DEFAULTS should make it easier to perform more SoC migrations.
Thanks for digging into this. I haven't seen any comments on the rpi conversion, so perhaps people could test that?
Well, it needs to be rebased on top of this series (which will be merged in the next -rc today).
On Mon, Mar 27, 2023 at 5:02 AM Simon Glass sjg@chromium.org wrote:
Hi Tom,
On Sat, 25 Mar 2023 at 09:58, Tom Rini trini@konsulko.com wrote:
Hey all,
I took a look at Simon's v3 series to fix the rk3399 bootstd migration, and it changed too much for everything else. I took about half of that series and then reworked a few things. Now only rk3399 platforms change at all and aside from bootcmd changes, the only thing is they now disable true/test/sysboot/showvar/false/exit commands as those were being pulled in from distro and now we don't set that flag. I think the way I changed how we enable BOOTSTD_DEFAULTS should make it easier to perform more SoC migrations.
Thanks for digging into this. I haven't seen any comments on the rpi conversion, so perhaps people could test that?
I was planning on looking at that once 2023.04 was out but TBH I have wasted so much time over the last few cycles dealing with regressions through a bunch of these series that I now have so little time for enhancements I now shy away. I know a lot of these series should improve things in the future but they don't feel like when there's unnecessary changes for things that are clearly untested.
There's also a lot of change for changes sake, for example the rockchips ATF binaries needed is called bl31.elf by the default output of the ATF build process, for others it's bl31.bin, binman for what ever reason has changed that to be atf-bl31, now I have to change the entire build process to be able to work out what is what on a board by board basis to be able to set the required variable to be able to specify the ATF where previously it "just worked (tm)"..... I suppose there is some perceived goal and improvement here but with both my "U-Boot device maintainer" and "distro maintainer" hats on, both of which I do in my own spare time, I currently fail to see it and I end up.
All that said, thank you Tom for picking up the pieces for something which should have been actually working when it landed.
Tested-by: Peter Robinson pbrobinson@gmail.com
Now I get to go and work out all of the rest of the mess!
Regards, Peter
Hi Peter,
On Tue, 28 Mar 2023 at 06:50, Peter Robinson pbrobinson@gmail.com wrote:
On Mon, Mar 27, 2023 at 5:02 AM Simon Glass sjg@chromium.org wrote:
Hi Tom,
On Sat, 25 Mar 2023 at 09:58, Tom Rini trini@konsulko.com wrote:
Hey all,
I took a look at Simon's v3 series to fix the rk3399 bootstd migration, and it changed too much for everything else. I took about half of that series and then reworked a few things. Now only rk3399 platforms change at all and aside from bootcmd changes, the only thing is they now disable true/test/sysboot/showvar/false/exit commands as those were being pulled in from distro and now we don't set that flag. I think the way I changed how we enable BOOTSTD_DEFAULTS should make it easier to perform more SoC migrations.
Thanks for digging into this. I haven't seen any comments on the rpi conversion, so perhaps people could test that?
I was planning on looking at that once 2023.04 was out but TBH I have wasted so much time over the last few cycles dealing with regressions through a bunch of these series that I now have so little time for enhancements I now shy away. I know a lot of these series should improve things in the future but they don't feel like when there's unnecessary changes for things that are clearly untested.
There's also a lot of change for changes sake, for example the rockchips ATF binaries needed is called bl31.elf by the default output of the ATF build process, for others it's bl31.bin, binman for what ever reason has changed that to be atf-bl31, now I have to change the entire build process to be able to work out what is what on a board by board basis to be able to set the required variable to be able to specify the ATF where previously it "just worked (tm)"..... I suppose
For this point, you could use the BL31 environment variable, which would allow the old file to be used. That is in the instructions for some boards.
Also, the change to atf-bl31 is because no particular filename is provided as a default, so we end up using the entry type. I suppose the problem is that there are two names in common use (bl31.bin and bl31.elf) and if we use the wrong one it won't boot. That is an unfortunate result of how things work with ATF. But in any case this is a decision for the SoC maintainer, who can provide a default filename if desired, in the binman description for that SoC.
there is some perceived goal and improvement here but with both my "U-Boot device maintainer" and "distro maintainer" hats on, both of which I do in my own spare time, I currently fail to see it and I end up.
Basically the benefit is that U-Boot is not full of loads of strange shell / Python scripts for each SoC type, it is possible (at least in principle) to figure out how to build an image without scanning the web for vendor-specific instructions, tools which are needed to build images are registered (binman tool -l) and can be fetched. Overall, having a data-driven approach to firmware packaging is vastly superior to a code-based approach, particularly as firmware fragments more and more. You can find more here [1]
All that said, thank you Tom for picking up the pieces for something which should have been actually working when it landed.
Tested-by: Peter Robinson pbrobinson@gmail.com
Now I get to go and work out all of the rest of the mess!
Yes, thank you Tom.
Thank you also Peter for your comments. I suspect a lot of people feel the same way.
From my perspective, these migrations can be exhausting, particularly
when drawn out over a long period of time. Removing SPL_FIT_GENERATOR started almost 3 years ago[2]. People continued adding new boards to it even until recently.
It is also often difficult to predict how things will turn out, even for people with many years of experience in software development. So we sometimes make wrong turns.
I would love to see more attention from SoC maintainers, to sit down with a coffee once a month and take a hard look at the state of the code they look after, what migrations are outstanding, etc.
The firmware world is changing rapidly. We need to be able to keep on top of the increasing complexity with new tools and techniques. Perhaps it will settle down at some point, but for now it is quite a daunting task and we all need to pitch in.
Regards, Simon
[1] https://u-boot.readthedocs.io/en/latest/develop/package/binman.html#motivati... [2] https://patchwork.ozlabs.org/project/uboot/patch/20200613205717.v2.42.I2428d...
On Mon, Mar 27, 2023 at 06:50:41PM +0100, Peter Robinson wrote:
On Mon, Mar 27, 2023 at 5:02 AM Simon Glass sjg@chromium.org wrote:
Hi Tom,
On Sat, 25 Mar 2023 at 09:58, Tom Rini trini@konsulko.com wrote:
Hey all,
I took a look at Simon's v3 series to fix the rk3399 bootstd migration, and it changed too much for everything else. I took about half of that series and then reworked a few things. Now only rk3399 platforms change at all and aside from bootcmd changes, the only thing is they now disable true/test/sysboot/showvar/false/exit commands as those were being pulled in from distro and now we don't set that flag. I think the way I changed how we enable BOOTSTD_DEFAULTS should make it easier to perform more SoC migrations.
Thanks for digging into this. I haven't seen any comments on the rpi conversion, so perhaps people could test that?
I was planning on looking at that once 2023.04 was out but TBH I have wasted so much time over the last few cycles dealing with regressions through a bunch of these series that I now have so little time for enhancements I now shy away. I know a lot of these series should improve things in the future but they don't feel like when there's unnecessary changes for things that are clearly untested.
I too am unhappy with how some of these have gone. The _intent_ here is that getting the current "boot generic distro" framework is complex / error prone, and we can do better. Unfortunately the first set of platforms to switch to this are Rockchip and I think there was overlap there with platforms that got broken at the end of the v2023.01 cycle to fix other platforms, and then those sets of platforms flipped early in v2023.04 and took until -rc2? to get resolved. Which was less than ideal.
There's also a lot of change for changes sake, for example the rockchips ATF binaries needed is called bl31.elf by the default output of the ATF build process, for others it's bl31.bin, binman for what ever reason has changed that to be atf-bl31, now I have to change the entire build process to be able to work out what is what on a board by board basis to be able to set the required variable to be able to specify the ATF where previously it "just worked (tm)"..... I suppose there is some perceived goal and improvement here but with both my "U-Boot device maintainer" and "distro maintainer" hats on, both of which I do in my own spare time, I currently fail to see it and I end up.
I wish I knew where to talk to with ATF / TF-A to get some agreed upon naming scheme going as one of the things that is very frustrating is getting the names and combinations of everything else that's required Just Right for every chip. And feedback that things aren't working is appreciated, since we do need to make things easier.
On Wed, Mar 29, 2023 at 3:54 PM Tom Rini trini@konsulko.com wrote:
On Mon, Mar 27, 2023 at 06:50:41PM +0100, Peter Robinson wrote:
On Mon, Mar 27, 2023 at 5:02 AM Simon Glass sjg@chromium.org wrote:
Hi Tom,
On Sat, 25 Mar 2023 at 09:58, Tom Rini trini@konsulko.com wrote:
Hey all,
I took a look at Simon's v3 series to fix the rk3399 bootstd migration, and it changed too much for everything else. I took about half of that series and then reworked a few things. Now only rk3399 platforms change at all and aside from bootcmd changes, the only thing is they now disable true/test/sysboot/showvar/false/exit commands as those were being pulled in from distro and now we don't set that flag. I think the way I changed how we enable BOOTSTD_DEFAULTS should make it easier to perform more SoC migrations.
Thanks for digging into this. I haven't seen any comments on the rpi conversion, so perhaps people could test that?
I was planning on looking at that once 2023.04 was out but TBH I have wasted so much time over the last few cycles dealing with regressions through a bunch of these series that I now have so little time for enhancements I now shy away. I know a lot of these series should improve things in the future but they don't feel like when there's unnecessary changes for things that are clearly untested.
I too am unhappy with how some of these have gone. The _intent_ here is that getting the current "boot generic distro" framework is complex / error prone, and we can do better. Unfortunately the first set of platforms to switch to this are Rockchip and I think there was overlap there with platforms that got broken at the end of the v2023.01 cycle to fix other platforms, and then those sets of platforms flipped early in v2023.04 and took until -rc2? to get resolved. Which was less than ideal.
There's also a lot of change for changes sake, for example the rockchips ATF binaries needed is called bl31.elf by the default output of the ATF build process, for others it's bl31.bin, binman for what ever reason has changed that to be atf-bl31, now I have to change the entire build process to be able to work out what is what on a board by board basis to be able to set the required variable to be able to specify the ATF where previously it "just worked (tm)"..... I suppose there is some perceived goal and improvement here but with both my "U-Boot device maintainer" and "distro maintainer" hats on, both of which I do in my own spare time, I currently fail to see it and I end up.
I wish I knew where to talk to with ATF / TF-A to get some agreed upon naming scheme going as one of the things that is very frustrating is getting the names and combinations of everything else that's required Just Right for every chip. And feedback that things aren't working is appreciated, since we do need to make things easier.
In all of the various make_fit_atf.py the various vendors specified them, this is the case for the rockchip one [1]. This is the case for the Allwinner boards [2] but the rockchip ports have missed this so it also should be fixed for GA.
A side point is that binman should not be storing firmware build specifics in the device tree which is a means of describing the hardware, This really needs to be fixed as it really isn't the right place for that sort of things. I suspect a file in arch/arm/mach-<SOC> is likely a better location, or if it's board specific in the board/ sub directory.
Peter
[1] https://source.denx.de/u-boot/u-boot/-/blob/v2023.01/arch/arm/mach-rockchip/... [2] https://source.denx.de/u-boot/u-boot/-/blob/master/arch/arm/dts/sunxi-u-boot...
Hi Peter,
On Thu, 30 Mar 2023 at 19:56, Peter Robinson pbrobinson@gmail.com wrote:
On Wed, Mar 29, 2023 at 3:54 PM Tom Rini trini@konsulko.com wrote:
On Mon, Mar 27, 2023 at 06:50:41PM +0100, Peter Robinson wrote:
On Mon, Mar 27, 2023 at 5:02 AM Simon Glass sjg@chromium.org wrote:
Hi Tom,
On Sat, 25 Mar 2023 at 09:58, Tom Rini trini@konsulko.com wrote:
Hey all,
I took a look at Simon's v3 series to fix the rk3399 bootstd migration, and it changed too much for everything else. I took about half of that series and then reworked a few things. Now only rk3399 platforms change at all and aside from bootcmd changes, the only thing is they now disable true/test/sysboot/showvar/false/exit commands as those were being pulled in from distro and now we don't set that flag. I think the way I changed how we enable BOOTSTD_DEFAULTS should make it easier to perform more SoC migrations.
Thanks for digging into this. I haven't seen any comments on the rpi conversion, so perhaps people could test that?
I was planning on looking at that once 2023.04 was out but TBH I have wasted so much time over the last few cycles dealing with regressions through a bunch of these series that I now have so little time for enhancements I now shy away. I know a lot of these series should improve things in the future but they don't feel like when there's unnecessary changes for things that are clearly untested.
I too am unhappy with how some of these have gone. The _intent_ here is that getting the current "boot generic distro" framework is complex / error prone, and we can do better. Unfortunately the first set of platforms to switch to this are Rockchip and I think there was overlap there with platforms that got broken at the end of the v2023.01 cycle to fix other platforms, and then those sets of platforms flipped early in v2023.04 and took until -rc2? to get resolved. Which was less than ideal.
There's also a lot of change for changes sake, for example the rockchips ATF binaries needed is called bl31.elf by the default output of the ATF build process, for others it's bl31.bin, binman for what ever reason has changed that to be atf-bl31, now I have to change the entire build process to be able to work out what is what on a board by board basis to be able to set the required variable to be able to specify the ATF where previously it "just worked (tm)"..... I suppose there is some perceived goal and improvement here but with both my "U-Boot device maintainer" and "distro maintainer" hats on, both of which I do in my own spare time, I currently fail to see it and I end up.
I wish I knew where to talk to with ATF / TF-A to get some agreed upon naming scheme going as one of the things that is very frustrating is getting the names and combinations of everything else that's required Just Right for every chip. And feedback that things aren't working is appreciated, since we do need to make things easier.
In all of the various make_fit_atf.py the various vendors specified them, this is the case for the rockchip one [1]. This is the case for the Allwinner boards [2] but the rockchip ports have missed this so it also should be fixed for GA.
A side point is that binman should not be storing firmware build specifics in the device tree which is a means of describing the hardware, This really needs to be fixed as it really isn't the right place for that sort of things. I suspect a file in arch/arm/mach-<SOC> is likely a better location, or if it's board specific in the board/ sub directory.
Sorry, I don't agree with that at all. We store configuration information in devicetree in firmware as this seems to be best format for it, particularly with the growing number of firmware components that need to share this information at runtime. The layout of firmware is an important part of the system. We are still figuring out the flows though. Also I have not attempted to upstream the binman binding. I am very open to ideas on how best to do that.
Regards, Simon
Peter
[1] https://source.denx.de/u-boot/u-boot/-/blob/v2023.01/arch/arm/mach-rockchip/... [2] https://source.denx.de/u-boot/u-boot/-/blob/master/arch/arm/dts/sunxi-u-boot...
Hi Simon,
I took a look at Simon's v3 series to fix the rk3399 bootstd migration, and it changed too much for everything else. I took about half of that series and then reworked a few things. Now only rk3399 platforms change at all and aside from bootcmd changes, the only thing is they now disable true/test/sysboot/showvar/false/exit commands as those were being pulled in from distro and now we don't set that flag. I think the way I changed how we enable BOOTSTD_DEFAULTS should make it easier to perform more SoC migrations.
Thanks for digging into this. I haven't seen any comments on the rpi conversion, so perhaps people could test that?
I was planning on looking at that once 2023.04 was out but TBH I have wasted so much time over the last few cycles dealing with regressions through a bunch of these series that I now have so little time for enhancements I now shy away. I know a lot of these series should improve things in the future but they don't feel like when there's unnecessary changes for things that are clearly untested.
I too am unhappy with how some of these have gone. The _intent_ here is that getting the current "boot generic distro" framework is complex / error prone, and we can do better. Unfortunately the first set of platforms to switch to this are Rockchip and I think there was overlap there with platforms that got broken at the end of the v2023.01 cycle to fix other platforms, and then those sets of platforms flipped early in v2023.04 and took until -rc2? to get resolved. Which was less than ideal.
There's also a lot of change for changes sake, for example the rockchips ATF binaries needed is called bl31.elf by the default output of the ATF build process, for others it's bl31.bin, binman for what ever reason has changed that to be atf-bl31, now I have to change the entire build process to be able to work out what is what on a board by board basis to be able to set the required variable to be able to specify the ATF where previously it "just worked (tm)"..... I suppose there is some perceived goal and improvement here but with both my "U-Boot device maintainer" and "distro maintainer" hats on, both of which I do in my own spare time, I currently fail to see it and I end up.
I wish I knew where to talk to with ATF / TF-A to get some agreed upon naming scheme going as one of the things that is very frustrating is getting the names and combinations of everything else that's required Just Right for every chip. And feedback that things aren't working is appreciated, since we do need to make things easier.
In all of the various make_fit_atf.py the various vendors specified them, this is the case for the rockchip one [1]. This is the case for the Allwinner boards [2] but the rockchip ports have missed this so it also should be fixed for GA.
Can you do a patch to fix this regression please and then specify the correct pieces in the binman section then?
A side point is that binman should not be storing firmware build specifics in the device tree which is a means of describing the hardware, This really needs to be fixed as it really isn't the right place for that sort of things. I suspect a file in arch/arm/mach-<SOC> is likely a better location, or if it's board specific in the board/ sub directory.
Sorry, I don't agree with that at all. We store configuration information in devicetree in firmware as this seems to be best format for it, particularly with the growing number of firmware components that need to share this information at runtime. The layout of firmware is an important part of the system. We are still figuring out the flows though. Also I have not attempted to upstream the binman binding. I am very open to ideas on how best to do that.
Rob what's your thoughts on the binman firmware build pieces being in device tree and the process on upstreaming the bindings?
Regards, Peter
[1] https://source.denx.de/u-boot/u-boot/-/blob/v2023.01/arch/arm/mach-rockchip/... [2] https://source.denx.de/u-boot/u-boot/-/blob/master/arch/arm/dts/sunxi-u-boot...
On Fri, Mar 31, 2023 at 09:39:53AM +0100, Peter Robinson wrote:
Hi Simon,
> I took a look at Simon's v3 series to fix the rk3399 bootstd migration, > and it changed too much for everything else. I took about half of that > series and then reworked a few things. Now only rk3399 platforms change > at all and aside from bootcmd changes, the only thing is they now > disable true/test/sysboot/showvar/false/exit commands as those were > being pulled in from distro and now we don't set that flag. I think the > way I changed how we enable BOOTSTD_DEFAULTS should make it easier to > perform more SoC migrations.
Thanks for digging into this. I haven't seen any comments on the rpi conversion, so perhaps people could test that?
I was planning on looking at that once 2023.04 was out but TBH I have wasted so much time over the last few cycles dealing with regressions through a bunch of these series that I now have so little time for enhancements I now shy away. I know a lot of these series should improve things in the future but they don't feel like when there's unnecessary changes for things that are clearly untested.
I too am unhappy with how some of these have gone. The _intent_ here is that getting the current "boot generic distro" framework is complex / error prone, and we can do better. Unfortunately the first set of platforms to switch to this are Rockchip and I think there was overlap there with platforms that got broken at the end of the v2023.01 cycle to fix other platforms, and then those sets of platforms flipped early in v2023.04 and took until -rc2? to get resolved. Which was less than ideal.
There's also a lot of change for changes sake, for example the rockchips ATF binaries needed is called bl31.elf by the default output of the ATF build process, for others it's bl31.bin, binman for what ever reason has changed that to be atf-bl31, now I have to change the entire build process to be able to work out what is what on a board by board basis to be able to set the required variable to be able to specify the ATF where previously it "just worked (tm)"..... I suppose there is some perceived goal and improvement here but with both my "U-Boot device maintainer" and "distro maintainer" hats on, both of which I do in my own spare time, I currently fail to see it and I end up.
I wish I knew where to talk to with ATF / TF-A to get some agreed upon naming scheme going as one of the things that is very frustrating is getting the names and combinations of everything else that's required Just Right for every chip. And feedback that things aren't working is appreciated, since we do need to make things easier.
In all of the various make_fit_atf.py the various vendors specified them, this is the case for the rockchip one [1]. This is the case for the Allwinner boards [2] but the rockchip ports have missed this so it also should be fixed for GA.
Can you do a patch to fix this regression please and then specify the correct pieces in the binman section then?
A side point is that binman should not be storing firmware build specifics in the device tree which is a means of describing the hardware, This really needs to be fixed as it really isn't the right place for that sort of things. I suspect a file in arch/arm/mach-<SOC> is likely a better location, or if it's board specific in the board/ sub directory.
Sorry, I don't agree with that at all. We store configuration information in devicetree in firmware as this seems to be best format for it, particularly with the growing number of firmware components that need to share this information at runtime. The layout of firmware is an important part of the system. We are still figuring out the flows though. Also I have not attempted to upstream the binman binding. I am very open to ideas on how best to do that.
Rob what's your thoughts on the binman firmware build pieces being in device tree and the process on upstreaming the bindings?
One question I have is, with your distribution person hat on, what do you think the right answer here is? It's not that U-Boot requires TF-A (and also other blobs depending on SoC) but rather that modern ARM SoCs are like x86-64 SoCs have been for ages and require multiple parts to get assembled "just so", to result in something that initializes enough of the system to get to loading something that can load the OS, and also provide assorted run time services. You have one set of fixups for amlogic SoCs, another for sunxi, yet another for imx8* and so on. And so long as the software does the job, some of those parts can be swapped for different ones. U-Boot has binman *stuff* because it's really useful for us to be able to take an output and test our part of the world. And since it's gotten more and more complex over the years, we've moved from handfuls of one-off tools to binman with the intention of making it easier, not harder, for the case of "I care about multiple vastly different SoCs".
Hi Peter,
On Fri, 31 Mar 2023 at 21:40, Peter Robinson pbrobinson@gmail.com wrote:
Hi Simon,
> I took a look at Simon's v3 series to fix the rk3399 bootstd migration, > and it changed too much for everything else. I took about half of that > series and then reworked a few things. Now only rk3399 platforms change > at all and aside from bootcmd changes, the only thing is they now > disable true/test/sysboot/showvar/false/exit commands as those were > being pulled in from distro and now we don't set that flag. I think the > way I changed how we enable BOOTSTD_DEFAULTS should make it easier to > perform more SoC migrations.
Thanks for digging into this. I haven't seen any comments on the rpi conversion, so perhaps people could test that?
I was planning on looking at that once 2023.04 was out but TBH I have wasted so much time over the last few cycles dealing with regressions through a bunch of these series that I now have so little time for enhancements I now shy away. I know a lot of these series should improve things in the future but they don't feel like when there's unnecessary changes for things that are clearly untested.
I too am unhappy with how some of these have gone. The _intent_ here is that getting the current "boot generic distro" framework is complex / error prone, and we can do better. Unfortunately the first set of platforms to switch to this are Rockchip and I think there was overlap there with platforms that got broken at the end of the v2023.01 cycle to fix other platforms, and then those sets of platforms flipped early in v2023.04 and took until -rc2? to get resolved. Which was less than ideal.
There's also a lot of change for changes sake, for example the rockchips ATF binaries needed is called bl31.elf by the default output of the ATF build process, for others it's bl31.bin, binman for what ever reason has changed that to be atf-bl31, now I have to change the entire build process to be able to work out what is what on a board by board basis to be able to set the required variable to be able to specify the ATF where previously it "just worked (tm)"..... I suppose there is some perceived goal and improvement here but with both my "U-Boot device maintainer" and "distro maintainer" hats on, both of which I do in my own spare time, I currently fail to see it and I end up.
I wish I knew where to talk to with ATF / TF-A to get some agreed upon naming scheme going as one of the things that is very frustrating is getting the names and combinations of everything else that's required Just Right for every chip. And feedback that things aren't working is appreciated, since we do need to make things easier.
In all of the various make_fit_atf.py the various vendors specified them, this is the case for the rockchip one [1]. This is the case for the Allwinner boards [2] but the rockchip ports have missed this so it also should be fixed for GA.
Can you do a patch to fix this regression please and then specify the correct pieces in the binman section then?
Yes I think this should be fixed.
We don't have any Rockchip maintainers / contributors on this thread. Would you like to start a new one, or add them to this thread?
A side point is that binman should not be storing firmware build specifics in the device tree which is a means of describing the hardware, This really needs to be fixed as it really isn't the right place for that sort of things. I suspect a file in arch/arm/mach-<SOC> is likely a better location, or if it's board specific in the board/ sub directory.
Sorry, I don't agree with that at all. We store configuration information in devicetree in firmware as this seems to be best format for it, particularly with the growing number of firmware components that need to share this information at runtime. The layout of firmware is an important part of the system. We are still figuring out the flows though. Also I have not attempted to upstream the binman binding. I am very open to ideas on how best to do that.
Rob what's your thoughts on the binman firmware build pieces being in device tree and the process on upstreaming the bindings?
It might be easier for Rob to comment on an actual proposal, which I have not done. It is on my radar though.
Regards, Simon
Regards, Peter
[1] https://source.denx.de/u-boot/u-boot/-/blob/v2023.01/arch/arm/mach-rockchip/... [2] https://source.denx.de/u-boot/u-boot/-/blob/master/arch/arm/dts/sunxi-u-boot...
On Fri, 24 Mar 2023 16:58:09 -0400, Tom Rini wrote:
Hey all,
I took a look at Simon's v3 series to fix the rk3399 bootstd migration, and it changed too much for everything else. I took about half of that series and then reworked a few things. Now only rk3399 platforms change at all and aside from bootcmd changes, the only thing is they now disable true/test/sysboot/showvar/false/exit commands as those were being pulled in from distro and now we don't set that flag. I think the way I changed how we enable BOOTSTD_DEFAULTS should make it easier to perform more SoC migrations.
[...]
Applied to u-boot/master, thanks!
participants (5)
-
Jonas Karlman -
Peter Robinson -
Simon Glass -
Tom Rini -
Vagrant Cascadian