summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* kbuild: remove unused cmd_gentimeconstMasahiro Yamada2018-12-251-4/+0
| | | | | | filechk_gentimeconst is always used instead of cmd_gentimeconst. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* kbuild: remove $(obj)/ prefixes in ./KbuildMasahiro Yamada2018-12-231-5/+5
| | | | | | | | Adding $(obj)/ to file paths relative from the top objtree looks a bit redundant to me. This commit has no functional change since $(obj) is '.' in this file. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* treewide: add intermediate .s files to targetsMasahiro Yamada2018-12-236-0/+10
| | | | | | Avoid unneeded recreation of these in the incremental build. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* treewide: remove explicit rules for *offsets.sMasahiro Yamada2018-12-236-25/+1
| | | | | | | These explicit rules are unneeded because scripts/Makefile.build provides a pattern rule to create %.s from %.c Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* firmware: refactor firmware/MakefileMasahiro Yamada2018-12-231-52/+32
| | | | | | | | | | | | | Clean up the Makefile. Equivalent *.gen.S files are still produced. - Use filechk to remove ugly wordsize_deps - Get FWNAME, FWSTR, ASM_WORD, ASM_ALIGN, and PROGBITS out of the recipe for readability - Remove 'mkdir' because filechk takes care of it Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* firmware: remove unnecessary patterns from .gitignoreMasahiro Yamada2018-12-221-5/+0
| | | | | | | | | Commit df85b2d767aa ("firmware: Restore support for built-in firmware") should not have restored this .gitignore blindly. *.gen.S is only pattern that should be ignored by git. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* scripts: remove unnecessary ihex2fw and check-lc_ctypes from .gitignoreMasahiro Yamada2018-12-221-2/+0
| | | | | | | | | | | | Commit c512d2544c68 ("gitignore: ignore scripts/ihex2fw") was unneeded. ihex2fw was generated in firmware/ instead of scripts/ at that time although ihex2fw.c was pushed back and forth between those directories in the past. check-lc_ctype was removed by commit cb43fb5775df ("docs: remove DocBook from the building system"). Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* um: remove unused filechk_gen_header in MakefileMasahiro Yamada2018-12-221-2/+0
| | | | | | | This is a leftover of commit ecba97d4aacf ("[PATCH] uml makefiles sanitized"). Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* scripts: add a tool to produce a compile_commands.json fileTom Roeder2018-12-191-0/+151
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The LLVM/Clang project provides many tools for analyzing C source code. Many of these tools are based on LibTooling (https://clang.llvm.org/docs/LibTooling.html), which depends on a database of compiler flags. The standard container for this database is compile_commands.json, which consists of a list of JSON objects, each with "directory", "file", and "command" fields. Some build systems, like cmake or bazel, produce this compilation information directly. Naturally, Makefiles don't. However, the kernel makefiles already create .<target>.o.cmd files that contain all the information needed to build a compile_commands.json file. So, this commit adds scripts/gen_compile_commands.py, which recursively searches through a directory for .<target>.o.cmd files and extracts appropriate compile commands from them. It writes a compile_commands.json file that LibTooling-based tools can use. By default, gen_compile_commands.py starts its search in its working directory and (over)writes compile_commands.json in the working directory. However, it also supports --output and --directory flags for out-of-tree use. Note that while gen_compile_commands.py enables the use of clang-based tools, it does not require the kernel to be compiled with clang. E.g., the following sequence of commands produces a compile_commands.json file that works correctly with LibTooling. make defconfig make scripts/gen_compile_commands.py Also note that this script is written to work correctly in both Python 2 and Python 3, so it does not specify the Python version in its first line. For an example of the utility of this script: after running gen_compile_commands.json on the latest kernel version, I was able to use Vim + the YouCompleteMe pluging + clangd to automatically jump to definitions and declarations. Obviously, cscope and ctags provide some of this functionality; the advantage of supporting LibTooling is that it opens the door to many other clang-based tools that understand the code directly and do not rely on regular expressions and heuristics. Tested: Built several recent kernel versions and ran the script against them, testing tools like clangd (for editor/LSP support) and clang-check (for static analysis). Also extracted some test .cmd files from a kernel build and wrote a test script to check that the script behaved correctly with all permutations of the --output and --directory flags. Signed-off-by: Tom Roeder <tmroeder@google.com> Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* kbuild: add -Werror=implicit-int flag unconditionallyMasahiro Yamada2018-12-171-4/+1
| | | | | | | | This flag is documented in the GCC 4.6 manual, and recognized by Clang as well. Let's rip off the cc-option switch. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com> Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
* kbuild: add -Werror=strict-prototypes flag unconditionallyMasahiro Yamada2018-12-171-4/+1
| | | | | | | | | | | -Wstrict-prototypes is added to KBUILD_CFLAGS first, then overridden by -Werror=strict-prototypes later. Let's add -Werror=strict-prototypes unconditionally because it is supported by GCC 4.6, and also by Clang. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com> Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
* kbuild: add -fno-PIE flag unconditionallyMasahiro Yamada2018-12-171-5/+2
| | | | | | | | This flag is documented in the GCC 4.6 manual, and recognized by Clang as well. Let's rip off the cc-option switch. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com> Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
* scripts: coccinelle: Correct warning messageJulia Lawall2018-12-171-6/+6
| | | | | | | | "Assignment" requires the assigned value before the place that value is stored into. Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr> Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* scripts: coccinelle: only suggest true/false in files that already use themJulia Lawall2018-12-171-10/+21
| | | | | | | | Some code may overall use 0 and 1, so don't introduce occasional uses of true and false in these cases. Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr> Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* kbuild: handle part-of-module correctly for *.ll and *.symtypesMasahiro Yamada2018-12-171-1/+1
| | | | | | | The single targets *.ll and *.symtypes have never been treated as a module. Fix it. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* kbuild: refactor part-of-moduleMasahiro Yamada2018-12-171-5/+3
| | | | | | Use $(foreach ...) to make it shorter. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* kbuild: refactor quiet_modtagMasahiro Yamada2018-12-171-6/+1
| | | | | | | part-of-module and quiet_modtag are set for the same targets. Define quiet_modtag based on part-of-module. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* kbuild: remove redundant quiet_modtag for $(obj-m)Masahiro Yamada2018-12-171-2/+0
| | | | | | | | | | | All objects in $(obj-m) are contained in $(real-obj-m) as well. It is true composite objects are only contained in $(obj-m), but [M] is hard-coded in quiet_cmd_link_multi-m. This line is redundant. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* kbuild: refactor Makefile.asm-genericMasahiro Yamada2018-12-172-22/+23
| | | | | | | | | | | | | | | | - Use conventional $(MAKE) $(asm-generic)=<dir> style for directory descending - Remove unneeded FORCE since "all" is a phony target - Remove unneeded "_dummy :=" assignment - Skip $(shell mkdir ...) when headers exist in the directory - Misc cleanups Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com> Acked-by: Sam Ravnborg <sam@ravnborg.org>
* user/Makefile: Fix typo and capitalization in comment sectionEugen Cazacu2018-12-111-2/+2
| | | | | Signed-off-by: Eugen Cazacu <green182day@hotmail.com> Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* kbuild: exploit parallel building for CONFIG_HEADERS_CHECKMasahiro Yamada2018-12-081-3/+4
| | | | | | | | | | When CONFIG_HEADERS_CHECK is enabled, the headers_check is executed as a serialized task in the vmlinux recipe. Make it independent of vmlinux so that parallel building can process the headers_check and other build targets simultaneously. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* kbuild: remove a special handling for *.agh in Makefile.headersinstMasahiro Yamada2018-12-081-1/+0
| | | | | | | | | | | | | scripts/Makefile.headersinst takes care of *.agh just for arch/cris/include/uapi/arch-v10/arch/sv_addr.agh because renaming exported headers is difficult (or impossible). This code is no longer necessary thanks to commit c690eddc2f3b ("CRIS: Drop support for the CRIS port"). Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* kbuild: move .SECONDARY special target to Kbuild.includeMasahiro Yamada2018-12-024-6/+5
| | | | | | | | | | | | | | | | | | | | | | | In commit 54a702f70589 ("kbuild: mark $(targets) as .SECONDARY and remove .PRECIOUS markers"), I missed one important feature of the .SECONDARY target: .SECONDARY with no prerequisites causes all targets to be treated as secondary. ... which agrees with the policy of Kbuild. Let's move it to scripts/Kbuild.include, with no prerequisites. Note: If an intermediate file is generated by $(call if_changed,...), you still need to add it to "targets" so its .*.cmd file is included. The arm/arm64 crypto files are generated by $(call cmd,shipped), so they do not need to be added to "targets", but need to be added to "clean-files" so "make clean" can properly clean them away. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* kbuild: remove redundant 'set -e' from cmd_* definesMasahiro Yamada2018-12-012-3/+0
| | | | | | | | | These three cmd_* are invoked in the $(call cmd,*) form. Now that 'set -e' moved to the 'cmd' macro, they do not need to explicitly give 'set -e'. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* kbuild: refactor if_changedMasahiro Yamada2018-12-011-2/+1
| | | | | | '@set -e; $(echo-cmd) $(cmd_$(1)' can be replaced with '$(cmd)'. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* kbuild: remove trailing semicolon from cmd_* passed to if_changed_ruleMasahiro Yamada2018-12-012-9/+9
| | | | | | | With the change of rule_cc_o_c / rule_as_o_S in the last commit, each command is executed in a separate subshell. Rip off unneeded semicolons. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* kbuild: change if_changed_rule for multi-line recipeMasahiro Yamada2018-12-012-19/+15
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The 'define' ... 'endef' directive is useful to confine a series of shell commands into a single macro: define foo [action1] [action2] [action3] endif Each action is executed in a separate subshell. However, rule_cc_o_c and rule_as_o_S in scripts/Makefile.build are written as follows (with a trailing semicolon in each cmd_*): define rule_cc_o_c [action1] ; \ [action2] ; \ [action3] ; endef All shell commands are concatenated with '; \' so that it looks like a single command from the Makefile point of view. This does not exploit the benefits of 'define' ... 'endef' form because a single shell command can be more simply written, like this: rule_cc_o_c = \ [action1] ; \ [action2] ; \ [action3] ; I guess the intention for the command concatenation was to let the '@set -e' in if_changed_rule cover all the commands. We can improve the readability by moving '@set -e' to the 'cmd' macro. The combo of $(call echo-cmd,*) $(cmd_*) in rule_cc_o_c and rule_as_o_S have been replaced with $(call cmd,*). The trailing back-slashes have been removed. Here is a note about the performance: the commands in rule_cc_o_c and rule_as_o_S were previously executed all together in a single subshell, but now each line in a separate subshell. This means Make will spawn extra subshells [1]. I measured the build performance for x86_64_defconfig + CONFIG_MODVERSIONS + CONFIG_TRIM_UNUSED_KSYMS and I saw slight performance regression, but I believe code readability and maintainability wins. [1] Precisely, GNU Make may optimize this by executing the command directly instead of forking a subshell, if no shell special characters are found in the command line and omitting the subshell will not change the behavior. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* kbuild: simplify dependency generation for CONFIG_TRIM_UNUSED_KSYMSMasahiro Yamada2018-12-016-69/+53
| | | | | | | | | | | | | | | | | | | | | | | | | My main motivation of this commit is to clean up scripts/Kbuild.include and scripts/Makefile.build. Currently, CONFIG_TRIM_UNUSED_KSYMS works with a tricky gimmick; possibly exported symbols are detected by letting $(CPP) replace EXPORT_SYMBOL* with a special string '=== __KSYM_*===', which is post-processed by sed, and passed to fixdep. The extra preprocessing is costly, and hacking cmd_and_fixdep is ugly. I came up with a new way to find exported symbols; insert a dummy symbol __ksym_marker_* to each potentially exported symbol. Those dummy symbols are picked up by $(NM), post-processed by sed, then appended to .*.cmd files. I collected the post-process part to a new shell script scripts/gen_ksymdeps.sh for readability. The dummy symbols are put into the .discard.* section so that the linker script rips them off the final vmlinux or modules. A nice side-effect is building with CONFIG_TRIM_UNUSED_KSYMS will be much faster. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com> Reviewed-by: Nicolas Pitre <nico@linaro.org>
* kbuild: refactor modversions build rulesMasahiro Yamada2018-12-011-37/+17
| | | | | | | | | | | | | | | Let $(CC) compile objects into normal files *.o instead of .tmp_*.o whether CONFIG_MODVERSIONS is enabled or not. With this, the input file for objtool is always *.o so objtool_o can go away. I guess the reason of using .tmp_*.o for intermediate objects was to avoid leaving incomplete *.o file (, whose timestamp says it is up-to-date) when the genksyms tool failed for some reasons. It no longer matters because any targets are deleted on errors since commit 9c2af1c7377a ("kbuild: add .DELETE_ON_ERROR special target"). Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* kbuild: remove redundant 'set -e' from sub_cmd_record_mcountMasahiro Yamada2018-12-011-1/+1
| | | | | | | This is executed inside the if_changed_rule, which already sets 'set -e'. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* kbuild: remove redundant 'set -e' from filechk_offsetsMasahiro Yamada2018-12-011-1/+1
| | | | | | The filechk macro in scripts/Kbuild.include already sets 'set -e'. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* kbuild: let fixdep directly write to .*.cmd filesMasahiro Yamada2018-12-013-19/+16
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Currently, fixdep writes dependencies to .*.tmp, which is renamed to .*.cmd after everything succeeds. This is a very safe way to avoid corrupted .*.cmd files. The if_changed_dep has carried this safety mechanism since it was added in 2002. If fixdep fails for some reasons or a user terminates the build while fixdep is running, the incomplete output from the fixdep could be troublesome. This is my insight about some bad scenarios: [1] If the compiler succeeds to generate *.o file, but fixdep fails to write necessary dependencies to .*.cmd file, Make will miss to rebuild the object when headers or CONFIG options are changed. In this case, fixdep should not generate .*.cmd file at all so that 'arg-check' will surely trigger the rebuild of the object. [2] A partially constructed .*.cmd file may not be a syntactically correct makefile. The next time Make runs, it would include it, then fail to parse it. Once this happens, 'make clean' is be the only way to fix it. In fact, [1] is no longer a problem since commit 9c2af1c7377a ("kbuild: add .DELETE_ON_ERROR special target"). Make deletes a target file on any failure in its recipe. Because fixdep is a part of the recipe of *.o target, if it fails, the *.o is deleted anyway. However, I am a bit worried about the slight possibility of [2]. So, here is a solution. Let fixdep directly write to a .*.cmd file, but allow makefiles to include it only when its corresponding target exists. This effectively reverts commit 2982c953570b ("kbuild: remove redundant $(wildcard ...) for cmd_files calculation"), and commit 00d78ab2ba75 ("kbuild: remove dead code in cmd_files calculation in top Makefile") because now we must check the presence of targets. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* kbuild: descend into scripts/gcc-plugins/ via scripts/MakefileMasahiro Yamada2018-12-013-10/+3
| | | | | | | | Now that 'archprepare' depends on 'scripts', Kbuild can descend into scripts/gcc-plugins in a more standard way. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com> Reviewed-by: Kees Cook <keescook@chromium.org>
* kbuild: make 'archprepare' depend on 'scripts'Masahiro Yamada2018-12-011-17/+15
| | | | | | | | | | | | | Before start descending, Kbuild needs to run 'prepare' and 'scripts', which has been orthogonal to each other. Going forward, let's consider 'scripts' is a part of the preparation. This will allow more cleanups. Move 'scripts' to the prerequisite of 'archprepare', where UML starts compiling target *.c files. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* kbuild: move modpost out of 'scripts' targetMasahiro Yamada2018-12-013-5/+3
| | | | | | | | | | | | I am eagar to build under the scripts/ directory only with $(HOSTCC), but scripts/mod/ highly depends on the $(CC) and target arch headers. That it why the 'scripts' target must depend on 'asm-generic', 'gcc-plugins', and $(autoksyms_h). Move it to the 'prepare0' stage. I know this is a cheesy workaround, but better than the current situation. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* kbuild: fix UML build error with CONFIG_GCC_PLUGINSMasahiro Yamada2018-12-013-25/+7
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | UML fails to build with CONFIG_GCC_PLUGINS=y. $ make -s ARCH=um mrproper $ make -s ARCH=um allmodconfig $ make ARCH=um UPD include/generated/uapi/linux/version.h WRAP arch/x86/include/generated/uapi/asm/bpf_perf_event.h WRAP arch/x86/include/generated/uapi/asm/poll.h WRAP arch/x86/include/generated/asm/dma-contiguous.h WRAP arch/x86/include/generated/asm/early_ioremap.h WRAP arch/x86/include/generated/asm/export.h WRAP arch/x86/include/generated/asm/mcs_spinlock.h WRAP arch/x86/include/generated/asm/mm-arch-hooks.h SYSTBL arch/x86/include/generated/asm/syscalls_32.h SYSHDR arch/x86/include/generated/asm/unistd_32_ia32.h SYSHDR arch/x86/include/generated/asm/unistd_64_x32.h SYSTBL arch/x86/include/generated/asm/syscalls_64.h SYSHDR arch/x86/include/generated/uapi/asm/unistd_32.h SYSHDR arch/x86/include/generated/uapi/asm/unistd_64.h SYSHDR arch/x86/include/generated/uapi/asm/unistd_x32.h HOSTCC scripts/unifdef CC arch/x86/um/user-offsets.s cc1: error: cannot load plugin ./scripts/gcc-plugins/cyc_complexity_plugin.so ./scripts/gcc-plugins/cyc_complexity_plugin.so: cannot open shared object file: No such file or directory cc1: error: cannot load plugin ./scripts/gcc-plugins/structleak_plugin.so ./scripts/gcc-plugins/structleak_plugin.so: cannot open shared object file: No such file or directory cc1: error: cannot load plugin ./scripts/gcc-plugins/latent_entropy_plugin.so ./scripts/gcc-plugins/latent_entropy_plugin.so: cannot open shared object file: No such file or directory cc1: error: cannot load plugin ./scripts/gcc-plugins/randomize_layout_plugin.so ./scripts/gcc-plugins/randomize_layout_plugin.so: cannot open shared object file: No such file or directory make[1]: *** [scripts/Makefile.build;119: arch/x86/um/user-offsets.s] Error 1 make: *** [arch/um/Makefile;152: arch/x86/um/user-offsets.s] Error 2 Reorder the preparation stage (with cleanups) to make sure gcc-plugins is built before descending to arch/x86/um/. Fixes: 6b90bd4ba40b ("GCC plugin infrastructure") Reported-by: kbuild test robot <lkp@intel.com> Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* modpost: move unresolved symbol checks to check_exports()Masahiro Yamada2018-12-011-15/+18
| | | | | | This will fit better in check_exports() than add_versions(). Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* modpost: merge module iterationsMasahiro Yamada2018-12-011-6/+1
| | | | | | | Probably, this is just a matter of the order of error/warning messages. Merge the two for-loops. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* modpost: refactor seen flag clearing in add_depends()Masahiro Yamada2018-12-011-6/+6
| | | | | | | | | | You do not need to iterate over all modules for resetting ->seen flag because add_depends() is only interested in modules that export symbols referenced from the given 'mod'. This also avoids shadowing the 'modules' parameter of add_depends(). Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* modpost: file2alias: check prototype of handlerMasahiro Yamada2018-12-011-4/+3
| | | | | | | | Use specific prototype instead of an opaque pointer so that the compiler can catch function prototype mismatch. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com> Reviewed-by: Mathieu Malaterre <malat@debian.org>
* modpost: file2alias: go back to simple devtable lookupMasahiro Yamada2018-12-011-95/+49
| | | | | | | | | | | | | | | Commit e49ce14150c6 ("modpost: use linker section to generate table.") was not so cool as we had expected first; it ended up with ugly section hacks when commit dd2a3acaecd7 ("mod/file2alias: make modpost compile on darwin again") came in. Given a certain degree of unknowledge about the link stage of host programs, I really want to see simple, stupid table lookup so that this works in the same way regardless of the underlying executable format. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com> Acked-by: Mathieu Malaterre <malat@debian.org>
* kbuild: fix single target build for external moduleMasahiro Yamada2018-12-011-8/+3
| | | | | | | | | | | | | | | | | | | | Building a single target in an external module fails due to missing .tmp_versions directory. For example, $ make -C /lib/modules/$(uname -r)/build M=$PWD foo.o will fail in the following way: CC [M] /home/masahiro/foo/foo.o /bin/sh: 1: cannot create /home/masahiro/foo/.tmp_versions/foo.mod: Directory nonexistent This is because $(cmd_crmodverdir) is executed only before building /, %/, %.ko single targets of external modules. Create .tmp_versions in the 'prepare' target. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* modpost: skip ELF local symbols during section mismatch checkPaul Walmsley2018-12-011-0/+12
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | During development of a serial console driver with a gcc 8.2.0 toolchain for RISC-V, the following modpost warning appeared: ---- WARNING: vmlinux.o(.data+0x19b10): Section mismatch in reference from the variable .LANCHOR1 to the function .init.text:sifive_serial_console_setup() The variable .LANCHOR1 references the function __init sifive_serial_console_setup() If the reference is valid then annotate the variable with __init* or __refdata (see linux/init.h) or name the variable: *_template, *_timer, *_sht, *_ops, *_probe, *_probe_one, *_console ---- ".LANCHOR1" is an ELF local symbol, automatically created by gcc's section anchor generation code: https://gcc.gnu.org/onlinedocs/gccint/Anchored-Addresses.html https://gcc.gnu.org/git/?p=gcc.git;a=blob;f=gcc/varasm.c;h=cd9591a45617464946dcf9a126dde277d9de9804;hb=9fb89fa845c1b2e0a18d85ada0b077c84508ab78#l7473 This was verified by compiling the kernel with -fno-section-anchors and observing that the ".LANCHOR1" ELF local symbol disappeared, and modpost no longer warned about the section mismatch. The serial driver code idiom triggering the warning is standard Linux serial driver practice that has a specific whitelist inclusion in modpost.c. I'm neither a modpost nor an ELF expert, but naively, it doesn't seem useful for modpost to report section mismatch warnings caused by ELF local symbols by default. Local symbols have compiler-generated names, and thus bypass modpost's whitelisting algorithm, which relies on the presence of a non-autogenerated symbol name. This increases the likelihood that false positive warnings will be generated (as in the above case). Thus, disable section mismatch reporting on ELF local symbols. The rationale here is similar to that of commit 2e3a10a1551d ("ARM: avoid ARM binutils leaking ELF local symbols") and of similar code already present in modpost.c: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/scripts/mod/modpost.c?h=v4.19-rc4&id=7876320f88802b22d4e2daf7eb027dd14175a0f8#n1256 This third version of the patch implements a suggestion from Masahiro Yamada <yamada.masahiro@socionext.com> to restructure the code as an additional pattern matching step inside secref_whitelist(), and further improves the patch description. Signed-off-by: Paul Walmsley <paul.walmsley@sifive.com> Signed-off-by: Paul Walmsley <paul@pwsan.com> Acked-by: Sam Ravnborg <sam@ravnborg.org> Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* kbuild: announce removal of SUBDIRS if usedMasahiro Yamada2018-12-015-21/+6
| | | | | | | | | | | | | | | | | | | | SUBDIRS has been kept as a backward compatibility since commit ("[PATCH] kbuild: external module support") in 2002. We do not need multiple ways to do the same thing, so I will remove SUBDIRS after the Linux 5.3 release. I cleaned up in-tree code, and updated the document so that nobody would try to use it. Meanwhile, display the following warning if SUBDIRS is used. Makefile:189: ================= WARNING ================ Makefile:190: 'SUBDIRS' will be removed after Linux 5.3 Makefile:191: Please use 'M=' or 'KBUILD_EXTMOD' instead Makefile:192: ========================================== Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com> Acked-by: Boris Brezillon <boris.brezillon@bootlin.com> # for scx200_docflash.c Acked-by: Guenter Roeck <linux@roeck-us.net> # for scx200_wdt.c
* modpost: drop unused command line switchesPaul Walmsley2018-11-211-1/+1
| | | | | | | | | | | | | | | | | | | | | | | Drop modpost command line switches that are no longer used by makefile.modpost, upon request from Sam Ravnborg <sam@ravnborg.org>, who wrote: modpost is not supposed to be used outside the kernel build. [...] I checked if there were any options supported by modpost that was not configurable in Makefile.modpost. And I could see that the -M and -K options in getopt() were leftovers. The code that used these option was dropped in: commit a8773769d1a1 ("Kbuild: clear marker out of modpost") Could you add a patch that delete these on top of what you already have. https://lore.kernel.org/lkml/20181020140835.GA3351@ravnborg.org/ Suggested-by: Sam Ravnborg <sam@ravnborg.org> Signed-off-by: Paul Walmsley <paul.walmsley@sifive.com> Signed-off-by: Paul Walmsley <paul@pwsan.com> Acked-by: Sam Ravnborg <sam@ravnborg.org> Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* scripts/setlocalversion: Improve -dirty check with git-status ↵Brian Norris2018-11-211-2/+10
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | --no-optional-locks git-diff-index does not refresh the index for you, so using it for a "-dirty" check can give misleading results. Commit 6147b1cf19651 ("scripts/setlocalversion: git: Make -dirty check more robust") tried to fix this by switching to git-status, but it overlooked the fact that git-status also writes to the .git directory of the source tree, which is definitely not kosher for an out-of-tree (O=) build. That is getting reverted. Fortunately, git-status now supports avoiding writing to the index via the --no-optional-locks flag, as of git 2.14. It still calculates an up-to-date index, but it avoids writing it out to the .git directory. So, let's retry the solution from commit 6147b1cf19651 using this new flag first, and if it fails, we assume this is an older version of git and just use the old git-diff-index method. It's hairy to get the 'grep -vq' (inverted matching) correct by stashing the output of git-status (you have to be careful about the difference betwen "empty stdin" and "blank line on stdin"), so just pipe the output directly to grep and use a regex that's good enough for both the git-status and git-diff-index version. Cc: Christian Kujau <lists@nerdbynature.de> Cc: Guenter Roeck <linux@roeck-us.net> Suggested-by: Alexander Kapshuk <alexander.kapshuk@gmail.com> Signed-off-by: Brian Norris <briannorris@chromium.org> Tested-by: Genki Sky <sky@genki.is> Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* modpost: validate symbol names also in find_elf_symbolSami Tolvanen2018-11-211-24/+26
| | | | | | | | | | | | | | | If an ARM mapping symbol shares an address with a valid symbol, find_elf_symbol can currently return the mapping symbol instead, as the symbol is not validated. This can result in confusing warnings: WARNING: vmlinux.o(.text+0x18f4028): Section mismatch in reference from the function set_reset_devices() to the variable .init.text:$x.0 This change adds a call to is_valid_name to find_elf_symbol, similarly to how it's already used in find_elf_symbol2. Signed-off-by: Sami Tolvanen <samitolvanen@google.com> Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* powerpc/boot: Set target when cross-compiling for clangJoel Stanley2018-11-141-0/+5
| | | | | | | | | | | | Clang needs to be told which target it is building for when cross compiling. Link: https://github.com/ClangBuiltLinux/linux/issues/259 Signed-off-by: Joel Stanley <joel@jms.id.au> Tested-by: Daniel Axtens <dja@axtens.net> # powerpc 64-bit BE Acked-by: Michael Ellerman <mpe@ellerman.id.au> Reviewed-by: Nick Desaulniers <ndesaulniers@google.com> Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* Makefile: Export clang toolchain variablesJoel Stanley2018-11-141-0/+1
| | | | | | | The powerpc makefile will use these in it's boot wrapper. Signed-off-by: Joel Stanley <joel@jms.id.au> Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
* kbuild: consolidate Clang compiler flagsMasahiro Yamada2018-11-141-7/+6
| | | | | | | | | | | Collect basic Clang options such as --target, --prefix, --gcc-toolchain, -no-integrated-as into a single variable CLANG_FLAGS so that it can be easily reused in other parts of Makefile. Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com> Reviewed-by: Nick Desaulniers <ndesaulniers@google.com> Tested-by: Nick Desaulniers <ndesaulniers@google.com> Acked-by: Greg Hackmann <ghackmann@google.com>
generated by cgit v1.2.3 (git 2.25.1) at 2024-11-26 09:03:31 +0000