get_src_version() strips 'o' or 'lto.o' from the end of the object file
path (so, postfixlen is 1 or 5), then adds 'mod'.

If you look at the code closely, mod->name already holds the base path
with the extension stripped.

Most of the code changes made by commit 7ac204b545f2 ("modpost: lto:
strip .lto from module names") was actually unneeded.

sumversion.c does not need strends(), so it can get back local in
modpost.c again.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Pull module updates from Jessica Yu:

 - Retire EXPORT_UNUSED_SYMBOL() and EXPORT_SYMBOL_GPL_FUTURE(). These
   export types were introduced between 2006 - 2008. All the of the
   unused symbols have been long removed and gpl future symbols were
   converted to gpl quite a long time ago, and I don't believe these
   export types have been used ever since. So, I think it should be safe
   to retire those export types now (Christoph Hellwig)

 - Refactor and clean up some aged code cruft in the module loader
   (Christoph Hellwig)

 - Build {,module_}kallsyms_on_each_symbol only when livepatching is
   enabled, as it is the only caller (Christoph Hellwig)

 - Unexport find_module() and module_mutex and fix the last module
   callers to not rely on these anymore. Make module_mutex internal to
   the module loader (Christoph Hellwig)

 - Harden ELF checks on module load and validate ELF structures before
   checking the module signature (Frank van der Linden)

 - Fix undefined symbol warning for clang (Fangrui Song)

 - Fix smatch warning (Dan Carpenter)

* tag 'modules-for-v5.12' of git://git.kernel.org/pub/scm/linux/kernel/git/jeyu/linux:
  module: potential uninitialized return in module_kallsyms_on_each_symbol()
  module: remove EXPORT_UNUSED_SYMBOL*
  module: remove EXPORT_SYMBOL_GPL_FUTURE
  module: move struct symsearch to module.c
  module: pass struct find_symbol_args to find_symbol
  module: merge each_symbol_section into find_symbol
  module: remove each_symbol_in_section
  module: mark module_mutex static
  kallsyms: only build {,module_}kallsyms_on_each_symbol when required
  kallsyms: refactor {,module_}kallsyms_on_each_symbol
  module: use RCU to synchronize find_module
  module: unexport find_module and module_mutex
  drm: remove drm_fb_helper_modinit
  powerpc/powernv: remove get_cxl_module
  module: harden ELF info handling
  module: Ignore _GLOBAL_OFFSET_TABLE_ when warning for undefined symbols

EXPORT_UNUSED_SYMBOL* is not actually used anywhere.  Remove the
unused functionality as we generally just remove unused code anyway.

Reviewed-by: Miroslav Benes <mbenes@suse.cz>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jessica Yu <jeyu@kernel.org>

As far as I can tell this has never been used at all, and certainly
not any time recently.

Reviewed-by: Miroslav Benes <mbenes@suse.cz>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jessica Yu <jeyu@kernel.org>

With LTO, everything is compiled into LLVM bitcode, so we have to link
each module into native code before modpost. Kbuild uses the .lto.o
suffix for these files, which also ends up in module information. This
change strips the unnecessary .lto suffix from the module name.

Suggested-by: Bill Wendling <morbo@google.com>
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Kees Cook <keescook@chromium.org>
Link: https://lore.kernel.org/r/20201211184633.3213045-11-samitolvanen@google.com

We have 3 log functions. fatal() is special because it lets modpost bail
out immediately. The difference between warn() and error() is the only
prefix parts ("WARNING:" vs "ERROR:").

In my understanding, the expected handling of error() is to propagate
the return code of the function to the exit code of modpost, as
check_exports() etc. already does. This is a good manner in general
because we should display as many error messages as possible in a
single run of modpost.

What is annoying about fatal() is that it kills modpost at the first
error. People would need to run Kbuild again and again until they fix
all errors.

But, unfortunately, people tend to do:
"This case should not be allowed. Let's replace warn() with fatal()."

One of the reasons is probably it is tedious to manually hoist the error
code to the main() function.

This commit refactors error() so any single call for it automatically
makes modpost return the error code.

I also added comments in modpost.h for warn(), error(), and fatal().

Please use fatal() only when you have a strong reason to do so.
For example:

  - Memory shortage (i.e. malloc() etc. has failed)
  - The ELF file is broken, and there is no point to continue parsing
  - Something really odd has happened

For general coding errors, please use error().

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Tested-by: Quentin Perret <qperret@google.com>

The log function names, warn(), merror(), fatal() are inconsistent.

Commit 2a11665945d5 ("kbuild: distinguish between errors and warnings
in modpost") intentionally chose merror() to avoid the conflict with
the library function error(). See man page of error(3).

But, we are already causing the conflict with warn() because it is also
a library function. See man page of warn(3). err() would be a problem
for the same reason.

The common technique to work around name conflicts is to use macros.
For example:

    /* in a header */
    #define error(fmt, ...)  __error(fmt, ##__VA_ARGS__)
    #define warn(fmt, ...)   __warn(fmt, ##__VA_ARGS__)

    /* function definition */
    void __error(const char *fmt, ...)
    {
            <our implementation>
    }

    void __warn(const char *fmt, ...)
    {
            <our implementation>
    }

In this way, we can implement our own warn() and error(), still we can
include <error.h> and <err.h> with no problem.

And, commit 93c95e526a4e ("modpost: rework and consolidate logging
interface") already did that.

Since the log functions are all macros, we can use error() without
causing "conflicting types" errors.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Align with the mmap / munmap APIs.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

new_module() conditionally strips the .o because the modname has .o
suffix when it is called from read_symbols(), but no .o when it is
called from read_dump().

It is clearer to strip .o in read_symbols().

I also used flexible-array for mod->name.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

The meaning of 'skip' is obscure since it does not explain
"what to skip".

mod->skip is set when it is vmlinux or the module info came from
a dump file.

So, mod->skip is equivalent to (mod->is_vmlinux || mod->from_dump).

For the check in write_namespace_deps_files(), mod->is_vmlinux is
unneeded because the -d option is not passed in the first pass of
modpost.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>