Re: [PATCH] powerpc/64: BE option to use ELFv2 ABI for big endian kernels
On Sat, Mar 07, 2020 at 10:58:54AM +1000, Nicholas Piggin wrote: > Segher Boessenkool's on March 5, 2020 8:55 pm: > > That name looks perfect to me. You'll have to update REs expecting the > > arch at the end (like /le$/), but you had to already I think? > > le$ is still okay for testing ppc64le, unless you wanted me to add the > -elfv2 suffix on there as well? If that was the case, for consistency > we'd also have to add -elfv1 for the BE v1 case. I was just going to add > -elfv2 for the new variant. I was thinking that, yes, but your plan is better alright (there aren't many supported configs at all, anyway!) Segher
Re: [PATCH] powerpc/64: BE option to use ELFv2 ABI for big endian kernels
Segher Boessenkool's on March 5, 2020 8:55 pm: > On Thu, Mar 05, 2020 at 01:34:22PM +1000, Nicholas Piggin wrote: >> Segher Boessenkool's on March 4, 2020 9:09 am: >> >> +override flavour := linux-ppc64v2 >> > >> > That isn't a good name, heh. This isn't "v2" of anything... Spell out >> > the name "ELFv2"? Or as "elfv2"? It is just a name after all, it is >> > version 1 in all three version fields in the ELF headers! >> >> Yeah okay. This part is only for some weird little perl asm generator >> script, but probably better to be careful. linux-ppc64-elfv2 ? > > That generator is from openssl, or inspired by it, it is everywhere. > So it is more important to get it right than it would seem at first > glance ;-) > > That name looks perfect to me. You'll have to update REs expecting the > arch at the end (like /le$/), but you had to already I think? le$ is still okay for testing ppc64le, unless you wanted me to add the -elfv2 suffix on there as well? If that was the case, for consistency we'd also have to add -elfv1 for the BE v1 case. I was just going to add -elfv2 for the new variant. Thanks, Nick
Re: [PATCH] powerpc/64: BE option to use ELFv2 ABI for big endian kernels
On Thu, Mar 05, 2020 at 01:34:22PM +1000, Nicholas Piggin wrote: > Segher Boessenkool's on March 4, 2020 9:09 am: > >> +override flavour := linux-ppc64v2 > > > > That isn't a good name, heh. This isn't "v2" of anything... Spell out > > the name "ELFv2"? Or as "elfv2"? It is just a name after all, it is > > version 1 in all three version fields in the ELF headers! > > Yeah okay. This part is only for some weird little perl asm generator > script, but probably better to be careful. linux-ppc64-elfv2 ? That generator is from openssl, or inspired by it, it is everywhere. So it is more important to get it right than it would seem at first glance ;-) That name looks perfect to me. You'll have to update REs expecting the arch at the end (like /le$/), but you had to already I think? Segher
Re: [PATCH] powerpc/64: BE option to use ELFv2 ABI for big endian kernels
Segher Boessenkool's on March 4, 2020 9:09 am: > Hi! > > On Tue, Mar 03, 2020 at 11:45:27AM +1000, Nicholas Piggin wrote: >> Provide an option to use ELFv2 ABI for big endian builds. This works on >> GCC and clang (since 2014). it is is not officially supported by the GNU >> toolchain, but it can give some useful advantages of the ELFv2 ABI for >> BE (e.g., less stack usage). Some distros build BE ELFv2 userspace. > > It is not officially supported in the sense that a) as a host config, > it does not exist *at all* (this isn't relevant for the kernel, it does > not use a libc or other libraries, of course); and b) as a target config, > it is not supported in the sense that no one tests it, so we cannot say > anything about what quality code it generates, if it works at all, etc. Yep. > But we *do* allow "-mbig -mabi=elfv2", it's just a chicken-and-egg > problem to have this properly tested. If someone would regularly test > it (incl. sending the test results to gcc-testresults@), I don't see why > it would not become a supported platform. > >> +override flavour := linux-ppc64v2 > > That isn't a good name, heh. This isn't "v2" of anything... Spell out > the name "ELFv2"? Or as "elfv2"? It is just a name after all, it is > version 1 in all three version fields in the ELF headers! Yeah okay. This part is only for some weird little perl asm generator script, but probably better to be careful. linux-ppc64-elfv2 ? > > > Anyway, looks like it will work, let's see where this goes :-) > > Reviewed-by: Segher Boessenkool > > > Segher >
Re: [PATCH] powerpc/64: BE option to use ELFv2 ABI for big endian kernels
Hi! On Tue, Mar 03, 2020 at 11:45:27AM +1000, Nicholas Piggin wrote: > Provide an option to use ELFv2 ABI for big endian builds. This works on > GCC and clang (since 2014). it is is not officially supported by the GNU > toolchain, but it can give some useful advantages of the ELFv2 ABI for > BE (e.g., less stack usage). Some distros build BE ELFv2 userspace. It is not officially supported in the sense that a) as a host config, it does not exist *at all* (this isn't relevant for the kernel, it does not use a libc or other libraries, of course); and b) as a target config, it is not supported in the sense that no one tests it, so we cannot say anything about what quality code it generates, if it works at all, etc. But we *do* allow "-mbig -mabi=elfv2", it's just a chicken-and-egg problem to have this properly tested. If someone would regularly test it (incl. sending the test results to gcc-testresults@), I don't see why it would not become a supported platform. > +override flavour := linux-ppc64v2 That isn't a good name, heh. This isn't "v2" of anything... Spell out the name "ELFv2"? Or as "elfv2"? It is just a name after all, it is version 1 in all three version fields in the ELF headers! Anyway, looks like it will work, let's see where this goes :-) Reviewed-by: Segher Boessenkool Segher
[PATCH] powerpc/64: BE option to use ELFv2 ABI for big endian kernels
Provide an option to use ELFv2 ABI for big endian builds. This works on
GCC and clang (since 2014). it is is not officially supported by the GNU
toolchain, but it can give some useful advantages of the ELFv2 ABI for
BE (e.g., less stack usage). Some distros build BE ELFv2 userspace.
Signed-off-by: Nicholas Piggin
---
arch/powerpc/Kconfig| 19 +++
arch/powerpc/Makefile | 15 ++-
arch/powerpc/boot/Makefile | 4
drivers/crypto/vmx/Makefile | 4
drivers/crypto/vmx/aesp8-ppc.pl | 2 +-
drivers/crypto/vmx/ppc-xlate.pl | 11 +++
6 files changed, 45 insertions(+), 10 deletions(-)
diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig
index 497b7d0b2d7e..31dd921a5145 100644
--- a/arch/powerpc/Kconfig
+++ b/arch/powerpc/Kconfig
@@ -146,6 +146,7 @@ config PPC
select ARCH_WEAK_RELEASE_ACQUIRE
select BINFMT_ELF
select BUILDTIME_TABLE_SORT
+ select BUILD_ELF_V2 if PPC64 && CPU_LITTLE_ENDIAN
select CLONE_BACKWARDS
select DCACHE_WORD_ACCESS if PPC64 && CPU_LITTLE_ENDIAN
select DYNAMIC_FTRACE if FUNCTION_TRACER
@@ -538,6 +539,24 @@ config KEXEC_FILE
config ARCH_HAS_KEXEC_PURGATORY
def_bool KEXEC_FILE
+config BUILD_ELF_V2
+ bool
+
+config BUILD_BIG_ENDIAN_ELF_V2
+ bool "Build big-endian kernel using ELFv2 ABI (EXPERIMENTAL)"
+ depends on PPC64 && CPU_BIG_ENDIAN && EXPERT
+ default n
+ select BUILD_ELF_V2
+ help
+ This builds the kernel image using the ELFv2 ABI, which has a
+ reduced stack overhead and faster function calls. This does not
+ affect the userspace ABIs.
+
+ ELFv2 is the standard ABI for little-endian, but for big-endian
+ this is an experimental option that is not well tested (kernel and
+ toolchain). This requires gcc 4.9 or newer and binutils 2.24 or
+ newer.
+
config RELOCATABLE
bool "Build a relocatable kernel"
depends on PPC64 || (FLATMEM && (44x || FSL_BOOKE))
diff --git a/arch/powerpc/Makefile b/arch/powerpc/Makefile
index f35730548e42..ae8036a0b169 100644
--- a/arch/powerpc/Makefile
+++ b/arch/powerpc/Makefile
@@ -92,10 +92,14 @@ endif
ifdef CONFIG_PPC64
ifndef CONFIG_CC_IS_CLANG
-cflags-$(CONFIG_CPU_BIG_ENDIAN)+= $(call cc-option,-mabi=elfv1)
-cflags-$(CONFIG_CPU_BIG_ENDIAN)+= $(call
cc-option,-mcall-aixdesc)
-aflags-$(CONFIG_CPU_BIG_ENDIAN)+= $(call cc-option,-mabi=elfv1)
-aflags-$(CONFIG_CPU_LITTLE_ENDIAN) += -mabi=elfv2
+ifdef CONFIG_BUILD_ELF_V2
+cflags-y += $(call cc-option,-mabi=elfv2,$(call
cc-option,-mcall-aixdesc))
+aflags-y += $(call cc-option,-mabi=elfv2)
+else
+cflags-y += $(call cc-option,-mabi=elfv1)
+cflags-y += $(call cc-option,-mcall-aixdesc)
+aflags-y += $(call cc-option,-mabi=elfv1)
+endif
endif
endif
@@ -144,7 +148,7 @@ endif
CFLAGS-$(CONFIG_PPC64) := $(call cc-option,-mtraceback=no)
ifndef CONFIG_CC_IS_CLANG
-ifdef CONFIG_CPU_LITTLE_ENDIAN
+ifdef CONFIG_BUILD_ELF_V2
CFLAGS-$(CONFIG_PPC64) += $(call cc-option,-mabi=elfv2,$(call
cc-option,-mcall-aixdesc))
AFLAGS-$(CONFIG_PPC64) += $(call cc-option,-mabi=elfv2)
else
@@ -153,6 +157,7 @@ CFLAGS-$(CONFIG_PPC64) += $(call
cc-option,-mcall-aixdesc)
AFLAGS-$(CONFIG_PPC64) += $(call cc-option,-mabi=elfv1)
endif
endif
+
CFLAGS-$(CONFIG_PPC64) += $(call cc-option,-mcmodel=medium,$(call
cc-option,-mminimal-toc))
CFLAGS-$(CONFIG_PPC64) += $(call cc-option,-mno-pointers-to-nested-functions)
diff --git a/arch/powerpc/boot/Makefile b/arch/powerpc/boot/Makefile
index 0556bf4fc9e9..137ff20b13f8 100644
--- a/arch/powerpc/boot/Makefile
+++ b/arch/powerpc/boot/Makefile
@@ -41,6 +41,10 @@ endif
BOOTCFLAGS += -isystem $(shell $(BOOTCC) -print-file-name=include)
+ifdef CONFIG_BUILD_ELF_V2
+BOOTCFLAGS += $(call cc-option,-mabi=elfv2)
+endif
+
ifdef CONFIG_CPU_BIG_ENDIAN
BOOTCFLAGS += -mbig-endian
else
diff --git a/drivers/crypto/vmx/Makefile b/drivers/crypto/vmx/Makefile
index 709670d2b553..8d79514eb474 100644
--- a/drivers/crypto/vmx/Makefile
+++ b/drivers/crypto/vmx/Makefile
@@ -5,8 +5,12 @@ vmx-crypto-objs := vmx.o aesp8-ppc.o ghashp8-ppc.o aes.o
aes_cbc.o aes_ctr.o aes
ifeq ($(CONFIG_CPU_LITTLE_ENDIAN),y)
override flavour := linux-ppc64le
else
+ifdef CONFIG_BUILD_ELF_V2
+override flavour := linux-ppc64v2
+else
override flavour := linux-ppc64
endif
+endif
quiet_cmd_perl = PERL $@
cmd_perl = $(PERL) $(<) $(flavour) > $(@)
diff --git a/drivers/crypto/vmx/aesp8-ppc.pl b/drivers/crypto/vmx/aesp8-ppc.pl
index db874367b602..6733a68f12ed 100644
--- a/drivers/crypto/vmx/aesp8-ppc.pl
+++ b/drivers/crypto/vmx/aesp8-ppc.pl
@@ -100,7 +100,7 @@ if ($flavour =~ /64/) {
$SHL="slwi";
}
