git.ipfire.org Git - thirdparty/kernel/stable-queue.git/blob

   1 From 50c02d93990e94bc6bacc1ce71e7895b40316b5a Mon Sep 17 00:00:00 2001
   2 From: Linus Torvalds <torvalds@linux-foundation.org>
   3 Date: Sun, 20 Apr 2025 15:30:53 -0700
   4 Subject: gcc-15: disable '-Wunterminated-string-initialization' entirely for now
   5
   6 commit 9d7a0577c9db35c4cc52db90bc415ea248446472 upstream.
   7
   8 From: Linus Torvalds <torvalds@linux-foundation.org>
   9
  10 I had left the warning around but as a non-fatal error to get my gcc-15
  11 builds going, but fixed up some of the most annoying warning cases so
  12 that it wouldn't be *too* verbose.
  13
  14 Because I like the _concept_ of the warning, even if I detested the
  15 implementation to shut it up.
  16
  17 It turns out the implementation to shut it up is even more broken than I
  18 thought, and my "shut up most of the warnings" patch just caused fatal
  19 errors on gcc-14 instead.
  20
  21 I had tested with clang, but when I upgrade my development environment,
  22 I try to do it on all machines because I hate having different systems
  23 to maintain, and hadn't realized that gcc-14 now had issues.
  24
  25 The ACPI case is literally why I wanted to have a *type* that doesn't
  26 trigger the warning (see commit d5d45a7f2619: "gcc-15: make
  27 'unterminated string initialization' just a warning"), instead of
  28 marking individual places as "__nonstring".
  29
  30 But gcc-14 doesn't like that __nonstring location that shut gcc-15 up,
  31 because it's on an array of char arrays, not on one single array:
  32
  33   drivers/acpi/tables.c:399:1: error: 'nonstring' attribute ignored on objects of type 'const char[][4]' [-Werror=attributes]
  34     399 | static const char table_sigs[][ACPI_NAMESEG_SIZE] __initconst __nonstring = {
  35         | ^~~~~~
  36
  37 and my attempts to nest it properly with a type had failed, because of
  38 how gcc doesn't like marking the types as having attributes, only
  39 symbols.
  40
  41 There may be some trick to it, but I was already annoyed by the bad
  42 attribute design, now I'm just entirely fed up with it.
  43
  44 I wish gcc had a proper way to say "this type is a *byte* array, not a
  45 string".
  46
  47 The obvious thing would be to distinguish between "char []" and an
  48 explicitly signed "unsigned char []" (as opposed to an implicitly
  49 unsigned char, which is typically an architecture-specific default, but
  50 for the kernel is universal thanks to '-funsigned-char').
  51
  52 But any "we can typedef a 8-bit type to not become a string just because
  53 it's an array" model would be fine.
  54
  55 But "__attribute__((nonstring))" is sadly not that sane model.
  56
  57 Reported-by: Chris Clayton <chris2553@googlemail.com>
  58 Fixes: 4b4bd8c50f48 ("gcc-15: acpi: sprinkle random '__nonstring' crumbles around")
  59 Fixes: d5d45a7f2619 ("gcc-15: make 'unterminated string initialization' just a warning")
  60 Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
  61 [nathan: drivers/acpi diff dropped due to lack of 4b4bd8c50f48 in stable]
  62 Signed-off-by: Nathan Chancellor <nathan@kernel.org>
  63 Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
  64 ---
  65  Makefile |    4 ++--
  66  1 file changed, 2 insertions(+), 2 deletions(-)
  67
  68 --- a/Makefile
  69 +++ b/Makefile
  70 @@ -1001,8 +1001,8 @@ KBUILD_CFLAGS += $(call cc-option, -fstr
  71  KBUILD_CFLAGS-$(CONFIG_CC_NO_STRINGOP_OVERFLOW) += $(call cc-option, -Wno-stringop-overflow)
  72  KBUILD_CFLAGS-$(CONFIG_CC_STRINGOP_OVERFLOW) += $(call cc-option, -Wstringop-overflow)
  73
  74 -#Currently, disable -Wunterminated-string-initialization as an error
  75 -KBUILD_CFLAGS += $(call cc-option, -Wno-error=unterminated-string-initialization)
  76 +#Currently, disable -Wunterminated-string-initialization as broken
  77 +KBUILD_CFLAGS += $(call cc-option, -Wno-unterminated-string-initialization)
  78
  79  # disable invalid "can't wrap" optimizations for signed / pointers
  80  KBUILD_CFLAGS  += -fno-strict-overflow