Skip to contentSkip to conversation

toolchain: granular choice for stack protector

PATCH Superseded Yann E. MORIN 3 participants 7 messages Dec 26, 2015 – Dec 27, 2015

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 toolchain: granular choice for stack protector

From: Steven Noonan <[email protected]>

Currently, we only support two levels of stach-smashing protection:
- entirely disabled,
- protect _all_ functions with -fstack-protector-all.

-fstack-protector-all tends to be far too aggressive and impacts performance too much to be worth on a real product.

Add a choice that allows us to select between different levels of stack-smashing protection:
- none
- basic (NEW)
- strong (NEW)
- all

The differences are documented in the GCC online documentation:

    https://gcc.gnu.org/onlinedocs/gcc-4.9.2/gcc/Optimize-Options.html

[yann.morin.1998 at free.fr:
- rebase
- add legacy handling
- SSP-strong depends on gcc >= 4.9
- slightly simple ifeq-block in package/Makefile.in
- keep the comment in the choice; add a comment shen strong is not available
- update commit log
]

---
Changes v1 -> v2:
- see commit log ;-) (Yann)

---
Note: I (Yann) have only slightly tested this patch. More testing is in order before we can apply this. Steven, care to see if it still fits your need? Thanks! :-)

---
 Config.in           | 50 +++++++++++++++++++++++++++++++++++++++++++++-----
 Config.in.legacy    |  8 ++++++++
 package/Makefile.in |  8 +++++++-
 3 files changed, 60 insertions(+), 6 deletions(-)
Config.in +45 -5
index 0be44d9..1e85d78 100644
@@ -522,12 +522,14 @@ config BR2_GOOGLE_BREAKPAD_INCLUDE_FILES
 
 endif
 
-config BR2_ENABLE_SSP
+choice
 	bool "build code with Stack Smashing Protection"
-	depends on BR2_TOOLCHAIN_HAS_SSP
+	default BR2_SSP_ALL if BR2_ENABLE_SSP # legacy
+	default BR2_SSP_STRONG if BR2_TOOLCHAIN_GCC_AT_LEAST_4_9
+	default BR2_SSP_REGULAR
 	help
-	  Enable stack smashing protection support using GCCs
-	  -fstack-protector-all option.
+	  Enable stack smashing protection support using GCC's
+	  -fstack-protector option family.
 
 	  See http://www.linuxfromscratch.org/hints/downloads/files/ssp.txt
 	  for details.
@@ -536,9 +538,47 @@ config BR2_ENABLE_SSP
 	  support. This is always the case for glibc and eglibc
 	  toolchain, but is optional in uClibc toolchains.
 
-comment "enabling Stack Smashing Protection requires support in the toolchain"
+config BR2_SSP_NONE
+	bool "None"
+	help
+	  Disable stack-smashing protection.
+
+comment "Stack Smashing Protection needs a toolchain w/ SSP"
 	depends on !BR2_TOOLCHAIN_HAS_SSP
 
+config BR2_SSP_REGULAR
+	bool "-fstack-protector"
+	depends on BR2_TOOLCHAIN_HAS_SSP
+	help
+	  Emit extra code to check for buffer overflows, such as stack
+	  smashing attacks. This is done by adding a guard variable to
+	  functions with vulnerable objects. This includes functions
+	  that call alloca, and functions with buffers larger than 8
+	  bytes. The guards are initialized when a function is entered
+	  and then checked when the function exits. If a guard check
+	  fails, an error message is printed and the program exits.
+
+config BR2_SSP_STRONG
+	bool "-fstack-protector-strong"
+	depends on BR2_TOOLCHAIN_HAS_SSP
+	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_HAS_SSP
+	depends on !BR2_TOOLCHAIN_GCC_AT_LEAST_4_9
+
+config BR2_SSP_ALL
+	bool "-fstack-protector-all"
+	depends on BR2_TOOLCHAIN_HAS_SSP
+	help
+	  Like -fstack-protector except that all functions are protected.
+
+endchoice
+
 choice
 	bool "libraries"
 	default BR2_SHARED_LIBS if BR2_BINFMT_SUPPORTS_SHARED
Config.in.legacy +8 -0
index 2628796..5d45d04 100644
@@ -145,6 +145,14 @@ endif
 ###############################################################################
 comment "Legacy options removed in 2016.02"
 
+# BR2_ENABLE_SSP is still referenced in Config.in (default in choice)
+config BR2_ENABLE_SSP
+	bool "Stack Smashing protection now has different levels"
+	help
+	  The protection offered by SSP can now be selected from different
+	  protection levels. Be sure to review the SSP level in the build
+	  options menu.
+
 config BR2_PACKAGE_DIRECTFB_CLE266
 	bool "cle266 driver for directfb removed"
 	select BR2_LEGACY
package/Makefile.in +7 -1
index 82a66c2..c5652af 100644
@@ -159,7 +159,13 @@ TARGET_CFLAGS += -msep-data
 TARGET_CXXFLAGS += -msep-data
 endif
 
-ifeq ($(BR2_ENABLE_SSP),y)
+ifeq ($(BR2_SSP_REGULAR),y)
+TARGET_CFLAGS += -fstack-protector
+TARGET_CXXFLAGS += -fstack-protector
+else ifeq ($(BR2_SSP_STRONG),y)
+TARGET_CFLAGS += -fstack-protector-strong
+TARGET_CXXFLAGS += -fstack-protector-strong
+else ifeq ($(BR2_SSP_ALL),y)
 TARGET_CFLAGS += -fstack-protector-all
 TARGET_CXXFLAGS += -fstack-protector-all
 endif
Signature · 1 line
1.9.1
#1 lore raw

PATCH v2 1/1 toolchain: granular choice for stack protector

LGTM. Also looks like a better implementation than what I've been using (like the BR2_TOOLCHAIN_GCC_AT_LEAST_4_9 usage). The next time I rebase my tree I'll include this version instead and see how things go (not expecting any difference in behavior based on my read of this, though).

Quoted 158 lines · Yann E. MORIN

From: Steven Noonan <[email protected]>

Currently, we only support two levels of stach-smashing protection:
- entirely disabled,
- protect _all_ functions with -fstack-protector-all.

-fstack-protector-all tends to be far too aggressive and impacts performance too much to be worth on a real product.

Add a choice that allows us to select between different levels of stack-smashing protection:
- none
- basic (NEW)
- strong (NEW)
- all

The differences are documented in the GCC online documentation:

    https://gcc.gnu.org/onlinedocs/gcc-4.9.2/gcc/Optimize-Options.html

[yann.morin.1998 at free.fr:
- rebase
- add legacy handling
- SSP-strong depends on gcc >= 4.9
- slightly simple ifeq-block in package/Makefile.in
- keep the comment in the choice; add a comment shen strong is not available
- update commit log
]

---
Changes v1 -> v2:
- see commit log ;-) (Yann)

---
Note: I (Yann) have only slightly tested this patch. More testing is in order before we can apply this. Steven, care to see if it still fits your need? Thanks! :-)

---
 Config.in           | 50 +++++++++++++++++++++++++++++++++++++++++++++-----
 Config.in.legacy    |  8 ++++++++
 package/Makefile.in |  8 +++++++-
 3 files changed, 60 insertions(+), 6 deletions(-)
Config.in +45 -5
index 0be44d9..1e85d78 100644
@@ -522,12 +522,14 @@ config BR2_GOOGLE_BREAKPAD_INCLUDE_FILES
 
 endif
 
-config BR2_ENABLE_SSP
+choice
        bool "build code with Stack Smashing Protection"
-       depends on BR2_TOOLCHAIN_HAS_SSP
+       default BR2_SSP_ALL if BR2_ENABLE_SSP # legacy
+       default BR2_SSP_STRONG if BR2_TOOLCHAIN_GCC_AT_LEAST_4_9
+       default BR2_SSP_REGULAR
        help
-         Enable stack smashing protection support using GCCs
-         -fstack-protector-all option.
+         Enable stack smashing protection support using GCC's
+         -fstack-protector option family.
 
          See http://www.linuxfromscratch.org/hints/downloads/files/ssp.txt
          for details.
@@ -536,9 +538,47 @@ config BR2_ENABLE_SSP
          support. This is always the case for glibc and eglibc
          toolchain, but is optional in uClibc toolchains.
 
-comment "enabling Stack Smashing Protection requires support in the toolchain"
+config BR2_SSP_NONE
+       bool "None"
+       help
+         Disable stack-smashing protection.
+
+comment "Stack Smashing Protection needs a toolchain w/ SSP"
        depends on !BR2_TOOLCHAIN_HAS_SSP
 
+config BR2_SSP_REGULAR
+       bool "-fstack-protector"
+       depends on BR2_TOOLCHAIN_HAS_SSP
+       help
+         Emit extra code to check for buffer overflows, such as stack
+         smashing attacks. This is done by adding a guard variable to
+         functions with vulnerable objects. This includes functions
+         that call alloca, and functions with buffers larger than 8
+         bytes. The guards are initialized when a function is entered
+         and then checked when the function exits. If a guard check
+         fails, an error message is printed and the program exits.
+
+config BR2_SSP_STRONG
+       bool "-fstack-protector-strong"
+       depends on BR2_TOOLCHAIN_HAS_SSP
+       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_HAS_SSP
+       depends on !BR2_TOOLCHAIN_GCC_AT_LEAST_4_9
+
+config BR2_SSP_ALL
+       bool "-fstack-protector-all"
+       depends on BR2_TOOLCHAIN_HAS_SSP
+       help
+         Like -fstack-protector except that all functions are protected.
+
+endchoice
+
 choice
        bool "libraries"
        default BR2_SHARED_LIBS if BR2_BINFMT_SUPPORTS_SHARED
Config.in.legacy +8 -0
index 2628796..5d45d04 100644
@@ -145,6 +145,14 @@ endif
 ###############################################################################
 comment "Legacy options removed in 2016.02"
 
+# BR2_ENABLE_SSP is still referenced in Config.in (default in choice)
+config BR2_ENABLE_SSP
+       bool "Stack Smashing protection now has different levels"
+       help
+         The protection offered by SSP can now be selected from different
+         protection levels. Be sure to review the SSP level in the build
+         options menu.
+
 config BR2_PACKAGE_DIRECTFB_CLE266
        bool "cle266 driver for directfb removed"
        select BR2_LEGACY
package/Makefile.in +7 -2
index 82a66c2..c5652af 100644
@@ -159,7 +159,13 @@ TARGET_CFLAGS += -msep-data
 TARGET_CXXFLAGS += -msep-data
 endif
 
-ifeq ($(BR2_ENABLE_SSP),y)
+ifeq ($(BR2_SSP_REGULAR),y)
+TARGET_CFLAGS += -fstack-protector
+TARGET_CXXFLAGS += -fstack-protector
+else ifeq ($(BR2_SSP_STRONG),y)
+TARGET_CFLAGS += -fstack-protector-strong
+TARGET_CXXFLAGS += -fstack-protector-strong
+else ifeq ($(BR2_SSP_ALL),y)
 TARGET_CFLAGS += -fstack-protector-all
 TARGET_CXXFLAGS += -fstack-protector-all
 endif
--

1.9.1

#2 lore raw

PATCH v2 1/1 toolchain: granular choice for stack protector

Steven, All,

On 2015-12-26 16:27 -0800, Steven Noonan spake thusly:

Quoted 2 lines

LGTM. Also looks like a better implementation than what I've been using (like the BR2_TOOLCHAIN_GCC_AT_LEAST_4_9 usage).

GCC_AT_LEAST_X_Y has been introduced after you sent your patch, so indeed you could not have used it. ;-)

Quoted 4 lines

The next time I
rebase my tree I'll include this version instead and see how things go (not expecting any difference in behavior based on my read of this, though).

Thanks! :-)

Regards,
Yann E. MORIN.

Quoted 159 lines · Yann E. MORIN
Quoted 158 lines

From: Steven Noonan <[email protected]>

Currently, we only support two levels of stach-smashing protection:
- entirely disabled,
- protect _all_ functions with -fstack-protector-all.

-fstack-protector-all tends to be far too aggressive and impacts performance too much to be worth on a real product.

Add a choice that allows us to select between different levels of stack-smashing protection:
- none
- basic (NEW)
- strong (NEW)
- all

The differences are documented in the GCC online documentation:

    https://gcc.gnu.org/onlinedocs/gcc-4.9.2/gcc/Optimize-Options.html

[yann.morin.1998 at free.fr:
- rebase
- add legacy handling
- SSP-strong depends on gcc >= 4.9
- slightly simple ifeq-block in package/Makefile.in
- keep the comment in the choice; add a comment shen strong is not available
- update commit log
]

---
Changes v1 -> v2:
- see commit log ;-) (Yann)

---
Note: I (Yann) have only slightly tested this patch. More testing is in order before we can apply this. Steven, care to see if it still fits your need? Thanks! :-)

---
 Config.in           | 50 +++++++++++++++++++++++++++++++++++++++++++++-----
 Config.in.legacy    |  8 ++++++++
 package/Makefile.in |  8 +++++++-
 3 files changed, 60 insertions(+), 6 deletions(-)
Config.in +45 -5
index 0be44d9..1e85d78 100644
@@ -522,12 +522,14 @@ config BR2_GOOGLE_BREAKPAD_INCLUDE_FILES
 
 endif
 
-config BR2_ENABLE_SSP
+choice
        bool "build code with Stack Smashing Protection"
-       depends on BR2_TOOLCHAIN_HAS_SSP
+       default BR2_SSP_ALL if BR2_ENABLE_SSP # legacy
+       default BR2_SSP_STRONG if BR2_TOOLCHAIN_GCC_AT_LEAST_4_9
+       default BR2_SSP_REGULAR
        help
-         Enable stack smashing protection support using GCCs
-         -fstack-protector-all option.
+         Enable stack smashing protection support using GCC's
+         -fstack-protector option family.
 
          See http://www.linuxfromscratch.org/hints/downloads/files/ssp.txt
          for details.
@@ -536,9 +538,47 @@ config BR2_ENABLE_SSP
          support. This is always the case for glibc and eglibc
          toolchain, but is optional in uClibc toolchains.
 
-comment "enabling Stack Smashing Protection requires support in the toolchain"
+config BR2_SSP_NONE
+       bool "None"
+       help
+         Disable stack-smashing protection.
+
+comment "Stack Smashing Protection needs a toolchain w/ SSP"
        depends on !BR2_TOOLCHAIN_HAS_SSP
 
+config BR2_SSP_REGULAR
+       bool "-fstack-protector"
+       depends on BR2_TOOLCHAIN_HAS_SSP
+       help
+         Emit extra code to check for buffer overflows, such as stack
+         smashing attacks. This is done by adding a guard variable to
+         functions with vulnerable objects. This includes functions
+         that call alloca, and functions with buffers larger than 8
+         bytes. The guards are initialized when a function is entered
+         and then checked when the function exits. If a guard check
+         fails, an error message is printed and the program exits.
+
+config BR2_SSP_STRONG
+       bool "-fstack-protector-strong"
+       depends on BR2_TOOLCHAIN_HAS_SSP
+       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_HAS_SSP
+       depends on !BR2_TOOLCHAIN_GCC_AT_LEAST_4_9
+
+config BR2_SSP_ALL
+       bool "-fstack-protector-all"
+       depends on BR2_TOOLCHAIN_HAS_SSP
+       help
+         Like -fstack-protector except that all functions are protected.
+
+endchoice
+
 choice
        bool "libraries"
        default BR2_SHARED_LIBS if BR2_BINFMT_SUPPORTS_SHARED
Config.in.legacy +8 -0
index 2628796..5d45d04 100644
@@ -145,6 +145,14 @@ endif
 ###############################################################################
 comment "Legacy options removed in 2016.02"
 
+# BR2_ENABLE_SSP is still referenced in Config.in (default in choice)
+config BR2_ENABLE_SSP
+       bool "Stack Smashing protection now has different levels"
+       help
+         The protection offered by SSP can now be selected from different
+         protection levels. Be sure to review the SSP level in the build
+         options menu.
+
 config BR2_PACKAGE_DIRECTFB_CLE266
        bool "cle266 driver for directfb removed"
        select BR2_LEGACY
package/Makefile.in +7 -2
index 82a66c2..c5652af 100644
@@ -159,7 +159,13 @@ TARGET_CFLAGS += -msep-data
 TARGET_CXXFLAGS += -msep-data
 endif
 
-ifeq ($(BR2_ENABLE_SSP),y)
+ifeq ($(BR2_SSP_REGULAR),y)
+TARGET_CFLAGS += -fstack-protector
+TARGET_CXXFLAGS += -fstack-protector
+else ifeq ($(BR2_SSP_STRONG),y)
+TARGET_CFLAGS += -fstack-protector-strong
+TARGET_CXXFLAGS += -fstack-protector-strong
+else ifeq ($(BR2_SSP_ALL),y)
 TARGET_CFLAGS += -fstack-protector-all
 TARGET_CXXFLAGS += -fstack-protector-all
 endif
--

1.9.1

Signature · 6 lines
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 223 225 172 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'
#3 lore raw

PATCH v2 1/1 toolchain: granular choice for stack protector

In reply to Yann E. MORIN

Yann,

Config.in config BR2_GOOGLE_BREAKPAD_INCLUDE_FILES · line 528

Quoted 6 lines · Yann E. MORIN
Config.in +3 -2
-config BR2_ENABLE_SSP
+choice
 	bool "build code with Stack Smashing Protection"
-	depends on BR2_TOOLCHAIN_HAS_SSP
+	default BR2_SSP_ALL if BR2_ENABLE_SSP # legacy
+	default BR2_SSP_STRONG if BR2_TOOLCHAIN_GCC_AT_LEAST_4_9

I don't understand this part. Why would we now defalut on BR2_SSP_STRONG if gcc >= 4.9 ?

I think we should keep the default of not having SSP enabled.

Best regards,

Thomas

Signature · 3 lines
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com
#4 lore raw

PATCH v2 1/1 toolchain: granular choice for stack protector

Thomas, All,

On 2015-12-27 11:19 +0100, Thomas Petazzoni spake thusly:

Quoted 10 lines · Yann E. MORIN

Config.in config BR2_GOOGLE_BREAKPAD_INCLUDE_FILES · line 528

Quoted 6 lines
Config.in +3 -2
-config BR2_ENABLE_SSP
+choice
 	bool "build code with Stack Smashing Protection"
-	depends on BR2_TOOLCHAIN_HAS_SSP
+	default BR2_SSP_ALL if BR2_ENABLE_SSP # legacy
+	default BR2_SSP_STRONG if BR2_TOOLCHAIN_GCC_AT_LEAST_4_9

I don't understand this part. Why would we now defalut on BR2_SSP_STRONG if gcc >= 4.9 ?

Hmmm... I did not think too much about that. I just "updated" the patch and kept its meaning as-is.

Quoted 1 line

I think we should keep the default of not having SSP enabled.

Yes, makes sense. At best, the change in the default should have been in another patch.

I'll fix and respin.

Regards,
Yann E. MORIN.

Signature · 6 lines
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 223 225 172 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'
#5 lore raw

PATCH v2 1/1 toolchain: granular choice for stack protector

Quoted 14 lines · Yann E. MORIN

Thomas, All,

On 2015-12-27 11:19 +0100, Thomas Petazzoni spake thusly:

Quoted 10 lines · Yann E. MORIN

Config.in config BR2_GOOGLE_BREAKPAD_INCLUDE_FILES · line 528

Quoted 6 lines
Config.in +3 -2
-config BR2_ENABLE_SSP
+choice
    bool "build code with Stack Smashing Protection"
-   depends on BR2_TOOLCHAIN_HAS_SSP
+   default BR2_SSP_ALL if BR2_ENABLE_SSP # legacy
+   default BR2_SSP_STRONG if BR2_TOOLCHAIN_GCC_AT_LEAST_4_9

I don't understand this part. Why would we now defalut on BR2_SSP_STRONG if gcc >= 4.9 ?

Quoted 7 lines

Hmmm... I did not think too much about that. I just "updated" the patch and kept its meaning as-is.

Quoted 1 line

I think we should keep the default of not having SSP enabled.

Yes, makes sense. At best, the change in the default should have been in another patch.

I could be wrong, but I *think* BuildRoot used to default to SSP on. If not, I don't know why I would have changed the default. I'm not a particularly big fan of SSP to begin with, so I'd have no incentive to default it enabled.

Quoted 4 lines

I'll fix and respin.

Regards,
Yann E. MORIN.

#6 lore raw

PATCH v2 1/1 toolchain: granular choice for stack protector

Steven, All,

On 2015-12-27 02:44 -0800, Steven Noonan spake thusly:

Quoted 18 lines · Yann E. MORIN
Quoted 17 lines

On 2015-12-27 11:19 +0100, Thomas Petazzoni spake thusly:

Quoted 10 lines · Yann E. MORIN

Config.in config BR2_GOOGLE_BREAKPAD_INCLUDE_FILES · line 528

Quoted 6 lines
Config.in +3 -2
-config BR2_ENABLE_SSP
+choice
    bool "build code with Stack Smashing Protection"
-   depends on BR2_TOOLCHAIN_HAS_SSP
+   default BR2_SSP_ALL if BR2_ENABLE_SSP # legacy
+   default BR2_SSP_STRONG if BR2_TOOLCHAIN_GCC_AT_LEAST_4_9

I don't understand this part. Why would we now defalut on BR2_SSP_STRONG if gcc >= 4.9 ?

Hmmm... I did not think too much about that. I just "updated" the patch and kept its meaning as-is.

Quoted 1 line

I think we should keep the default of not having SSP enabled.

Quoted 7 lines
Quoted 2 lines

Yes, makes sense. At best, the change in the default should have been in another patch.

I could be wrong, but I *think* BuildRoot used to default to SSP on. If not, I don't know why I would have changed the default. I'm not a particularly big fan of SSP to begin with, so I'd have no incentive to default it enabled.

OK, so I will definitely remove the new defaults before I respin.

Thanks for the feedback! :-)

Regards,
Yann E. MORIN.

Quoted 4 lines
Quoted 4 lines

I'll fix and respin.

Regards,
Yann E. MORIN.

Signature · 6 lines
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 223 225 172 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'