[U-Boot] [RFC PATCH] ARM: create .secure_stack section only for PSCI
Jon Master reports that QEMU refuses to load a U-Boot image built with CONFIG_ARMV7_NONSEC, but without CONFIG_ARMV7_PSCI since commit 5a3aae68c74e ("ARM: armv7: guard memory reserve for PSCI with #ifdef CONFIG_ARMV7_PSCI").
It looks like only PSCI that needs the Secure stack, so move the #ifdef to guard the whole of .secure_stack allocation in order not to create the empty section.
Signed-off-by: Masahiro Yamada yamada.masahiro@socionext.com Reported-by: Jon Masters jcm@redhat.com Link: http://patchwork.ozlabs.org/patch/664025/ ---
With this commit, the SECURE_MAX_SIZE check will go inside the #ifdef CONFIG_ARMV7_PSCI, so this patch is probably wrong.
I am CCing Chen-Yu Tsai. He mostly expanded this linker script for PSCI work, so I hope he can suggest the correct way for fixing this problem.
arch/arm/cpu/u-boot.lds | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)
diff --git a/arch/arm/cpu/u-boot.lds b/arch/arm/cpu/u-boot.lds index 0a5fae6..37d4c60 100644 --- a/arch/arm/cpu/u-boot.lds +++ b/arch/arm/cpu/u-boot.lds @@ -74,6 +74,7 @@ SECTIONS *(._secure.data) }
+#ifdef CONFIG_ARMV7_PSCI .secure_stack ALIGN(ADDR(.secure_data) + SIZEOF(.secure_data), CONSTANT(COMMONPAGESIZE)) (NOLOAD) : #ifdef __ARMV7_PSCI_STACK_IN_RAM @@ -83,10 +84,10 @@ SECTIONS #endif { KEEP(*(.__secure_stack_start)) -#ifdef CONFIG_ARMV7_PSCI + /* Skip addreses for stack */ . = . + CONFIG_ARMV7_PSCI_NR_CPUS * ARM_PSCI_STACK_SIZE; -#endif + /* Align end of stack section to page boundary */ . = ALIGN(CONSTANT(COMMONPAGESIZE));
@@ -109,6 +110,8 @@ SECTIONS . = LOADADDR(.secure_stack); #endif
+#endif + .__secure_end : AT(ADDR(.__secure_end)) { *(.__secure_end) LONG(0x1d1071c); /* Must output something to reset LMA */
On Mon, Sep 26, 2016 at 1:21 PM, Masahiro Yamada yamada.masahiro@socionext.com wrote:
Jon Master reports that QEMU refuses to load a U-Boot image built with CONFIG_ARMV7_NONSEC, but without CONFIG_ARMV7_PSCI since commit 5a3aae68c74e ("ARM: armv7: guard memory reserve for PSCI with #ifdef CONFIG_ARMV7_PSCI").
It looks like only PSCI that needs the Secure stack, so move the #ifdef to guard the whole of .secure_stack allocation in order not to create the empty section.
Signed-off-by: Masahiro Yamada yamada.masahiro@socionext.com Reported-by: Jon Masters jcm@redhat.com Link: http://patchwork.ozlabs.org/patch/664025/
With this commit, the SECURE_MAX_SIZE check will go inside the #ifdef CONFIG_ARMV7_PSCI, so this patch is probably wrong.
I wonder if you could move the SECURE_MAX_SIZE check outside of the secure_stack section. I might have put it where it is because of some issues, but I can't remember.
I am CCing Chen-Yu Tsai. He mostly expanded this linker script for PSCI work, so I hope he can suggest the correct way for fixing this problem.
The patch looks good. Though I wonder if you need to guard the __secure_stack_start and __secure_stack_end symbols in arch/arm/lib/sections.c as well. Otherwise they might end up in the data section?
Regards ChenYu
arch/arm/cpu/u-boot.lds | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)
diff --git a/arch/arm/cpu/u-boot.lds b/arch/arm/cpu/u-boot.lds index 0a5fae6..37d4c60 100644 --- a/arch/arm/cpu/u-boot.lds +++ b/arch/arm/cpu/u-boot.lds @@ -74,6 +74,7 @@ SECTIONS *(._secure.data) }
+#ifdef CONFIG_ARMV7_PSCI .secure_stack ALIGN(ADDR(.secure_data) + SIZEOF(.secure_data), CONSTANT(COMMONPAGESIZE)) (NOLOAD) : #ifdef __ARMV7_PSCI_STACK_IN_RAM @@ -83,10 +84,10 @@ SECTIONS #endif { KEEP(*(.__secure_stack_start)) -#ifdef CONFIG_ARMV7_PSCI
/* Skip addreses for stack */ . = . + CONFIG_ARMV7_PSCI_NR_CPUS * ARM_PSCI_STACK_SIZE;-#endif
/* Align end of stack section to page boundary */ . = ALIGN(CONSTANT(COMMONPAGESIZE));@@ -109,6 +110,8 @@ SECTIONS . = LOADADDR(.secure_stack); #endif
+#endif
.__secure_end : AT(ADDR(.__secure_end)) { *(.__secure_end) LONG(0x1d1071c); /* Must output something to reset LMA */-- 1.9.1
On Mon, Sep 26, 2016 at 02:21:30PM +0900, Masahiro Yamada wrote:
Jon Master reports that QEMU refuses to load a U-Boot image built with CONFIG_ARMV7_NONSEC, but without CONFIG_ARMV7_PSCI since commit 5a3aae68c74e ("ARM: armv7: guard memory reserve for PSCI with #ifdef CONFIG_ARMV7_PSCI").
It looks like only PSCI that needs the Secure stack, so move the #ifdef to guard the whole of .secure_stack allocation in order not to create the empty section.
Signed-off-by: Masahiro Yamada yamada.masahiro@socionext.com Reported-by: Jon Masters jcm@redhat.com Link: http://patchwork.ozlabs.org/patch/664025/
We're close I think, but this breaks boards like ls1021aqds_nor_lpuart
On Mon, Sep 26, 2016 at 02:21:30PM +0900, Masahiro Yamada wrote:
Jon Master reports that QEMU refuses to load a U-Boot image built with CONFIG_ARMV7_NONSEC, but without CONFIG_ARMV7_PSCI since commit 5a3aae68c74e ("ARM: armv7: guard memory reserve for PSCI with #ifdef CONFIG_ARMV7_PSCI").
It looks like only PSCI that needs the Secure stack, so move the #ifdef to guard the whole of .secure_stack allocation in order not to create the empty section.
Signed-off-by: Masahiro Yamada yamada.masahiro@socionext.com Reported-by: Jon Masters jcm@redhat.com Link: http://patchwork.ozlabs.org/patch/664025/
Applied to u-boot/master, thanks!
... and no, I don't know how / why this caused build failures before, but it did and I replicated it at the time too. But I just got a clean build everywhere. And I've added vexpress-a15 and vexpress-a9 via QEMU to my test setup now.
participants (3)
-
Chen-Yu Tsai -
Masahiro Yamada -
Tom Rini