public inbox for libc-alpha@sourceware.org
 help / color / mirror / Atom feed
From: Siddhesh Poyarekar <siddhesh@sourceware.org>
To: libc-alpha@sourceware.org
Cc: fweimer@redhat.com
Subject: [PATCH 3/3] Build libc-start with stack protector for SHARED
Date: Wed, 10 Mar 2021 15:44:00 +0530	[thread overview]
Message-ID: <20210310101400.3904724-4-siddhesh@sourceware.org> (raw)
In-Reply-To: <20210310101400.3904724-1-siddhesh@sourceware.org>

This does not change the emitted code since __libc_start_main does not
return, but is important for formal flags compliance.

This also cleans up the cosmetic inconsistency in the stack protector
flags in csu, especially the incorrect value of STACK_PROTECTOR_LEVEL.
---
 Makeconfig   |  8 ++++++++
 csu/Makefile | 22 ++++++++++++----------
 elf/Makefile |  4 ----
 3 files changed, 20 insertions(+), 14 deletions(-)

diff --git a/Makeconfig b/Makeconfig
index 0a4811b5e5..01f8638c2e 100644
--- a/Makeconfig
+++ b/Makeconfig
@@ -856,6 +856,14 @@ ifneq ($(stack-protector),)
 +stack-protector=$(stack-protector)
 endif
 
+# Some routines are unsafe to build with stack-protection since they're called
+# before the stack check guard is set up.  Provide a way to disable stack
+# protector.  The first argument is the extension (.o, .os, .oS) and the second
+# is a list of routines that this path should be applied to.
+define elide-stack-protector
+$(if $(filter $(@F),$(patsubst %,%$(1),$(2))), $(no-stack-protector))
+endef
+
 # This is the program that generates makefile dependencies from C source files.
 # The -MP flag tells GCC >= 3.2 (which we now require) to produce dummy
 # targets for headers so that removed headers don't break the build.
diff --git a/csu/Makefile b/csu/Makefile
index e587434be8..3054329cea 100644
--- a/csu/Makefile
+++ b/csu/Makefile
@@ -45,18 +45,20 @@ install-lib = $(start-installed-name) g$(start-installed-name) $(csu-dummies)
 # code is compiled with special flags.
 tests =
 
-CFLAGS-.o += $(no-stack-protector)
-CFLAGS-.op += $(no-stack-protector)
-CFLAGS-.os += $(no-stack-protector)
-
-# Dummy object not actually used for anything.  It is linked into
-# crt1.o nevertheless, which in turn is statically linked into
+# static-reloc.os is a dummy object not actually used for anything.  It is
+# linked into crt1.o nevertheless, which in turn is statically linked into
 # applications, so that build flags matter.
 # See <https://sourceware.org/ml/libc-alpha/2018-07/msg00101.html>.
-# NB: Using $(stack-protector) in this way causes a wrong definition
-# STACK_PROTECTOR_LEVEL due to the preceding $(no-stack-protector),
-# but it does not matter for this source file.
-CFLAGS-static-reloc.os += $(stack-protector)
+#
+# libc-start.os is safe to be built with stack protector since
+# __libc_start_main is called after stack canary setup is done.
+ssp-safe.os = static-reloc libc-start
+
+CFLAGS-.o += $(call elide-stack-protector,.o,$(routines))
+CFLAGS-.op += $(call elide-stack-protector,.op,$(routines))
+CFLAGS-.oS += $(call elide-stack-protector,.oS,$(routines))
+CFLAGS-.os += $(call elide-stack-protector,.os,$(filter-out \
+						 $(ssp-safe.os),$(routines)))
 
 ifeq (yes,$(build-shared))
 extra-objs += S$(start-installed-name) gmon-start.os
diff --git a/elf/Makefile b/elf/Makefile
index b06bf6ca20..285d9f2f3c 100644
--- a/elf/Makefile
+++ b/elf/Makefile
@@ -83,10 +83,6 @@ endif
 # Also compile all routines in the static library that are elided from
 # the shared libc because they are in libc.a in the same way.
 
-define elide-stack-protector
-$(if $(filter $(@F),$(patsubst %,%$(1),$(2))), $(no-stack-protector))
-endef
-
 CFLAGS-.o += $(call elide-stack-protector,.o,$(elide-routines.os))
 CFLAGS-.op += $(call elide-stack-protector,.op,$(elide-routines.os))
 CFLAGS-.os += $(call elide-stack-protector,.os,$(all-rtld-routines))
-- 
2.29.2


  parent reply	other threads:[~2021-03-10 10:14 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-03-10 10:13 [PATCH v2 0/3] Clean up stack-protector-all build Siddhesh Poyarekar
2021-03-10 10:13 ` [PATCH 1/3] Add inhibit_stack_protector to ifuncmain9 [BZ #25680] Siddhesh Poyarekar
2021-03-10 12:50   ` Adhemerval Zanella
2021-03-12  8:52     ` Siddhesh Poyarekar
2021-03-12 17:34       ` Adhemerval Zanella
2021-03-12 19:55         ` H.J. Lu
2021-03-13  6:01           ` Siddhesh Poyarekar
2021-03-13 13:44             ` H.J. Lu
2021-03-13  6:07         ` Siddhesh Poyarekar
2021-03-15 13:58           ` Adhemerval Zanella
2021-03-15 16:34             ` Siddhesh Poyarekar
2021-03-15 13:59   ` Adhemerval Zanella
2021-03-10 10:13 ` [PATCH 2/3] Build get-cpuid-feature-leaf.c without stack-protector [BZ #27555] Siddhesh Poyarekar
2021-03-15 13:59   ` Adhemerval Zanella
2021-03-10 10:14 ` Siddhesh Poyarekar [this message]
2021-03-15 14:00   ` [PATCH 3/3] Build libc-start with stack protector for SHARED Adhemerval Zanella
  -- strict thread matches above, loose matches on Subject: below --
2021-03-09 18:12 [PATCH 0/3] Clean up stack-protector-all build Siddhesh Poyarekar
2021-03-09 18:12 ` [PATCH 3/3] Build libc-start with stack protector for SHARED Siddhesh Poyarekar
2021-03-10  8:00   ` Florian Weimer

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20210310101400.3904724-4-siddhesh@sourceware.org \
    --to=siddhesh@sourceware.org \
    --cc=fweimer@redhat.com \
    --cc=libc-alpha@sourceware.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).