Re: [U-Boot] [PATCH v3 0/17] sandbox: Generic board support and other improvements
On Fri, Apr 26, 2013 at 06:10:15AM -0700, Simon Glass wrote: Hi Tom, On Mon, Apr 22, 2013 at 9:08 AM, Simon Glass s...@chromium.org wrote: Hi Tom, On Mon, Apr 22, 2013 at 8:18 AM, Tom Rini tr...@ti.com wrote: On Sat, Apr 20, 2013 at 11:42:35AM -0700, Simon Glass wrote: This series adds generic board support to sandbox and switches to use this always. With sandbox it was noticed that turning CONFIG_SYS_GENERIC_BOARD off can cause a build failure if a previous autoconf.mk exists which indicates that generic board is not supported, so a patch is provided to fix this. It is useful to convert a pointer into an 'address' in the sandbox RAM buffer - the opposite of map_sysmem(). This is added in this series and used in several places. With sandbox it is easier to read a file from the host than to use the CONFIG_OF_SEPARATE option, since this option requires knowledge of the executable image structure which is not really appropriate on the host system. A new CONFIG_OF_HOSTFILE provides this. A few related FDT changes are included in this series also. The -c option is enhanced to support passing entire scripts to sandbox. This is useful when writing non-trivial test code. Most of these patches were previously submitted as part of the verified boot effort. This series collects the independent sandbox-related patches together to make it easier to review. THe whole series is marked as version 3 for this reason. For the series, Reviewed-by: Tom Rini tr...@ti.com And I'd say 3/4/5 should be squashed into one patch, but it's your arch so I'l defer if you think it adds bisect value or similar to do it in that manner. I did that so that it could be kind-of an example of how this can be done for an arch, given that I am not planning to convert the rest. By removing the dead code in a separate step it seemed a bit clearer to me. But it's fine either way - I will squash it and resend. I have put this series in patchwork as: http://patchwork.ozlabs.org/bundle/sjg/sandbox/ This has now been applied to u-boot/master, thanks! and below is a pull request if you want to take that instead. I did not go through and add your Reviewed-by to each patch. Am I supposed to do that? No, that pain is supposed to help spur us into improving patchwork or the new tool we talked about back at LSM. -- Tom signature.asc Description: Digital signature ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] [PATCH v3 0/17] sandbox: Generic board support and other improvements
Hi Tom, On Mon, Apr 22, 2013 at 9:08 AM, Simon Glass s...@chromium.org wrote: Hi Tom, On Mon, Apr 22, 2013 at 8:18 AM, Tom Rini tr...@ti.com wrote: On Sat, Apr 20, 2013 at 11:42:35AM -0700, Simon Glass wrote: This series adds generic board support to sandbox and switches to use this always. With sandbox it was noticed that turning CONFIG_SYS_GENERIC_BOARD off can cause a build failure if a previous autoconf.mk exists which indicates that generic board is not supported, so a patch is provided to fix this. It is useful to convert a pointer into an 'address' in the sandbox RAM buffer - the opposite of map_sysmem(). This is added in this series and used in several places. With sandbox it is easier to read a file from the host than to use the CONFIG_OF_SEPARATE option, since this option requires knowledge of the executable image structure which is not really appropriate on the host system. A new CONFIG_OF_HOSTFILE provides this. A few related FDT changes are included in this series also. The -c option is enhanced to support passing entire scripts to sandbox. This is useful when writing non-trivial test code. Most of these patches were previously submitted as part of the verified boot effort. This series collects the independent sandbox-related patches together to make it easier to review. THe whole series is marked as version 3 for this reason. For the series, Reviewed-by: Tom Rini tr...@ti.com And I'd say 3/4/5 should be squashed into one patch, but it's your arch so I'l defer if you think it adds bisect value or similar to do it in that manner. I did that so that it could be kind-of an example of how this can be done for an arch, given that I am not planning to convert the rest. By removing the dead code in a separate step it seemed a bit clearer to me. But it's fine either way - I will squash it and resend. I have put this series in patchwork as: http://patchwork.ozlabs.org/bundle/sjg/sandbox/ and below is a pull request if you want to take that instead. I did not go through and add your Reviewed-by to each patch. Am I supposed to do that? The following changes since commit d10f68ae47b67acab8b110b5c605dde4197a1820: Prepare v2013.04 (2013-04-19 10:25:43 -0400) are available in the git repository at: git://git.denx.de/u-boot-x86.git sandbox for you to fetch changes up to 3a95570c609a35f60d4c798948f83f8427544ff2: sandbox: config: Enable CONFIG_FIT and CONFIG_CMD_FIT (2013-04-26 06:00:06 -0700) Simon Glass (15): Trigger generic board error only when building sandbox: Provide a way to map from host RAM to U-Boot RAM sandbox: Switch over to generic board sandbox: Add CONFIG_OF_HOSTFILE to read FDT from host file fdt: Add a parameter to fdt_valid() Add getenv_hex() to return an environment variable as hex fdt: Allow fdt command to check and update control FDT sandbox: fdt: Support fdt command for sandbox fdt: Skip checking FDT if the pointer is NULL Revert fdt- Tell the FDT library where the device tree is sandbox: Allow -c argument to provide a command list sandbox: Support 'source' command fs: Add support for saving data to filesystems sandbox: fs: Add support for saving files to host filesystem sandbox: config: Enable CONFIG_FIT and CONFIG_CMD_FIT Makefile | 10 ++ arch/sandbox/config.mk| 5 +- arch/sandbox/cpu/cpu.c| 5 + arch/sandbox/cpu/start.c | 9 +- arch/sandbox/include/asm/io.h | 8 ++ arch/sandbox/include/asm/state.h | 1 + arch/sandbox/include/asm/u-boot.h | 22 +-- arch/sandbox/lib/Makefile | 1 - arch/sandbox/lib/board.c | 285 -- board/sandbox/sandbox/sandbox.c | 2 +- common/board_f.c | 106 -- common/board_r.c | 8 +- common/cmd_fdt.c | 87 common/cmd_nvedit.c | 15 ++ common/cmd_sandbox.c | 18 ++- common/cmd_source.c | 11 +- common/main.c | 8 -- config.mk | 4 +- doc/README.fdt-control| 6 +- fs/fs.c | 75 ++ fs/sandbox/sandboxfs.c| 33 + include/asm-generic/sections.h| 2 +- include/common.h | 18 +++ include/configs/sandbox.h | 11 +- include/fs.h | 2 + include/sandboxfs.h | 1 + lib/fdtdec.c | 5 +- 27 files changed, 383 insertions(+), 375 deletions(-) delete mode 100644 arch/sandbox/lib/board.c Regards, Simon ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] [PATCH v3 0/17] sandbox: Generic board support and other improvements
On Sat, Apr 20, 2013 at 11:42:35AM -0700, Simon Glass wrote: This series adds generic board support to sandbox and switches to use this always. With sandbox it was noticed that turning CONFIG_SYS_GENERIC_BOARD off can cause a build failure if a previous autoconf.mk exists which indicates that generic board is not supported, so a patch is provided to fix this. It is useful to convert a pointer into an 'address' in the sandbox RAM buffer - the opposite of map_sysmem(). This is added in this series and used in several places. With sandbox it is easier to read a file from the host than to use the CONFIG_OF_SEPARATE option, since this option requires knowledge of the executable image structure which is not really appropriate on the host system. A new CONFIG_OF_HOSTFILE provides this. A few related FDT changes are included in this series also. The -c option is enhanced to support passing entire scripts to sandbox. This is useful when writing non-trivial test code. Most of these patches were previously submitted as part of the verified boot effort. This series collects the independent sandbox-related patches together to make it easier to review. THe whole series is marked as version 3 for this reason. For the series, Reviewed-by: Tom Rini tr...@ti.com And I'd say 3/4/5 should be squashed into one patch, but it's your arch so I'l defer if you think it adds bisect value or similar to do it in that manner. -- Tom signature.asc Description: Digital signature ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] [PATCH v3 0/17] sandbox: Generic board support and other improvements
Hi Tom, On Mon, Apr 22, 2013 at 8:18 AM, Tom Rini tr...@ti.com wrote: On Sat, Apr 20, 2013 at 11:42:35AM -0700, Simon Glass wrote: This series adds generic board support to sandbox and switches to use this always. With sandbox it was noticed that turning CONFIG_SYS_GENERIC_BOARD off can cause a build failure if a previous autoconf.mk exists which indicates that generic board is not supported, so a patch is provided to fix this. It is useful to convert a pointer into an 'address' in the sandbox RAM buffer - the opposite of map_sysmem(). This is added in this series and used in several places. With sandbox it is easier to read a file from the host than to use the CONFIG_OF_SEPARATE option, since this option requires knowledge of the executable image structure which is not really appropriate on the host system. A new CONFIG_OF_HOSTFILE provides this. A few related FDT changes are included in this series also. The -c option is enhanced to support passing entire scripts to sandbox. This is useful when writing non-trivial test code. Most of these patches were previously submitted as part of the verified boot effort. This series collects the independent sandbox-related patches together to make it easier to review. THe whole series is marked as version 3 for this reason. For the series, Reviewed-by: Tom Rini tr...@ti.com And I'd say 3/4/5 should be squashed into one patch, but it's your arch so I'l defer if you think it adds bisect value or similar to do it in that manner. I did that so that it could be kind-of an example of how this can be done for an arch, given that I am not planning to convert the rest. By removing the dead code in a separate step it seemed a bit clearer to me. But it's fine either way - I will squash it and resend. -- Tom Regards, Simon ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
[U-Boot] [PATCH v3 0/17] sandbox: Generic board support and other improvements
This series adds generic board support to sandbox and switches to use this always. With sandbox it was noticed that turning CONFIG_SYS_GENERIC_BOARD off can cause a build failure if a previous autoconf.mk exists which indicates that generic board is not supported, so a patch is provided to fix this. It is useful to convert a pointer into an 'address' in the sandbox RAM buffer - the opposite of map_sysmem(). This is added in this series and used in several places. With sandbox it is easier to read a file from the host than to use the CONFIG_OF_SEPARATE option, since this option requires knowledge of the executable image structure which is not really appropriate on the host system. A new CONFIG_OF_HOSTFILE provides this. A few related FDT changes are included in this series also. The -c option is enhanced to support passing entire scripts to sandbox. This is useful when writing non-trivial test code. Most of these patches were previously submitted as part of the verified boot effort. This series collects the independent sandbox-related patches together to make it easier to review. THe whole series is marked as version 3 for this reason. Changes in v3: - Rebase on master, to take account of generic board - Add CONFIG_OF_HOSTFILE support to generic board instead of sandbox - Remove #ifdefs which are not needed now that we have generic global_data Changes in v2: - Use gd-arch.ram_buf instead of gd-ram_buf (now that generic board is in) - Fix typo os defined - is defined - Use gd-arch.ram_buf instead of gd-ram_buf (now that generic board is in) - Fix checkpatch warnings about split strings - fdt_valid() sets the FDT pointer to NULL on error, to simplify callers - Allow the control FDT to be set even if there is currently no control FDT - Correct bug in setting control FDT - Fix checkpatch checks about parenthesis alignment - Rebase to use updated fdt_valid() function - Fix checkpatch checks about parenthesis alignment - Revert the whole change including the set_working_fdt_addr() part - Move sandbox's command list patch from a later series Simon Glass (17): Trigger generic board error only when building sandbox: Provide a way to map from host RAM to U-Boot RAM sandbox: Add support for generic board sandbox: Use generic board init sandbox: Remove old board init code sandbox: Add CONFIG_OF_HOSTFILE to read FDT from host file fdt: Add a parameter to fdt_valid() Add getenv_hex() to return an environment variable as hex fdt: Allow fdt command to check and update control FDT sandbox: fdt: Support fdt command for sandbox fdt: Skip checking FDT if the pointer is NULL Revert fdt- Tell the FDT library where the device tree is sandbox: Allow -c argument to provide a command list sandbox: Support 'source' command fs: Add support for saving data to filesystems sandbox: fs: Add support for saving files to host filesystem sandbox: config: Enable CONFIG_FIT and CONFIG_CMD_FIT Makefile | 10 ++ arch/sandbox/config.mk| 5 +- arch/sandbox/cpu/cpu.c| 5 + arch/sandbox/cpu/start.c | 9 +- arch/sandbox/include/asm/io.h | 8 ++ arch/sandbox/include/asm/state.h | 1 + arch/sandbox/include/asm/u-boot.h | 22 +-- arch/sandbox/lib/Makefile | 1 - arch/sandbox/lib/board.c | 285 -- board/sandbox/sandbox/sandbox.c | 2 +- common/board_f.c | 106 -- common/board_r.c | 8 +- common/cmd_fdt.c | 87 common/cmd_nvedit.c | 15 ++ common/cmd_sandbox.c | 18 ++- common/cmd_source.c | 11 +- common/main.c | 8 -- config.mk | 4 +- doc/README.fdt-control| 6 +- fs/fs.c | 75 ++ fs/sandbox/sandboxfs.c| 33 + include/asm-generic/sections.h| 2 +- include/common.h | 18 +++ include/configs/sandbox.h | 11 +- include/fs.h | 2 + include/sandboxfs.h | 1 + lib/fdtdec.c | 5 +- 27 files changed, 383 insertions(+), 375 deletions(-) delete mode 100644 arch/sandbox/lib/board.c -- 1.8.2.1 ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
