starfive-tech/u-boot.git/tools/patman, branch rtthread_AMP

patman: Avoid blank lines between tags

2021-08-08T17:27:27+00:00

In some cases 'patman status' leaves a blank line between the sign-off and the tags it collects from patchwork. Fix this and add a test. Signed-off-by: Simon Glass

patman: add warning for invalid tag

2021-08-01T15:05:24+00:00

Add a error in patman tool when the commit message contents an invalid tag "Serie-.*" instead of "Series-.*". Signed-off-by: Patrick Delaunay Reviewed-by: Simon Glass

patman: Use bytearray instead of string

2021-07-21T16:27:35+00:00

If the process outputs a lot of data on stdout this can be quite slow, since the bytestring is regenerated each time. Use a bytearray instead. Signed-off-by: Simon Glass

binman: Put compressed data into separate files

2021-07-21T16:27:35+00:00

At present compression uses the same temporary file for all invocations. With multithreading this causes the data to become corrupted. Use a different filename each time. Signed-off-by: Simon Glass

tools: Fix default target compile tools in Python tools

2021-07-06T16:38:03+00:00

In commit 1e4687aa47ed ("binman: Use target-specific tools when cross-compiling"), a utility function was implemented to get preferred compilation tools using environment variables like CC and CROSS_COMPILE. Although it intended to provide custom default tools (same as those in the global Makefile) when no relevant variables were set (for example using "gcc" for "cc"), it is only doing so when CROSS_COMPILE is set and returning the literal name of the tool otherwise. Remove the check for an empty CROSS_COMPILE, which makes the function use it as an empty prefix to the custom defaults and return the intended executables. Fixes: 1e4687aa47ed ("binman: Use target-specific tools when cross-compiling") Signed-off-by: Alper Nebi Yasak

patman: Parse checkpatch by message instead of by line

2021-04-29T10:23:39+00:00

Parse each empty-line-delimited message separately. This saves having to deal with all the different line content styles, we only care about the header ERROR | WARNING | NOTE... Also make checkpatch print line information for a uboot specific warning. Signed-off-by: Evan Benn Reviewed-by: Simon Glass

patman: Assume we always have pygit2 for tests

2021-04-29T10:23:39+00:00

Given that we have tests that require pygit2 and it can be installed like any other python module, fail much more loudly if it is missing. Cc: Simon Glass Signed-off-by: Tom Rini

checkpatch: Add warnings for using strn(cat|cpy)

2021-04-12T21:44:55+00:00

strn(cat|cpy) has a bad habit of not nul-terminating the destination, resulting in constructions like strncpy(foo, bar, sizeof(foo) - 1); foo[sizeof(foo) - 1] = '\0'; However, it is very easy to forget about this behavior and accidentally leave a string unterminated. This has shown up in some recent coverity scans [1, 2] (including code recently touched by yours truly). Fortunately, the guys at OpenBSD came up with strl(cat|cpy), which always nul-terminate strings. These functions are already in U-Boot, so we should encourage new code to use them instead of strn(cat|cpy). [1] https://lists.denx.de/pipermail/u-boot/2021-March/442888.html [2] https://lists.denx.de/pipermail/u-boot/2021-January/438073.html Signed-off-by: Sean Anderson Reviewed-by: Simon Glass

patman: Continue on if warnings are found outside a commit

2021-04-06T04:33:10+00:00

While we cannot know which commit the warning relates to, this should not be fatal. Print the warning and carry on. Signed-off-by: Simon Glass

binman: Use standard filenames for SPL/TPL devicetree

2021-03-26T04:03:09+00:00

At present, before any entry expansion is done (such as a 'files' entry expanding out to individual entries for each file it contains), we check the binman definition (i.e. '/binman' node) to find out what devicetree files are used in the images. This is a pain, since the definition may change during expansion. For example if there is no u-boot-spl-dtb entry in the definition at the start, we assume that the SPL devicetree is not used. But if an entry later expands to include this, then we don't notice. In fact the flexibility provided by the current approach of checking the definition is not really useful. We know that we can have SPL and TPL devicetrees. We know the pathname to each, so we can simply check if the files are present. If they are present, we can prepare them and update them regardless of whether they are actually used. If they are not present, we cannot prepare/update them anyway, i.e. an error will be generated. Simplify state.Prepare() so it uses a hard-coded list of devicetree files. Note that state.PrepareFromLoadedData() is left untouched, since in that case we have a complete definition from the loaded file, but cannot of course rely on the devicetree files that created it still being present. So in that case we still check the image defitions. Signed-off-by: Simon Glass