On 3/3/20 3:44 PM, Egeyar Bagcioglu wrote:
Hello,
Hi.
I would like to propose the second version of the patches which introduce a compile option --record-gcc-command-line. When passed to gcc, it saves the command line invoking gcc into the produced object file. The option makes it trivial to trace back with which command a file was compiled and by which version of the gcc. It helps with debugging, reproducing bugs and repeating the build process. The reviews addressed in this version include indentation changes, error handling and corner case coverage pointed out by Segher Boessenkool in the first two patches; while the new third patch is another corner case (lto) coverage requested by Martin Liska.
Great, thanks for working on that. Based on the mentioned disadvantages of -frecord-gcc-switches, where each option entry is put into a mergeable section and so one can't disambiguate in between multiple command line options, I would suggest to replace -frecord-gcc-switches functionality with what --record-gcc-command-line does. Moreover, looking at clang: https://reviews.llvm.org/D54487, their current implementation of -frecord-gcc-switches does the very same: $ clang main.c -frecord-gcc-switches -c $ readelf -p .GCC.command.line main.o String dump of section '.GCC.command.line': [ 1] /usr/bin/clang-9.0.1 main.c -frecord-command-line -c One additional advantage is that driver options also catch macros (like -D_FORTIFY_SOURCE), which is also what clang does. It's question to Nick who originally implemented the option. Nick, are you fine with that? If I'm correct you're using the annobin plugin and this can help you to replace it?
Although we discussed after the submission of the first version that there are several other options performing similar tasks, I believe we established that there is still a need for this specific functionality. Therefore, I am skipping in this email the comparison between this option and the existing options with similarities. This functionality operates as the following: It saves gcc's argv into a temporary file, and passes --record-gcc-command-line <tempfilename> to cc1 or cc1plus. The functionality of the backend is implemented via a hook. This patch includes an example implementation of the hook for elf targets: elf_record_gcc_command_line function. This function reads the given file and writes gcc's version and the command line into a mergeable string section, .GCC.command.line. It creates one entry per invocation. By doing so, it makes it clear which options were used together in a single gcc invocation, even after linking. Here is an *example usage* of the option: [egeyar@localhost save-commandline]$ gcc main.c --record-gcc-command-line [egeyar@localhost save-commandline]$ readelf -p .GCC.command.line a.out String dump of section '.GCC.command.line': [ 0] 10.0.1 20200227 (experimental) : gcc main.c --record-gcc-command-line The following is a *second example* calling g++ with -save-temps and a repetition of options, where --save-temps saves the intermediate file, main.cmdline in this case. You can see that the options are recorded unprocessed: [egeyar@localhost save-commandline]$ g++ main.c -save-temps --record-gcc-command-line -O0 -O2 -O3 -DFORTIFY=2 --record-gcc-command-line [egeyar@localhost save-commandline]$ readelf -p .GCC.command.line a.out String dump of section '.GCC.command.line': [ 0] 10.0.1 20200227 (experimental) : g++ main.c -save-temps --record-gcc-command-line -O0 -O2 -O3 -DFORTIFY=2 --record-gcc-command-line The first patch of this three-patch-series only extends the testsuite machinery, while the second patch implements the functionality and adds a test case for it. The third patch that alters libiberty is to make sure the .GCC.command.line section in LTO objects survive the linking and appear in the linked object.
One small nit: I would not send in 3 parts, as the first and last one are just one hunks. Martin
In addition to the new test case, I built binutils as my test case after passing this option to CFLAGS. The added .GCC.command.line section of ld.bfd listed many compile commands as expected. Tested on x86_64-pc-linux-gnu. Please review the patches, let me know what you think and apply if appropriate. Regards Egeyar Egeyar Bagcioglu (3): Introduce dg-require-target-object-format Introduce the gcc option --record-gcc-command-line Keep .GCC.command.line sections of LTO objetcs. gcc/common.opt | 4 +++ gcc/config/elfos.h | 5 +++ gcc/doc/tm.texi | 22 ++++++++++++ gcc/doc/tm.texi.in | 4 +++ gcc/gcc.c | 41 ++++++++++++++++++++++ gcc/gcc.h | 1 + gcc/target.def | 30 ++++++++++++++++ gcc/target.h | 3 ++ .../c-c++-common/record-gcc-command-line.c | 8 +++++ gcc/testsuite/lib/target-supports-dg.exp | 11 ++++++ gcc/toplev.c | 13 +++++++ gcc/varasm.c | 40 +++++++++++++++++++++ libiberty/simple-object.c | 3 ++ 13 files changed, 185 insertions(+) create mode 100644 gcc/testsuite/c-c++-common/record-gcc-command-line.c
