Skip to contentSkip to conversation

package/glibc: move version string to Kconfig

PATCH New Romain Naour 2 participants 7 messages Jul 20, 2020 – Jul 23, 2020 RISC-V

lore
Share & tools
lore
#0 lore raw

PATCH 1/4 package/glibc: move version string to Kconfig

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

Currently, the libc version is shared by glibc and localedef. Syncing is usually not strictly mandatory, but we had breakage in the past, when glibc 2.27 changed locale handling, so that a localedef from 2.26 or earlier would generate locale definitions unsuitable for glibc 2.27. So, the safe bet is to keep both in sync.

Keeping them in sync is currently done manually, with hints as comments in glibc.mk and localedef.mk, and this is error-prone.

By moving the version string to Kconfig, we can ensure the version is always defined before we scan the packages, and thus it will be easier to share between glibc and localedef, at which point we'll no longer need the manual sync.

However, the matter is made a little bit more complex by the fact that some architectures require a forked version of glibc. For those, in theory, we should also be using a localedef that is in sync with them. However, that would require that we also carry a patch so we can build only localedef for those versions as well.

In practice, those versions are close-enough to upstream that the localedef from upstream is compatible with those versions. Besides, the current behaviour is that we are using upstream localedef for those versions already, so we keep doing that.

So we do not move the site to Kconfig.

So, we will always need to know the upstream version string, even when usign one of the forked trees. We use a Kconfig trick: a default for a symbol can be another symbol. We define the upstream version in its own symbol, which we use as fallback default for the generic version symbol, when not specific fork is needed.

---
 package/glibc/Config.in | 17 +++++++++++++++++
 package/glibc/glibc.mk  |  8 +-------
 2 files changed, 18 insertions(+), 7 deletions(-)
package/glibc/Config.in +17 -0
index 5ecd058145..1c5b33a0dd 100644
@@ -17,3 +17,20 @@ config BR2_PACKAGE_GLIBC_UTILS
 	  ldconfig, ldd and locale glibc utilities for the target.
 
 endif # BR2_TOOLCHAIN_BUILDROOT_GLIBC
+
+# Keep version string outside of the conditional and keep the upstream
+# version last and unconditional.
+
+# Generate upstream version string using:
+#   git describe --match 'glibc-*' --abbrev=40 origin/release/MAJOR.MINOR/master | cut -d '-' -f 2-
+# When updating the version, please also update localedef
+config BR2_GLIBC_VERSION_UPSTREAM
+	string
+	default "2.30-67-g4748829f86a458b76642f3e98b1d80f7b868e427"
+
+config BR2_GLIBC_VERSION
+	string
+	default "arc-2020.03-release"                               if BR2_arc
+	default "7630ed2fa60caea98f500e4a7a51b88f9bf1e176"          if BR2_csky
+	default "06983fe52cfe8e4779035c27e8cc5d2caab31531"          if BR2_RISCV_32
+	default BR2_GLIBC_VERSION_UPSTREAM
package/glibc/glibc.mk +1 -7
index 1881cf0cdc..1801ac7744 100644
@@ -4,20 +4,14 @@
 #
 ################################################################################
 
+GLIBC_VERSION = $(call qstrip,$(BR2_GLIBC_VERSION))
 ifeq ($(BR2_arc),y)
-GLIBC_VERSION =  arc-2020.03-release
 GLIBC_SITE = $(call github,foss-for-synopsys-dwc-arc-processors,glibc,$(GLIBC_VERSION))
 else ifeq ($(BR2_RISCV_32),y)
-GLIBC_VERSION = 06983fe52cfe8e4779035c27e8cc5d2caab31531
 GLIBC_SITE = $(call github,riscv,riscv-glibc,$(GLIBC_VERSION))
 else ifeq ($(BR2_csky),y)
-GLIBC_VERSION = 7630ed2fa60caea98f500e4a7a51b88f9bf1e176
 GLIBC_SITE = $(call github,c-sky,glibc,$(GLIBC_VERSION))
 else
-# Generate version string using:
-#   git describe --match 'glibc-*' --abbrev=40 origin/release/MAJOR.MINOR/master | cut -d '-' -f 2-
-# When updating the version, please also update localedef
-GLIBC_VERSION = 2.30-67-g4748829f86a458b76642f3e98b1d80f7b868e427
 # Upstream doesn't officially provide an https download link.
 # There is one (https://sourceware.org/git/glibc.git) but it's not reliable,
 # sometimes the connection times out. So use an unofficial github mirror.
Signature · 1 line
2.25.4
#1 lore raw

PATCH 2/4 package/localedef: use upstream glibc version

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

Since the upstream version and site are always defined in Kconfig now, they are always available, so we can use them to define the localedef version and site. We also change the localedef hash file to be a symlink to the glibc one.

Now, there is no manual action to do to keep localedef and glibc in sync for the internal toolchain.

External toolchains will use the upstream glibc version (as is currently the case already); with some luck, it is good enough if the external toolchain is not too old (2.26 or older may have issues, but there's nothing we can do as we can't know what version is used).

Additionally, for some architectures, the internall toolchain uses a forked glibc tree. In theory, we would also want to use the localedef from those trees, but that would mean we would have to carry a patch for each, to be able to build only localedef. This is highly inpractical.

However, those architectures are close enought to upstream that upstream localedef is compatible. Besides, the current behaviour is already to use an upstream localedef for those architectures anyway, so we keep doing that.

---
 package/glibc/Config.in                                   | 3 +--
 .../localedef.hash                                        | 8 +-------
 package/localedef/localedef.mk                            | 6 ++----
 3 files changed, 4 insertions(+), 13 deletions(-)
 mode change 100644 => 120000 package/localedef/2.30-67-g4748829f86a458b76642f3e98b1d80f7b868e427/localedef.hash
package/glibc/Config.in +1 -2
index 1c5b33a0dd..9901ad8537 100644
@@ -19,11 +19,10 @@ config BR2_PACKAGE_GLIBC_UTILS
 endif # BR2_TOOLCHAIN_BUILDROOT_GLIBC
 
 # Keep version string outside of the conditional and keep the upstream
-# version last and unconditional.
+# version last and unconditional, because it is shared with localedef.
 
 # Generate upstream version string using:
 #   git describe --match 'glibc-*' --abbrev=40 origin/release/MAJOR.MINOR/master | cut -d '-' -f 2-
-# When updating the version, please also update localedef
 config BR2_GLIBC_VERSION_UPSTREAM
 	string
 	default "2.30-67-g4748829f86a458b76642f3e98b1d80f7b868e427"
package/localedef/2.30-67-g4748829f86a458b76642f3e98b1d80f7b868e427/localedef.hash deleted +0 -7
deleted file mode 100644
index 6677d32db9..0000000000
@@ -1,7 +0,0 @@
-# Locally calculated (fetched from Github)
-sha256  4462f56696332efbc5b0c2f86d7aa75a2a02c3d44bc4345fa42b5bab1225de5c  glibc-2.30-67-g4748829f86a458b76642f3e98b1d80f7b868e427.tar.gz
-
-# Hashes for license files
-sha256  8177f97513213526df2cf6184d8ff986c675afb514d4e68a404010521b880643  COPYING
-sha256  dc626520dcd53a22f727af3ee42c770e56c97a64fe3adb063799d8ab032fe551  COPYING.LIB
-sha256  35bdb41dc0bcb10702ddacbd51ec4c0fe6fb3129f734e8c85fc02e4d3eb0ce3f  LICENSES
package/localedef/2.30-67-g4748829f86a458b76642f3e98b1d80f7b868e427/localedef.hash new file +1 -0
new file mode 120000
index 0000000000..7e125009ee
@@ -0,0 +1 @@
+../../glibc/2.30-67-g4748829f86a458b76642f3e98b1d80f7b868e427/glibc.hash
\ No newline at end of file
package/localedef/localedef.mk +2 -4
index f6e26b5666..4c7b442d72 100644
@@ -4,10 +4,8 @@
 #
 ################################################################################
 
-# Use the same VERSION and SITE as target glibc
-# As in glibc.mk, generate version string using:
-#   git describe --match 'glibc-*' --abbrev=40 origin/release/MAJOR.MINOR/master | cut -d '-' -f 2-
-LOCALEDEF_VERSION = 2.30-67-g4748829f86a458b76642f3e98b1d80f7b868e427
+# Use the same VERSION and SITE as upstream glibc
+LOCALEDEF_VERSION = $(call qstrip,$(BR2_GLIBC_VERSION_UPSTREAM))
 LOCALEDEF_SOURCE = glibc-$(LOCALEDEF_VERSION).tar.gz
 LOCALEDEF_SITE = $(call github,bminor,glibc,$(LOCALEDEF_VERSION))
 HOST_LOCALEDEF_DL_SUBDIR = glibc
Signature · 1 line
2.25.4
#2 lore raw

PATCH 2/4 package/localedef: use upstream glibc version

Hello,

Quoted 20 lines · On Mon, 20 Jul 2020 18:47:21 +0200 Romain Naour <[email protected]> wrote:

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

Since the upstream version and site are always defined in Kconfig now, they are always available, so we can use them to define the localedef version and site. We also change the localedef hash file to be a symlink to the glibc one.

Now, there is no manual action to do to keep localedef and glibc in sync for the internal toolchain.

External toolchains will use the upstream glibc version (as is currently the case already); with some luck, it is good enough if the external toolchain is not too old (2.26 or older may have issues, but there's nothing we can do as we can't know what version is used).

Additionally, for some architectures, the internall toolchain uses a forked glibc tree. In theory, we would also want to use the localedef from those trees, but that would mean we would have to carry a patch for each, to be able to build only localedef. This is highly inpractical.

Quoted 7 lines

However, those architectures are close enought to upstream that upstream localedef is compatible. Besides, the current behaviour is already to use an upstream localedef for those architectures anyway, so we keep doing that.

Unfortunately, this breaks building host-localedef for external toolchains. Indeed, package/glibc/Config.in is only included in the Config.in machinery when the internal toolchain backend is used. I'm not sure what is the best way to solve this.

Another concern is that this change doesn't make it as "transparent" as it seems to update the glibc version: the idea was that changing BR2_GLIBC_VERSION_UPSTREAM would be sufficient, and localedef would automatically use that new version. Unfortunately, this doesn't work due to the hash file of localedef:

package/localedef/2.30-67-g4748829f86a458b76642f3e98b1d80f7b868e427/localedef.hash -> ../../glibc/2.30-67-g4748829f86a458b76642f3e98b1d80f7b868e427/glibc.hash

This symlink anyway needs to be updated in localedef whenever BR2_GLIBC_VERSION_UPSTREAM is changed.

With PATCH 3/4 in this series, it becomes worse as another symlink is needed for the other glibc version that we would support.

Another concern with localedef, though not related to this series, is that with external toolchains, you don't know which glibc version is used in the toolchain. We may need a post-2.27 localedef, or a pre-2.27 localedef. I'm not sure how to solve that though. Random thought: for the internal toolchains, have a host-glibc package that builds localedef for the host, so that its version is naturally in sync with the target glibc. For external toolchains, have the host-localedef package with a version choice between different upstream glibc versions, and it's up to the user to select the right thing ?

Overall, this is quite complicated, and it's annoying to keep glibc 2.31 unsupported because of this. To me, it feels like glibc 2.31 has been around for quite a while now. Why not just make the jump, keep a single glibc version that we bump to 2.31 ?

Thomas

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

PATCH 2/4 package/localedef: use upstream glibc version

Hello,

Le 21/07/2020 ? 21:27, Thomas Petazzoni a ?crit?:

Quoted 30 lines

Hello,

Quoted 25 lines · On Mon, 20 Jul 2020 18:47:21 +0200 Romain Naour <[email protected]> wrote:

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

Since the upstream version and site are always defined in Kconfig now, they are always available, so we can use them to define the localedef version and site. We also change the localedef hash file to be a symlink to the glibc one.

Now, there is no manual action to do to keep localedef and glibc in sync for the internal toolchain.

External toolchains will use the upstream glibc version (as is currently the case already); with some luck, it is good enough if the external toolchain is not too old (2.26 or older may have issues, but there's nothing we can do as we can't know what version is used).

Additionally, for some architectures, the internall toolchain uses a forked glibc tree. In theory, we would also want to use the localedef from those trees, but that would mean we would have to carry a patch for each, to be able to build only localedef. This is highly inpractical.

However, those architectures are close enought to upstream that upstream localedef is compatible. Besides, the current behaviour is already to use an upstream localedef for those architectures anyway, so we keep doing that.

Quoted 7 lines
Quoted 2 lines

Unfortunately, this breaks building host-localedef for external toolchains. Indeed, package/glibc/Config.in is only included in the Config.in machinery when the internal toolchain backend is used. I'm not sure what is the best way to solve this.

Indeed...

Quoted 14 lines

Another concern is that this change doesn't make it as "transparent" as it seems to update the glibc version: the idea was that changing BR2_GLIBC_VERSION_UPSTREAM would be sufficient, and localedef would automatically use that new version. Unfortunately, this doesn't work due to the hash file of localedef:

package/localedef/2.30-67-g4748829f86a458b76642f3e98b1d80f7b868e427/localedef.hash -> ../../glibc/2.30-67-g4748829f86a458b76642f3e98b1d80f7b868e427/glibc.hash

This symlink anyway needs to be updated in localedef whenever BR2_GLIBC_VERSION_UPSTREAM is changed.

With PATCH 3/4 in this series, it becomes worse as another symlink is needed for the other glibc version that we would support.

Quoted 9 lines

Another concern with localedef, though not related to this series, is that with external toolchains, you don't know which glibc version is used in the toolchain. We may need a post-2.27 localedef, or a pre-2.27 localedef. I'm not sure how to solve that though. Random thought: for the internal toolchains, have a host-glibc package that builds localedef for the host, so that its version is naturally in sync with the target glibc. For external toolchains, have the host-localedef package with a version choice between different upstream glibc versions, and it's up to the user to select the right thing ?

I'm not sure how we can check if the localedef version is compatible with the external toolchain glibc version (add a check in toolchain/helpers.mk?). We only have BR2_TOOLCHAIN_USES_GLIBC to know it's a glibc toolchain.

I'm not sure we want to introduce BR2_TOOLCHAIN_GLIBC_AT_LEAST_x_xx just to build the right localedef version.

Quoted 5 lines

Overall, this is quite complicated, and it's annoying to keep glibc 2.31 unsupported because of this. To me, it feels like glibc 2.31 has been around for quite a while now. Why not just make the jump, keep a single glibc version that we bump to 2.31 ?

This was the initial question:
http://lists.busybox.net/pipermail/buildroot/2020-April/279015.html

Initially I didn't take into acount the host-localedef problem, In previous series version host-localedef was at the same version as the latest upstream glibc.

My concern is about having only one glibc version choice (the latest) to build the internal toolchain. It's pity to block on host-localedef problem not related to the glibc version choice.

Note, glibc 2.32 will be released next week (2020-08-01). Maybe we should merge the glibc version bump without introducing the version choice now and wait for glibc 2.32 version bump series.

Best regards,
Romain

Quoted 3 lines

Thomas

#4 lore raw

PATCH 2/4 package/localedef: use upstream glibc version

Hello Romain,

Quoted 10 lines · On Thu, 23 Jul 2020 00:05:32 +0200 Romain Naour <[email protected]> wrote:
Quoted 9 lines

Another concern with localedef, though not related to this series, is that with external toolchains, you don't know which glibc version is used in the toolchain. We may need a post-2.27 localedef, or a pre-2.27 localedef. I'm not sure how to solve that though. Random thought: for the internal toolchains, have a host-glibc package that builds localedef for the host, so that its version is naturally in sync with the target glibc. For external toolchains, have the host-localedef package with a version choice between different upstream glibc versions, and it's up to the user to select the right thing ?

Quoted 6 lines

I'm not sure how we can check if the localedef version is compatible with the external toolchain glibc version (add a check in toolchain/helpers.mk?). We only have BR2_TOOLCHAIN_USES_GLIBC to know it's a glibc toolchain.

I'm not sure we want to introduce BR2_TOOLCHAIN_GLIBC_AT_LEAST_x_xx just to build the right localedef version.

My idea was to leave that up to the user, i.e when BR2_GENERATE_LOCALE=y && BR2_TOOLCHAIN_USES_GLIBC=y, we ask the user to select which version of localedef should be used. It's not the nicest from a user experience point of view, but it's probably the simplest solution for us.

Quoted 11 lines
Quoted 4 lines

Overall, this is quite complicated, and it's annoying to keep glibc 2.31 unsupported because of this. To me, it feels like glibc 2.31 has been around for quite a while now. Why not just make the jump, keep a single glibc version that we bump to 2.31 ?

This was the initial question:
http://lists.busybox.net/pipermail/buildroot/2020-April/279015.html

Initially I didn't take into acount the host-localedef problem, In previous series version host-localedef was at the same version as the latest upstream glibc.

Quoted 7 lines

My concern is about having only one glibc version choice (the latest) to build the internal toolchain. It's pity to block on host-localedef problem not related to the glibc version choice.

Note, glibc 2.32 will be released next week (2020-08-01). Maybe we should merge the glibc version bump without introducing the version choice now and wait for glibc 2.32 version bump series.

Yes, glibc 2.31 has been around for a while, let's switch to it as our default glibc version, so that in Buildroot 2020.08, we use glibc 2.31. And in parallel, we can continue to work on this glibc/localedef issue.

The localedef version issue exists regardless of whether we support two glibc versions in the internal toolchain backend: right now if someone uses an external toolchain with a pre-2.27 glibc, the host-localedef we provide will be too recent and not compatible...

Best regards,

Thomas

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

PATCH 3/4 package/glibc: introduce version choice, add 2.31.x

In reply to Romain Naour

Glibc version choice was removed by b1ca91c59db6 (glibc: remove version choice) but experience has proven that glibc can take some time after its release to get stable or for softwares requiring changes to get an upstream patch (or new release). This stabilization time can take more than our 3 months development cycle, so let's re-introduce a version choice for glibc package to avoid releasing Buildroot with unfixed packages.

Unlike gcc, binutils and gdb, when a new version is introduced we use it as the default as soon as we are reasonably confident that most build issues have been sorted out. Also we only support two versions, the latest release and the previous release.

Since we follow the upstream stable branch of glibc, we need to clarify that this is not the version from the initial release archive. Make it explicit from the menuconfig by using "2.30-x (stable branch)".

http://lists.busybox.net/pipermail/buildroot/2020-April/279015.html

For glibc 2.31.x:

- update LICENSES file hash due to url change:
"Prefer https to http for gnu.org and fsf.org URLs"

- apply the patch suggested by Florian Weimer fixing a build issue for riscv64 with kernel headers < 5.0:
0001-riscv64-flush-icache-fix-glibc-2.31-toolchain-with-k.patch

For host-localedef 2.31.x:

- remove upstream patch for localedef:
0003-localedef-Use-initializer-for-flexible-array-member-.patch

Note that this version bump required some patches applied on several packages (already applied):

    [Busybox] 13f2d688a24f47446af236829bd6ca194d5aea5b
    [openssh] bad75bca315dbd2c69f8a9cb02fa9f27636e3d48

There are still some known issues, that will be addressed in a followup patch:
[gcc] disable libsanitizer with gcc 7.5

Since the reintroduction of the version choice of glibc, we don't switch to this version by default yet.

See:
https://sourceware.org/legacy-ml/libc-announce/2020/msg00001.html

Tested by toolchain builder:
https://gitlab.com/kubu93/toolchains-builder/pipelines/129551000

[yann.morin.1998 at free.fr:
- meld the choice addtion with the version addition
- hide the prompt for non-upstream forks
- symlink localedef hash file so it's always in sync
- tweak and massage the commit log
]

---
 ...ache-fix-glibc-2.31-toolchain-with-k.patch | 51 +++++++++++++++
 .../glibc.hash                                |  7 ++
 package/glibc/Config.in                       | 18 ++++-
 ...ACK-only-build-and-install-localedef.patch | 65 +++++++++++++++++++
 ...y-on-GCC-to-4.8-and-binutils-to-2.24.patch | 64 ++++++++++++++++++
 .../localedef.hash                            |  1 +
 6 files changed, 205 insertions(+), 1 deletion(-)
 create mode 100644 package/glibc/2.31-44-g3a44844c97a8ca1d0798c45b5252e6a49f9cf8e1/0001-riscv64-flush-icache-fix-glibc-2.31-toolchain-with-k.patch
 create mode 100644 package/glibc/2.31-44-g3a44844c97a8ca1d0798c45b5252e6a49f9cf8e1/glibc.hash
 create mode 100644 package/localedef/2.31-44-g3a44844c97a8ca1d0798c45b5252e6a49f9cf8e1/0001-HACK-only-build-and-install-localedef.patch
 create mode 100644 package/localedef/2.31-44-g3a44844c97a8ca1d0798c45b5252e6a49f9cf8e1/0002-relax-dependency-on-GCC-to-4.8-and-binutils-to-2.24.patch
 create mode 120000 package/localedef/2.31-44-g3a44844c97a8ca1d0798c45b5252e6a49f9cf8e1/localedef.hash
package/glibc/2.31-44-g3a44844c97a8ca1d0798c45b5252e6a49f9cf8e1/0001-riscv64-flush-icache-fix-glibc-2.31-toolchain-with-k.patch new file +51 -0
new file mode 100644
index 0000000000..e33f955f87
@@ -0,0 +1,51 @@
+From 5142becd7662fea48c1313364dadc4c17e0e4fb6 Mon Sep 17 00:00:00 2001
+From: Romain Naour <[email protected]>
+Date: Wed, 25 Mar 2020 11:25:23 +0100
+Subject: [PATCH] riscv64/flush-icache: fix glibc 2.31 toolchain with kernel
+ headers < 5.0
+
+Without this patch, the risc64 port needs@least a kernel headers >= 5.0
+
+Otherwise glibc fail to build with:
+
+../sysdeps/unix/sysv/linux/riscv/rv64/arch-syscall.h:193:33: error: expected
+declaration specifiers or '...' before numeric constant
+  193 | #define __NR_riscv_flush_icache 259
+      |                                 ^~~
+In file included from ../sysdeps/unix/sysv/linux/riscv/flush-icache.c:25:
+/home/naourr/buildroot/test/toolchain/host/riscv64-buildroot-linux-gnu/sysroot/usr/include/asm/syscalls.h:29:36:
+error: unknown type name 'sys_riscv_flush_icache'
+   29 | __SYSCALL(__NR_riscv_flush_icache, sys_riscv_flush_icache)
+      |                                    ^~~~~~~~~~~~~~~~~~~~~~
+
+See: https://gitlab.com/kubu93/toolchains-builder/-/jobs/422726962
+
+The syscalls.h header wasn't originally written as a UAPI header, including
+<sys/syscall.h> should now be sufficient.
+
+https://sourceware.org/ml/libc-alpha/2020-02/msg00018.html
+
+Signed-off-by: Romain Naour <[email protected]>
+---
+ sysdeps/unix/sysv/linux/riscv/flush-icache.c | 5 -----
+ 1 file changed, 5 deletions(-)
+
+diff --git a/sysdeps/unix/sysv/linux/riscv/flush-icache.c b/sysdeps/unix/sysv/linux/riscv/flush-icache.c
+index 72caeb190f..ef33582bdf 100644
+--- a/sysdeps/unix/sysv/linux/riscv/flush-icache.c
++++ b/sysdeps/unix/sysv/linux/riscv/flush-icache.c
+@@ -21,11 +21,6 @@
+ #include <stdlib.h>
+ #include <atomic.h>
+ #include <sys/cachectl.h>
+-#if __has_include (<asm/syscalls.h>)
+-# include <asm/syscalls.h>
+-#else
+-# include <asm/unistd.h>
+-#endif
+ #include <sys/syscall.h>
+ 
+ typedef int (*func_type) (void *, void *, unsigned long int);
+-- 
+2.21.1
+
package/glibc/2.31-44-g3a44844c97a8ca1d0798c45b5252e6a49f9cf8e1/glibc.hash new file +7 -0
new file mode 100644
index 0000000000..27f04a38b9
@@ -0,0 +1,7 @@
+# Locally calculated (fetched from Github)
+sha256  09b57f576cd98333bc8a001e1e412652bcc75464a81684b39e3f71f3b4ecda51  glibc-2.31-44-g3a44844c97a8ca1d0798c45b5252e6a49f9cf8e1.tar.gz
+
+# Hashes for license files
+sha256  8177f97513213526df2cf6184d8ff986c675afb514d4e68a404010521b880643  COPYING
+sha256  dc626520dcd53a22f727af3ee42c770e56c97a64fe3adb063799d8ab032fe551  COPYING.LIB
+sha256  b33d0bd9f685b46853548814893a6135e74430d12f6d94ab3eba42fc591f83bc  LICENSES
package/glibc/Config.in +17 -1
index 9901ad8537..c9c1493167 100644
@@ -10,6 +10,21 @@ config BR2_PACKAGE_GLIBC
 	help
 	  https://www.gnu.org/software/libc/
 
+choice
+	prompt "glibc version"
+	# Following architectures are not using an upstream version:
+	depends on !BR2_arc
+	depends on !BR2_csky
+	depends on !BR2_RISCV_32
+
+config BR2_GLIBC_VERSION_2_30_X
+	bool "2.30-x (stable branch)"
+
+config BR2_GLIBC_VERSION_2_31_X
+	bool "2.31-x (stable branch)"
+
+endchoice
+
 config BR2_PACKAGE_GLIBC_UTILS
 	bool "Install glibc utilities"
 	help
@@ -25,7 +40,8 @@ endif # BR2_TOOLCHAIN_BUILDROOT_GLIBC
 #   git describe --match 'glibc-*' --abbrev=40 origin/release/MAJOR.MINOR/master | cut -d '-' -f 2-
 config BR2_GLIBC_VERSION_UPSTREAM
 	string
-	default "2.30-67-g4748829f86a458b76642f3e98b1d80f7b868e427"
+	default "2.30-67-g4748829f86a458b76642f3e98b1d80f7b868e427" if BR2_GLIBC_VERSION_2_30_X
+	default "2.31-44-g3a44844c97a8ca1d0798c45b5252e6a49f9cf8e1"
 
 config BR2_GLIBC_VERSION
 	string
package/localedef/2.31-44-g3a44844c97a8ca1d0798c45b5252e6a49f9cf8e1/0001-HACK-only-build-and-install-localedef.patch new file +65 -0

65 lines not on this page · raw message

package/localedef/2.31-44-g3a44844c97a8ca1d0798c45b5252e6a49f9cf8e1/0002-relax-dependency-on-GCC-to-4.8-and-binutils-to-2.24.patch new file +64 -0

64 lines not on this page · raw message

package/localedef/2.31-44-g3a44844c97a8ca1d0798c45b5252e6a49f9cf8e1/localedef.hash new file +1 -0
new file mode 120000
index 0000000000..209dc30a19
@@ -0,0 +1 @@
+../../glibc/2.31-44-g3a44844c97a8ca1d0798c45b5252e6a49f9cf8e1/glibc.hash
\ No newline@end of file
Signature · 1 line
2.25.4
#6 lore raw

PATCH 4/4 package/gcc: disable libsanitizer with gcc 7.5 and glibc 2.31 toolchain

In reply to Romain Naour

From: Romain Naour <[email protected]>

libsanitizer in gcc 7.x fails to build with glibc 2.31. Fixing it would require backporting an upstream commit:

    https://gcc.gnu.org/git/?p=gcc.git;a=commitdiff;h=4abc46b51af5751d657764d0c44b8a4aeed06302

However, the backport is not trivial, as there are a lot of conflicts.

Disable libsanitizer since the gcc 7.5 branch is now closed (unmaintained) and it's not a trivial merge.

---
 package/gcc/gcc.mk | 10 ++++++++++
 1 file changed, 10 insertions(+)
package/gcc/gcc.mk +10 -0
index d8260bc87b..87e613f18c 100644
@@ -119,6 +119,16 @@ else
 HOST_GCC_COMMON_CONF_OPTS += --disable-libquadmath
 endif
 
+# Disable libsanitizer due to a build issue with gcc 7.5 and glibc 2.31.
+# It would require to backport the following upstream commit
+# https://gcc.gnu.org/git/?p=gcc.git;a=commitdiff;h=4abc46b51af5751d657764d0c44b8a4aeed06302
+# but it conflict with gcc 7.5 libsanitizer code.
+# Disable libsanitizer since the gcc 7.5 branch is now closed
+# (unmaintained) and it's not a trivial merge.
+ifeq ($(BR2_GLIBC_VERSION_2_31_X)$(BR2_GCC_VERSION_7_X),yy)
+HOST_GCC_COMMON_CONF_OPTS += --disable-libsanitizer
+endif
+
 # libsanitizer requires wordexp, not in default uClibc config. Also
 # doesn't build properly with musl.
 ifeq ($(BR2_TOOLCHAIN_BUILDROOT_UCLIBC)$(BR2_TOOLCHAIN_BUILDROOT_MUSL),y)
Signature · 1 line
2.25.4