Merge branch 'kbuild-next-speedups' into kbuild-for-next

* kbuild-next-speedups:
  kbuild: Mark the initcall ordering recipe as recursive
  kbuild: Mark the gzip recipes as recursive
  kconfig: Improve ergonomics around disabling warnings with cc-option
  scripts/Kconfig.toolchain: Drop compiler symbols from dependencies
  kbuild: keep .modinfo when vmlinux is linked with --gc-sections
  kbuild: compress the kernel with pigz if available
  kbuild: avoid re-running compiler and linker probes
  kbuild: move the toolchain checks into scripts/Kconfig.toolchain
  kallsyms: reimplement mksysmap in C
  elf-parse: add section flags, symbol binding and mapping helpers
  kbuild: only emit vmlinux relocations when required
  kbuild: do not sort nm output where the order is irrelevant
  sorttable: parse nm output correctly
  kallsyms: output binary data to speed output and kallsyms assembly
  kallsyms: index symbols by token to speed up table compression
  kbuild: do not allocate .modinfo in vmlinux

Signed-off-by: Nathan Chancellor <nathan@kernel.org>

1 files changed, 14 insertions, 1 deletions

diff --git a/scripts/sorttable.c b/scripts/sorttable.c
index d7b50581c732..88e49a5c4251 100644
--- a/scripts/sorttable.c
+++ b/scripts/sorttable.c
@@ -322,10 +322,23 @@ static int parse_symbols(const char *fname)
 		return -1;
 	}
 
-	while (fscanf(fp, "%16s %16s %c %*s\n", addr_str, size_str, &type) == 3) {
+	while (fscanf(fp, "%16s %16s %c%*[^\n]", addr_str, size_str, &type) == 3) {
 		uint64_t addr;
 		uint64_t size;
 
+		/*
+		 * nm outputs size-less entries with a missing 2nd value, which
+		 * means fscanf() just misread its fields.
+		 *
+		 * These don't matter as a call site cannot be in an empty
+		 * function, so just skip them.
+		 *
+		 * nm pads address and size to the same width, so if their
+		 * widths differ, this is a misread.
+		 */
+		if (strlen(size_str) != strlen(addr_str))
+			continue;
+
 		/* Only care about functions */
 		if (type != 't' && type != 'T' && type != 'W')
 			continue;