toolchain: add gcc 10 entry
PATCH 1/5 toolchain: add gcc 10 entry
In order to add gcc 10 support for internal and external toolchain in follow-up commits, introduce BR2_TOOLCHAIN_GCC_AT_LEAST_10 symbol.
--- toolchain/Config.in | 5 +++++ 1 file changed, 5 insertions(+)
toolchain/Config.in +5 -0
@@ -591,10 +591,15 @@ config BR2_TOOLCHAIN_GCC_AT_LEAST_9 bool select BR2_TOOLCHAIN_GCC_AT_LEAST_8 +config BR2_TOOLCHAIN_GCC_AT_LEAST_10 + bool + select BR2_TOOLCHAIN_GCC_AT_LEAST_9 + # This order guarantees that the highest version is set, as kconfig # stops affecting a value on the first matching default. config BR2_TOOLCHAIN_GCC_AT_LEAST string + default "10" if BR2_TOOLCHAIN_GCC_AT_LEAST_10 default "9" if BR2_TOOLCHAIN_GCC_AT_LEAST_9 default "8" if BR2_TOOLCHAIN_GCC_AT_LEAST_8 default "7" if BR2_TOOLCHAIN_GCC_AT_LEAST_7
Signature · 1 line
2.25.4
PATCH 2/5 package/gcc: add support for gcc 10
Keep the same patch as for gcc 9.x fixing a parallel build issue:58ecbbc3ef18c43ae1c02a5c4bf30aa7ef2d8092.
Keep libzstd support disabled for now, it could be used to LTO bytecode.
https://gcc.gnu.org/gcc-10/changes.html
https://gcc.gnu.org/gcc-10/porting_to.html
--- ...move-SELFTEST_DEPS-before-including-.patch | 79 +++++++++++++++++++ package/gcc/Config.in.host | 14 ++++ package/gcc/gcc.hash | 2 + package/gcc/gcc.mk | 3 +- 4 files changed, 97 insertions(+), 1 deletion(-) create mode 100644 package/gcc/10.1.0/0001-gcc-Makefile.in-move-SELFTEST_DEPS-before-including-.patch
package/gcc/10.1.0/0001-gcc-Makefile.in-move-SELFTEST_DEPS-before-including-.patch new file +79 -0
@@ -0,0 +1,79 @@ +From f05b73275515c4a1a70e28e06229bf682ac74385 Mon Sep 17 00:00:00 2001 +From: Romain Naour <[email protected]> +Date: Thu, 21 May 2020 15:58:02 +0200 +Subject: [PATCH] gcc/Makefile.in: move SELFTEST_DEPS before including language + makefile fragments + +As reported by several Buildroot users [1][2][3], the gcc build +may fail while running selftests makefile target. + +The problem only occurs when ccache is used with gcc 9 and 10, +probably due to a race condition. + +While debuging with "make -p" we can notice that s-selftest-c target +contain only "cc1" as dependency instead of cc1 and SELFTEST_DEPS [4]. + + s-selftest-c: cc1 + +While the build is failing, the s-selftest-c dependencies recipe is +still running and reported as a bug by make. + + "Dependencies recipe running (THIS IS A BUG)." + +A change [5] in gcc 9 seems to introduce the problem since we can't +reproduce this problem with gcc 8. + +As suggested by Yann E. MORIN [6], move SELFTEST_DEPS before +including language makefile fragments. + +With the fix applied, the s-seltest-c dependency contains +SELFTEST_DEPS value. + + s-selftest-c: cc1 xgcc specs stmp-int-hdrs ../../gcc/testsuite/selftests + +[1] http://lists.busybox.net/pipermail/buildroot/2020-May/282171.html +[2] http://lists.busybox.net/pipermail/buildroot/2020-May/282766.html +[3] https://github.com/cirosantilli/linux-kernel-module-cheat/issues/108 +[4] https://gcc.gnu.org/git/?p=gcc.git;a=blob;f=gcc/c/Make-lang.in;h=bfae6fd2549c4f728816cd355fa9739dcc08fcde;hb=033eb5671769a4c681a44aad08a454e667e08502#l120 +[5] https://gcc.gnu.org/git/?p=gcc.git;a=commitdiff;h=033eb5671769a4c681a44aad08a454e667e08502 +[6] http://lists.busybox.net/pipermail/buildroot/2020-May/283213.html + +Signed-off-by: Romain Naour <[email protected]> +Cc: Ben Dakin-Norris <[email protected]> +Cc: Maxim Kochetkov <[email protected]> +Cc: Thomas Petazzoni <[email protected]> +Cc: Yann E. MORIN <[email protected]> +Cc: David Malcolm <[email protected]> +--- +This patch should be backported to gcc 10 and gcc 9. +--- + gcc/Makefile.in | 6 ++++-- + 1 file changed, 4 insertions(+), 2 deletions(-) + +diff --git a/gcc/Makefile.in b/gcc/Makefile.in +index 543b477ff18..c8611e35c70 100644 +--- a/gcc/Makefile.in ++++ b/gcc/Makefile.in +@@ -1725,6 +1725,10 @@ $(FULL_DRIVER_NAME): ./xgcc + rm -f $@ + $(LN_S) $< $@ + ++# SELFTEST_DEPS need to be set before including language makefile fragments. ++# Otherwise $(SELFTEST_DEPS) is empty when used from various <LANG>/Make-lang.in. ++SELFTEST_DEPS = $(GCC_PASSES) stmp-int-hdrs $(srcdir)/testsuite/selftests ++ + # + # Language makefile fragments. + +@@ -2001,8 +2005,6 @@ DEVNULL=$(if $(findstring mingw,$(build)),nul,/dev/null) + SELFTEST_FLAGS = -nostdinc $(DEVNULL) -S -o $(DEVNULL) \ + -fself-test=$(srcdir)/testsuite/selftests + +-SELFTEST_DEPS = $(GCC_PASSES) stmp-int-hdrs $(srcdir)/testsuite/selftests +- + # Run the selftests during the build once we have a driver and the frontend, + # so that self-test failures are caught as early as possible. + # Use "s-selftest-FE" to ensure that we only run the selftests if the +-- +2.25.4 +
package/gcc/Config.in.host +14 -0
@@ -50,6 +50,19 @@ config BR2_GCC_VERSION_9_X depends on !BR2_csky select BR2_TOOLCHAIN_GCC_AT_LEAST_9 +config BR2_GCC_VERSION_10_X + bool "gcc 10.x" + # or1k needs binutils >= 2.32 + depends on !(BR2_or1k && BR2_BINUTILS_VERSION_2_31_X) + # powerpc spe support has been deprecated since gcc 8.x. + # https://gcc.gnu.org/ml/gcc/2018-04/msg00102.html + depends on !BR2_powerpc_SPE + # C-SKY sk610 needs abiv1, which is not supported in + # upstream gcc. C-SKY gcc upstream support not tested + # with upstream binutils and glibc. + depends on !BR2_csky + select BR2_TOOLCHAIN_GCC_AT_LEAST_10 + endchoice # libcilkrts was introduced in gcc 4.9 and removed in gcc 8.x
@@ -75,6 +88,7 @@ config BR2_GCC_VERSION default "7.5.0" if BR2_GCC_VERSION_7_X default "8.4.0" if BR2_GCC_VERSION_8_X default "9.3.0" if BR2_GCC_VERSION_9_X + default "10.1.0" if BR2_GCC_VERSION_10_X default "arc-2020.03-release" if BR2_GCC_VERSION_ARC default "48152afb96c59733d5bc79e3399bb7b3d4b44266" if BR2_GCC_VERSION_CSKY
package/gcc/gcc.hash +2 -0
@@ -4,6 +4,8 @@ sha512 fe716cc19f2e3255d3a8b1b8290777bf769c6d98e6e0b07b81a3d6ad43f8af74cb170dfa sha512 6de904f552a02de33b11ef52312bb664396efd7e1ce3bbe37bfad5ef617f133095b3767b4804bc7fe78df335cb53bc83f1ac055baed40979ce4c2c3e46b70280 gcc-8.4.0.tar.xz # From ftp://gcc.gnu.org/pub/gcc/releases/gcc-9.3.0/sha512.sum sha512 4b9e3639eef6e623747a22c37a904b4750c93b6da77cf3958d5047e9b5ebddb7eebe091cc16ca0a227c0ecbd2bf3b984b221130f269a97ee4cc18f9cf6c444de gcc-9.3.0.tar.xz +# From ftp://gcc.gnu.org/pub/gcc/releases/gcc-10.1.0/sha512.sum +sha512 0cb2a74c793face751f42bc580960b00e2bfea785872a0a2155f1f1dbfaa248f9591b67f4322db0f096f8844aca9243bc02732bda106c3b6e43b02bb67eb3096 gcc-10.1.0.tar.xz # Locally calculated (fetched from Github) sha512 09ad77fce757d77f2db49cd049b78861abfa5c1c6c3be76228815ec2b15810c1985525c48b0300e83e88f3fa33dee0062f34790cc8b6bc2fa6b0301595acf42b gcc-arc-2020.03-release.tar.gz
package/gcc/gcc.mk +2 -1
@@ -82,7 +82,8 @@ HOST_GCC_COMMON_CONF_OPTS = \ --with-mpc=$(HOST_DIR) \ --with-mpfr=$(HOST_DIR) \ --with-pkgversion="Buildroot $(BR2_VERSION_FULL)" \ - --with-bugurl="http://bugs.buildroot.net/" + --with-bugurl="http://bugs.buildroot.net/" \ + --without-zstd # Don't build documentation. It takes up extra space / build time, # and sometimes needs specific makeinfo versions to work
Signature · 1 line
2.25.4
PATCH 2/5 package/gcc: add support for gcc 10
Quoted 4 lines · On Tue, 9 Jun 2020 00:13:44 +0200 Romain Naour <[email protected]> wrote:
Keep the same patch as for gcc 9.x fixing a parallel build issue:
58ecbbc3ef18c43ae1c02a5c4bf30aa7ef2d8092.Keep libzstd support disabled for now, it could be used to LTO bytecode.
"... it could be used for LTO bytecode compression", otherwise it doesn't make a lot of sense.
I've applied, but I have some comments/questions below.
package/gcc/Config.in.host config BR2_GCC_VERSION_9_X · line 56
Quoted 4 lines
package/gcc/Config.in.host +4
-0+config BR2_GCC_VERSION_10_X + bool "gcc 10.x" + # or1k needs binutils >= 2.32 + depends on !(BR2_or1k && BR2_BINUTILS_VERSION_2_31_X)
Why are we doing this here, instead of in the binutils package ? This looks strange.
package/gcc/Config.in.host config BR2_GCC_VERSION_9_X · line 59
Quoted 3 lines
package/gcc/Config.in.host +3
-0+ # powerpc spe support has been deprecated since gcc 8.x. + # https://gcc.gnu.org/ml/gcc/2018-04/msg00102.html + depends on !BR2_powerpc_SPE
Should we now drop support for PowerPC SPE ?
package/gcc/Config.in.host config BR2_GCC_VERSION_9_X · line 63
Quoted 4 lines
package/gcc/Config.in.host +4
-0+ # C-SKY sk610 needs abiv1, which is not supported in + # upstream gcc. C-SKY gcc upstream support not tested + # with upstream binutils and glibc. + depends on !BR2_csky
We haven't had any news/contribution for C-SKY people for quite a while. I know in recent SoC they have used RISC-V cores, but they were supposed to continue making C-SKY based SoCs. I am fearing that this looks like yet another abandoned CPU architecture :-/
Best regards,
Thomas
Signature · 3 lines
Thomas Petazzoni, CTO, Bootlin Embedded Linux and Kernel engineering https://bootlin.com
PATCH 2/5 package/gcc: add support for gcc 10
Hi Thomas,
Le 24/06/2020 ? 21:55, Thomas Petazzoni a ?crit?:
Quoted 13 lines
Quoted 4 lines · On Tue, 9 Jun 2020 00:13:44 +0200 Romain Naour <[email protected]> wrote:
Keep the same patch as for gcc 9.x fixing a parallel build issue:
58ecbbc3ef18c43ae1c02a5c4bf30aa7ef2d8092.Keep libzstd support disabled for now, it could be used to LTO bytecode.
"... it could be used for LTO bytecode compression", otherwise it doesn't make a lot of sense.
I've applied, but I have some comments/questions below.
Quoted 7 lines
package/gcc/Config.in.host config BR2_GCC_VERSION_9_X · line 56
Quoted 4 lines
package/gcc/Config.in.host +4
-0+config BR2_GCC_VERSION_10_X + bool "gcc 10.x" + # or1k needs binutils >= 2.32 + depends on !(BR2_or1k && BR2_BINUTILS_VERSION_2_31_X)Why are we doing this here, instead of in the binutils package ? This looks strange.
This dependency was added by [1] when we had a gcc 5 fork for or1k.
[1]
https://git.buildroot.net/buildroot/commit/?id=e0ba09768eb27872e8685754952601bdac9ac75b
It here because it's gcc that require a recent enough binutils version. Maybe we should consider adding BR2_TOOLCHAIN_BINUTILS_AT_LEAST_xxxx. I think we already have a problem with new cpu variant that require the latest binutils version.
Quoted 6 lines
package/gcc/Config.in.host config BR2_GCC_VERSION_9_X · line 59
Quoted 4 lines
package/gcc/Config.in.host +3
-0+ # powerpc spe support has been deprecated since gcc 8.x. + # https://gcc.gnu.org/ml/gcc/2018-04/msg00102.html + depends on !BR2_powerpc_SPEShould we now drop support for PowerPC SPE ?
If every body is ok, then we can drop it.
Quoted 10 lines
package/gcc/Config.in.host config BR2_GCC_VERSION_9_X · line 63
Quoted 5 lines
package/gcc/Config.in.host +4
-0+ # C-SKY sk610 needs abiv1, which is not supported in + # upstream gcc. C-SKY gcc upstream support not tested + # with upstream binutils and glibc. + depends on !BR2_cskyWe haven't had any news/contribution for C-SKY people for quite a while. I know in recent SoC they have used RISC-V cores, but they were supposed to continue making C-SKY based SoCs. I am fearing that this looks like yet another abandoned CPU architecture :-/
I'm not sure about this, their gitlab is still active:
https://gitlab.com/c-sky/buildroot/-/pipelines
IIRC we had some news from Guo Ren about the upstream qemu support for csky cpus
http://lists.busybox.net/pipermail/buildroot/2020-May/283411.html
Best regards,
Romain
Quoted 5 lines
Best regards,
Thomas
PATCH 3/5 arch: add BR2_ARCH_NEEDS_GCC_AT_LEAST_10
In reply to Romain NaourThis new symbol will be used by architectures introduced with gcc 10 and by external toolchains based on gcc 10.
[1] https://gcc.gnu.org/gcc-10/changes.html
--- arch/Config.in | 4 ++++ package/gcc/Config.in.host | 1 + 2 files changed, 5 insertions(+)
arch/Config.in +4 -0
@@ -308,6 +308,10 @@ config BR2_ARCH_NEEDS_GCC_AT_LEAST_9 bool select BR2_ARCH_NEEDS_GCC_AT_LEAST_8 +config BR2_ARCH_NEEDS_GCC_AT_LEAST_10 + bool + select BR2_ARCH_NEEDS_GCC_AT_LEAST_9 + # The following string values are defined by the individual # Config.in.$ARCH files config BR2_ARCH
package/gcc/Config.in.host +1 -0
@@ -39,6 +39,7 @@ config BR2_GCC_VERSION_8_X config BR2_GCC_VERSION_9_X bool "gcc 9.x" + depends on !BR2_ARCH_NEEDS_GCC_AT_LEAST_10 # or1k needs binutils >= 2.32 depends on !(BR2_or1k && BR2_BINUTILS_VERSION_2_31_X) # powerpc spe support has been deprecated since gcc 8.x.
Signature · 1 line
2.25.4
Thomas Petazzoni
AppliedPATCH 3/5 arch: add BR2_ARCH_NEEDS_GCC_AT_LEAST_10
Quoted 10 lines · On Tue, 9 Jun 2020 00:13:45 +0200 Romain Naour <[email protected]> wrote:
This new symbol will be used by architectures introduced with gcc 10 and by external toolchains based on gcc 10.
[1] https://gcc.gnu.org/gcc-10/changes.html
Signed-off-by Romain Naour <[email protected]>--- arch/Config.in | 4 ++++ package/gcc/Config.in.host | 1 + 2 files changed, 5 insertions(+)
Applied to master, thanks.
Thomas
Signature · 3 lines
Thomas Petazzoni, CTO, Bootlin Embedded Linux and Kernel engineering https://bootlin.com
PATCH 4/5 toolchain-external: add gcc 10 entry
In reply to Romain NaourThis patch allows to use an external toolchain based on gcc 10.
--- .../toolchain-external-custom/Config.in.options | 4 ++++ 1 file changed, 4 insertions(+)
toolchain/toolchain-external/toolchain-external-custom/Config.in.options +4 -0
@@ -31,6 +31,10 @@ choice Set to the gcc version that is used by your external toolchain. +config BR2_TOOLCHAIN_EXTERNAL_GCC_10 + bool "10.x" + select BR2_TOOLCHAIN_GCC_AT_LEAST_10 + config BR2_TOOLCHAIN_EXTERNAL_GCC_9 bool "9.x" select BR2_TOOLCHAIN_GCC_AT_LEAST_9
Signature · 1 line
2.25.4
Thomas Petazzoni
AppliedPATCH 4/5 toolchain-external: add gcc 10 entry
Quoted 6 lines · On Tue, 9 Jun 2020 00:13:46 +0200 Romain Naour <[email protected]> wrote:
This patch allows to use an external toolchain based on gcc 10.
Signed-off-by Romain Naour <[email protected]>--- .../toolchain-external-custom/Config.in.options | 4 ++++ 1 file changed, 4 insertions(+)
Applied to master, thanks.
Thomas
Signature · 3 lines
Thomas Petazzoni, CTO, Bootlin Embedded Linux and Kernel engineering https://bootlin.com
PATCH 5/5 package/gcc: switch to gcc 9.x as the default
In reply to Romain NaourEven if gcc 8 is still maintained for some time (gcc 8.5 is pending), switch to gcc 9.x since it has been released since 2019-05-03 and gcc 10.x is available since 2020-05-07.
We have been having toolchains in the autobuilders with gcc 9.x for a while, so the vast majority of the problems should have already been solved.
--- package/gcc/Config.in.host | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
package/gcc/Config.in.host +1 -1
@@ -4,7 +4,7 @@ choice prompt "GCC compiler Version" default BR2_GCC_VERSION_ARC if BR2_arc default BR2_GCC_VERSION_CSKY if BR2_csky - default BR2_GCC_VERSION_8_X + default BR2_GCC_VERSION_9_X help Select the version of gcc you wish to use.
Signature · 1 line
2.25.4
Thomas Petazzoni
AppliedPATCH 5/5 package/gcc: switch to gcc 9.x as the default
Quoted 4 lines · On Tue, 9 Jun 2020 00:13:47 +0200 Romain Naour <[email protected]> wrote:
Even if gcc 8 is still maintained for some time (gcc 8.5 is pending), switch to gcc 9.x since it has been released since 2019-05-03 and gcc 10.x is available since 2020-05-07.
Quoted 8 lines
We have been having toolchains in the autobuilders with gcc 9.x for a while, so the vast majority of the problems should have already been solved.
Signed-off-by Romain Naour <[email protected]>--- package/gcc/Config.in.host | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
Applied to master, thanks.
Following this, we should drop gcc 7.x support. However, one defconfig (roseapplepi) still uses gcc 7.x. I've notified Peter Korsgaard on IRC, and I am doing it here as well: if this defconfig is not updated to a newer kernel that builds with gcc >= 8.x, we will have to drop support for this defconfig.
Thomas
Signature · 3 lines
Thomas Petazzoni, CTO, Bootlin Embedded Linux and Kernel engineering https://bootlin.com
PATCH 5/5 package/gcc: switch to gcc 9.x as the default
Hi Thomas,
Le 24/06/2020 ? 21:56, Thomas Petazzoni a ?crit?:
Quoted 16 lines
Quoted 4 lines · On Tue, 9 Jun 2020 00:13:47 +0200 Romain Naour <[email protected]> wrote:
Even if gcc 8 is still maintained for some time (gcc 8.5 is pending), switch to gcc 9.x since it has been released since 2019-05-03 and gcc 10.x is available since 2020-05-07.
Quoted 8 lines
We have been having toolchains in the autobuilders with gcc 9.x for a while, so the vast majority of the problems should have already been solved.
Signed-off-by Romain Naour <[email protected]>--- package/gcc/Config.in.host | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
Quoted 7 lines
Applied to master, thanks.
Following this, we should drop gcc 7.x support. However, one defconfig (roseapplepi) still uses gcc 7.x. I've notified Peter Korsgaard on IRC, and I am doing it here as well: if this defconfig is not updated to a newer kernel that builds with gcc >= 8.x, we will have to drop support for this defconfig.
Yes, I would like to avoid the mess like for beagleboneblack defconfig for qt5 when it was using gcc 5 only.
Best regards,
Romain
Quoted 3 lines
Thomas
PATCH 5/5 package/gcc: switch to gcc 9.x as the default
In reply to Thomas PetazzoniQuoted 18 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
"Thomas" == Thomas Petazzoni <[email protected]> writes:
Quoted 4 lines · On Tue, 9 Jun 2020 00:13:47 +0200 Romain Naour <[email protected]> wrote:
Even if gcc 8 is still maintained for some time (gcc 8.5 is pending), switch to gcc 9.x since it has been released since 2019-05-03 and gcc 10.x is available since 2020-05-07.
Quoted 8 lines
We have been having toolchains in the autobuilders with gcc 9.x for a while, so the vast majority of the problems should have already been solved.
Signed-off-by Romain Naour <[email protected]>---
package/gcc/Config.in.host | 2 +-1 file changed, 1 insertion(+), 1 deletion(-)
Quoted 7 lines
Applied to master, thanks.
Following this, we should drop gcc 7.x support. However, one defconfig (roseapplepi) still uses gcc 7.x. I've notified Peter Korsgaard on IRC, and I am doing it here as well: if this defconfig is not updated to a newer kernel that builds with gcc >= 8.x, we will have to drop support for this defconfig.
I'll take a look at updating it next weekend, otherwise we can just drop it.
Signature · 1 line
Bye, Peter Korsgaard
PATCH 5/5 package/gcc: switch to gcc 9.x as the default
In reply to Thomas PetazzoniQuoted 18 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
"Thomas" == Thomas Petazzoni <[email protected]> writes:
Quoted 4 lines · On Tue, 9 Jun 2020 00:13:47 +0200 Romain Naour <[email protected]> wrote:
Even if gcc 8 is still maintained for some time (gcc 8.5 is pending), switch to gcc 9.x since it has been released since 2019-05-03 and gcc 10.x is available since 2020-05-07.
Quoted 8 lines
We have been having toolchains in the autobuilders with gcc 9.x for a while, so the vast majority of the problems should have already been solved.
Signed-off-by Romain Naour <[email protected]>---
package/gcc/Config.in.host | 2 +-1 file changed, 1 insertion(+), 1 deletion(-)
Quoted 7 lines
Applied to master, thanks.
Following this, we should drop gcc 7.x support. However, one defconfig (roseapplepi) still uses gcc 7.x. I've notified Peter Korsgaard on IRC, and I am doing it here as well: if this defconfig is not updated to a newer kernel that builds with gcc >= 8.x, we will have to drop support for this defconfig.
This has now been done (in next), so we can drop gcc 7.
Signature · 1 line
Bye, Peter Korsgaard
Thomas Petazzoni
AppliedPATCH 1/5 toolchain: add gcc 10 entry
In reply to Romain NaourQuoted 7 lines · On Tue, 9 Jun 2020 00:13:43 +0200 Romain Naour <[email protected]> wrote:
In order to add gcc 10 support for internal and external toolchain in follow-up commits, introduce
BR2_TOOLCHAIN_GCC_AT_LEAST_10symbol.Signed-off-by Romain Naour <[email protected]>--- toolchain/Config.in | 5 +++++ 1 file changed, 5 insertions(+)
Applied to master, thanks.
Thomas
Signature · 3 lines
Thomas Petazzoni, CTO, Bootlin Embedded Linux and Kernel engineering https://bootlin.com