Skip to contentSkip to conversation

toolchain: prepare to pass more additional CFLAGS via the wrapper

PATCH New yann.morin at orange.com 4 participants 18 messages Mar 12, 2019 – Aug 5, 2019 1 message not in the archive

lore
Share & tools
lore
#1 lore raw

PATCH v2 1/5 toolchain: prepare to pass more additional CFLAGS via the wrapper

In reply to a message not in the archive

From: "Yann E. MORIN" <[email protected]>

Currently, we pass the user-supplied so-called target optimisation flags to the wrapper.

We're going to have additional such CFLAGS to pass, so push-back the formatting loop to quote the options at the last moment.

Cc Arnout Vandecappelle <[email protected]>
Cc Thomas Petazzoni <[email protected]>
Cc Thomas De Schampheleire <[email protected]>
---
 toolchain/toolchain-wrapper.mk | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)
toolchain/toolchain-wrapper.mk +5 -3
index 613f5f6c56..c9e6fd3f3d 100644
@@ -16,11 +16,13 @@ endif
 TOOLCHAIN_WRAPPER_ARGS = $($(PKG)_TOOLCHAIN_WRAPPER_ARGS)
 TOOLCHAIN_WRAPPER_ARGS += -DBR_SYSROOT='"$(STAGING_SUBDIR)"'
 
+TOOLCHAIN_WRAPPER_OPTS = \
+	$(call qstrip,$(BR2_TARGET_OPTIMIZATION))
+
 # We create a list like '"-mfoo", "-mbar", "-mbarfoo"' so that each flag is a
 # separate argument when used in execv() by the toolchain wrapper.
-TOOLCHAIN_WRAPPER_OPTS = \
-	$(foreach f,$(call qstrip,$(BR2_TARGET_OPTIMIZATION)),"$(f)"$(comma))
-TOOLCHAIN_WRAPPER_ARGS += -DBR_ADDITIONAL_CFLAGS='$(TOOLCHAIN_WRAPPER_OPTS)'
+TOOLCHAIN_WRAPPER_ARGS += \
+	-DBR_ADDITIONAL_CFLAGS='$(foreach f,$(TOOLCHAIN_WRAPPER_OPTS),"$(f)"$(comma))'
 
 ifeq ($(BR2_CCACHE),y)
 TOOLCHAIN_WRAPPER_ARGS += -DBR_CCACHE
Signature · 15 lines
2.17.1


_________________________________________________________________________________________________________________________

Ce message et ses pieces jointes peuvent contenir des informations confidentielles ou privilegiees et ne doivent donc
pas etre diffuses, exploites ou copies sans autorisation. Si vous avez recu ce message par erreur, veuillez le signaler
a l'expediteur et le detruire ainsi que les pieces jointes. Les messages electroniques etant susceptibles d'alteration,
Orange decline toute responsabilite si ce message a ete altere, deforme ou falsifie. Merci.

This message and its attachments may contain confidential or privileged information that may be protected by law;
they should not be distributed, used or copied without authorisation.
If you have received this email in error, please notify the sender and delete this message and its attachments.
As emails may be altered, Orange is not liable for messages that have been modified, changed or falsified.
Thank you.
#3 lore raw

PATCH v2 2/5 toolchain: set the ssp gcc option in kconfig

In reply to a message not in the archive

From: "Yann E. MORIN" <[email protected]>

Currently, we repeat all the SSP level selection deep down to the toolchain wrapper itself, where we eventually translate it to the actual SSP option to use. This is a bit redundant.

Additionally, we will want to check that the toolchain actually supports that option (for those toolchain where it was backported).

So, move the translation into kconfig, and add the qstrip'ed value to the additional flags passed to the wrapper. Add it before user-supplied opitons, to keep the previous behaviour (and allow anyone crazy-enough to override it with BR2_TARGET_OPTIMIZATION).

Cc Matt Weber <[email protected]>
Cc Thomas Petazzoni <[email protected]>
Cc Thomas De Schampheleire <[email protected]>

---
Changes v1 -> v2:
- use TOOLCHAIN_WRAPPER_OPTS (Arnout)

---
 Config.in                      | 6 ++++++
 toolchain/toolchain-wrapper.c  | 9 ---------
 toolchain/toolchain-wrapper.mk | 9 +--------
 3 files changed, 7 insertions(+), 17 deletions(-)
Config.in +6 -0
index d58d8dc04a..757ad1ca40 100644
@@ -764,6 +764,12 @@ config BR2_SSP_ALL
 
 endchoice
 
+config BR2_SSP_OPTION
+	string
+	default "-fstack-protector"        if BR2_SSP_REGULAR
+	default "-fstack-protector-strong" if BR2_SSP_STRONG
+	default "-fstack-protector-all"    if BR2_SSP_ALL
+
 comment "Stack Smashing Protection needs a toolchain w/ SSP"
 	depends on !BR2_TOOLCHAIN_HAS_SSP
 
toolchain/toolchain-wrapper.c +0 -9
index e9c5cd9d32..c73a0cc079 100644
@@ -98,15 +98,6 @@ static char *predef_args[] = {
 #if defined(BR_MIPS_TARGET_BIG_ENDIAN) || defined(BR_ARC_TARGET_BIG_ENDIAN)
 	"-EB",
 #endif
-#ifdef BR_SSP_REGULAR
-	"-fstack-protector",
-#endif
-#ifdef BR_SSP_STRONG
-	"-fstack-protector-strong",
-#endif
-#ifdef BR_SSP_ALL
-	"-fstack-protector-all",
-#endif
 #ifdef BR_ADDITIONAL_CFLAGS
 	BR_ADDITIONAL_CFLAGS
 #endif
toolchain/toolchain-wrapper.mk +1 -8
index c9e6fd3f3d..ca66fa7ba4 100644
@@ -17,6 +17,7 @@ TOOLCHAIN_WRAPPER_ARGS = $($(PKG)_TOOLCHAIN_WRAPPER_ARGS)
 TOOLCHAIN_WRAPPER_ARGS += -DBR_SYSROOT='"$(STAGING_SUBDIR)"'
 
 TOOLCHAIN_WRAPPER_OPTS = \
+	$(call qstrip,$(BR2_SSP_OPTION)) \
 	$(call qstrip,$(BR2_TARGET_OPTIMIZATION))
 
 # We create a list like '"-mfoo", "-mbar", "-mbarfoo"' so that each flag is a
@@ -53,14 +54,6 @@ else ifeq ($(BR2_RELRO_FULL),y)
 TOOLCHAIN_WRAPPER_ARGS += -DBR2_RELRO_FULL
 endif
 
-ifeq ($(BR2_SSP_REGULAR),y)
-TOOLCHAIN_WRAPPER_ARGS += -DBR_SSP_REGULAR
-else ifeq ($(BR2_SSP_STRONG),y)
-TOOLCHAIN_WRAPPER_ARGS += -DBR_SSP_STRONG
-else ifeq ($(BR2_SSP_ALL),y)
-TOOLCHAIN_WRAPPER_ARGS += -DBR_SSP_ALL
-endif
-
 define TOOLCHAIN_WRAPPER_BUILD
 	$(HOSTCC) $(HOST_CFLAGS) $(TOOLCHAIN_WRAPPER_ARGS) \
 		-s -Wl,--hash-style=$(TOOLCHAIN_WRAPPER_HASH_STYLE) \
Signature · 15 lines
2.17.1


_________________________________________________________________________________________________________________________

Ce message et ses pieces jointes peuvent contenir des informations confidentielles ou privilegiees et ne doivent donc
pas etre diffuses, exploites ou copies sans autorisation. Si vous avez recu ce message par erreur, veuillez le signaler
a l'expediteur et le detruire ainsi que les pieces jointes. Les messages electroniques etant susceptibles d'alteration,
Orange decline toute responsabilite si ce message a ete altere, deforme ou falsifie. Merci.

This message and its attachments may contain confidential or privileged information that may be protected by law;
they should not be distributed, used or copied without authorisation.
If you have received this email in error, please notify the sender and delete this message and its attachments.
As emails may be altered, Orange is not liable for messages that have been modified, changed or falsified.
Thank you.
#6 lore raw

PATCH v2 3/5 toolchain: check the SSP option is known

In reply to a message not in the archive

From: "Yann E. MORIN" <[email protected]>

Some toolchain vendors may have backported those options to older gcc versions, and we have no way to know, so we have to check that the user's selection is acceptable.

Extend the macro that currently checks for SSP in the toolchain, with a new test that the actual SSP option is recognised and accepted.

Note that the SSP option is either totaly empty, or an already-quoted string, so we can safely and easily assign it to a shell variable to test and use it.

Note that we do not introduce BR2_TOOLCHAIN_HAS_SSP_STRONG, because:

- our internal toolchain infra only supports gcc >= 4.9, so it has

    SSP strong;

- of the external pre-built toolchains, only the codesourcery-arm one has a gcc-4.8 which lacks SSP strong, all the others have a gcc >= 4.9;

- we'd still have to do the actual check for custom external toolchains anyway.

So, we're not adding BR2_TOOLCHAIN_HAS_SSP_STRONG just for a single case.

Cc Matt Weber <[email protected]>
Cc Thomas Petazzoni <[email protected]>
Cc Thomas De Schampheleire <[email protected]>
Cc Arnout Vandecappelle <[email protected]>

---
Changes v1 -> v2:
- expand the commit log to explain why we're not adding BR2_TOOLCHAIN_HAS_SSP_STRONG (Arnout)

---
 toolchain/helpers.mk                                   | 8 ++++++++
 toolchain/toolchain-external/pkg-toolchain-external.mk | 2 +-
 2 files changed, 9 insertions(+), 1 deletion(-)
toolchain/helpers.mk +8 -0
index e5520c00c3..ba097e83cf 100644
@@ -415,6 +415,7 @@ check_unusable_toolchain = \
 # Check if the toolchain has SSP (stack smashing protector) support
 #
 # $1: cross-gcc path
+# $2: gcc ssp option
 #
 check_toolchain_ssp = \
 	__CROSS_CC=$(strip $1) ; \
@@ -427,6 +428,13 @@ check_toolchain_ssp = \
 		echo "SSP support not available in this toolchain, please disable BR2_TOOLCHAIN_EXTERNAL_HAS_SSP" ; \
 		exit 1 ; \
 	fi ; \
+	__SSP_OPTION=$(2); \
+	if [ -n "$${__SSP_OPTION}" ] ; then \
+		if ! echo 'void main(){}' | $${__CROSS_CC} -Werror $${__SSP_OPTION} -x c - -o $(BUILD_DIR)/.br-toolchain-test.tmp >/dev/null 2>&1 ; then \
+			echo "SSP option $${__SSP_OPTION} not available in this toolchain, please select another SSP level" ; \
+			exit 1 ; \
+		fi; \
+	fi; \
 	rm -f $(BUILD_DIR)/.br-toolchain-test.tmp*
 
 #
toolchain/toolchain-external/pkg-toolchain-external.mk +1 -1
index db3570d96f..00cbd7b17a 100644
@@ -549,7 +549,7 @@ define $(2)_CONFIGURE_CMDS
 	else \
 		$$(call check_glibc,$$$${SYSROOT_DIR}) ; \
 	fi
-	$$(Q)$$(call check_toolchain_ssp,$$(TOOLCHAIN_EXTERNAL_CC))
+	$$(Q)$$(call check_toolchain_ssp,$$(TOOLCHAIN_EXTERNAL_CC),$(BR2_SSP_OPTION))
 endef
 
 $(2)_TOOLCHAIN_WRAPPER_ARGS += $$(TOOLCHAIN_EXTERNAL_TOOLCHAIN_WRAPPER_ARGS)
Signature · 15 lines
2.17.1


_________________________________________________________________________________________________________________________

Ce message et ses pieces jointes peuvent contenir des informations confidentielles ou privilegiees et ne doivent donc
pas etre diffuses, exploites ou copies sans autorisation. Si vous avez recu ce message par erreur, veuillez le signaler
a l'expediteur et le detruire ainsi que les pieces jointes. Les messages electroniques etant susceptibles d'alteration,
Orange decline toute responsabilite si ce message a ete altere, deforme ou falsifie. Merci.

This message and its attachments may contain confidential or privileged information that may be protected by law;
they should not be distributed, used or copied without authorisation.
If you have received this email in error, please notify the sender and delete this message and its attachments.
As emails may be altered, Orange is not liable for messages that have been modified, changed or falsified.
Thank you.
#12 lore raw

PATCH v2 4/5 toolchain: -fstack-protector-strong can be back-ported

In reply to a message not in the archive

From: "Yann E. MORIN" <[email protected]>

Currently, use of -fstack-protector-strong is only available for gcc starting with 4.9, on the assumption that it appeared with that version.

Although this is true, it happens that quite a few vendors will have back-ported -fstack-protector-strong to older gcc versions (at least 4.8 seen in the wild).

Remove the guard against gcc>=4.9, and expand the help text.

Cc Matt Weber <[email protected]>
Cc Thomas Petazzoni <[email protected]>
Cc Thomas De Schampheleire <[email protected]>

---
Notes:

We could have changed the guard to something like:

    depends on BR2_TOOLCHAIN_GCC_AT_LEAST_4_9 || BR2_TOOLCHAIN_EXTERNAL

However, the latest gcc we support in the internal toolchain *is* gcc-4.9, so the condition would have always been true. Hence, we just drop the condition.

---
 Config.in | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
Config.in +3 -3
index 757ad1ca40..d5a0460f98 100644
@@ -746,14 +746,14 @@ config BR2_SSP_REGULAR
 
 config BR2_SSP_STRONG
 	bool "-fstack-protector-strong"
-	depends on BR2_TOOLCHAIN_GCC_AT_LEAST_4_9
 	help
 	  Like -fstack-protector but includes additional functions to be
 	  protected - those that have local array definitions, or have
 	  references to local frame addresses.
 
-comment "Stack Smashing Protection strong needs a toolchain w/ gcc >= 4.9"
-	depends on !BR2_TOOLCHAIN_GCC_AT_LEAST_4_9
+	  -fstack-protector-strong officially appeared in gcc 4.9, but
+	  some vendors have backported -fstack-protector-strong to older
+	  versions of gcc.
 
 config BR2_SSP_ALL
 	bool "-fstack-protector-all"
Signature · 15 lines
2.17.1


_________________________________________________________________________________________________________________________

Ce message et ses pieces jointes peuvent contenir des informations confidentielles ou privilegiees et ne doivent donc
pas etre diffuses, exploites ou copies sans autorisation. Si vous avez recu ce message par erreur, veuillez le signaler
a l'expediteur et le detruire ainsi que les pieces jointes. Les messages electroniques etant susceptibles d'alteration,
Orange decline toute responsabilite si ce message a ete altere, deforme ou falsifie. Merci.

This message and its attachments may contain confidential or privileged information that may be protected by law;
they should not be distributed, used or copied without authorisation.
If you have received this email in error, please notify the sender and delete this message and its attachments.
As emails may be altered, Orange is not liable for messages that have been modified, changed or falsified.
Thank you.
#15 lore raw

PATCH v2 5/5 toolchain: allow PIC/PIE without RELRO

In reply to a message not in the archive

From: "Yann E. MORIN" <[email protected]>

In commit 7484c1c3b806 (toolchain/toolchain-wrapper: add BR2_RELRO_), we added the PIC/PIE flags, but based on the RELRO_FULL condition.

It is however totally possible to do a PIC/PIE executable without RELRO_FULL, as it is also valid to do a PIC/PIE build with RELRO_PARTIAL.

Add a new option that now governs the PIC/PIE flags.

Note: it is unknown if RELRO_FULL really needs PIC/PIE or not, so we keep the current situation, where RELRO-FULL forces PIC/PIE compilation. Decoupling can come later from an interested party.

Cc Matt Weber <[email protected]>
Cc Thomas Petazzoni <[email protected]>
Cc Thomas De Schampheleire <[email protected]>
---
 Config.in                      | 8 ++++++++
 toolchain/toolchain-wrapper.c  | 2 +-
 toolchain/toolchain-wrapper.mk | 4 ++++
 3 files changed, 13 insertions(+), 1 deletion(-)
Config.in +8 -0
index d5a0460f98..31fea3ab34 100644
@@ -712,6 +712,13 @@ endmenu
 
 comment "Security Hardening Options"
 
+config BR2_PIC_PIE
+	bool "Build code with PIC/PIE"
+	depends on BR2_SHARED_LIBS
+	help
+	  Generate Position-Independent Code (PIC) and link
+	  Position-Independent Executables (PIE).
+
 choice
 	bool "Stack Smashing Protection"
 	default BR2_SSP_ALL if BR2_ENABLE_SSP # legacy
@@ -794,6 +801,7 @@ config BR2_RELRO_PARTIAL
 
 config BR2_RELRO_FULL
 	bool "Full"
+	select BR2_PIC_PIE
 	help
 	  This option includes the partial configuration, but also marks
 	  the GOT as read-only at the cost of initialization time during
toolchain/toolchain-wrapper.c +1 -1
index c73a0cc079..7a4b9c4007 100644
@@ -367,7 +367,7 @@ int main(int argc, char **argv)
 		*cur++ = "-Wno-builtin-macro-redefined";
 	}
 
-#ifdef BR2_RELRO_FULL
+#ifdef BR2_PIC_PIE
 	/* Patterned after Fedora/Gentoo hardening approaches.
 	 * https://fedoraproject.org/wiki/Changes/Harden_All_Packages
 	 * https://wiki.gentoo.org/wiki/Hardened/Toolchain#Position_Independent_Executables_.28PIEs.29
toolchain/toolchain-wrapper.mk +4 -0
index ca66fa7ba4..3c42146cea 100644
@@ -48,6 +48,10 @@ ifeq ($(BR2_CCACHE_USE_BASEDIR),y)
 TOOLCHAIN_WRAPPER_ARGS += -DBR_CCACHE_BASEDIR='"$(BASE_DIR)"'
 endif
 
+ifeq ($(BR2_PIC_PIE),y)
+TOOLCHAIN_WRAPPER_ARGS += -DBR2_PIC_PIE
+endif
+
 ifeq ($(BR2_RELRO_PARTIAL),y)
 TOOLCHAIN_WRAPPER_ARGS += -DBR2_RELRO_PARTIAL
 else ifeq ($(BR2_RELRO_FULL),y)
Signature · 15 lines
2.17.1


_________________________________________________________________________________________________________________________

Ce message et ses pieces jointes peuvent contenir des informations confidentielles ou privilegiees et ne doivent donc
pas etre diffuses, exploites ou copies sans autorisation. Si vous avez recu ce message par erreur, veuillez le signaler
a l'expediteur et le detruire ainsi que les pieces jointes. Les messages electroniques etant susceptibles d'alteration,
Orange decline toute responsabilite si ce message a ete altere, deforme ou falsifie. Merci.

This message and its attachments may contain confidential or privileged information that may be protected by law;
they should not be distributed, used or copied without authorisation.
If you have received this email in error, please notify the sender and delete this message and its attachments.
As emails may be altered, Orange is not liable for messages that have been modified, changed or falsified.
Thank you.
#4 lore raw

PATCH v2 2/5 toolchain: set the ssp gcc option in kconfig

In reply to yann.morin at orange.com

Yann,

Quoted 36 lines · On Tue, Mar 12, 2019 at 7:10 AM <[email protected]> wrote:

From: "Yann E. MORIN" <[email protected]>

Currently, we repeat all the SSP level selection deep down to the toolchain wrapper itself, where we eventually translate it to the actual SSP option to use. This is a bit redundant.

Additionally, we will want to check that the toolchain actually supports that option (for those toolchain where it was backported).

So, move the translation into kconfig, and add the qstrip'ed value to the additional flags passed to the wrapper. Add it before user-supplied opitons, to keep the previous behaviour (and allow anyone crazy-enough to override it with BR2_TARGET_OPTIMIZATION).

Cc Matt Weber <[email protected]>
Cc Thomas Petazzoni <[email protected]>
Cc Thomas De Schampheleire <[email protected]>

---
Changes v1 -> v2:
- use TOOLCHAIN_WRAPPER_OPTS (Arnout)

---
 Config.in                      | 6 ++++++
 toolchain/toolchain-wrapper.c  | 9 ---------
 toolchain/toolchain-wrapper.mk | 9 +--------
 3 files changed, 7 insertions(+), 17 deletions(-)
Config.in +0 -0
index d58d8dc04a..757ad1ca40 100644
@@ -764,6 +764,12 @@ config BR2_SSP_ALL
 
 endchoice

Config.in config BR2_SSP_ALL · line 775

Quoted 10 lines
Config.in +6 -0
 
+config BR2_SSP_OPTION
+       string
+       default "-fstack-protector"        if BR2_SSP_REGULAR
+       default "-fstack-protector-strong" if BR2_SSP_STRONG
+       default "-fstack-protector-all"    if BR2_SSP_ALL
+
 comment "Stack Smashing Protection needs a toolchain w/ SSP"
        depends on !BR2_TOOLCHAIN_HAS_SSP
 

Much cleaner then when I pulled SSP into the wrapper :-)

#16 lore raw

PATCH v2 5/5 toolchain: allow PIC/PIE without RELRO

In reply to yann.morin at orange.com

Yann,

Quoted 44 lines · On Tue, Mar 12, 2019 at 7:09 AM <[email protected]> wrote:

From: "Yann E. MORIN" <[email protected]>

In commit 7484c1c3b806 (toolchain/toolchain-wrapper: add BR2_RELRO_), we added the PIC/PIE flags, but based on the RELRO_FULL condition.

It is however totally possible to do a PIC/PIE executable without RELRO_FULL, as it is also valid to do a PIC/PIE build with RELRO_PARTIAL.

Add a new option that now governs the PIC/PIE flags.

Note: it is unknown if RELRO_FULL really needs PIC/PIE or not, so we keep the current situation, where RELRO-FULL forces PIC/PIE compilation. Decoupling can come later from an interested party.

Cc Matt Weber <[email protected]>
Cc Thomas Petazzoni <[email protected]>
Cc Thomas De Schampheleire <[email protected]>
---
 Config.in                      | 8 ++++++++
 toolchain/toolchain-wrapper.c  | 2 +-
 toolchain/toolchain-wrapper.mk | 4 ++++
 3 files changed, 13 insertions(+), 1 deletion(-)
Config.in +7 -0
index d5a0460f98..31fea3ab34 100644
@@ -712,6 +712,13 @@ endmenu
 
 comment "Security Hardening Options"
 
+config BR2_PIC_PIE
+       bool "Build code with PIC/PIE"
+       depends on BR2_SHARED_LIBS
+       help
+         Generate Position-Independent Code (PIC) and link
+         Position-Independent Executables (PIE).
+
 choice
        bool "Stack Smashing Protection"
        default BR2_SSP_ALL if BR2_ENABLE_SSP # legacy

Config.in config BR2_RELRO_PARTIAL · line 804

Quoted 4 lines
Config.in +1 -0
@@ -794,6 +801,7 @@ config BR2_RELRO_PARTIAL
 
 config BR2_RELRO_FULL
        bool "Full"
+       select BR2_PIC_PIE

In the previous email chain it was being discussed if PIC/PIE was required for full RELRO. Like you guys mentioned, I believe it was always just lumped into the configuration the other distros called "full". However, I'm pretty sure they are independent and you could have full RELRO without PIC/PIE (I did not test this theory, just checked some docs and the theory holds). I'd be on the fence if we should remove this select and keep the BR2_PIC_PIE as something you'd need to independently set if you want it with BR2_RELRO_FULL. I guess, better to make that sort of change now then after more time has gone by.

toolchain/toolchain-wrapper.c int main(int argc, char **argv) · line 370

Quoted 13 lines
Config.in +0 -0
        help
          This option includes the partial configuration, but also marks
          the GOT as read-only at the cost of initialization time during
toolchain/toolchain-wrapper.c +1 -1
index c73a0cc079..7a4b9c4007 100644
@@ -367,7 +367,7 @@ int main(int argc, char **argv)
                *cur++ = "-Wno-builtin-macro-redefined";
        }
 
-#ifdef BR2_RELRO_FULL
+#ifdef BR2_PIC_PIE
#17 lore raw

PATCH v2 5/5 toolchain: allow PIC/PIE without RELRO

Matt, All,

On 2019-03-12 08:50 -0500, Matthew Weber spake thusly:

Quoted 4 lines · On Tue, Mar 12, 2019 at 7:09 AM <[email protected]> wrote:
Quoted 3 lines

From: "Yann E. MORIN" <[email protected]>
In commit 7484c1c3b806 (toolchain/toolchain-wrapper: add BR2_RELRO_), we added the PIC/PIE flags, but based on the RELRO_FULL condition.

[--SNIP--]

Quoted 10 lines
Quoted 3 lines

config BR2_RELRO_FULL
bool "Full"
+ select BR2_PIC_PIE

In the previous email chain it was being discussed if PIC/PIE was required for full RELRO. Like you guys mentioned, I believe it was always just lumped into the configuration the other distros called "full". However, I'm pretty sure they are independent and you could have full RELRO without PIC/PIE (I did not test this theory, just checked some docs and the theory holds).

Arnout did test it, and it indeed works.

Quoted 5 lines

I'd be on the fence if we
should remove this select and keep the BR2_PIC_PIE as something you'd need to independently set if you want it with BR2_RELRO_FULL. I guess, better to make that sort of change now then after more time has gone by.

I did not do that in this patch, to introduce the minimal disruption possible. I.e. a configuration which had relro-full will still get PIC/PIE. This patch just adds the possiblity to do PIC/PIE without relro-full.

If we really want to decouple the two, then I think we should do that in a separate patch, which just drops this new select, to allow relro-full without PIC/PIE.

I'll do that in a followup patch.

Regards,
Yann E. MORIN.

Quoted 19 lines

toolchain/toolchain-wrapper.c int main(int argc, char **argv) · line 370

Quoted 13 lines

help
This option includes the partial configuration, but also marks the GOT as read-only at the cost of initialization time during

toolchain/toolchain-wrapper.c +1 -1
index c73a0cc079..7a4b9c4007 100644
@@ -367,7 +367,7 @@ int main(int argc, char **argv)
                *cur++ = "-Wno-builtin-macro-redefined";
        }
 
-#ifdef BR2_RELRO_FULL
+#ifdef BR2_PIC_PIE

_______________________________________________
buildroot mailing list
buildroot at busybox.net
http://lists.busybox.net/mailman/listinfo/buildroot

Signature · 21 lines
                                        ____________
.-----------------.--------------------:       _    :------------------.
|  Yann E. MORIN  | Real-Time Embedded |    __/ )   | /"\ ASCII RIBBON |
| +33 534.541.179 | Software  Designer |  _/ - /'   | \ / CAMPAIGN     |
| +33 638.411.245 '--------------------: (_    `--, |  X  AGAINST      |
|      yann.morin (at) orange.com      |_="    ,--' | / \ HTML MAIL    |
'--------------------------------------:______/_____:------------------'


_________________________________________________________________________________________________________________________

Ce message et ses pieces jointes peuvent contenir des informations confidentielles ou privilegiees et ne doivent donc
pas etre diffuses, exploites ou copies sans autorisation. Si vous avez recu ce message par erreur, veuillez le signaler
a l'expediteur et le detruire ainsi que les pieces jointes. Les messages electroniques etant susceptibles d'alteration,
Orange decline toute responsabilite si ce message a ete altere, deforme ou falsifie. Merci.

This message and its attachments may contain confidential or privileged information that may be protected by law;
they should not be distributed, used or copied without authorisation.
If you have received this email in error, please notify the sender and delete this message and its attachments.
As emails may be altered, Orange is not liable for messages that have been modified, changed or falsified.
Thank you.
#2 lore raw

PATCH v2 1/5 toolchain: prepare to pass more additional CFLAGS via the wrapper

In reply to yann.morin at orange.com
Quoted 5 lines · yann.morin at orange.com

From: "Yann E. MORIN" <[email protected]>

Currently, we pass the user-supplied so-called target optimisation flags to the wrapper.

Quoted 8 lines

We're going to have additional such CFLAGS to pass, so push-back the formatting loop to quote the options at the last moment.

Cc Arnout Vandecappelle <[email protected]>
Cc Thomas Petazzoni <[email protected]>
Cc Thomas De Schampheleire <[email protected]>

Applied to master, thanks.

Regards,
Arnout

Quoted 26 lines
---
 toolchain/toolchain-wrapper.mk | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)
toolchain/toolchain-wrapper.mk +5 -3
index 613f5f6c56..c9e6fd3f3d 100644
@@ -16,11 +16,13 @@ endif
 TOOLCHAIN_WRAPPER_ARGS = $($(PKG)_TOOLCHAIN_WRAPPER_ARGS)
 TOOLCHAIN_WRAPPER_ARGS += -DBR_SYSROOT='"$(STAGING_SUBDIR)"'
 
+TOOLCHAIN_WRAPPER_OPTS = \
+	$(call qstrip,$(BR2_TARGET_OPTIMIZATION))
+
 # We create a list like '"-mfoo", "-mbar", "-mbarfoo"' so that each flag is a
 # separate argument when used in execv() by the toolchain wrapper.
-TOOLCHAIN_WRAPPER_OPTS = \
-	$(foreach f,$(call qstrip,$(BR2_TARGET_OPTIMIZATION)),"$(f)"$(comma))
-TOOLCHAIN_WRAPPER_ARGS += -DBR_ADDITIONAL_CFLAGS='$(TOOLCHAIN_WRAPPER_OPTS)'
+TOOLCHAIN_WRAPPER_ARGS += \
+	-DBR_ADDITIONAL_CFLAGS='$(foreach f,$(TOOLCHAIN_WRAPPER_OPTS),"$(f)"$(comma))'
 
 ifeq ($(BR2_CCACHE),y)
 TOOLCHAIN_WRAPPER_ARGS += -DBR_CCACHE
#5 lore raw

PATCH v2 2/5 toolchain: set the ssp gcc option in kconfig

In reply to yann.morin at orange.com
Quoted 14 lines · yann.morin at orange.com

From: "Yann E. MORIN" <[email protected]>

Currently, we repeat all the SSP level selection deep down to the toolchain wrapper itself, where we eventually translate it to the actual SSP option to use. This is a bit redundant.

Additionally, we will want to check that the toolchain actually supports that option (for those toolchain where it was backported).

So, move the translation into kconfig, and add the qstrip'ed value to the additional flags passed to the wrapper. Add it before user-supplied opitons, to keep the previous behaviour (and allow anyone crazy-enough to override it with BR2_TARGET_OPTIMIZATION).

Quoted 4 lines
Cc Matt Weber <[email protected]>
Cc Thomas Petazzoni <[email protected]>
Cc Thomas De Schampheleire <[email protected]>

Applied to master, thanks.

Regards,
Arnout

Quoted 75 lines

---
Changes v1 -> v2:
- use TOOLCHAIN_WRAPPER_OPTS (Arnout)

---
 Config.in                      | 6 ++++++
 toolchain/toolchain-wrapper.c  | 9 ---------
 toolchain/toolchain-wrapper.mk | 9 +--------
 3 files changed, 7 insertions(+), 17 deletions(-)
Config.in +6 -0
index d58d8dc04a..757ad1ca40 100644
@@ -764,6 +764,12 @@ config BR2_SSP_ALL
 
 endchoice
 
+config BR2_SSP_OPTION
+	string
+	default "-fstack-protector"        if BR2_SSP_REGULAR
+	default "-fstack-protector-strong" if BR2_SSP_STRONG
+	default "-fstack-protector-all"    if BR2_SSP_ALL
+
 comment "Stack Smashing Protection needs a toolchain w/ SSP"
 	depends on !BR2_TOOLCHAIN_HAS_SSP
 
toolchain/toolchain-wrapper.c +0 -9
index e9c5cd9d32..c73a0cc079 100644
@@ -98,15 +98,6 @@ static char *predef_args[] = {
 #if defined(BR_MIPS_TARGET_BIG_ENDIAN) || defined(BR_ARC_TARGET_BIG_ENDIAN)
 	"-EB",
 #endif
-#ifdef BR_SSP_REGULAR
-	"-fstack-protector",
-#endif
-#ifdef BR_SSP_STRONG
-	"-fstack-protector-strong",
-#endif
-#ifdef BR_SSP_ALL
-	"-fstack-protector-all",
-#endif
 #ifdef BR_ADDITIONAL_CFLAGS
 	BR_ADDITIONAL_CFLAGS
 #endif
toolchain/toolchain-wrapper.mk +1 -8
index c9e6fd3f3d..ca66fa7ba4 100644
@@ -17,6 +17,7 @@ TOOLCHAIN_WRAPPER_ARGS = $($(PKG)_TOOLCHAIN_WRAPPER_ARGS)
 TOOLCHAIN_WRAPPER_ARGS += -DBR_SYSROOT='"$(STAGING_SUBDIR)"'
 
 TOOLCHAIN_WRAPPER_OPTS = \
+	$(call qstrip,$(BR2_SSP_OPTION)) \
 	$(call qstrip,$(BR2_TARGET_OPTIMIZATION))
 
 # We create a list like '"-mfoo", "-mbar", "-mbarfoo"' so that each flag is a
@@ -53,14 +54,6 @@ else ifeq ($(BR2_RELRO_FULL),y)
 TOOLCHAIN_WRAPPER_ARGS += -DBR2_RELRO_FULL
 endif
 
-ifeq ($(BR2_SSP_REGULAR),y)
-TOOLCHAIN_WRAPPER_ARGS += -DBR_SSP_REGULAR
-else ifeq ($(BR2_SSP_STRONG),y)
-TOOLCHAIN_WRAPPER_ARGS += -DBR_SSP_STRONG
-else ifeq ($(BR2_SSP_ALL),y)
-TOOLCHAIN_WRAPPER_ARGS += -DBR_SSP_ALL
-endif
-
 define TOOLCHAIN_WRAPPER_BUILD
 	$(HOSTCC) $(HOST_CFLAGS) $(TOOLCHAIN_WRAPPER_ARGS) \
 		-s -Wl,--hash-style=$(TOOLCHAIN_WRAPPER_HASH_STYLE) \
#7 lore raw

PATCH v2 3/5 toolchain: check the SSP option is known

In reply to yann.morin at orange.com
Quoted 56 lines · yann.morin at orange.com

From: "Yann E. MORIN" <[email protected]>

Some toolchain vendors may have backported those options to older gcc versions, and we have no way to know, so we have to check that the user's selection is acceptable.

Extend the macro that currently checks for SSP in the toolchain, with a new test that the actual SSP option is recognised and accepted.

Note that the SSP option is either totaly empty, or an already-quoted string, so we can safely and easily assign it to a shell variable to test and use it.

Note that we do not introduce BR2_TOOLCHAIN_HAS_SSP_STRONG, because:

- our internal toolchain infra only supports gcc >= 4.9, so it has

    SSP strong;

- of the external pre-built toolchains, only the codesourcery-arm one has a gcc-4.8 which lacks SSP strong, all the others have a gcc >= 4.9;

- we'd still have to do the actual check for custom external toolchains anyway.

So, we're not adding BR2_TOOLCHAIN_HAS_SSP_STRONG just for a single case.

Cc Matt Weber <[email protected]>
Cc Thomas Petazzoni <[email protected]>
Cc Thomas De Schampheleire <[email protected]>
Cc Arnout Vandecappelle <[email protected]>

---
Changes v1 -> v2:
- expand the commit log to explain why we're not adding BR2_TOOLCHAIN_HAS_SSP_STRONG (Arnout)

---
 toolchain/helpers.mk                                   | 8 ++++++++
 toolchain/toolchain-external/pkg-toolchain-external.mk | 2 +-
 2 files changed, 9 insertions(+), 1 deletion(-)
toolchain/helpers.mk +1 -0
index e5520c00c3..ba097e83cf 100644
@@ -415,6 +415,7 @@ check_unusable_toolchain = \
 # Check if the toolchain has SSP (stack smashing protector) support
 #
 # $1: cross-gcc path
+# $2: gcc ssp option
 #
 check_toolchain_ssp = \
 	__CROSS_CC=$(strip $1) ; \

toolchain/helpers.mk check_toolchain_ssp = \ · line 431

Quoted 4 lines
toolchain/helpers.mk +1 -0
@@ -427,6 +428,13 @@ check_toolchain_ssp = \
 		echo "SSP support not available in this toolchain, please disable BR2_TOOLCHAIN_EXTERNAL_HAS_SSP" ; \
 		exit 1 ; \
 	fi ; \
+	__SSP_OPTION=$(2); \

Is there any reason to pass this as an argument rather than using
$(BR2_SSP_OPTION) directly?

The same goes for $(TOOLCHAIN_EXTERNAL_CC), but that ship has sailed.

Regards,
Arnout

Quoted 23 lines
toolchain/helpers.mk +6 -0
+	if [ -n "$${__SSP_OPTION}" ] ; then \
+		if ! echo 'void main(){}' | $${__CROSS_CC} -Werror $${__SSP_OPTION} -x c - -o $(BUILD_DIR)/.br-toolchain-test.tmp >/dev/null 2>&1 ; then \
+			echo "SSP option $${__SSP_OPTION} not available in this toolchain, please select another SSP level" ; \
+			exit 1 ; \
+		fi; \
+	fi; \
 	rm -f $(BUILD_DIR)/.br-toolchain-test.tmp*
 
 #
toolchain/toolchain-external/pkg-toolchain-external.mk +1 -1
index db3570d96f..00cbd7b17a 100644
@@ -549,7 +549,7 @@ define $(2)_CONFIGURE_CMDS
 	else \
 		$$(call check_glibc,$$$${SYSROOT_DIR}) ; \
 	fi
-	$$(Q)$$(call check_toolchain_ssp,$$(TOOLCHAIN_EXTERNAL_CC))
+	$$(Q)$$(call check_toolchain_ssp,$$(TOOLCHAIN_EXTERNAL_CC),$(BR2_SSP_OPTION))
 endef
 
 $(2)_TOOLCHAIN_WRAPPER_ARGS += $$(TOOLCHAIN_EXTERNAL_TOOLCHAIN_WRAPPER_ARGS)
#8 lore raw

PATCH v2 3/5 toolchain: check the SSP option is known

Arnout, All,

On 2019-03-13 00:25 +0100, Arnout Vandecappelle spake thusly:

Quoted 2 lines · yann.morin at orange.com
Quoted 1 line

From: "Yann E. MORIN" <[email protected]>

[--SNIP--]

Quoted 3 lines
Quoted 3 lines

Note that the SSP option is either totaly empty, or an already-quoted string, so we can safely and easily assign it to a shell variable to test and use it.

Please note this paragraph above, and...

[--SNIP--]

Quoted 8 lines

toolchain/helpers.mk check_toolchain_ssp = \ · line 431

Quoted 5 lines
toolchain/helpers.mk +1 -0
@@ -427,6 +428,13 @@ check_toolchain_ssp = \
 		echo "SSP support not available in this toolchain, please disable BR2_TOOLCHAIN_EXTERNAL_HAS_SSP" ; \
 		exit 1 ; \
 	fi ; \
+	__SSP_OPTION=$(2); \

Is there any reason to pass this as an argument rather than using
$(BR2_SSP_OPTION) directly?

... correlate it with your question here.

There are two cases:

- no SSP BR2_SSP_OPTION is not defined, so it is not even the
empty quoted string ""

- some SSP BR2_SSP_OPTION is a non-empty quoted string.

So...

Quoted 1 line

toolchain/helpers.mk check_toolchain_ssp = \ · line 432

Quoted 1 line
toolchain/helpers.mk +1 -0
+	if [ -n "$${__SSP_OPTION}" ] ; then \

... this test would have to use the qstrip-then-quote-anyway dance.

And finally, I tried to mimick the existing code, which lead to the current patch: pass the option as a parameter to the macro, and use an intermediate shell variable.

Regards,
Yann E. MORIN.

Quoted 22 lines
Quoted 22 lines
toolchain/helpers.mk +5 -0
+		if ! echo 'void main(){}' | $${__CROSS_CC} -Werror $${__SSP_OPTION} -x c - -o $(BUILD_DIR)/.br-toolchain-test.tmp >/dev/null 2>&1 ; then \
+			echo "SSP option $${__SSP_OPTION} not available in this toolchain, please select another SSP level" ; \
+			exit 1 ; \
+		fi; \
+	fi; \
 	rm -f $(BUILD_DIR)/.br-toolchain-test.tmp*
 
 #
toolchain/toolchain-external/pkg-toolchain-external.mk +1 -1
index db3570d96f..00cbd7b17a 100644
@@ -549,7 +549,7 @@ define $(2)_CONFIGURE_CMDS
 	else \
 		$$(call check_glibc,$$$${SYSROOT_DIR}) ; \
 	fi
-	$$(Q)$$(call check_toolchain_ssp,$$(TOOLCHAIN_EXTERNAL_CC))
+	$$(Q)$$(call check_toolchain_ssp,$$(TOOLCHAIN_EXTERNAL_CC),$(BR2_SSP_OPTION))
 endef
 
 $(2)_TOOLCHAIN_WRAPPER_ARGS += $$(TOOLCHAIN_EXTERNAL_TOOLCHAIN_WRAPPER_ARGS)
Signature · 21 lines
                                        ____________
.-----------------.--------------------:       _    :------------------.
|  Yann E. MORIN  | Real-Time Embedded |    __/ )   | /"\ ASCII RIBBON |
| +33 534.541.179 | Software  Designer |  _/ - /'   | \ / CAMPAIGN     |
| +33 638.411.245 '--------------------: (_    `--, |  X  AGAINST      |
|      yann.morin (at) orange.com      |_="    ,--' | / \ HTML MAIL    |
'--------------------------------------:______/_____:------------------'


_________________________________________________________________________________________________________________________

Ce message et ses pieces jointes peuvent contenir des informations confidentielles ou privilegiees et ne doivent donc
pas etre diffuses, exploites ou copies sans autorisation. Si vous avez recu ce message par erreur, veuillez le signaler
a l'expediteur et le detruire ainsi que les pieces jointes. Les messages electroniques etant susceptibles d'alteration,
Orange decline toute responsabilite si ce message a ete altere, deforme ou falsifie. Merci.

This message and its attachments may contain confidential or privileged information that may be protected by law;
they should not be distributed, used or copied without authorisation.
If you have received this email in error, please notify the sender and delete this message and its attachments.
As emails may be altered, Orange is not liable for messages that have been modified, changed or falsified.
Thank you.
#9 lore raw

PATCH v2 3/5 toolchain: check the SSP option is known

In reply to yann.morin at orange.com

Hello,

Quoted 6 lines · On Tue, 12 Mar 2019 13:09:34 +0100 <[email protected]> wrote:

# $1: cross-gcc path

toolchain/helpers.mk +1 -0
+# $2: gcc ssp option
 #
 check_toolchain_ssp = \
 	__CROSS_CC=$(strip $1) ; \

toolchain/helpers.mk check_toolchain_ssp = \ · line 437

Quoted 10 lines
toolchain/helpers.mk +7 -0
@@ -427,6 +428,13 @@ check_toolchain_ssp = \
 		echo "SSP support not available in this toolchain, please disable BR2_TOOLCHAIN_EXTERNAL_HAS_SSP" ; \
 		exit 1 ; \
 	fi ; \
+	__SSP_OPTION=$(2); \
+	if [ -n "$${__SSP_OPTION}" ] ; then \
+		if ! echo 'void main(){}' | $${__CROSS_CC} -Werror $${__SSP_OPTION} -x c - -o $(BUILD_DIR)/.br-toolchain-test.tmp >/dev/null 2>&1 ; then \
+			echo "SSP option $${__SSP_OPTION} not available in this toolchain, please select another SSP level" ; \
+			exit 1 ; \
+		fi; \
+	fi; \

Does it really make sense to check for both -fstack-protector (in the code that already exists) *and* BR2_SSP_OPTION ?

Wouldn't it make more sense to just check for BR2_SSP_OPTION ?

Thomas

Signature · 3 lines
Thomas Petazzoni, CTO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
#10 lore raw

PATCH v2 3/5 toolchain: check the SSP option is known

Thomas, All,

On 2019-03-26 20:26 +0100, Thomas Petazzoni spake thusly:

Quoted 23 lines
Quoted 6 lines · On Tue, 12 Mar 2019 13:09:34 +0100 <[email protected]> wrote:

# $1: cross-gcc path

toolchain/helpers.mk +1 -0
+# $2: gcc ssp option
 #
 check_toolchain_ssp = \
 	__CROSS_CC=$(strip $1) ; \

toolchain/helpers.mk check_toolchain_ssp = \ · line 437

Quoted 10 lines
toolchain/helpers.mk +7 -0
@@ -427,6 +428,13 @@ check_toolchain_ssp = \
 		echo "SSP support not available in this toolchain, please disable BR2_TOOLCHAIN_EXTERNAL_HAS_SSP" ; \
 		exit 1 ; \
 	fi ; \
+	__SSP_OPTION=$(2); \
+	if [ -n "$${__SSP_OPTION}" ] ; then \
+		if ! echo 'void main(){}' | $${__CROSS_CC} -Werror $${__SSP_OPTION} -x c - -o $(BUILD_DIR)/.br-toolchain-test.tmp >/dev/null 2>&1 ; then \
+			echo "SSP option $${__SSP_OPTION} not available in this toolchain, please select another SSP level" ; \
+			exit 1 ; \
+		fi; \
+	fi; \

Does it really make sense to check for both -fstack-protector (in the code that already exists) *and* BR2_SSP_OPTION ?

Quoted 1 line

Wouldn't it make more sense to just check for BR2_SSP_OPTION ?

My reasoning behind this additional test, is that there are two things we need to test:

* does the toolchain have SSP support (or not) as claimed by the user?

* does the toolchain actually support the SSP level requested by the user?

There are packages that, wrongly or rightfully so, are not concerned with te SSP level, but just the fact that there is SSP or not.

So, I decided to keep the existing test, and expand it. It means that we may indeed check for -fstack-protector twice.

Regards,
Yann E. MORIN.

Signature · 21 lines
                                        ____________
.-----------------.--------------------:       _    :------------------.
|  Yann E. MORIN  | Real-Time Embedded |    __/ )   | /"\ ASCII RIBBON |
| +33 534.541.179 | Software  Designer |  _/ - /'   | \ / CAMPAIGN     |
| +33 638.411.245 '--------------------: (_    `--, |  X  AGAINST      |
|      yann.morin (at) orange.com      |_="    ,--' | / \ HTML MAIL    |
'--------------------------------------:______/_____:------------------'


_________________________________________________________________________________________________________________________

Ce message et ses pieces jointes peuvent contenir des informations confidentielles ou privilegiees et ne doivent donc
pas etre diffuses, exploites ou copies sans autorisation. Si vous avez recu ce message par erreur, veuillez le signaler
a l'expediteur et le detruire ainsi que les pieces jointes. Les messages electroniques etant susceptibles d'alteration,
Orange decline toute responsabilite si ce message a ete altere, deforme ou falsifie. Merci.

This message and its attachments may contain confidential or privileged information that may be protected by law;
they should not be distributed, used or copied without authorisation.
If you have received this email in error, please notify the sender and delete this message and its attachments.
As emails may be altered, Orange is not liable for messages that have been modified, changed or falsified.
Thank you.
#11 lore raw

PATCH v2 3/5 toolchain: check the SSP option is known

In reply to yann.morin at orange.com
Quoted 25 lines · yann.morin at orange.com

From: "Yann E. MORIN" <[email protected]>

Some toolchain vendors may have backported those options to older gcc versions, and we have no way to know, so we have to check that the user's selection is acceptable.

Extend the macro that currently checks for SSP in the toolchain, with a new test that the actual SSP option is recognised and accepted.

Note that the SSP option is either totaly empty, or an already-quoted string, so we can safely and easily assign it to a shell variable to test and use it.

Note that we do not introduce BR2_TOOLCHAIN_HAS_SSP_STRONG, because:

- our internal toolchain infra only supports gcc >= 4.9, so it has

    SSP strong;

- of the external pre-built toolchains, only the codesourcery-arm one has a gcc-4.8 which lacks SSP strong, all the others have a gcc >= 4.9;

- we'd still have to do the actual check for custom external toolchains anyway.

Quoted 8 lines

So, we're not adding BR2_TOOLCHAIN_HAS_SSP_STRONG just for a single case.

Cc Matt Weber <[email protected]>
Cc Thomas Petazzoni <[email protected]>
Cc Thomas De Schampheleire <[email protected]>
Cc Arnout Vandecappelle <[email protected]>

You replied to all concerns, so I've applied as-is.

Regards,
Arnout

#13 lore raw

PATCH v2 4/5 toolchain: -fstack-protector-strong can be back-ported

In reply to yann.morin at orange.com
Quoted 19 lines · yann.morin at orange.com

From: "Yann E. MORIN" <[email protected]>

Currently, use of -fstack-protector-strong is only available for gcc starting with 4.9, on the assumption that it appeared with that version.

Although this is true, it happens that quite a few vendors will have back-ported -fstack-protector-strong to older gcc versions (at least 4.8 seen in the wild).

Remove the guard against gcc>=4.9, and expand the help text.

Cc Matt Weber <[email protected]>
Cc Thomas Petazzoni <[email protected]>
Cc Thomas De Schampheleire <[email protected]>

---
Notes:

Quoted 6 lines

We could have changed the guard to something like:

    depends on BR2_TOOLCHAIN_GCC_AT_LEAST_4_9 || BR2_TOOLCHAIN_EXTERNAL

However, the latest gcc we support in the internal toolchain *is* gcc-4.9, so the condition would have always been true. Hence, we just drop the condition.

This note is in fact interesting, so I've added it to the commit message.
However, it would have been more appropriate (if we would have had the condition) to make it

depends on BR2_TOOLCHAIN_GCC_AT_LEAST_4_9 || BR2_TOOLCHAIN_EXTERNAL_CUSTOM

(because we know our predefined external toolchains <4.9 *don't* suport this option). In fact, we only have one 4.8 predefined external toolchain: Sourcery ARM. So for that one toolchain, it's not worth adding that condition. The problem will anyway be caught by the check you just added, because it's also executed for predefined external toolchains.

Regards,
Arnout

Quoted 27 lines
---
 Config.in | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
Config.in +3 -3
index 757ad1ca40..d5a0460f98 100644
@@ -746,14 +746,14 @@ config BR2_SSP_REGULAR
 
 config BR2_SSP_STRONG
 	bool "-fstack-protector-strong"
-	depends on BR2_TOOLCHAIN_GCC_AT_LEAST_4_9
 	help
 	  Like -fstack-protector but includes additional functions to be
 	  protected - those that have local array definitions, or have
 	  references to local frame addresses.
 
-comment "Stack Smashing Protection strong needs a toolchain w/ gcc >= 4.9"
-	depends on !BR2_TOOLCHAIN_GCC_AT_LEAST_4_9
+	  -fstack-protector-strong officially appeared in gcc 4.9, but
+	  some vendors have backported -fstack-protector-strong to older
+	  versions of gcc.
 
 config BR2_SSP_ALL
 	bool "-fstack-protector-all"
#18 lore raw

PATCH v2 5/5 toolchain: allow PIC/PIE without RELRO

In reply to yann.morin at orange.com
Quoted 10 lines · yann.morin at orange.com

From: "Yann E. MORIN" <[email protected]>

In commit 7484c1c3b806 (toolchain/toolchain-wrapper: add BR2_RELRO_), we added the PIC/PIE flags, but based on the RELRO_FULL condition.

It is however totally possible to do a PIC/PIE executable without RELRO_FULL, as it is also valid to do a PIC/PIE build with RELRO_PARTIAL.

Add a new option that now governs the PIC/PIE flags.

Quoted 8 lines

Note: it is unknown if RELRO_FULL really needs PIC/PIE or not, so we keep the current situation, where RELRO-FULL forces PIC/PIE compilation. Decoupling can come later from an interested party.

Cc Matt Weber <[email protected]>
Cc Thomas Petazzoni <[email protected]>
Cc Thomas De Schampheleire <[email protected]>

Applied to master, thanks.

Regards,
Arnout

#14 lore raw

PATCH v2 4/5 toolchain: -fstack-protector-strong can be back-ported

In reply to Arnout Vandecappelle

Arnout, All,

On 2019-08-03 23:16 +0200, Arnout Vandecappelle spake thusly:

Quoted 27 lines · yann.morin at orange.com
Quoted 19 lines

From: "Yann E. MORIN" <[email protected]>

Currently, use of -fstack-protector-strong is only available for gcc starting with 4.9, on the assumption that it appeared with that version.

Although this is true, it happens that quite a few vendors will have back-ported -fstack-protector-strong to older gcc versions (at least 4.8 seen in the wild).

Remove the guard against gcc>=4.9, and expand the help text.

Cc Matt Weber <[email protected]>
Cc Thomas Petazzoni <[email protected]>
Cc Thomas De Schampheleire <[email protected]>

---
Notes:

Quoted 6 lines

We could have changed the guard to something like:

    depends on BR2_TOOLCHAIN_GCC_AT_LEAST_4_9 || BR2_TOOLCHAIN_EXTERNAL

However, the latest gcc we support in the internal toolchain *is* gcc-4.9, so the condition would have always been true. Hence, we just drop the condition.

Quoted 5 lines

This note is in fact interesting, so I've added it to the commit message.
However, it would have been more appropriate (if we would have had the condition) to make it

depends on BR2_TOOLCHAIN_GCC_AT_LEAST_4_9 || BR2_TOOLCHAIN_EXTERNAL_CUSTOM

Actually, this would no longer be true, since we recently merged support for preconfigured toolchains from br2-external trees, and such toolchain may well be a gcc-4.8 or earlier with back-ported SSP strong.

Regards,
Yann E. MORIN.

Quoted 36 lines

(because we know our predefined external toolchains <4.9 *don't* suport this option). In fact, we only have one 4.8 predefined external toolchain: Sourcery ARM. So for that one toolchain, it's not worth adding that condition. The problem will anyway be caught by the check you just added, because it's also executed for predefined external toolchains.

Regards,
Arnout

Quoted 27 lines
---
 Config.in | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
Config.in +3 -3
index 757ad1ca40..d5a0460f98 100644
@@ -746,14 +746,14 @@ config BR2_SSP_REGULAR
 
 config BR2_SSP_STRONG
 	bool "-fstack-protector-strong"
-	depends on BR2_TOOLCHAIN_GCC_AT_LEAST_4_9
 	help
 	  Like -fstack-protector but includes additional functions to be
 	  protected - those that have local array definitions, or have
 	  references to local frame addresses.
 
-comment "Stack Smashing Protection strong needs a toolchain w/ gcc >= 4.9"
-	depends on !BR2_TOOLCHAIN_GCC_AT_LEAST_4_9
+	  -fstack-protector-strong officially appeared in gcc 4.9, but
+	  some vendors have backported -fstack-protector-strong to older
+	  versions of gcc.
 
 config BR2_SSP_ALL
 	bool "-fstack-protector-all"
Signature · 21 lines
                                        ____________
.-----------------.--------------------:       _    :------------------.
|  Yann E. MORIN  | Real-Time Embedded |    __/ )   | /"\ ASCII RIBBON |
| +33 534.541.179 | Software  Designer |  _/ - /'   | \ / CAMPAIGN     |
| +33 638.411.245 '--------------------: (_    `--, |  X  AGAINST      |
|      yann.morin (at) orange.com      |_="    ,--' | / \ HTML MAIL    |
'--------------------------------------:______/_____:------------------'


_________________________________________________________________________________________________________________________

Ce message et ses pieces jointes peuvent contenir des informations confidentielles ou privilegiees et ne doivent donc
pas etre diffuses, exploites ou copies sans autorisation. Si vous avez recu ce message par erreur, veuillez le signaler
a l'expediteur et le detruire ainsi que les pieces jointes. Les messages electroniques etant susceptibles d'alteration,
Orange decline toute responsabilite si ce message a ete altere, deforme ou falsifie. Merci.

This message and its attachments may contain confidential or privileged information that may be protected by law;
they should not be distributed, used or copied without authorisation.
If you have received this email in error, please notify the sender and delete this message and its attachments.
As emails may be altered, Orange is not liable for messages that have been modified, changed or falsified.
Thank you.