toolchain-external: remove support for glibc before 2.17
cover toolchain-external: remove support for glibc before 2.17
As discussed[1] recently on the list it is time to remove support for glibc older than 2.17. This does not mean that toolchains with old glibc will stop working. It only means that we stop fixing older glibc build issues. In particular, the need to link with librt for clock_* system calls.
The only significant user visible change in this series is the removal of CodeSourcery toolchains for x86 and sh.
Two patches update a comment section and the manual to take into account the removal of toolchain in this series and in previous commits. A few more per-package patches remove local patches for the glibc librt issue.
The first patch is an unrelated fix to the external toolchain section in the manual.
[1] http://patchwork.ozlabs.org/patch/773384/
Baruch Siach (14):
docs/manual: update external toolchain advantages list
toolchain: remove CodeSourcery x86 toolchain
toolchain: remove CodeSourcery sh toolchain
package: remove CodeSourcery sh dependencies
toolchain-external: update list of toolchains docs/manual: drop mention of removed external toolchains
swupdate: don't link with librt
aiccu: don't link with librt
libv4l: don't link with librt
e2fsprogs: don't link with librt
liboping: no need to check for clock_gettime
openipmi: no need to check for clock_gettime
lirc-tools: no need to check for clock_gettime
softether: no need to check for clock_gettime
Config.in.legacy | 18 +++++++
docs/manual/configure.txt | 16 +++---
package/aiccu/0003-clock-suite-functions-fix.patch | 18 -------
...her-header.patch => 0003-if-ether-header.patch} | 0
package/domoticz/Config.in | 3 --
.../0002-MCONFIG.in-handle-CLOCK_GETTIME_LIB.patch | 35 -------------
...001-configure-also-check-for-clockgettime.patch | 39 ---------------
package/liboping/liboping.mk | 2 -
...3-utils-qv4l2-Makefile.am-link-with-librt.patch | 29 -----------
package/libv4l/libv4l.mk | 3 +-
package/liquid-dsp/Config.in | 1 -
...onfigure-check-for-clock_gettime-in-librt.patch | 54 --------------------
...-use-of-functions-killed-in-kernel-4.8.0.patch} | 0
package/lirc-tools/lirc-tools.mk | 2 -
...c-detect-lrt-requirement-for-clock_gettim.patch | 57 ----------------------
package/prboom/Config.in | 2 -
package/qt/Config.in | 7 +--
package/softether/0006-librt.patch | 33 -------------
...onfig.patch => 0006-uclibc-ai-addrconfig.patch} | 0
package/swupdate/swupdate.config | 2 +-
toolchain/toolchain-external/Config.in | 12 +----
.../toolchain-external/pkg-toolchain-external.mk | 3 +-
.../toolchain-external-codesourcery-sh/Config.in | 25 ----------
.../Config.in.options | 9 ----
.../toolchain-external-codesourcery-sh.hash | 3 --
.../toolchain-external-codesourcery-sh.mk | 13 -----
.../toolchain-external-codesourcery-x86/Config.in | 29 -----------
.../Config.in.options | 9 ----
.../toolchain-external-codesourcery-x86.hash | 3 --
.../toolchain-external-codesourcery-x86.mk | 12 -----
30 files changed, 31 insertions(+), 408 deletions(-)
delete mode 100644 package/aiccu/0003-clock-suite-functions-fix.patch
rename package/aiccu/{0004-if-ether-header.patch => 0003-if-ether-header.patch} (100%)
delete mode 100644 package/e2fsprogs/0002-MCONFIG.in-handle-CLOCK_GETTIME_LIB.patch
delete mode 100644 package/liboping/0001-configure-also-check-for-clockgettime.patch
delete mode 100644 package/libv4l/0003-utils-qv4l2-Makefile.am-link-with-librt.patch
delete mode 100644 package/lirc-tools/0002-configure-check-for-clock_gettime-in-librt.patch
rename package/lirc-tools/{0003-lircd-Remove-use-of-functions-killed-in-kernel-4.8.0.patch => 0002-lircd-Remove-use-of-functions-killed-in-kernel-4.8.0.patch} (100%)
delete mode 100644 package/openipmi/0002-configure.ac-detect-lrt-requirement-for-clock_gettim.patch
delete mode 100644 package/softether/0006-librt.patch
rename package/softether/{0007-uclibc-ai-addrconfig.patch => 0006-uclibc-ai-addrconfig.patch} (100%)
delete mode 100644 toolchain/toolchain-external/toolchain-external-codesourcery-sh/Config.in
delete mode 100644 toolchain/toolchain-external/toolchain-external-codesourcery-sh/Config.in.options
delete mode 100644 toolchain/toolchain-external/toolchain-external-codesourcery-sh/toolchain-external-codesourcery-sh.hash
delete mode 100644 toolchain/toolchain-external/toolchain-external-codesourcery-sh/toolchain-external-codesourcery-sh.mk
delete mode 100644 toolchain/toolchain-external/toolchain-external-codesourcery-x86/Config.in
delete mode 100644 toolchain/toolchain-external/toolchain-external-codesourcery-x86/Config.in.options
delete mode 100644 toolchain/toolchain-external/toolchain-external-codesourcery-x86/toolchain-external-codesourcery-x86.hash
delete mode 100644 toolchain/toolchain-external/toolchain-external-codesourcery-x86/toolchain-external-codesourcery-x86.mkSignature · 1 line
2.11.0
PATCH 1/14 docs/manual: update external toolchain advantages list
Buildroot can build non-uClibc toolchains internally for quite some time now. Update the manual text.
--- docs/manual/configure.txt | 2 -- 1 file changed, 2 deletions(-)
docs/manual/configure.txt +0 -2
@@ -221,8 +221,6 @@ Advantages of this backend: often very significant in the overall build time of an embedded Linux system. -* Not limited to uClibc: glibc and eglibc toolchains are supported. - Drawbacks of this backend: * If your pre-built external toolchain has a bug, may be hard to get a
Signature · 1 line
2.11.0
PATCH 1/14 docs/manual: update external toolchain advantages list
Baruch, All,
On 2017-06-16 06:32 +0300, Baruch Siach spake thusly:
Quoted 4 lines
Buildroot can build non-uClibc toolchains internally for quite some time now. Update the manual text.
Signed-off-by Baruch Siach <[email protected]>
Regards,
Yann E. MORIN.
Quoted 24 lines
--- docs/manual/configure.txt | 2 -- 1 file changed, 2 deletions(-)docs/manual/configure.txt +0
-2@@ -221,8 +221,6 @@ Advantages of this backend: often very significant in the overall build time of an embedded Linux system. -* Not limited to uClibc: glibc and eglibc toolchains are supported. - Drawbacks of this backend: * If your pre-built external toolchain has a bug, may be hard to get aSignature · 6 lines
2.11.0 _______________________________________________ buildroot mailing list buildroot at busybox.net http://lists.busybox.net/mailman/listinfo/buildroot
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. | '------------------------------^-------^------------------^--------------------'
PATCH 2/14 toolchain: remove CodeSourcery x86 toolchain
In reply to Baruch SiachSince glibc 2.17, executable link command need not include the -lrt option for clock_* system calls. As a result, over time less and less software packages bother to check whether to toolchain needs -lrt. We are now at a point where maintainers refuse to add this complexity into their build system. This requires Buildroot to carry patches fixing this issue indefinitely.
glibc 2.17 is now 4.5 years old. There is no reason to use an older version with current software.
This commit removes the predefined profile for CodeSourcery x86 toolchain that is based on glibc 2.16. One may still use the custom external toolchain support in Buildroot to get this toolchain back, and deal with any build issues that this toolchain causes.
--- Config.in.legacy | 9 +++++++ toolchain/toolchain-external/Config.in | 6 ++--- .../toolchain-external-codesourcery-x86/Config.in | 29 ---------------------- .../Config.in.options | 9 ------- .../toolchain-external-codesourcery-x86.hash | 3 --- .../toolchain-external-codesourcery-x86.mk | 12 --------- 6 files changed, 11 insertions(+), 57 deletions(-) delete mode 100644 toolchain/toolchain-external/toolchain-external-codesourcery-x86/Config.in delete mode 100644 toolchain/toolchain-external/toolchain-external-codesourcery-x86/Config.in.options delete mode 100644 toolchain/toolchain-external/toolchain-external-codesourcery-x86/toolchain-external-codesourcery-x86.hash delete mode 100644 toolchain/toolchain-external/toolchain-external-codesourcery-x86/toolchain-external-codesourcery-x86.mk
Config.in.legacy +9 -0
@@ -145,6 +145,15 @@ endif ############################################################################### comment "Legacy options removed in 2017.08" +config BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_X86 + bool "x86 Sourcery toolchain has been removed" + select BR2_LEGACY + help + The Sourcery CodeBench toolchain for the x86 architecture has + been removed, since it uses glibc older than 2.17 that requires + -lrt to link executables using clock_* system calls. This makes + this toolchain difficult to maintain over time. + config BR2_GCC_VERSION_4_8_X bool "gcc 4.8.x support removed" select BR2_LEGACY
toolchain/toolchain-external/Config.in +2 -4
@@ -36,9 +36,8 @@ source "toolchain/toolchain-external/toolchain-external-codesourcery-niosII/Conf # SH4a source "toolchain/toolchain-external/toolchain-external-codesourcery-sh/Config.in" -# x86/x86_64 (use amd64 toolchain by default for AMD64) +# x86_64 source "toolchain/toolchain-external/toolchain-external-codesourcery-amd64/Config.in" -source "toolchain/toolchain-external/toolchain-external-codesourcery-x86/Config.in" # Kept last, so it remains the non-default choice, unless there isn't # any available toolchain profile for the currently selected
@@ -130,9 +129,8 @@ source "toolchain/toolchain-external/toolchain-external-codesourcery-niosII/Conf # SH4a source "toolchain/toolchain-external/toolchain-external-codesourcery-sh/Config.in.options" -# x86/x86_64 +# x86_64 source "toolchain/toolchain-external/toolchain-external-codesourcery-amd64/Config.in.options" -source "toolchain/toolchain-external/toolchain-external-codesourcery-x86/Config.in.options" # Custom toolchains source "toolchain/toolchain-external/toolchain-external-custom/Config.in.options"
toolchain/toolchain-external/toolchain-external-codesourcery-x86/Config.in deleted +0 -29
@@ -1,29 +0,0 @@ -config BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_X86 - bool "Sourcery CodeBench x86/x86_64 2012.09" - depends on BR2_i386 || BR2_x86_64 - depends on BR2_HOSTARCH = "x86_64" || BR2_HOSTARCH = "x86" - depends on !BR2_STATIC_LIBS - depends on !BR2_x86_jaguar - depends on !BR2_x86_steamroller - select BR2_TOOLCHAIN_EXTERNAL_GLIBC - select BR2_TOOLCHAIN_HAS_SSP - select BR2_TOOLCHAIN_HAS_NATIVE_RPC - select BR2_INSTALL_LIBSTDCPP - select BR2_HOSTARCH_NEEDS_IA32_LIBS - select BR2_TOOLCHAIN_HEADERS_AT_LEAST_3_5 - select BR2_TOOLCHAIN_GCC_AT_LEAST_4_7 - help - Sourcery CodeBench toolchain for the x86/x86_64 - architectures, from Mentor Graphics. It uses gcc 4.7.2, - binutils 2.23.51, glibc 2.16, gdb 7.4.50 and kernel headers - 3.5.4. It has support for the following variants: - - Intel Pentium 4, glibc, 32 bits - Default for x86, nothing special to do. - - Intel Atom, glibc, 32 bits - Select an Atom core - - Intel Xeon, glibc, 64 bits - Default for x86_64, nothing special to do. - - Intel Core 2, glibc, 64 bits - Select a Core 2 core - Other architecture variants (beyond Pentium-4/Xeon) are - supported as well, but glibc is not optimised for it.
toolchain/toolchain-external/toolchain-external-codesourcery-x86/Config.in.options deleted +0 -9
@@ -1,9 +0,0 @@ -if BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_X86 - -config BR2_TOOLCHAIN_EXTERNAL_PREFIX - default "i686-pc-linux-gnu" - -config BR2_PACKAGE_PROVIDES_TOOLCHAIN_EXTERNAL - default "toolchain-external-codesourcery-x86" - -endif
toolchain/toolchain-external/toolchain-external-codesourcery-x86/toolchain-external-codesourcery-x86.hash deleted +0 -3
@@ -1,3 +0,0 @@ -# Locally calculated -sha256 ea804cf02014369da52abc4f64e91e96bde2dd2230aca96109459013d4545458 ia32-2012.09-62-i686-pc-linux-gnu-i386-linux.tar.bz2 -sha256 1a9519e415a1e6892c760bf21f7e98f3a633a9d1c5bb8781a96d338e4dd62717 ia32-2012.09-62-i686-pc-linux-gnu.src.tar.bz2
toolchain/toolchain-external/toolchain-external-codesourcery-x86/toolchain-external-codesourcery-x86.mk deleted +0 -12
@@ -1,12 +0,0 @@ -################################################################################ -# -# toolchain-external-codesourcery-x86 -# -################################################################################ - -TOOLCHAIN_EXTERNAL_CODESOURCERY_X86_SITE = https://sourcery.mentor.com/public/gnu_toolchain/$(TOOLCHAIN_EXTERNAL_PREFIX) -TOOLCHAIN_EXTERNAL_CODESOURCERY_X86_VERSION = 2012.09-62 -TOOLCHAIN_EXTERNAL_CODESOURCERY_X86_SOURCE = ia32-$(TOOLCHAIN_EXTERNAL_CODESOURCERY_X86_VERSION)-$(TOOLCHAIN_EXTERNAL_PREFIX)-i386-linux.tar.bz2 -TOOLCHAIN_EXTERNAL_CODESOURCERY_X86_ACTUAL_SOURCE_TARBALL = ia32-$(TOOLCHAIN_EXTERNAL_CODESOURCERY_X86_VERSION)-$(TOOLCHAIN_EXTERNAL_PREFIX).src.tar.bz2 - -$(eval $(toolchain-external-package))
Signature · 1 line
2.11.0
PATCH 2/14 toolchain: remove CodeSourcery x86 toolchain
Baruch, All,
On 2017-06-16 06:32 +0300, Baruch Siach spake thusly:
Quoted 5 lines
Since glibc 2.17, executable link command need not include the -lrt option for clock_* system calls. As a result, over time less and less software packages bother to check whether to toolchain needs -lrt. We are now at a point where maintainers refuse to add this complexity into their build system. This requires Buildroot to carry patches fixing this issue indefinitely.
What upstreams are refusing to fix that? Do you have references to share?
Quoted 2 lines
glibc 2.17 is now 4.5 years old. There is no reason to use an older version with current software.
So we can now consider the cut-off at what we consider legacy is 4.5 years old. Fine with me.
Quoted 6 lines
This commit removes the predefined profile for CodeSourcery x86 toolchain that is based on glibc 2.16. One may still use the custom external toolchain support in Buildroot to get this toolchain back, and deal with any build issues that this toolchain causes.
Signed-off-by Baruch Siach <[email protected]>
Regards,
Yann E. MORIN.
Quoted 143 lines
--- Config.in.legacy | 9 +++++++ toolchain/toolchain-external/Config.in | 6 ++--- .../toolchain-external-codesourcery-x86/Config.in | 29 ---------------------- .../Config.in.options | 9 ------- .../toolchain-external-codesourcery-x86.hash | 3 --- .../toolchain-external-codesourcery-x86.mk | 12 --------- 6 files changed, 11 insertions(+), 57 deletions(-) delete mode 100644 toolchain/toolchain-external/toolchain-external-codesourcery-x86/Config.in delete mode 100644 toolchain/toolchain-external/toolchain-external-codesourcery-x86/Config.in.options delete mode 100644 toolchain/toolchain-external/toolchain-external-codesourcery-x86/toolchain-external-codesourcery-x86.hash delete mode 100644 toolchain/toolchain-external/toolchain-external-codesourcery-x86/toolchain-external-codesourcery-x86.mkConfig.in.legacy +9
-0@@ -145,6 +145,15 @@ endif ############################################################################### comment "Legacy options removed in 2017.08" +config BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_X86 + bool "x86 Sourcery toolchain has been removed" + select BR2_LEGACY + help + The Sourcery CodeBench toolchain for the x86 architecture has + been removed, since it uses glibc older than 2.17 that requires + -lrt to link executables using clock_* system calls. This makes + this toolchain difficult to maintain over time. + config BR2_GCC_VERSION_4_8_X bool "gcc 4.8.x support removed" select BR2_LEGACYtoolchain/toolchain-external/Config.in +2
-4@@ -36,9 +36,8 @@ source "toolchain/toolchain-external/toolchain-external-codesourcery-niosII/Conf # SH4a source "toolchain/toolchain-external/toolchain-external-codesourcery-sh/Config.in" -# x86/x86_64 (use amd64 toolchain by default for AMD64) +# x86_64 source "toolchain/toolchain-external/toolchain-external-codesourcery-amd64/Config.in" -source "toolchain/toolchain-external/toolchain-external-codesourcery-x86/Config.in" # Kept last, so it remains the non-default choice, unless there isn't # any available toolchain profile for the currently selected@@ -130,9 +129,8 @@ source "toolchain/toolchain-external/toolchain-external-codesourcery-niosII/Conf # SH4a source "toolchain/toolchain-external/toolchain-external-codesourcery-sh/Config.in.options" -# x86/x86_64 +# x86_64 source "toolchain/toolchain-external/toolchain-external-codesourcery-amd64/Config.in.options" -source "toolchain/toolchain-external/toolchain-external-codesourcery-x86/Config.in.options" # Custom toolchains source "toolchain/toolchain-external/toolchain-external-custom/Config.in.options"toolchain/toolchain-external/toolchain-external-codesourcery-x86/Config.in deleted +0
-29@@ -1,29 +0,0 @@ -config BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_X86 - bool "Sourcery CodeBench x86/x86_64 2012.09" - depends on BR2_i386 || BR2_x86_64 - depends on BR2_HOSTARCH = "x86_64" || BR2_HOSTARCH = "x86" - depends on !BR2_STATIC_LIBS - depends on !BR2_x86_jaguar - depends on !BR2_x86_steamroller - select BR2_TOOLCHAIN_EXTERNAL_GLIBC - select BR2_TOOLCHAIN_HAS_SSP - select BR2_TOOLCHAIN_HAS_NATIVE_RPC - select BR2_INSTALL_LIBSTDCPP - select BR2_HOSTARCH_NEEDS_IA32_LIBS - select BR2_TOOLCHAIN_HEADERS_AT_LEAST_3_5 - select BR2_TOOLCHAIN_GCC_AT_LEAST_4_7 - help - Sourcery CodeBench toolchain for the x86/x86_64 - architectures, from Mentor Graphics. It uses gcc 4.7.2, - binutils 2.23.51, glibc 2.16, gdb 7.4.50 and kernel headers - 3.5.4. It has support for the following variants: - - Intel Pentium 4, glibc, 32 bits - Default for x86, nothing special to do. - - Intel Atom, glibc, 32 bits - Select an Atom core - - Intel Xeon, glibc, 64 bits - Default for x86_64, nothing special to do. - - Intel Core 2, glibc, 64 bits - Select a Core 2 core - Other architecture variants (beyond Pentium-4/Xeon) are - supported as well, but glibc is not optimised for it.toolchain/toolchain-external/toolchain-external-codesourcery-x86/Config.in.options deleted +0
-9@@ -1,9 +0,0 @@ -if BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_X86 - -config BR2_TOOLCHAIN_EXTERNAL_PREFIX - default "i686-pc-linux-gnu" - -config BR2_PACKAGE_PROVIDES_TOOLCHAIN_EXTERNAL - default "toolchain-external-codesourcery-x86" - -endiftoolchain/toolchain-external/toolchain-external-codesourcery-x86/toolchain-external-codesourcery-x86.hash deleted +0
-3@@ -1,3 +0,0 @@ -# Locally calculated -sha256 ea804cf02014369da52abc4f64e91e96bde2dd2230aca96109459013d4545458 ia32-2012.09-62-i686-pc-linux-gnu-i386-linux.tar.bz2 -sha256 1a9519e415a1e6892c760bf21f7e98f3a633a9d1c5bb8781a96d338e4dd62717 ia32-2012.09-62-i686-pc-linux-gnu.src.tar.bz2toolchain/toolchain-external/toolchain-external-codesourcery-x86/toolchain-external-codesourcery-x86.mk deleted +0
-12@@ -1,12 +0,0 @@ -################################################################################ -# -# toolchain-external-codesourcery-x86 -# -################################################################################ - -TOOLCHAIN_EXTERNAL_CODESOURCERY_X86_SITE = https://sourcery.mentor.com/public/gnu_toolchain/$(TOOLCHAIN_EXTERNAL_PREFIX) -TOOLCHAIN_EXTERNAL_CODESOURCERY_X86_VERSION = 2012.09-62 -TOOLCHAIN_EXTERNAL_CODESOURCERY_X86_SOURCE = ia32-$(TOOLCHAIN_EXTERNAL_CODESOURCERY_X86_VERSION)-$(TOOLCHAIN_EXTERNAL_PREFIX)-i386-linux.tar.bz2 -TOOLCHAIN_EXTERNAL_CODESOURCERY_X86_ACTUAL_SOURCE_TARBALL = ia32-$(TOOLCHAIN_EXTERNAL_CODESOURCERY_X86_VERSION)-$(TOOLCHAIN_EXTERNAL_PREFIX).src.tar.bz2 - -$(eval $(toolchain-external-package))Signature · 6 lines
2.11.0 _______________________________________________ buildroot mailing list buildroot at busybox.net http://lists.busybox.net/mailman/listinfo/buildroot
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. | '------------------------------^-------^------------------^--------------------'
PATCH 2/14 toolchain: remove CodeSourcery x86 toolchain
Hi Yann,
Quoted 11 lines · Yann E. MORIN
Baruch, All,
On 2017-06-16 06:32 +0300, Baruch Siach spake thusly:
Quoted 5 lines
Since glibc 2.17, executable link command need not include the -lrt option for clock_* system calls. As a result, over time less and less software packages bother to check whether to toolchain needs -lrt. We are now at a point where maintainers refuse to add this complexity into their build system. This requires Buildroot to carry patches fixing this issue indefinitely.
What upstreams are refusing to fix that? Do you have references to share?
See http://patchwork.ozlabs.org/patch/773384/ (swupdate).
I put the link in the series cover. Should add it here as well?
baruch
Quoted 15 lines
Quoted 3 lines
glibc 2.17 is now 4.5 years old. There is no reason to use an older version with current software.
So we can now consider the cut-off at what we consider legacy is 4.5 years old. Fine with me.
Quoted 6 lines
This commit removes the predefined profile for CodeSourcery x86 toolchain that is based on glibc 2.16. One may still use the custom external toolchain support in Buildroot to get this toolchain back, and deal with any build issues that this toolchain causes.
Signed-off-by Baruch Siach <[email protected]>Reviewed-by "Yann E. MORIN" <[email protected]>
Signature · 3 lines
http://baruch.siach.name/blog/ ~. .~ Tk Open Systems
=}------------------------------------------------ooO--U--Ooo------------{=
- baruch at tkos.co.il - tel: +972.52.368.4656, http://www.tkos.co.il -PATCH 2/14 toolchain: remove CodeSourcery x86 toolchain
Baruch, All,
On 2017-06-18 12:14 +0300, Baruch Siach spake thusly:
Quoted 10 lines · Yann E. MORIN
Quoted 9 lines
Baruch, All,
On 2017-06-16 06:32 +0300, Baruch Siach spake thusly:
Quoted 5 lines
Since glibc 2.17, executable link command need not include the -lrt option for clock_* system calls. As a result, over time less and less software packages bother to check whether to toolchain needs -lrt. We are now at a point where maintainers refuse to add this complexity into their build system. This requires Buildroot to carry patches fixing this issue indefinitely.
Quoted 5 lines
Quoted 2 lines
What upstreams are refusing to fix that? Do you have references to share?
See http://patchwork.ozlabs.org/patch/773384/ (swupdate). I put the link in the series cover. Should add it here as well?
Ah, I missed it in the cover letter. Thanks. I don't think it is important that you repeat it here.
Regards,
Yann E. MORIN.
Quoted 22 lines
baruch
Quoted 15 lines
Quoted 3 lines
glibc 2.17 is now 4.5 years old. There is no reason to use an older version with current software.
So we can now consider the cut-off at what we consider legacy is 4.5 years old. Fine with me.
Quoted 6 lines
This commit removes the predefined profile for CodeSourcery x86 toolchain that is based on glibc 2.16. One may still use the custom external toolchain support in Buildroot to get this toolchain back, and deal with any build issues that this toolchain causes.
Signed-off-by Baruch Siach <[email protected]>Reviewed-by "Yann E. MORIN" <[email protected]>Signature · 3 lines
http://baruch.siach.name/blog/ ~. .~ Tk Open Systems =}------------------------------------------------ooO--U--Ooo------------{= - baruch at tkos.co.il - tel: +972.52.368.4656, http://www.tkos.co.il -
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. | '------------------------------^-------^------------------^--------------------'
PATCH 3/14 toolchain: remove CodeSourcery sh toolchain
In reply to Baruch SiachSince glibc 2.17, executable link command need not include the -lrt option for clock_* system calls. As a result, over time less and less software packages bother to check whether to toolchain needs -lrt. We are now at a point where maintainers refuse to add this complexity into their build system. This requires Buildroot to carry patches fixing this issue indefinitely.
glibc 2.17 is now 4.5 years old. There is no reason to use an older version with current software.
This commit removes the predefined profile for CodeSourcery sh toolchain that is based on glibc 2.16. One may still use the custom external toolchain support in Buildroot to get this toolchain back, and deal with any build issues that this toolchain causes.
--- Config.in.legacy | 9 ++++++++ toolchain/toolchain-external/Config.in | 6 ------ .../toolchain-external-codesourcery-sh/Config.in | 25 ---------------------- .../Config.in.options | 9 -------- .../toolchain-external-codesourcery-sh.hash | 3 --- .../toolchain-external-codesourcery-sh.mk | 13 ----------- 6 files changed, 9 insertions(+), 56 deletions(-) delete mode 100644 toolchain/toolchain-external/toolchain-external-codesourcery-sh/Config.in delete mode 100644 toolchain/toolchain-external/toolchain-external-codesourcery-sh/Config.in.options delete mode 100644 toolchain/toolchain-external/toolchain-external-codesourcery-sh/toolchain-external-codesourcery-sh.hash delete mode 100644 toolchain/toolchain-external/toolchain-external-codesourcery-sh/toolchain-external-codesourcery-sh.mk
Config.in.legacy +9 -0
@@ -145,6 +145,15 @@ endif ############################################################################### comment "Legacy options removed in 2017.08" +config BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_SH + bool "SH Sourcery toolchain has been removed" + select BR2_LEGACY + help + The Sourcery CodeBench toolchain for the sh architecture has + been removed, since it uses glibc older than 2.17 that requires + -lrt to link executables using clock_* system calls. This makes + this toolchain difficult to maintain over time. + config BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_X86 bool "x86 Sourcery toolchain has been removed" select BR2_LEGACY
toolchain/toolchain-external/Config.in +0 -6
@@ -33,9 +33,6 @@ source "toolchain/toolchain-external/toolchain-external-codescape-mti-mips/Confi # NIOSII source "toolchain/toolchain-external/toolchain-external-codesourcery-niosII/Config.in" -# SH4a -source "toolchain/toolchain-external/toolchain-external-codesourcery-sh/Config.in" - # x86_64 source "toolchain/toolchain-external/toolchain-external-codesourcery-amd64/Config.in"
@@ -126,9 +123,6 @@ source "toolchain/toolchain-external/toolchain-external-codescape-mti-mips/Confi # NIOSII source "toolchain/toolchain-external/toolchain-external-codesourcery-niosII/Config.in.options" -# SH4a -source "toolchain/toolchain-external/toolchain-external-codesourcery-sh/Config.in.options" - # x86_64 source "toolchain/toolchain-external/toolchain-external-codesourcery-amd64/Config.in.options"
toolchain/toolchain-external/toolchain-external-codesourcery-sh/Config.in deleted +0 -25
@@ -1,25 +0,0 @@ -config BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_SH - bool "Sourcery CodeBench SH 2012.09" - depends on BR2_sh4a || BR2_sh4aeb - depends on BR2_HOSTARCH = "x86_64" || BR2_HOSTARCH = "x86" - depends on !BR2_STATIC_LIBS - select BR2_TOOLCHAIN_EXTERNAL_GLIBC - select BR2_TOOLCHAIN_HAS_SSP - select BR2_TOOLCHAIN_HAS_NATIVE_RPC - select BR2_INSTALL_LIBSTDCPP - select BR2_HOSTARCH_NEEDS_IA32_LIBS - select BR2_TOOLCHAIN_HEADERS_AT_LEAST_3_5 - select BR2_TOOLCHAIN_GCC_AT_LEAST_4_7 - help - Sourcery CodeBench toolchain for the SuperH architecture, - from Mentor Graphics. It uses gcc 4.7.2, binutils 2.23.51, - glibc 2.16, uClibc 0.9.30, gdb 7.4.50 and kernel headers - 3.5.4. It has support for the following variants: - - SH4A, glibc, little endian - Default. - - SH4A, glibc, big endian - Add -mb to BR2_TARGET_OPTIMIZATION - - SH4A, uClibc, little endian - Not usable in Buildroot yet. - - SH4A, uClibc, big endian - Not usable in Buildroot yet.
toolchain/toolchain-external/toolchain-external-codesourcery-sh/Config.in.options deleted +0 -9
@@ -1,9 +0,0 @@ -if BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_SH - -config BR2_TOOLCHAIN_EXTERNAL_PREFIX - default "sh-linux-gnu" - -config BR2_PACKAGE_PROVIDES_TOOLCHAIN_EXTERNAL - default "toolchain-external-codesourcery-sh" - -endif
toolchain/toolchain-external/toolchain-external-codesourcery-sh/toolchain-external-codesourcery-sh.hash deleted +0 -3
@@ -1,3 +0,0 @@ -# Locally calculated -sha256 59d6766fde244931aa52db01433d5acd051998762a931121c5fc109536a1a802 renesas-2012.09-61-sh-linux-gnu-i686-pc-linux-gnu.tar.bz2 -sha256 e2e58c10e52395d5d35157e35f85233f713c6f9223a652dfc56194cfd2eed004 renesas-2012.09-61-sh-linux-gnu.src.tar.bz2
toolchain/toolchain-external/toolchain-external-codesourcery-sh/toolchain-external-codesourcery-sh.mk deleted +0 -13
@@ -1,13 +0,0 @@ -################################################################################ -# -# toolchain-external-sourcery-sh -# -################################################################################ - -TOOLCHAIN_EXTERNAL_CODESOURCERY_SH_VERSION = 2012.09-61 - -TOOLCHAIN_EXTERNAL_CODESOURCERY_SH_SITE = https://sourcery.mentor.com/public/gnu_toolchain/$(TOOLCHAIN_EXTERNAL_PREFIX) -TOOLCHAIN_EXTERNAL_CODESOURCERY_SH_SOURCE = renesas-$(TOOLCHAIN_EXTERNAL_CODESOURCERY_SH_VERSION)-$(TOOLCHAIN_EXTERNAL_PREFIX)-i686-pc-linux-gnu.tar.bz2 -TOOLCHAIN_EXTERNAL_CODESOURCERY_SH_ACTUAL_SOURCE_TARBALL = renesas-$(TOOLCHAIN_EXTERNAL_CODESOURCERY_SH_VERSION)-$(TOOLCHAIN_EXTERNAL_PREFIX).src.tar.bz2 - -$(eval $(toolchain-external-package))
Signature · 1 line
2.11.0
PATCH 3/14 toolchain: remove CodeSourcery sh toolchain
Baruch, All,
On 2017-06-16 06:32 +0300, Baruch Siach spake thusly:
Quoted 9 lines
Since glibc 2.17, executable link command need not include the -lrt option for clock_* system calls. As a result, over time less and less software packages bother to check whether to toolchain needs -lrt. We are now at a point where maintainers refuse to add this complexity into their build system. This requires Buildroot to carry patches fixing this issue indefinitely.
glibc 2.17 is now 4.5 years old. There is no reason to use an older version with current software.
Quoted 6 lines
This commit removes the predefined profile for CodeSourcery sh toolchain that is based on glibc 2.16. One may still use the custom external toolchain support in Buildroot to get this toolchain back, and deal with any build issues that this toolchain causes.
Signed-off-by Baruch Siach <[email protected]>
Regards,
Yann E. MORIN.
Quoted 138 lines
--- Config.in.legacy | 9 ++++++++ toolchain/toolchain-external/Config.in | 6 ------ .../toolchain-external-codesourcery-sh/Config.in | 25 ---------------------- .../Config.in.options | 9 -------- .../toolchain-external-codesourcery-sh.hash | 3 --- .../toolchain-external-codesourcery-sh.mk | 13 ----------- 6 files changed, 9 insertions(+), 56 deletions(-) delete mode 100644 toolchain/toolchain-external/toolchain-external-codesourcery-sh/Config.in delete mode 100644 toolchain/toolchain-external/toolchain-external-codesourcery-sh/Config.in.options delete mode 100644 toolchain/toolchain-external/toolchain-external-codesourcery-sh/toolchain-external-codesourcery-sh.hash delete mode 100644 toolchain/toolchain-external/toolchain-external-codesourcery-sh/toolchain-external-codesourcery-sh.mkConfig.in.legacy +9
-0@@ -145,6 +145,15 @@ endif ############################################################################### comment "Legacy options removed in 2017.08" +config BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_SH + bool "SH Sourcery toolchain has been removed" + select BR2_LEGACY + help + The Sourcery CodeBench toolchain for the sh architecture has + been removed, since it uses glibc older than 2.17 that requires + -lrt to link executables using clock_* system calls. This makes + this toolchain difficult to maintain over time. + config BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_X86 bool "x86 Sourcery toolchain has been removed" select BR2_LEGACYtoolchain/toolchain-external/Config.in +0
-6@@ -33,9 +33,6 @@ source "toolchain/toolchain-external/toolchain-external-codescape-mti-mips/Confi # NIOSII source "toolchain/toolchain-external/toolchain-external-codesourcery-niosII/Config.in" -# SH4a -source "toolchain/toolchain-external/toolchain-external-codesourcery-sh/Config.in" - # x86_64 source "toolchain/toolchain-external/toolchain-external-codesourcery-amd64/Config.in"@@ -126,9 +123,6 @@ source "toolchain/toolchain-external/toolchain-external-codescape-mti-mips/Confi # NIOSII source "toolchain/toolchain-external/toolchain-external-codesourcery-niosII/Config.in.options" -# SH4a -source "toolchain/toolchain-external/toolchain-external-codesourcery-sh/Config.in.options" - # x86_64 source "toolchain/toolchain-external/toolchain-external-codesourcery-amd64/Config.in.options"toolchain/toolchain-external/toolchain-external-codesourcery-sh/Config.in deleted +0
-25@@ -1,25 +0,0 @@ -config BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_SH - bool "Sourcery CodeBench SH 2012.09" - depends on BR2_sh4a || BR2_sh4aeb - depends on BR2_HOSTARCH = "x86_64" || BR2_HOSTARCH = "x86" - depends on !BR2_STATIC_LIBS - select BR2_TOOLCHAIN_EXTERNAL_GLIBC - select BR2_TOOLCHAIN_HAS_SSP - select BR2_TOOLCHAIN_HAS_NATIVE_RPC - select BR2_INSTALL_LIBSTDCPP - select BR2_HOSTARCH_NEEDS_IA32_LIBS - select BR2_TOOLCHAIN_HEADERS_AT_LEAST_3_5 - select BR2_TOOLCHAIN_GCC_AT_LEAST_4_7 - help - Sourcery CodeBench toolchain for the SuperH architecture, - from Mentor Graphics. It uses gcc 4.7.2, binutils 2.23.51, - glibc 2.16, uClibc 0.9.30, gdb 7.4.50 and kernel headers - 3.5.4. It has support for the following variants: - - SH4A, glibc, little endian - Default. - - SH4A, glibc, big endian - Add -mb to BR2_TARGET_OPTIMIZATION - - SH4A, uClibc, little endian - Not usable in Buildroot yet. - - SH4A, uClibc, big endian - Not usable in Buildroot yet.toolchain/toolchain-external/toolchain-external-codesourcery-sh/Config.in.options deleted +0
-9@@ -1,9 +0,0 @@ -if BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_SH - -config BR2_TOOLCHAIN_EXTERNAL_PREFIX - default "sh-linux-gnu" - -config BR2_PACKAGE_PROVIDES_TOOLCHAIN_EXTERNAL - default "toolchain-external-codesourcery-sh" - -endiftoolchain/toolchain-external/toolchain-external-codesourcery-sh/toolchain-external-codesourcery-sh.hash deleted +0
-3@@ -1,3 +0,0 @@ -# Locally calculated -sha256 59d6766fde244931aa52db01433d5acd051998762a931121c5fc109536a1a802 renesas-2012.09-61-sh-linux-gnu-i686-pc-linux-gnu.tar.bz2 -sha256 e2e58c10e52395d5d35157e35f85233f713c6f9223a652dfc56194cfd2eed004 renesas-2012.09-61-sh-linux-gnu.src.tar.bz2toolchain/toolchain-external/toolchain-external-codesourcery-sh/toolchain-external-codesourcery-sh.mk deleted +0
-13@@ -1,13 +0,0 @@ -################################################################################ -# -# toolchain-external-sourcery-sh -# -################################################################################ - -TOOLCHAIN_EXTERNAL_CODESOURCERY_SH_VERSION = 2012.09-61 - -TOOLCHAIN_EXTERNAL_CODESOURCERY_SH_SITE = https://sourcery.mentor.com/public/gnu_toolchain/$(TOOLCHAIN_EXTERNAL_PREFIX) -TOOLCHAIN_EXTERNAL_CODESOURCERY_SH_SOURCE = renesas-$(TOOLCHAIN_EXTERNAL_CODESOURCERY_SH_VERSION)-$(TOOLCHAIN_EXTERNAL_PREFIX)-i686-pc-linux-gnu.tar.bz2 -TOOLCHAIN_EXTERNAL_CODESOURCERY_SH_ACTUAL_SOURCE_TARBALL = renesas-$(TOOLCHAIN_EXTERNAL_CODESOURCERY_SH_VERSION)-$(TOOLCHAIN_EXTERNAL_PREFIX).src.tar.bz2 - -$(eval $(toolchain-external-package))Signature · 6 lines
2.11.0 _______________________________________________ buildroot mailing list buildroot at busybox.net http://lists.busybox.net/mailman/listinfo/buildroot
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. | '------------------------------^-------^------------------^--------------------'
PATCH 4/14 package: remove CodeSourcery sh dependencies
In reply to Baruch SiachThe CodeSourcery sh toolchain has been removed. Drop negative dependencies on that toolchain.
--- package/domoticz/Config.in | 3 --- package/liquid-dsp/Config.in | 1 - package/prboom/Config.in | 2 -- package/qt/Config.in | 7 ++----- 4 files changed, 2 insertions(+), 11 deletions(-)
package/domoticz/Config.in +0 -3
@@ -7,8 +7,6 @@ config BR2_PACKAGE_DOMOTICZ depends on BR2_INSTALL_LIBSTDCPP depends on BR2_USE_WCHAR depends on BR2_PACKAGE_LUA_5_2 || BR2_PACKAGE_LUA_5_3 - # These result in internal compiler error - depends on !BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_SH select BR2_PACKAGE_BOOST select BR2_PACKAGE_BOOST_DATE_TIME select BR2_PACKAGE_BOOST_SYSTEM
@@ -30,7 +28,6 @@ config BR2_PACKAGE_DOMOTICZ comment "domoticz needs lua >= 5.2 and a toolchain w/ C++, threads, wchar, dynamic library" depends on BR2_USE_MMU depends on BR2_TOOLCHAIN_HAS_SYNC_4 - depends on !BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_SH depends on !BR2_INSTALL_LIBSTDCPP || !BR2_TOOLCHAIN_HAS_THREADS || \ !BR2_USE_WCHAR || BR2_STATIC_LIBS || \ !(BR2_PACKAGE_LUA_5_2 || BR2_PACKAGE_LUA_5_3)
package/liquid-dsp/Config.in +0 -1
@@ -18,7 +18,6 @@ if BR2_PACKAGE_LIQUID_DSP config BR2_PACKAGE_LIQUID_DSP_FAST bool "optimize for speed over accuracy" - depends on !BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_SH help Optimize for speed over accuracy.
package/prboom/Config.in +0 -2
@@ -1,7 +1,5 @@ config BR2_PACKAGE_PRBOOM bool "prboom" - # These result in internal compiler error - depends on !BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_SH select BR2_PACKAGE_SDL select BR2_PACKAGE_SDL_MIXER select BR2_PACKAGE_SDL_NET
package/qt/Config.in +2 -5
@@ -444,11 +444,8 @@ config BR2_PACKAGE_QT_ARCH_SUPPORTS_WEBKIT default y if BR2_arc || BR2_arm || BR2_armeb || BR2_i386 || BR2_microblazeel || \ BR2_microblazebe || BR2_mips || BR2_mipsel || \ (BR2_mips64 || BR2_mips64el) && !BR2_MIPS_NABI32 || BR2_powerpc || \ - BR2_powerpc64 || BR2_powerpc64le || BR2_x86_64 - # The CodeSourcery SuperH toolchain fails to build Webkit, - # with an assertion failure in binutils. - default y if (BR2_sh4 || BR2_sh4eb || BR2_sh4a || BR2_sh4aeb) && \ - !BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_SH + BR2_powerpc64 || BR2_powerpc64le || BR2_x86_64 || \ + BR2_sh4 || BR2_sh4eb || BR2_sh4a || BR2_sh4aeb depends on BR2_TOOLCHAIN_HAS_SYNC_4 config BR2_PACKAGE_QT_WEBKIT
Signature · 1 line
2.11.0
PATCH 4/14 package: remove CodeSourcery sh dependencies
Baruch, All,
On 2017-06-16 06:32 +0300, Baruch Siach spake thusly:
Quoted 4 lines
The CodeSourcery sh toolchain has been removed. Drop negative dependencies on that toolchain.
Signed-off-by Baruch Siach <[email protected]>
Regards,
Yann E. MORIN.
Quoted 77 lines
--- package/domoticz/Config.in | 3 --- package/liquid-dsp/Config.in | 1 - package/prboom/Config.in | 2 -- package/qt/Config.in | 7 ++----- 4 files changed, 2 insertions(+), 11 deletions(-)package/domoticz/Config.in +0
-3@@ -7,8 +7,6 @@ config BR2_PACKAGE_DOMOTICZ depends on BR2_INSTALL_LIBSTDCPP depends on BR2_USE_WCHAR depends on BR2_PACKAGE_LUA_5_2 || BR2_PACKAGE_LUA_5_3 - # These result in internal compiler error - depends on !BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_SH select BR2_PACKAGE_BOOST select BR2_PACKAGE_BOOST_DATE_TIME select BR2_PACKAGE_BOOST_SYSTEM@@ -30,7 +28,6 @@ config BR2_PACKAGE_DOMOTICZ comment "domoticz needs lua >= 5.2 and a toolchain w/ C++, threads, wchar, dynamic library" depends on BR2_USE_MMU depends on BR2_TOOLCHAIN_HAS_SYNC_4 - depends on !BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_SH depends on !BR2_INSTALL_LIBSTDCPP || !BR2_TOOLCHAIN_HAS_THREADS || \ !BR2_USE_WCHAR || BR2_STATIC_LIBS || \ !(BR2_PACKAGE_LUA_5_2 || BR2_PACKAGE_LUA_5_3)package/liquid-dsp/Config.in +0
-1@@ -18,7 +18,6 @@ if BR2_PACKAGE_LIQUID_DSP config BR2_PACKAGE_LIQUID_DSP_FAST bool "optimize for speed over accuracy" - depends on !BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_SH help Optimize for speed over accuracy.package/prboom/Config.in +0
-2@@ -1,7 +1,5 @@ config BR2_PACKAGE_PRBOOM bool "prboom" - # These result in internal compiler error - depends on !BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_SH select BR2_PACKAGE_SDL select BR2_PACKAGE_SDL_MIXER select BR2_PACKAGE_SDL_NETpackage/qt/Config.in +2
-5@@ -444,11 +444,8 @@ config BR2_PACKAGE_QT_ARCH_SUPPORTS_WEBKIT default y if BR2_arc || BR2_arm || BR2_armeb || BR2_i386 || BR2_microblazeel || \ BR2_microblazebe || BR2_mips || BR2_mipsel || \ (BR2_mips64 || BR2_mips64el) && !BR2_MIPS_NABI32 || BR2_powerpc || \ - BR2_powerpc64 || BR2_powerpc64le || BR2_x86_64 - # The CodeSourcery SuperH toolchain fails to build Webkit, - # with an assertion failure in binutils. - default y if (BR2_sh4 || BR2_sh4eb || BR2_sh4a || BR2_sh4aeb) && \ - !BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_SH + BR2_powerpc64 || BR2_powerpc64le || BR2_x86_64 || \ + BR2_sh4 || BR2_sh4eb || BR2_sh4a || BR2_sh4aeb depends on BR2_TOOLCHAIN_HAS_SYNC_4 config BR2_PACKAGE_QT_WEBKITSignature · 6 lines
2.11.0 _______________________________________________ buildroot mailing list buildroot at busybox.net http://lists.busybox.net/mailman/listinfo/buildroot
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. | '------------------------------^-------^------------------^--------------------'
PATCH 5/14 toolchain-external: update list of toolchains
In reply to Baruch SiachRemove mention of toolchains the we don't have.
--- toolchain/toolchain-external/pkg-toolchain-external.mk | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)
toolchain/toolchain-external/pkg-toolchain-external.mk +1 -2
@@ -13,10 +13,9 @@ # * Toolchains provided by Linaro for the ARM and AArch64 # architectures # * Sourcery CodeBench toolchains (from Mentor Graphics) for the ARM, -# MIPS, PowerPC, x86, x86_64 and NIOS 2 architectures. For the MIPS +# MIPS, PowerPC, x86_64 and NIOS 2 architectures. For the MIPS # toolchain, the -muclibc variant isn't supported yet, only the # default glibc-based variant is. -# * Xilinx toolchains for the Microblaze architecture # * Synopsys DesignWare toolchains for ARC cores # # The basic principle is the following
Signature · 1 line
2.11.0
PATCH 5/14 toolchain-external: update list of toolchains
Baruch, All,
On 2017-06-16 06:32 +0300, Baruch Siach spake thusly:
Quoted 3 lines
Remove mention of toolchains the we don't have.
Signed-off-by Baruch Siach <[email protected]>
Regards,
Yann E. MORIN.
Quoted 27 lines
--- toolchain/toolchain-external/pkg-toolchain-external.mk | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)toolchain/toolchain-external/pkg-toolchain-external.mk +1
-2@@ -13,10 +13,9 @@ # * Toolchains provided by Linaro for the ARM and AArch64 # architectures # * Sourcery CodeBench toolchains (from Mentor Graphics) for the ARM, -# MIPS, PowerPC, x86, x86_64 and NIOS 2 architectures. For the MIPS +# MIPS, PowerPC, x86_64 and NIOS 2 architectures. For the MIPS # toolchain, the -muclibc variant isn't supported yet, only the # default glibc-based variant is. -# * Xilinx toolchains for the Microblaze architecture # * Synopsys DesignWare toolchains for ARC cores # # The basic principle is the followingSignature · 6 lines
2.11.0 _______________________________________________ buildroot mailing list buildroot at busybox.net http://lists.busybox.net/mailman/listinfo/buildroot
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. | '------------------------------^-------^------------------^--------------------'
PATCH 6/14 docs/manual: drop mention of removed external toolchains
In reply to Baruch SiachThe CodeSourcery x86 and sh, the ADI Blackfin, and the Xilinx Mircoblaze external toolchain profiles have all been removed. Update the manual.
--- docs/manual/configure.txt | 14 ++++++-------- 1 file changed, 6 insertions(+), 8 deletions(-)
docs/manual/configure.txt +6 -8
@@ -146,19 +146,17 @@ cross-compilation toolchains. Buildroot knows about a number of well-known cross-compilation toolchains (from http://www.linaro.org[Linaro] for ARM, http://www.mentor.com/embedded-software/sourcery-tools/sourcery-codebench/editions/lite-edition/[Sourcery -CodeBench] for ARM, x86, x86-64, PowerPC, MIPS and SuperH, -https://blackfin.uclinux.org/gf/project/toolchain[Blackfin toolchains -from Analog Devices], etc.) and is capable of downloading them -automatically, or it can be pointed to a custom toolchain, either -available for download or installed locally. +CodeBench] for ARM, x86-64, PowerPC, and MIPS, and is capable of +downloading them automatically, or it can be pointed to a custom +toolchain, either available for download or installed locally. Then, you have three solutions to use an external toolchain: * Use a predefined external toolchain profile, and let Buildroot download, extract and install the toolchain. Buildroot already knows - about a few CodeSourcery, Linaro, Blackfin and Xilinx toolchains. - Just select the toolchain profile in +Toolchain+ from the - available ones. This is definitely the easiest solution. + about a few CodeSourcery and Linaro toolchains. Just select the + toolchain profile in +Toolchain+ from the available ones. This is + definitely the easiest solution. * Use a predefined external toolchain profile, but instead of having Buildroot download and extract the toolchain, you can tell Buildroot
Signature · 1 line
2.11.0
PATCH 6/14 docs/manual: drop mention of removed external toolchains
Baruch, All,
On 2017-06-16 06:32 +0300, Baruch Siach spake thusly:
Quoted 4 lines
The CodeSourcery x86 and sh, the ADI Blackfin, and the Xilinx Mircoblaze external toolchain profiles have all been removed. Update the manual.
Signed-off-by Baruch Siach <[email protected]>
Regards,
Yann E. MORIN.
Quoted 41 lines
--- docs/manual/configure.txt | 14 ++++++-------- 1 file changed, 6 insertions(+), 8 deletions(-)docs/manual/configure.txt +6
-8@@ -146,19 +146,17 @@ cross-compilation toolchains. Buildroot knows about a number of well-known cross-compilation toolchains (from http://www.linaro.org[Linaro] for ARM, http://www.mentor.com/embedded-software/sourcery-tools/sourcery-codebench/editions/lite-edition/[Sourcery -CodeBench] for ARM, x86, x86-64, PowerPC, MIPS and SuperH, -https://blackfin.uclinux.org/gf/project/toolchain[Blackfin toolchains -from Analog Devices], etc.) and is capable of downloading them -automatically, or it can be pointed to a custom toolchain, either -available for download or installed locally. +CodeBench] for ARM, x86-64, PowerPC, and MIPS, and is capable of +downloading them automatically, or it can be pointed to a custom +toolchain, either available for download or installed locally. Then, you have three solutions to use an external toolchain: * Use a predefined external toolchain profile, and let Buildroot download, extract and install the toolchain. Buildroot already knows - about a few CodeSourcery, Linaro, Blackfin and Xilinx toolchains. - Just select the toolchain profile in +Toolchain+ from the - available ones. This is definitely the easiest solution. + about a few CodeSourcery and Linaro toolchains. Just select the + toolchain profile in +Toolchain+ from the available ones. This is + definitely the easiest solution. * Use a predefined external toolchain profile, but instead of having Buildroot download and extract the toolchain, you can tell BuildrootSignature · 6 lines
2.11.0 _______________________________________________ buildroot mailing list buildroot at busybox.net http://lists.busybox.net/mailman/listinfo/buildroot
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. | '------------------------------^-------^------------------^--------------------'
PATCH 7/14 swupdate: don't link with librt
In reply to Baruch SiachBuildroot no longer supports toolchains with glibc older than 2.17, so there is no need to link with librt for clock_* system calls.
--- package/swupdate/swupdate.config | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
package/swupdate/swupdate.config +1 -1
@@ -26,7 +26,7 @@ CONFIG_CROSS_COMPILE="" CONFIG_SYSROOT="" CONFIG_EXTRA_CFLAGS="" CONFIG_EXTRA_LDFLAGS="" -CONFIG_EXTRA_LDLIBS="rt" +CONFIG_EXTRA_LDLIBS="" # # Debugging Options
Signature · 1 line
2.11.0
PATCH 7/14 swupdate: don't link with librt
Baruch, All,
On 2017-06-16 06:32 +0300, Baruch Siach spake thusly:
Quoted 5 lines
Buildroot no longer supports toolchains with glibc older than 2.17, so there is no need to link with librt for clock_* system calls.
Cc J?rg Krause <[email protected]>Signed-off-by Baruch Siach <[email protected]>
Regards,
Yann E. MORIN.
Quoted 24 lines
--- package/swupdate/swupdate.config | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)package/swupdate/swupdate.config +1
-1@@ -26,7 +26,7 @@ CONFIG_CROSS_COMPILE="" CONFIG_SYSROOT="" CONFIG_EXTRA_CFLAGS="" CONFIG_EXTRA_LDFLAGS="" -CONFIG_EXTRA_LDLIBS="rt" +CONFIG_EXTRA_LDLIBS="" # # Debugging OptionsSignature · 6 lines
2.11.0 _______________________________________________ buildroot mailing list buildroot at busybox.net http://lists.busybox.net/mailman/listinfo/buildroot
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. | '------------------------------^-------^------------------^--------------------'
PATCH 8/14 aiccu: don't link with librt
In reply to Baruch SiachBuildroot no longer supports toolchains with glibc older than 2.17, so there is no need to link with librt for clock_* system calls.
---
package/aiccu/0003-clock-suite-functions-fix.patch | 18 ------------------
...f-ether-header.patch => 0003-if-ether-header.patch} | 0
2 files changed, 18 deletions(-)
delete mode 100644 package/aiccu/0003-clock-suite-functions-fix.patch
rename package/aiccu/{0004-if-ether-header.patch => 0003-if-ether-header.patch} (100%)package/aiccu/0003-clock-suite-functions-fix.patch deleted +0 -18
@@ -1,18 +0,0 @@ -aiccu: fix undefined reference to clock_* functions - -Since glibc 2.17 the clock_* suite functions are available directly in -the main C library. For previous versions we still need -lrt to link. - -Signed-off-by: Vicente Olivert Riera <[email protected]> - ---- aiccu/unix-console/Makefile.orig 2014-01-14 10:17:20.204602513 +0000 -+++ aiccu/unix-console/Makefile 2014-01-14 10:17:31.435777359 +0000 -@@ -42,7 +42,7 @@ CFLAGS += -D AICCU_CONSOLE - # Currently defaultly builds only on Linux, but other platforms might easily also support it - ifeq ($(shell uname | grep -c "Linux"),1) - CFLAGS += -D AICCU_GNUTLS --LDFLAGS += -lgnutls -+LDFLAGS += -lgnutls -lrt - endif - - # Linux
package/aiccu/0003-if-ether-header.patch renamed from package/aiccu/0004-if-ether-header.patch +0 -0
Signature · 1 line
2.11.0
PATCH 8/14 aiccu: don't link with librt
Baruch, All,
On 2017-06-16 06:32 +0300, Baruch Siach spake thusly:
Quoted 5 lines
Buildroot no longer supports toolchains with glibc older than 2.17, so there is no need to link with librt for clock_* system calls.
Cc Michael Rommel <[email protected]>Signed-off-by Baruch Siach <[email protected]>
Regards,
Yann E. MORIN.
Quoted 42 lines
--- package/aiccu/0003-clock-suite-functions-fix.patch | 18 ------------------ ...f-ether-header.patch => 0003-if-ether-header.patch} | 0 2 files changed, 18 deletions(-) delete mode 100644 package/aiccu/0003-clock-suite-functions-fix.patch rename package/aiccu/{0004-if-ether-header.patch => 0003-if-ether-header.patch} (100%)package/aiccu/0003-clock-suite-functions-fix.patch deleted +0
-18@@ -1,18 +0,0 @@ -aiccu: fix undefined reference to clock_* functions - -Since glibc 2.17 the clock_* suite functions are available directly in -the main C library. For previous versions we still need -lrt to link. - -Signed-off-by: Vicente Olivert Riera <[email protected]> - ---- aiccu/unix-console/Makefile.orig 2014-01-14 10:17:20.204602513 +0000 -+++ aiccu/unix-console/Makefile 2014-01-14 10:17:31.435777359 +0000 -@@ -42,7 +42,7 @@ CFLAGS += -D AICCU_CONSOLE - # Currently defaultly builds only on Linux, but other platforms might easily also support it - ifeq ($(shell uname | grep -c "Linux"),1) - CFLAGS += -D AICCU_GNUTLS --LDFLAGS += -lgnutls -+LDFLAGS += -lgnutls -lrt - endif - - # Linuxpackage/aiccu/0003-if-ether-header.patch renamed from package/aiccu/0004-if-ether-header.patch +0
-0Signature · 6 lines
2.11.0 _______________________________________________ buildroot mailing list buildroot at busybox.net http://lists.busybox.net/mailman/listinfo/buildroot
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. | '------------------------------^-------^------------------^--------------------'
PATCH 9/14 libv4l: don't link with librt
In reply to Baruch SiachBuildroot no longer supports toolchains with glibc older than 2.17, so there is no need to link with librt for clock_* system calls.
The following patches are not renumbered. The noise is too high.
--- ...3-utils-qv4l2-Makefile.am-link-with-librt.patch | 29 ---------------------- package/libv4l/libv4l.mk | 3 +-- 2 files changed, 1 insertion(+), 31 deletions(-) delete mode 100644 package/libv4l/0003-utils-qv4l2-Makefile.am-link-with-librt.patch
package/libv4l/0003-utils-qv4l2-Makefile.am-link-with-librt.patch deleted +0 -29
@@ -1,29 +0,0 @@ -From 66d1d752976f4b45a5faff2bc68fc2fa9d435c71 Mon Sep 17 00:00:00 2001 -From: Thomas Petazzoni <[email protected]> -Date: Sun, 14 May 2017 22:04:55 +0200 -Subject: [PATCH] utils/qv4l2/Makefile.am: link with librt - -qv4l2 uses clock_gettime(), so it should link with librt to be -compatible with old versions of glibc. - -Signed-off-by: Thomas Petazzoni <[email protected]> ---- - utils/qv4l2/Makefile.am | 2 +- - 1 file changed, 1 insertion(+), 1 deletion(-) - -diff --git a/utils/qv4l2/Makefile.am b/utils/qv4l2/Makefile.am -index fd58486..a991d66 100644 ---- a/utils/qv4l2/Makefile.am -+++ b/utils/qv4l2/Makefile.am -@@ -7,7 +7,7 @@ qv4l2_SOURCES = qv4l2.cpp general-tab.cpp ctrl-tab.cpp vbi-tab.cpp capture-win.c - v4l2-tpg-core.c v4l2-tpg-colors.c - nodist_qv4l2_SOURCES = moc_qv4l2.cpp moc_general-tab.cpp moc_capture-win.cpp moc_vbi-tab.cpp qrc_qv4l2.cpp - qv4l2_LDADD = ../../lib/libv4l2/libv4l2.la ../../lib/libv4lconvert/libv4lconvert.la \ -- ../libv4l2util/libv4l2util.la ../libmedia_dev/libmedia_dev.la -+ ../libv4l2util/libv4l2util.la ../libmedia_dev/libmedia_dev.la -lrt - qv4l2_CPPFLAGS = -I../common - - if WITH_QTGL --- -2.7.4 -
package/libv4l/libv4l.mk +1 -2
@@ -11,8 +11,6 @@ LIBV4L_INSTALL_STAGING = YES LIBV4L_DEPENDENCIES = host-pkgconf LIBV4L_CONF_OPTS = --disable-doxygen-doc -# 0003-utils-qv4l2-Makefile.am-link-with-librt.patch, host-gettext -# needed for autoreconf to work # below patches requires autoreconf: # 0004-configure.ac-clarify-configure-summary.patch # 0005-configure.ac-revisit-v4l2-ctl-compliance-using-libv4.patch
@@ -21,6 +19,7 @@ LIBV4L_CONF_OPTS = --disable-doxygen-doc # 0008-configure.ac-fix-build-of-v4l-utils-on-uclinux.patch # 0009-configure.ac-add-USE_LIBV4L-to-summary.patch LIBV4L_AUTORECONF = YES +# host-gettext needed for autoreconf to work LIBV4L_DEPENDENCIES += host-gettext # fix uclibc-ng configure/compile
Signature · 1 line
2.11.0
PATCH 9/14 libv4l: don't link with librt
Baruch, All,
On 2017-06-16 06:32 +0300, Baruch Siach spake thusly:
Quoted 4 lines
Buildroot no longer supports toolchains with glibc older than 2.17, so there is no need to link with librt for clock_* system calls.
The following patches are not renumbered. The noise is too high.
If you use -M and/or -C when doing the patches, then git will detect renames and/or copies, and will not actually send the diff, just the rename.
Quoted 1 line
Signed-off-by Baruch Siach <[email protected]>
Regards,
Yann E. MORIN.
Quoted 69 lines
--- ...3-utils-qv4l2-Makefile.am-link-with-librt.patch | 29 ---------------------- package/libv4l/libv4l.mk | 3 +-- 2 files changed, 1 insertion(+), 31 deletions(-) delete mode 100644 package/libv4l/0003-utils-qv4l2-Makefile.am-link-with-librt.patchpackage/libv4l/0003-utils-qv4l2-Makefile.am-link-with-librt.patch deleted +0
-29@@ -1,29 +0,0 @@ -From 66d1d752976f4b45a5faff2bc68fc2fa9d435c71 Mon Sep 17 00:00:00 2001 -From: Thomas Petazzoni <[email protected]> -Date: Sun, 14 May 2017 22:04:55 +0200 -Subject: [PATCH] utils/qv4l2/Makefile.am: link with librt - -qv4l2 uses clock_gettime(), so it should link with librt to be -compatible with old versions of glibc. - -Signed-off-by: Thomas Petazzoni <[email protected]> ---- - utils/qv4l2/Makefile.am | 2 +- - 1 file changed, 1 insertion(+), 1 deletion(-) - -diff --git a/utils/qv4l2/Makefile.am b/utils/qv4l2/Makefile.am -index fd58486..a991d66 100644 ---- a/utils/qv4l2/Makefile.am -+++ b/utils/qv4l2/Makefile.am -@@ -7,7 +7,7 @@ qv4l2_SOURCES = qv4l2.cpp general-tab.cpp ctrl-tab.cpp vbi-tab.cpp capture-win.c - v4l2-tpg-core.c v4l2-tpg-colors.c - nodist_qv4l2_SOURCES = moc_qv4l2.cpp moc_general-tab.cpp moc_capture-win.cpp moc_vbi-tab.cpp qrc_qv4l2.cpp - qv4l2_LDADD = ../../lib/libv4l2/libv4l2.la ../../lib/libv4lconvert/libv4lconvert.la \ -- ../libv4l2util/libv4l2util.la ../libmedia_dev/libmedia_dev.la -+ ../libv4l2util/libv4l2util.la ../libmedia_dev/libmedia_dev.la -lrt - qv4l2_CPPFLAGS = -I../common - - if WITH_QTGL --- -2.7.4 -package/libv4l/libv4l.mk +1
-2@@ -11,8 +11,6 @@ LIBV4L_INSTALL_STAGING = YES LIBV4L_DEPENDENCIES = host-pkgconf LIBV4L_CONF_OPTS = --disable-doxygen-doc -# 0003-utils-qv4l2-Makefile.am-link-with-librt.patch, host-gettext -# needed for autoreconf to work # below patches requires autoreconf: # 0004-configure.ac-clarify-configure-summary.patch # 0005-configure.ac-revisit-v4l2-ctl-compliance-using-libv4.patch@@ -21,6 +19,7 @@ LIBV4L_CONF_OPTS = --disable-doxygen-doc # 0008-configure.ac-fix-build-of-v4l-utils-on-uclinux.patch # 0009-configure.ac-add-USE_LIBV4L-to-summary.patch LIBV4L_AUTORECONF = YES +# host-gettext needed for autoreconf to work LIBV4L_DEPENDENCIES += host-gettext # fix uclibc-ng configure/compileSignature · 6 lines
2.11.0 _______________________________________________ buildroot mailing list buildroot at busybox.net http://lists.busybox.net/mailman/listinfo/buildroot
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. | '------------------------------^-------^------------------^--------------------'
PATCH 9/14 libv4l: don't link with librt
Hi Yann,
Quoted 11 lines · Yann E. MORIN
Baruch, All,
On 2017-06-16 06:32 +0300, Baruch Siach spake thusly:
Quoted 4 lines
Buildroot no longer supports toolchains with glibc older than 2.17, so there is no need to link with librt for clock_* system calls.
The following patches are not renumbered. The noise is too high.
If you use -M and/or -C when doing the patches, then git will detect renames and/or copies, and will not actually send the diff, just the rename.
The noise is mostly in the libv4l.mk comment. It would hide the actual significant change there. Should I do it anyway?
baruch
Quoted 71 lines
Quoted 1 line
Signed-off-by Baruch Siach <[email protected]>Reviewed-by "Yann E. MORIN" <[email protected]>Regards,
Yann E. MORIN.Quoted 64 lines
--- ...3-utils-qv4l2-Makefile.am-link-with-librt.patch | 29 ---------------------- package/libv4l/libv4l.mk | 3 +-- 2 files changed, 1 insertion(+), 31 deletions(-) delete mode 100644 package/libv4l/0003-utils-qv4l2-Makefile.am-link-with-librt.patchpackage/libv4l/0003-utils-qv4l2-Makefile.am-link-with-librt.patch deleted +0
-29@@ -1,29 +0,0 @@ -From 66d1d752976f4b45a5faff2bc68fc2fa9d435c71 Mon Sep 17 00:00:00 2001 -From: Thomas Petazzoni <[email protected]> -Date: Sun, 14 May 2017 22:04:55 +0200 -Subject: [PATCH] utils/qv4l2/Makefile.am: link with librt - -qv4l2 uses clock_gettime(), so it should link with librt to be -compatible with old versions of glibc. - -Signed-off-by: Thomas Petazzoni <[email protected]> ---- - utils/qv4l2/Makefile.am | 2 +- - 1 file changed, 1 insertion(+), 1 deletion(-) - -diff --git a/utils/qv4l2/Makefile.am b/utils/qv4l2/Makefile.am -index fd58486..a991d66 100644 ---- a/utils/qv4l2/Makefile.am -+++ b/utils/qv4l2/Makefile.am -@@ -7,7 +7,7 @@ qv4l2_SOURCES = qv4l2.cpp general-tab.cpp ctrl-tab.cpp vbi-tab.cpp capture-win.c - v4l2-tpg-core.c v4l2-tpg-colors.c - nodist_qv4l2_SOURCES = moc_qv4l2.cpp moc_general-tab.cpp moc_capture-win.cpp moc_vbi-tab.cpp qrc_qv4l2.cpp - qv4l2_LDADD = ../../lib/libv4l2/libv4l2.la ../../lib/libv4lconvert/libv4lconvert.la \ -- ../libv4l2util/libv4l2util.la ../libmedia_dev/libmedia_dev.la -+ ../libv4l2util/libv4l2util.la ../libmedia_dev/libmedia_dev.la -lrt - qv4l2_CPPFLAGS = -I../common - - if WITH_QTGL --- -2.7.4 -package/libv4l/libv4l.mk +1
-2@@ -11,8 +11,6 @@ LIBV4L_INSTALL_STAGING = YES LIBV4L_DEPENDENCIES = host-pkgconf LIBV4L_CONF_OPTS = --disable-doxygen-doc -# 0003-utils-qv4l2-Makefile.am-link-with-librt.patch, host-gettext -# needed for autoreconf to work # below patches requires autoreconf: # 0004-configure.ac-clarify-configure-summary.patch # 0005-configure.ac-revisit-v4l2-ctl-compliance-using-libv4.patch@@ -21,6 +19,7 @@ LIBV4L_CONF_OPTS = --disable-doxygen-doc # 0008-configure.ac-fix-build-of-v4l-utils-on-uclinux.patch # 0009-configure.ac-add-USE_LIBV4L-to-summary.patch LIBV4L_AUTORECONF = YES +# host-gettext needed for autoreconf to work LIBV4L_DEPENDENCIES += host-gettext # fix uclibc-ng configure/compileSignature · 1 line
2.11.0
Signature · 3 lines
http://baruch.siach.name/blog/ ~. .~ Tk Open Systems
=}------------------------------------------------ooO--U--Ooo------------{=
- baruch at tkos.co.il - tel: +972.52.368.4656, http://www.tkos.co.il -PATCH 9/14 libv4l: don't link with librt
Baruch, All,
On 2017-06-18 12:34 +0300, Baruch Siach spake thusly:
Quoted 7 lines · Yann E. MORIN
Quoted 6 lines
Baruch, All,
On 2017-06-16 06:32 +0300, Baruch Siach spake thusly:
Quoted 3 lines
Buildroot no longer supports toolchains with glibc older than 2.17, so there is no need to link with librt for clock_* system calls.
Quoted 8 lines
Quoted 5 lines
Quoted 1 line
The following patches are not renumbered. The noise is too high.
If you use -M and/or -C when doing the patches, then git will detect renames and/or copies, and will not actually send the diff, just the rename.
The noise is mostly in the libv4l.mk comment. It would hide the actual significant change there.
OK, I see what you meant, now.
Quoted 1 line
Should I do it anyway?
I won't ask for it! ;-)
Regards,
Yann E. MORIN.
Quoted 78 lines
baruch
Quoted 71 lines
Quoted 1 line
Signed-off-by Baruch Siach <[email protected]>Reviewed-by "Yann E. MORIN" <[email protected]>Regards,
Yann E. MORIN.Quoted 64 lines
--- ...3-utils-qv4l2-Makefile.am-link-with-librt.patch | 29 ---------------------- package/libv4l/libv4l.mk | 3 +-- 2 files changed, 1 insertion(+), 31 deletions(-) delete mode 100644 package/libv4l/0003-utils-qv4l2-Makefile.am-link-with-librt.patchpackage/libv4l/0003-utils-qv4l2-Makefile.am-link-with-librt.patch deleted +0
-29@@ -1,29 +0,0 @@ -From 66d1d752976f4b45a5faff2bc68fc2fa9d435c71 Mon Sep 17 00:00:00 2001 -From: Thomas Petazzoni <[email protected]> -Date: Sun, 14 May 2017 22:04:55 +0200 -Subject: [PATCH] utils/qv4l2/Makefile.am: link with librt - -qv4l2 uses clock_gettime(), so it should link with librt to be -compatible with old versions of glibc. - -Signed-off-by: Thomas Petazzoni <[email protected]> ---- - utils/qv4l2/Makefile.am | 2 +- - 1 file changed, 1 insertion(+), 1 deletion(-) - -diff --git a/utils/qv4l2/Makefile.am b/utils/qv4l2/Makefile.am -index fd58486..a991d66 100644 ---- a/utils/qv4l2/Makefile.am -+++ b/utils/qv4l2/Makefile.am -@@ -7,7 +7,7 @@ qv4l2_SOURCES = qv4l2.cpp general-tab.cpp ctrl-tab.cpp vbi-tab.cpp capture-win.c - v4l2-tpg-core.c v4l2-tpg-colors.c - nodist_qv4l2_SOURCES = moc_qv4l2.cpp moc_general-tab.cpp moc_capture-win.cpp moc_vbi-tab.cpp qrc_qv4l2.cpp - qv4l2_LDADD = ../../lib/libv4l2/libv4l2.la ../../lib/libv4lconvert/libv4lconvert.la \ -- ../libv4l2util/libv4l2util.la ../libmedia_dev/libmedia_dev.la -+ ../libv4l2util/libv4l2util.la ../libmedia_dev/libmedia_dev.la -lrt - qv4l2_CPPFLAGS = -I../common - - if WITH_QTGL --- -2.7.4 -package/libv4l/libv4l.mk +1
-2@@ -11,8 +11,6 @@ LIBV4L_INSTALL_STAGING = YES LIBV4L_DEPENDENCIES = host-pkgconf LIBV4L_CONF_OPTS = --disable-doxygen-doc -# 0003-utils-qv4l2-Makefile.am-link-with-librt.patch, host-gettext -# needed for autoreconf to work # below patches requires autoreconf: # 0004-configure.ac-clarify-configure-summary.patch # 0005-configure.ac-revisit-v4l2-ctl-compliance-using-libv4.patch@@ -21,6 +19,7 @@ LIBV4L_CONF_OPTS = --disable-doxygen-doc # 0008-configure.ac-fix-build-of-v4l-utils-on-uclinux.patch # 0009-configure.ac-add-USE_LIBV4L-to-summary.patch LIBV4L_AUTORECONF = YES +# host-gettext needed for autoreconf to work LIBV4L_DEPENDENCIES += host-gettext # fix uclibc-ng configure/compileSignature · 1 line
2.11.0Signature · 3 lines
http://baruch.siach.name/blog/ ~. .~ Tk Open Systems =}------------------------------------------------ooO--U--Ooo------------{= - baruch at tkos.co.il - tel: +972.52.368.4656, http://www.tkos.co.il -
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. | '------------------------------^-------^------------------^--------------------'
PATCH 10/14 e2fsprogs: don't link with librt
In reply to Baruch SiachBuildroot no longer supports toolchains with glibc older than 2.17, so there is no need to link with librt for clock_* system calls.
--- .../0002-MCONFIG.in-handle-CLOCK_GETTIME_LIB.patch | 35 ---------------------- 1 file changed, 35 deletions(-) delete mode 100644 package/e2fsprogs/0002-MCONFIG.in-handle-CLOCK_GETTIME_LIB.patch
package/e2fsprogs/0002-MCONFIG.in-handle-CLOCK_GETTIME_LIB.patch deleted +0 -35
@@ -1,35 +0,0 @@ -From b5baf15b36107925186d89af9097b4b8e2d7cad6 Mon Sep 17 00:00:00 2001 -From: Thomas Petazzoni <[email protected]> -Date: Wed, 21 Sep 2016 08:38:21 +0200 -Subject: [PATCH] MCONFIG.in: handle CLOCK_GETTIME_LIB - -Commit 62bc243f1bfe623312a350e4cd0d2e17a69aac45 ("fuse2fs: might need -librt") added a check for the clock_gettime() to verify whether linking -with the librt library is needed or not to use this function. - -However, this commit forgot the change in MCONFIG.in that allows to -define the make variable CLOCK_GETTIME_LIB from the CLOCK_GETTIME_LIB -variable substituted by the configure script. Due to this, from a make -perspective $(CLOCK_GETTIME_LIB) is always empty, and we get link -failure on platforms where clock_gettime() is available in librt. - -Signed-off-by: Thomas Petazzoni <[email protected]> ---- - MCONFIG.in | 1 + - 1 file changed, 1 insertion(+) - -diff --git a/MCONFIG.in b/MCONFIG.in -index e82963d..9ba9ce2 100644 ---- a/MCONFIG.in -+++ b/MCONFIG.in -@@ -125,6 +125,7 @@ LIBSUPPORT = $(LIBINTL) $(LIB)/libsupport at STATIC_LIB_EXT@ - LIBBLKID = @LIBBLKID@ @PRIVATE_LIBS_CMT@ $(LIBUUID) - LIBINTL = @LIBINTL@ - SYSLIBS = @LIBS@ -+CLOCK_GETTIME_LIB = @CLOCK_GETTIME_LIB@ - DEPLIBSS = $(LIB)/libss at LIB_EXT@ - DEPLIBCOM_ERR = $(LIB)/libcom_err at LIB_EXT@ - DEPLIBUUID = @DEPLIBUUID@ --- -2.7.4 -
Signature · 1 line
2.11.0
PATCH 10/14 e2fsprogs: don't link with librt
Baruch, All,
On 2017-06-16 06:32 +0300, Baruch Siach spake thusly:
Quoted 4 lines
Buildroot no longer supports toolchains with glibc older than 2.17, so there is no need to link with librt for clock_* system calls.
Signed-off-by Baruch Siach <[email protected]>
Regards,
Yann E. MORIN.
Quoted 53 lines
--- .../0002-MCONFIG.in-handle-CLOCK_GETTIME_LIB.patch | 35 ---------------------- 1 file changed, 35 deletions(-) delete mode 100644 package/e2fsprogs/0002-MCONFIG.in-handle-CLOCK_GETTIME_LIB.patchpackage/e2fsprogs/0002-MCONFIG.in-handle-CLOCK_GETTIME_LIB.patch deleted +0
-35@@ -1,35 +0,0 @@ -From b5baf15b36107925186d89af9097b4b8e2d7cad6 Mon Sep 17 00:00:00 2001 -From: Thomas Petazzoni <[email protected]> -Date: Wed, 21 Sep 2016 08:38:21 +0200 -Subject: [PATCH] MCONFIG.in: handle CLOCK_GETTIME_LIB - -Commit 62bc243f1bfe623312a350e4cd0d2e17a69aac45 ("fuse2fs: might need -librt") added a check for the clock_gettime() to verify whether linking -with the librt library is needed or not to use this function. - -However, this commit forgot the change in MCONFIG.in that allows to -define the make variable CLOCK_GETTIME_LIB from the CLOCK_GETTIME_LIB -variable substituted by the configure script. Due to this, from a make -perspective $(CLOCK_GETTIME_LIB) is always empty, and we get link -failure on platforms where clock_gettime() is available in librt. - -Signed-off-by: Thomas Petazzoni <[email protected]> ---- - MCONFIG.in | 1 + - 1 file changed, 1 insertion(+) - -diff --git a/MCONFIG.in b/MCONFIG.in -index e82963d..9ba9ce2 100644 ---- a/MCONFIG.in -+++ b/MCONFIG.in -@@ -125,6 +125,7 @@ LIBSUPPORT = $(LIBINTL) $(LIB)/libsupport at STATIC_LIB_EXT@ - LIBBLKID = @LIBBLKID@ @PRIVATE_LIBS_CMT@ $(LIBUUID) - LIBINTL = @LIBINTL@ - SYSLIBS = @LIBS@ -+CLOCK_GETTIME_LIB = @CLOCK_GETTIME_LIB@ - DEPLIBSS = $(LIB)/libss at LIB_EXT@ - DEPLIBCOM_ERR = $(LIB)/libcom_err at LIB_EXT@ - DEPLIBUUID = @DEPLIBUUID@ --- -2.7.4 -Signature · 6 lines
2.11.0 _______________________________________________ buildroot mailing list buildroot at busybox.net http://lists.busybox.net/mailman/listinfo/buildroot
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. | '------------------------------^-------^------------------^--------------------'
PATCH 10/14 e2fsprogs: don't link with librt
Hi Yann,
Quoted 7 lines · Yann E. MORIN
On 2017-06-16 06:32 +0300, Baruch Siach spake thusly:
Quoted 4 lines
Buildroot no longer supports toolchains with glibc older than 2.17, so there is no need to link with librt for clock_* system calls.
Signed-off-by Baruch Siach <[email protected]>Reviewed-by "Yann E. MORIN" <[email protected]>
Thanks.
On a second thought, e2fsprogs is also a host package. We don't want to abandon pre glibc 2.17 hosts just yet, do we?
baruch
Quoted 43 lines
Quoted 43 lines
--- .../0002-MCONFIG.in-handle-CLOCK_GETTIME_LIB.patch | 35 ---------------------- 1 file changed, 35 deletions(-) delete mode 100644 package/e2fsprogs/0002-MCONFIG.in-handle-CLOCK_GETTIME_LIB.patchpackage/e2fsprogs/0002-MCONFIG.in-handle-CLOCK_GETTIME_LIB.patch deleted +0
-32@@ -1,35 +0,0 @@ -From b5baf15b36107925186d89af9097b4b8e2d7cad6 Mon Sep 17 00:00:00 2001 -From: Thomas Petazzoni <[email protected]> -Date: Wed, 21 Sep 2016 08:38:21 +0200 -Subject: [PATCH] MCONFIG.in: handle CLOCK_GETTIME_LIB - -Commit 62bc243f1bfe623312a350e4cd0d2e17a69aac45 ("fuse2fs: might need -librt") added a check for the clock_gettime() to verify whether linking -with the librt library is needed or not to use this function. - -However, this commit forgot the change in MCONFIG.in that allows to -define the make variable CLOCK_GETTIME_LIB from the CLOCK_GETTIME_LIB -variable substituted by the configure script. Due to this, from a make -perspective $(CLOCK_GETTIME_LIB) is always empty, and we get link -failure on platforms where clock_gettime() is available in librt. - -Signed-off-by: Thomas Petazzoni <[email protected]> ---- - MCONFIG.in | 1 + - 1 file changed, 1 insertion(+) - -diff --git a/MCONFIG.in b/MCONFIG.in -index e82963d..9ba9ce2 100644 ---- a/MCONFIG.in -+++ b/MCONFIG.in -@@ -125,6 +125,7 @@ LIBSUPPORT = $(LIBINTL) $(LIB)/libsupport at STATIC_LIB_EXT@ - LIBBLKID = @LIBBLKID@ @PRIVATE_LIBS_CMT@ $(LIBUUID) - LIBINTL = @LIBINTL@ - SYSLIBS = @LIBS@ -+CLOCK_GETTIME_LIB = @CLOCK_GETTIME_LIB@ - DEPLIBSS = $(LIB)/libss at LIB_EXT@ - DEPLIBCOM_ERR = $(LIB)/libcom_err at LIB_EXT@ - DEPLIBUUID = @DEPLIBUUID@
Signature · 3 lines
http://baruch.siach.name/blog/ ~. .~ Tk Open Systems
=}------------------------------------------------ooO--U--Ooo------------{=
- baruch at tkos.co.il - tel: +972.52.368.4656, http://www.tkos.co.il -PATCH 10/14 e2fsprogs: don't link with librt
Baruch, All,
On 2017-06-18 13:13 +0300, Baruch Siach spake thusly:
Quoted 5 lines · Yann E. MORIN
Quoted 4 lines
On 2017-06-16 06:32 +0300, Baruch Siach spake thusly:
Quoted 3 lines
Buildroot no longer supports toolchains with glibc older than 2.17, so there is no need to link with librt for clock_* system calls.
Quoted 8 lines
Quoted 3 lines
Quoted 1 line
Signed-off-by Baruch Siach <[email protected]>Reviewed-by "Yann E. MORIN" <[email protected]>Thanks.
On a second thought, e2fsprogs is also a host package. We don't want to abandon pre glibc 2.17 hosts just yet, do we?
Indeed no.
But the patch was made to fix a build failure for fuse2fs, which we explicitly disable for the host variant.
So, we really can drop this patch.
Regards,
Yann E. MORIN.
Quoted 50 lines
baruch
Quoted 43 lines
Quoted 43 lines
--- .../0002-MCONFIG.in-handle-CLOCK_GETTIME_LIB.patch | 35 ---------------------- 1 file changed, 35 deletions(-) delete mode 100644 package/e2fsprogs/0002-MCONFIG.in-handle-CLOCK_GETTIME_LIB.patchpackage/e2fsprogs/0002-MCONFIG.in-handle-CLOCK_GETTIME_LIB.patch deleted +0
-32@@ -1,35 +0,0 @@ -From b5baf15b36107925186d89af9097b4b8e2d7cad6 Mon Sep 17 00:00:00 2001 -From: Thomas Petazzoni <[email protected]> -Date: Wed, 21 Sep 2016 08:38:21 +0200 -Subject: [PATCH] MCONFIG.in: handle CLOCK_GETTIME_LIB - -Commit 62bc243f1bfe623312a350e4cd0d2e17a69aac45 ("fuse2fs: might need -librt") added a check for the clock_gettime() to verify whether linking -with the librt library is needed or not to use this function. - -However, this commit forgot the change in MCONFIG.in that allows to -define the make variable CLOCK_GETTIME_LIB from the CLOCK_GETTIME_LIB -variable substituted by the configure script. Due to this, from a make -perspective $(CLOCK_GETTIME_LIB) is always empty, and we get link -failure on platforms where clock_gettime() is available in librt. - -Signed-off-by: Thomas Petazzoni <[email protected]> ---- - MCONFIG.in | 1 + - 1 file changed, 1 insertion(+) - -diff --git a/MCONFIG.in b/MCONFIG.in -index e82963d..9ba9ce2 100644 ---- a/MCONFIG.in -+++ b/MCONFIG.in -@@ -125,6 +125,7 @@ LIBSUPPORT = $(LIBINTL) $(LIB)/libsupport at STATIC_LIB_EXT@ - LIBBLKID = @LIBBLKID@ @PRIVATE_LIBS_CMT@ $(LIBUUID) - LIBINTL = @LIBINTL@ - SYSLIBS = @LIBS@ -+CLOCK_GETTIME_LIB = @CLOCK_GETTIME_LIB@ - DEPLIBSS = $(LIB)/libss at LIB_EXT@ - DEPLIBCOM_ERR = $(LIB)/libcom_err at LIB_EXT@ - DEPLIBUUID = @DEPLIBUUID@Signature · 3 lines
http://baruch.siach.name/blog/ ~. .~ Tk Open Systems =}------------------------------------------------ooO--U--Ooo------------{= - baruch at tkos.co.il - tel: +972.52.368.4656, http://www.tkos.co.il -
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. | '------------------------------^-------^------------------^--------------------'
PATCH 10/14 e2fsprogs: don't link with librt
Hi Yann,
Quoted 15 lines · Yann E. MORIN
On 2017-06-18 13:13 +0300, Baruch Siach spake thusly:
Quoted 5 lines · Yann E. MORIN
Quoted 4 lines
On 2017-06-16 06:32 +0300, Baruch Siach spake thusly:
Quoted 3 lines
Buildroot no longer supports toolchains with glibc older than 2.17, so there is no need to link with librt for clock_* system calls.
Quoted 8 lines
Quoted 3 lines
Quoted 1 line
Signed-off-by Baruch Siach <[email protected]>Reviewed-by "Yann E. MORIN" <[email protected]>Thanks.
On a second thought, e2fsprogs is also a host package. We don't want to abandon pre glibc 2.17 hosts just yet, do we?
Quoted 6 lines
Indeed no.
But the patch was made to fix a build failure for fuse2fs, which we explicitly disable for the host variant.
So, we really can drop this patch.
Right. However the softether librt patch is applicable[1] to the host version. I have marked that one (no. 14 in this series) as Rejected.
Thanks,
baruch
[1] http://autobuild.buildroot.net/results/48f/48f778a891e0bf6a60ed2e4bec057c338ec25ec9/
Signature · 3 lines
http://baruch.siach.name/blog/ ~. .~ Tk Open Systems
=}------------------------------------------------ooO--U--Ooo------------{=
- baruch at tkos.co.il - tel: +972.52.368.4656, http://www.tkos.co.il -PATCH 11/14 liboping: no need to check for clock_gettime
In reply to Baruch SiachBuildroot no longer supports toolchains with glibc older than 2.17, so there is no need to check whether librt is required for clock_* system calls.
--- ...001-configure-also-check-for-clockgettime.patch | 39 ---------------------- package/liboping/liboping.mk | 2 -- 2 files changed, 41 deletions(-) delete mode 100644 package/liboping/0001-configure-also-check-for-clockgettime.patch
package/liboping/0001-configure-also-check-for-clockgettime.patch deleted +0 -39
@@ -1,39 +0,0 @@ -From a45f3610beba1f7e6e1a038e3a1fe4150057c262 Mon Sep 17 00:00:00 2001 -From: "Yann E. MORIN" <[email protected]> -Date: Tue, 5 Jul 2016 15:31:28 +0200 -Subject: [PATCH] configure: also check for clockgettime() - -clock_gettime() is also in -lrt so we also need to check for it. - -Signed-off-by: "Yann E. MORIN" <[email protected]> ---- - configure.ac | 10 +++++++--- - 1 file changed, 7 insertions(+), 3 deletions(-) - -diff --git a/configure.ac b/configure.ac -index f83f5ab..3b6ae09 100644 ---- a/configure.ac -+++ b/configure.ac -@@ -199,12 +199,16 @@ fi - - AC_SUBST(LIBOPING_PC_LIBS_PRIVATE) - --nanosleep_needs_rt="no" -+needs_rt="no" - AC_CHECK_FUNCS(nanosleep, [], - AC_CHECK_LIB(rt, nanosleep, -- [nanosleep_needs_rt="yes"], -+ [needs_rt="yes"], - AC_MSG_ERROR(cannot find nanosleep))) --AM_CONDITIONAL(BUILD_WITH_LIBRT, test "x$nanosleep_needs_rt" = "xyes") -+AC_CHECK_FUNCS(clock_gettime, [], -+ AC_CHECK_LIB(rt, clock_gettime, -+ [needs_rt="yes"], -+ AC_MSG_ERROR(cannot find clock_gettime))) -+AM_CONDITIONAL(BUILD_WITH_LIBRT, test "x$needs_rt" = "xyes") - - with_ncurses="no" - AC_CHECK_HEADERS(ncursesw/ncurses.h ncurses.h, [with_ncurses="yes"], []) --- -2.7.4 -
package/liboping/liboping.mk +0 -2
@@ -13,6 +13,4 @@ LIBOPING_CONF_OPTS = --without-perl-bindings LIBOPING_LICENSE = LGPL-2.1+, GPL-2.0 LIBOPING_LICENSE_FILES = COPYING -LIBOPING_AUTORECONF = YES - $(eval $(autotools-package))
Signature · 1 line
2.11.0
PATCH 11/14 liboping: no need to check for clock_gettime
Baruch, All,
On 2017-06-16 06:32 +0300, Baruch Siach spake thusly:
Quoted 5 lines
Buildroot no longer supports toolchains with glibc older than 2.17, so there is no need to check whether librt is required for clock_* system calls.
Cc Gustavo Zacarias <[email protected]>Signed-off-by Baruch Siach <[email protected]>
Removing one of my patches? I'm all for it! :-)
Regards,
Yann E. MORIN.
Quoted 69 lines
--- ...001-configure-also-check-for-clockgettime.patch | 39 ---------------------- package/liboping/liboping.mk | 2 -- 2 files changed, 41 deletions(-) delete mode 100644 package/liboping/0001-configure-also-check-for-clockgettime.patchpackage/liboping/0001-configure-also-check-for-clockgettime.patch deleted +0
-39@@ -1,39 +0,0 @@ -From a45f3610beba1f7e6e1a038e3a1fe4150057c262 Mon Sep 17 00:00:00 2001 -From: "Yann E. MORIN" <[email protected]> -Date: Tue, 5 Jul 2016 15:31:28 +0200 -Subject: [PATCH] configure: also check for clockgettime() - -clock_gettime() is also in -lrt so we also need to check for it. - -Signed-off-by: "Yann E. MORIN" <[email protected]> ---- - configure.ac | 10 +++++++--- - 1 file changed, 7 insertions(+), 3 deletions(-) - -diff --git a/configure.ac b/configure.ac -index f83f5ab..3b6ae09 100644 ---- a/configure.ac -+++ b/configure.ac -@@ -199,12 +199,16 @@ fi - - AC_SUBST(LIBOPING_PC_LIBS_PRIVATE) - --nanosleep_needs_rt="no" -+needs_rt="no" - AC_CHECK_FUNCS(nanosleep, [], - AC_CHECK_LIB(rt, nanosleep, -- [nanosleep_needs_rt="yes"], -+ [needs_rt="yes"], - AC_MSG_ERROR(cannot find nanosleep))) --AM_CONDITIONAL(BUILD_WITH_LIBRT, test "x$nanosleep_needs_rt" = "xyes") -+AC_CHECK_FUNCS(clock_gettime, [], -+ AC_CHECK_LIB(rt, clock_gettime, -+ [needs_rt="yes"], -+ AC_MSG_ERROR(cannot find clock_gettime))) -+AM_CONDITIONAL(BUILD_WITH_LIBRT, test "x$needs_rt" = "xyes") - - with_ncurses="no" - AC_CHECK_HEADERS(ncursesw/ncurses.h ncurses.h, [with_ncurses="yes"], []) --- -2.7.4 -package/liboping/liboping.mk +0
-2@@ -13,6 +13,4 @@ LIBOPING_CONF_OPTS = --without-perl-bindings LIBOPING_LICENSE = LGPL-2.1+, GPL-2.0 LIBOPING_LICENSE_FILES = COPYING -LIBOPING_AUTORECONF = YES - $(eval $(autotools-package))Signature · 6 lines
2.11.0 _______________________________________________ buildroot mailing list buildroot at busybox.net http://lists.busybox.net/mailman/listinfo/buildroot
Signature · 6 lines
.-----------------.--------------------.------------------.--------------------. | Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: | | +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___ | | +33 223 225 172 `------------.-------: X AGAINST | \e/ There is no | | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v conspiracy. | '------------------------------^-------^------------------^--------------------'
PATCH 12/14 openipmi: no need to check for clock_gettime
In reply to Baruch SiachBuildroot no longer supports toolchains with glibc older than 2.17, so there is no need to check whether librt is required for clock_* system calls.
--- ...c-detect-lrt-requirement-for-clock_gettim.patch | 57 ---------------------- 1 file changed, 57 deletions(-) delete mode 100644 package/openipmi/0002-configure.ac-detect-lrt-requirement-for-clock_gettim.patch
package/openipmi/0002-configure.ac-detect-lrt-requirement-for-clock_gettim.patch deleted +0 -57
@@ -1,57 +0,0 @@ -From d3be03464be7fa3844dfe0837a7e2c920c107836 Mon Sep 17 00:00:00 2001 -From: Baruch Siach <[email protected]> -Date: Thu, 23 Jul 2015 13:35:35 +0300 -Subject: [PATCH] configure.ac: detect -lrt requirement for clock_gettime - -glibc versions before 2.17 require link with -lrt. - -Upstream status: sent, stuck in moderator's queue - -Signed-off-by: Baruch Siach <[email protected]> ---- - configure.ac | 8 ++++++++ - unix/Makefile.am | 4 ++-- - 2 files changed, 10 insertions(+), 2 deletions(-) - -diff --git a/configure.ac b/configure.ac -index 18401f161d45..6b350a41a916 100644 ---- a/configure.ac -+++ b/configure.ac -@@ -854,6 +854,14 @@ OPENIPMI_SWIG_SO="${PYTHON_POSIX_SO}:\$(top_builddir)/lib/.libs/libOpenIPMI.so:\ - AC_SUBST(OPENIPMI_SWIG_LIBS) - AC_SUBST(OPENIPMI_SWIG_SO) - -+LIB_CLOCK_GETTIME= -+clk_saved_libs=$LIBS -+AC_SEARCH_LIBS([clock_gettime], [rt], -+ [test "$ac_cv_search_clock_gettime" = "none required" || -+ LIB_CLOCK_GETTIME=$ac_cv_search_clock_gettime]) -+AC_SUBST([LIB_CLOCK_GETTIME]) -+LIBS=$clk_saved_libs -+ - # Everything from here to AC_OUTPUT is for libedit - - # AC_PROG_LIBTOOL runs AC_CANONICAL_HOST -diff --git a/unix/Makefile.am b/unix/Makefile.am -index a108bc679c9a..fb60242d5d3b 100644 ---- a/unix/Makefile.am -+++ b/unix/Makefile.am -@@ -9,13 +9,13 @@ lib_LTLIBRARIES = libOpenIPMIposix.la libOpenIPMIpthread.la - - libOpenIPMIpthread_la_SOURCES = posix_thread_os_hnd.c selector.c - libOpenIPMIpthread_la_LIBADD = -lpthread $(GDBM_LIB) \ -- $(top_builddir)/utils/libOpenIPMIutils.la -+ $(top_builddir)/utils/libOpenIPMIutils.la $(LIB_CLOCK_GETTIME) - libOpenIPMIpthread_la_LDFLAGS = -rdynamic -version-info $(LD_VERSION) \ - -Wl,-Map -Wl,libOpenIPMIpthread.map - - libOpenIPMIposix_la_SOURCES = posix_os_hnd.c selector.c - libOpenIPMIposix_la_LIBADD = $(top_builddir)/utils/libOpenIPMIutils.la \ -- $(GDBM_LIB) -+ $(GDBM_LIB) $(LIB_CLOCK_GETTIME) - libOpenIPMIposix_la_LDFLAGS = -rdynamic -version-info $(LD_VERSION) \ - -Wl,-Map -Wl,libOpenIPMIposix.map - --- -2.1.4 -
Signature · 1 line
2.11.0
PATCH 12/14 openipmi: no need to check for clock_gettime
Baruch, All,
On 2017-06-16 06:32 +0300, Baruch Siach spake thusly:
Quoted 2 lines
Buildroot no longer supports toolchains with glibc older than 2.17, so there is no need to check whether librt is required for clock_* system calls.
To be noted: upstream has fixed that recently:
https://sourceforge.net/p/openipmi/code/ci/488c5dbfe54cb0da26a61c0b892a4106ac87a878/
Sad that they waited two years to fix it, and that they did not use your patch... :-(
Quoted 1 line
Signed-off-by Baruch Siach <[email protected]>
Regards,
Yann E. MORIN.
Quoted 75 lines
--- ...c-detect-lrt-requirement-for-clock_gettim.patch | 57 ---------------------- 1 file changed, 57 deletions(-) delete mode 100644 package/openipmi/0002-configure.ac-detect-lrt-requirement-for-clock_gettim.patchpackage/openipmi/0002-configure.ac-detect-lrt-requirement-for-clock_gettim.patch deleted +0
-57@@ -1,57 +0,0 @@ -From d3be03464be7fa3844dfe0837a7e2c920c107836 Mon Sep 17 00:00:00 2001 -From: Baruch Siach <[email protected]> -Date: Thu, 23 Jul 2015 13:35:35 +0300 -Subject: [PATCH] configure.ac: detect -lrt requirement for clock_gettime - -glibc versions before 2.17 require link with -lrt. - -Upstream status: sent, stuck in moderator's queue - -Signed-off-by: Baruch Siach <[email protected]> ---- - configure.ac | 8 ++++++++ - unix/Makefile.am | 4 ++-- - 2 files changed, 10 insertions(+), 2 deletions(-) - -diff --git a/configure.ac b/configure.ac -index 18401f161d45..6b350a41a916 100644 ---- a/configure.ac -+++ b/configure.ac -@@ -854,6 +854,14 @@ OPENIPMI_SWIG_SO="${PYTHON_POSIX_SO}:\$(top_builddir)/lib/.libs/libOpenIPMI.so:\ - AC_SUBST(OPENIPMI_SWIG_LIBS) - AC_SUBST(OPENIPMI_SWIG_SO) - -+LIB_CLOCK_GETTIME= -+clk_saved_libs=$LIBS -+AC_SEARCH_LIBS([clock_gettime], [rt], -+ [test "$ac_cv_search_clock_gettime" = "none required" || -+ LIB_CLOCK_GETTIME=$ac_cv_search_clock_gettime]) -+AC_SUBST([LIB_CLOCK_GETTIME]) -+LIBS=$clk_saved_libs -+ - # Everything from here to AC_OUTPUT is for libedit - - # AC_PROG_LIBTOOL runs AC_CANONICAL_HOST -diff --git a/unix/Makefile.am b/unix/Makefile.am -index a108bc679c9a..fb60242d5d3b 100644 ---- a/unix/Makefile.am -+++ b/unix/Makefile.am -@@ -9,13 +9,13 @@ lib_LTLIBRARIES = libOpenIPMIposix.la libOpenIPMIpthread.la - - libOpenIPMIpthread_la_SOURCES = posix_thread_os_hnd.c selector.c - libOpenIPMIpthread_la_LIBADD = -lpthread $(GDBM_LIB) \ -- $(top_builddir)/utils/libOpenIPMIutils.la -+ $(top_builddir)/utils/libOpenIPMIutils.la $(LIB_CLOCK_GETTIME) - libOpenIPMIpthread_la_LDFLAGS = -rdynamic -version-info $(LD_VERSION) \ - -Wl,-Map -Wl,libOpenIPMIpthread.map - - libOpenIPMIposix_la_SOURCES = posix_os_hnd.c selector.c - libOpenIPMIposix_la_LIBADD = $(top_builddir)/utils/libOpenIPMIutils.la \ -- $(GDBM_LIB) -+ $(GDBM_LIB) $(LIB_CLOCK_GETTIME) - libOpenIPMIposix_la_LDFLAGS = -rdynamic -version-info $(LD_VERSION) \ - -Wl,-Map -Wl,libOpenIPMIposix.map - --- -2.1.4 -Signature · 6 lines
2.11.0 _______________________________________________ buildroot mailing list buildroot at busybox.net http://lists.busybox.net/mailman/listinfo/buildroot
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. | '------------------------------^-------^------------------^--------------------'
PATCH 13/14 lirc-tools: no need to check for clock_gettime
In reply to Baruch SiachBuildroot no longer supports toolchains with glibc older than 2.17, so there is no need to check whether librt is required for clock_* system calls.
---
...onfigure-check-for-clock_gettime-in-librt.patch | 54 ----------------------
...-use-of-functions-killed-in-kernel-4.8.0.patch} | 0
package/lirc-tools/lirc-tools.mk | 2 -
3 files changed, 56 deletions(-)
delete mode 100644 package/lirc-tools/0002-configure-check-for-clock_gettime-in-librt.patch
rename package/lirc-tools/{0003-lircd-Remove-use-of-functions-killed-in-kernel-4.8.0.patch => 0002-lircd-Remove-use-of-functions-killed-in-kernel-4.8.0.patch} (100%)package/lirc-tools/0002-configure-check-for-clock_gettime-in-librt.patch deleted +0 -54
@@ -1,54 +0,0 @@ -From e654ae139cec42a6f1b5684261787d0c241cfd3b Mon Sep 17 00:00:00 2001 -From: Thomas Petazzoni <[email protected]> -Date: Fri, 19 Aug 2016 15:25:48 +0200 -Subject: [PATCH] configure: check for clock_gettime in librt - -Signed-off-by: Thomas Petazzoni <[email protected]> ---- - configure.ac | 3 +++ - daemons/Makefile.am | 2 +- - plugins/Makefile.am | 1 + - 3 files changed, 5 insertions(+), 1 deletion(-) - -diff --git a/configure.ac b/configure.ac -index 73340c7..466e638 100644 ---- a/configure.ac -+++ b/configure.ac -@@ -67,6 +67,9 @@ AC_CHECK_FUNCS(gethostname gettimeofday mkfifo select socket strdup \ - AC_SEARCH_LIBS([dlopen], [dl dld], [], [ - AC_MSG_ERROR([unable to find the dlopen() function]) - ]) -+# glibc < 2.17 needs librt for clock_gettime() -+AC_CHECK_LIB(rt, clock_gettime, LIBCLOCK_GETTIME="-lrt") -+AC_SUBST(LIBCLOCK_GETTIME) - AC_CHECK_FUNCS(daemon) - if test "$ac_cv_func_daemon" != yes; then - daemon="" -diff --git a/daemons/Makefile.am b/daemons/Makefile.am -index 5625627..85a28f3 100644 ---- a/daemons/Makefile.am -+++ b/daemons/Makefile.am -@@ -21,7 +21,7 @@ sbin_PROGRAMS += lircd-uinput - endif - - lircd_SOURCES = lircd.cpp --lircd_LDADD = ../lib/liblirc.la -+lircd_LDADD = ../lib/liblirc.la @LIBCLOCK_GETTIME@ - - lircd_uinput_SOURCES = lircd-uinput.cpp - lircd_uinput_LDADD = ../lib/liblirc.la -diff --git a/plugins/Makefile.am b/plugins/Makefile.am -index ddff01d..45c012a 100644 ---- a/plugins/Makefile.am -+++ b/plugins/Makefile.am -@@ -127,6 +127,7 @@ zotac_la_SOURCES = zotac.c - - plugin_LTLIBRARIES += mplay.la - mplay_la_SOURCES = mplay.c -+mplay_la_LIBADD = @LIBCLOCK_GETTIME@ - endif - - $(srcdir)/pluginlist.am: --- -2.7.4 -
package/lirc-tools/0002-lircd-Remove-use-of-functions-killed-in-kernel-4.8.0.patch renamed from package/lirc-tools/0003-lircd-Remove-use-of-functions-killed-in-kernel-4.8.0.patch +0 -0
package/lirc-tools/lirc-tools.mk +0 -2
@@ -11,8 +11,6 @@ LIRC_TOOLS_LICENSE = GPL-2.0+ LIRC_TOOLS_LICENSE_FILES = COPYING LIRC_TOOLS_DEPENDENCIES = host-libxslt host-pkgconf host-python3 LIRC_TOOLS_INSTALL_STAGING = YES -# 0002-configure-check-for-clock_gettime-in-librt.patch -LIRC_TOOLS_AUTORECONF = YES LIRC_TOOLS_CONF_ENV = XSLTPROC=yes LIRC_TOOLS_CONF_OPTS = --without-x
Signature · 1 line
2.11.0
PATCH 13/14 lirc-tools: no need to check for clock_gettime
Baruch, All,
On 2017-06-16 06:32 +0300, Baruch Siach spake thusly:
Quoted 5 lines
Buildroot no longer supports toolchains with glibc older than 2.17, so there is no need to check whether librt is required for clock_* system calls.
Cc Rhys Williams <[email protected]>Signed-off-by Baruch Siach <[email protected]>
Regards,
Yann E. MORIN.
Quoted 92 lines
--- ...onfigure-check-for-clock_gettime-in-librt.patch | 54 ---------------------- ...-use-of-functions-killed-in-kernel-4.8.0.patch} | 0 package/lirc-tools/lirc-tools.mk | 2 - 3 files changed, 56 deletions(-) delete mode 100644 package/lirc-tools/0002-configure-check-for-clock_gettime-in-librt.patch rename package/lirc-tools/{0003-lircd-Remove-use-of-functions-killed-in-kernel-4.8.0.patch => 0002-lircd-Remove-use-of-functions-killed-in-kernel-4.8.0.patch} (100%)package/lirc-tools/0002-configure-check-for-clock_gettime-in-librt.patch deleted +0
-54@@ -1,54 +0,0 @@ -From e654ae139cec42a6f1b5684261787d0c241cfd3b Mon Sep 17 00:00:00 2001 -From: Thomas Petazzoni <[email protected]> -Date: Fri, 19 Aug 2016 15:25:48 +0200 -Subject: [PATCH] configure: check for clock_gettime in librt - -Signed-off-by: Thomas Petazzoni <[email protected]> ---- - configure.ac | 3 +++ - daemons/Makefile.am | 2 +- - plugins/Makefile.am | 1 + - 3 files changed, 5 insertions(+), 1 deletion(-) - -diff --git a/configure.ac b/configure.ac -index 73340c7..466e638 100644 ---- a/configure.ac -+++ b/configure.ac -@@ -67,6 +67,9 @@ AC_CHECK_FUNCS(gethostname gettimeofday mkfifo select socket strdup \ - AC_SEARCH_LIBS([dlopen], [dl dld], [], [ - AC_MSG_ERROR([unable to find the dlopen() function]) - ]) -+# glibc < 2.17 needs librt for clock_gettime() -+AC_CHECK_LIB(rt, clock_gettime, LIBCLOCK_GETTIME="-lrt") -+AC_SUBST(LIBCLOCK_GETTIME) - AC_CHECK_FUNCS(daemon) - if test "$ac_cv_func_daemon" != yes; then - daemon="" -diff --git a/daemons/Makefile.am b/daemons/Makefile.am -index 5625627..85a28f3 100644 ---- a/daemons/Makefile.am -+++ b/daemons/Makefile.am -@@ -21,7 +21,7 @@ sbin_PROGRAMS += lircd-uinput - endif - - lircd_SOURCES = lircd.cpp --lircd_LDADD = ../lib/liblirc.la -+lircd_LDADD = ../lib/liblirc.la @LIBCLOCK_GETTIME@ - - lircd_uinput_SOURCES = lircd-uinput.cpp - lircd_uinput_LDADD = ../lib/liblirc.la -diff --git a/plugins/Makefile.am b/plugins/Makefile.am -index ddff01d..45c012a 100644 ---- a/plugins/Makefile.am -+++ b/plugins/Makefile.am -@@ -127,6 +127,7 @@ zotac_la_SOURCES = zotac.c - - plugin_LTLIBRARIES += mplay.la - mplay_la_SOURCES = mplay.c -+mplay_la_LIBADD = @LIBCLOCK_GETTIME@ - endif - - $(srcdir)/pluginlist.am: --- -2.7.4 -package/lirc-tools/0002-lircd-Remove-use-of-functions-killed-in-kernel-4.8.0.patch renamed from package/lirc-tools/0003-lircd-Remove-use-of-functions-killed-in-kernel-4.8.0.patch +0
-0package/lirc-tools/lirc-tools.mk +0
-2@@ -11,8 +11,6 @@ LIRC_TOOLS_LICENSE = GPL-2.0+ LIRC_TOOLS_LICENSE_FILES = COPYING LIRC_TOOLS_DEPENDENCIES = host-libxslt host-pkgconf host-python3 LIRC_TOOLS_INSTALL_STAGING = YES -# 0002-configure-check-for-clock_gettime-in-librt.patch -LIRC_TOOLS_AUTORECONF = YES LIRC_TOOLS_CONF_ENV = XSLTPROC=yes LIRC_TOOLS_CONF_OPTS = --without-xSignature · 6 lines
2.11.0 _______________________________________________ buildroot mailing list buildroot at busybox.net http://lists.busybox.net/mailman/listinfo/buildroot
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. | '------------------------------^-------^------------------^--------------------'
PATCH 14/14 softether: no need to check for clock_gettime
In reply to Baruch SiachBuildroot no longer supports toolchains with glibc older than 2.17, so there is no need to check whether librt is required for clock_* system calls.
---
package/softether/0006-librt.patch | 33 ----------------------
...onfig.patch => 0006-uclibc-ai-addrconfig.patch} | 0
2 files changed, 33 deletions(-)
delete mode 100644 package/softether/0006-librt.patch
rename package/softether/{0007-uclibc-ai-addrconfig.patch => 0006-uclibc-ai-addrconfig.patch} (100%)package/softether/0006-librt.patch deleted +0 -33
@@ -1,33 +0,0 @@ -Downloaded from -https://github.com/dajhorn/SoftEtherVPN/commit/9a8538a6f2287fe8bebd0a7185de408e31e5a7df - -Fixes -http://autobuild.buildroot.net/results/48f/48f778a891e0bf6a60ed2e4bec057c338ec25ec9/ -and many others - -Signed-off-by: Bernd Kuhls <[email protected]> - - -From 9a8538a6f2287fe8bebd0a7185de408e31e5a7df Mon Sep 17 00:00:00 2001 -From: Darik Horn <[email protected]> -Date: Thu, 5 Jun 2014 21:02:02 -0400 -Subject: [PATCH] Add an explicit autoconf check for librt. - -The realtime extensions library is not automatically linked on Debian ARM -platforms, so do an AC_CHECK_LIB for it in the configuration macro. ---- - configure.ac | 1 + - 1 file changed, 1 insertion(+) - -diff --git a/configure.ac b/configure.ac -index 165623a..6f85393 100644 ---- a/configure.ac -+++ b/configure.ac -@@ -60,6 +60,7 @@ CC="$PTHREAD_CC" - CFLAGS="$CFLAGS $PTHREAD_CFLAGS" - LIBS="$PTHREAD_LIBS $LIBS" - -+AC_CHECK_LIB([rt],[clock_gettime]) - - AX_CHECK_OPENSSL([ - AC_SUBST(OPENSSL_LIBS)
package/softether/0006-uclibc-ai-addrconfig.patch renamed from package/softether/0007-uclibc-ai-addrconfig.patch +0 -0
Signature · 1 line
2.11.0
PATCH 14/14 softether: no need to check for clock_gettime
Baruch, All,
On 2017-06-16 06:32 +0300, Baruch Siach spake thusly:
Quoted 5 lines
Buildroot no longer supports toolchains with glibc older than 2.17, so there is no need to check whether librt is required for clock_* system calls.
Cc Bernd Kuhls <[email protected]>Signed-off-by Baruch Siach <[email protected]>
Regards,
Yann E. MORIN.
Quoted 57 lines
--- package/softether/0006-librt.patch | 33 ---------------------- ...onfig.patch => 0006-uclibc-ai-addrconfig.patch} | 0 2 files changed, 33 deletions(-) delete mode 100644 package/softether/0006-librt.patch rename package/softether/{0007-uclibc-ai-addrconfig.patch => 0006-uclibc-ai-addrconfig.patch} (100%)package/softether/0006-librt.patch deleted +0
-33@@ -1,33 +0,0 @@ -Downloaded from -https://github.com/dajhorn/SoftEtherVPN/commit/9a8538a6f2287fe8bebd0a7185de408e31e5a7df - -Fixes -http://autobuild.buildroot.net/results/48f/48f778a891e0bf6a60ed2e4bec057c338ec25ec9/ -and many others - -Signed-off-by: Bernd Kuhls <[email protected]> - - -From 9a8538a6f2287fe8bebd0a7185de408e31e5a7df Mon Sep 17 00:00:00 2001 -From: Darik Horn <[email protected]> -Date: Thu, 5 Jun 2014 21:02:02 -0400 -Subject: [PATCH] Add an explicit autoconf check for librt. - -The realtime extensions library is not automatically linked on Debian ARM -platforms, so do an AC_CHECK_LIB for it in the configuration macro. ---- - configure.ac | 1 + - 1 file changed, 1 insertion(+) - -diff --git a/configure.ac b/configure.ac -index 165623a..6f85393 100644 ---- a/configure.ac -+++ b/configure.ac -@@ -60,6 +60,7 @@ CC="$PTHREAD_CC" - CFLAGS="$CFLAGS $PTHREAD_CFLAGS" - LIBS="$PTHREAD_LIBS $LIBS" - -+AC_CHECK_LIB([rt],[clock_gettime]) - - AX_CHECK_OPENSSL([ - AC_SUBST(OPENSSL_LIBS)package/softether/0006-uclibc-ai-addrconfig.patch renamed from package/softether/0007-uclibc-ai-addrconfig.patch +0
-0Signature · 6 lines
2.11.0 _______________________________________________ buildroot mailing list buildroot at busybox.net http://lists.busybox.net/mailman/listinfo/buildroot
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. | '------------------------------^-------^------------------^--------------------'
cover toolchain-external: remove support for glibc before 2.17
In reply to Baruch SiachHello,
Quoted 7 lines · Baruch Siach
Baruch Siach (14):
docs/manual: update external toolchain advantages list
toolchain: remove CodeSourcery x86 toolchain
toolchain: remove CodeSourcery sh toolchain
package: remove CodeSourcery sh dependencies
toolchain-external: update list of toolchains docs/manual: drop mention of removed external toolchains
Quoted 8 lines
swupdate: don't link with librt
aiccu: don't link with librt
libv4l: don't link with librt
e2fsprogs: don't link with librt
liboping: no need to check forclock_gettime
openipmi: no need to check forclock_gettime
lirc-tools: no need to check forclock_gettime
softether: no need to check forclock_gettime
Series applied, except the last patch, which is needed since softether is also built on the host.
Thanks a lot!
Thomas
Signature · 3 lines
Thomas Petazzoni, CTO, Free Electrons Embedded Linux and Kernel engineering http://free-electrons.com
cover toolchain-external: remove support for glibc before 2.17
In reply to Baruch SiachHi Baruch,
Quoted 17 lines · Baruch Siach
As discussed[1] recently on the list it is time to remove support for glibc older than 2.17. This does not mean that toolchains with old glibc will stop working. It only means that we stop fixing older glibc build issues. In particular, the need to link with librt for clock_* system calls.
The only significant user visible change in this series is the removal of CodeSourcery toolchains for x86 and sh.
Two patches update a comment section and the manual to take into account the removal of toolchain in this series and in previous commits. A few more per-package patches remove local patches for the glibc librt issue.
The first patch is an unrelated fix to the external toolchain section in the manual.
Quoted 14 lines
Baruch Siach (14):
docs/manual: update external toolchain advantages list
toolchain: remove CodeSourcery x86 toolchain
toolchain: remove CodeSourcery sh toolchain
package: remove CodeSourcery sh dependencies
toolchain-external: update list of toolchains docs/manual: drop mention of removed external toolchains
swupdate: don't link with librt
aiccu: don't link with librt
libv4l: don't link with librt
e2fsprogs: don't link with librt
liboping: no need to check forclock_gettime
openipmi: no need to check forclock_gettime
lirc-tools: no need to check forclock_gettime
After all of these have been applied, "git grep -e -lrt" still yields quite a
few results that at first sight could also be removed. Was that intentional or did you just miss them? Perhaps also grep for clock_gettime.
Regards,
Arnout
Signature · 6 lines
Arnout Vandecappelle arnout at mind be Senior Embedded Software Architect +32-16-286500 Essensium/Mind http://www.mind.be G.Geenslaan 9, 3001 Leuven, Belgium BE 872 984 063 RPR Leuven LinkedIn profile: http://www.linkedin.com/in/arnoutvandecappelle GPG fingerprint: 7493 020B C7E3 8618 8DEC 222C 82EB F404 F9AC 0DDF
cover toolchain-external: remove support for glibc before 2.17
Hi Arnout,
Peter, a question for you below.
Quoted 33 lines · Arnout Vandecappelle
Quoted 17 lines · Baruch Siach
As discussed[1] recently on the list it is time to remove support for glibc older than 2.17. This does not mean that toolchains with old glibc will stop working. It only means that we stop fixing older glibc build issues. In particular, the need to link with librt for clock_* system calls.
The only significant user visible change in this series is the removal of CodeSourcery toolchains for x86 and sh.
Two patches update a comment section and the manual to take into account the removal of toolchain in this series and in previous commits. A few more per-package patches remove local patches for the glibc librt issue.
The first patch is an unrelated fix to the external toolchain section in the manual.
Quoted 14 lines
Baruch Siach (14):
docs/manual: update external toolchain advantages list
toolchain: remove CodeSourcery x86 toolchain
toolchain: remove CodeSourcery sh toolchain
package: remove CodeSourcery sh dependencies
toolchain-external: update list of toolchains docs/manual: drop mention of removed external toolchains
swupdate: don't link with librt
aiccu: don't link with librt
libv4l: don't link with librt
e2fsprogs: don't link with librt
liboping: no need to check forclock_gettime
openipmi: no need to check forclock_gettime
lirc-tools: no need to check forclock_gettime
Quoted 3 lines
After all of these have been applied, "git grep -e -lrt" still yields quite a
few results that at first sight could also be removed. Was that intentional or did you just miss them? Perhaps also grep forclock_gettime.
Here is a summary of all files matching the '-lrt\|clock_gettime' RE pattern.
package/blktrace/0001-iowatcher-link-with-lrt.patch
Needed for posix_spawn*().
package/canfestival/0001-install-pkgconfig-module-for-canfestival.patch
Small part of of the pkgconfig support patch.
package/eudev/eudev.mk
Needed for mq_getattr()
package/faketime/0001-Disable-the-non-null-compare-warning-error.patch
Patch context.
package/flac/0001-configure.ac-relax-linux-OS-detection.patch
Patch description.
package/flashbench/flashbench.mk
Replicate upstream -lrt.
package/gstreamer/gst-fsl-plugins/0004-Fix-libgstfsl-linkage-add-lrt-for-shared-memory-func.patch
Needed for shm_open().
package/haveged/haveged.mk
configure option name.
package/iozone/0001-targets.patch
Needed for AIO.
package/libasplib/0001-asplib_CPUTimer.h-Fix-time.h-include.patch
Patch description.
package/libv4l/0007-configure.ac-add-disable-libv4l-option.patch
Patch context.
package/mariadb/0001-add-extra-check-for-librt.patch
Patch context.
package/multicat/0001-Fix-missing-pthread.patch
Patch context.
package/qemu/qemu.mk
Needed for timer_gettime()
package/qt5/qt5base/qmake.conf
Maybe redundant -lrt. I'm not too familiar with the Qt5 build system to know. Peter (Seiderer)?
package/softether/0006-librt.patch
Host package.
package/squeezelite/0001-Makefile-allow-passing-CFLAGS-and-LDFLAGS.patch
Replicate upstream libs of libraries.
package/squid/squid.mk
Needed for shm_open().
package/syslog-ng/syslog-ng.mk
Needed for mq_*().
package/vlc/0008-fix-fallback-code-and-add-required-realtime-library.patch
Needed for uClibc posix_spawnp().
package/wiringpi/0001-Adjust-for-buildroot-build.patch
Patch context.
package/x11r7/xserver_xorg-server/1.19.3/0002-configure.ac-Fix-check-for-CLOCK_MONOTONIC.patch
Patch context.
All in all only the qt5 match might be removed. I would appreciate an advice from someone familiar with the Qt build system.
baruch
Signature · 3 lines
http://baruch.siach.name/blog/ ~. .~ Tk Open Systems
=}------------------------------------------------ooO--U--Ooo------------{=
- baruch at tkos.co.il - tel: +972.52.368.4656, http://www.tkos.co.il -cover toolchain-external: remove support for glibc before 2.17
Hi,
Baruch Siach wrote,
Quoted 3 lines
package/vlc/0008-fix-fallback-code-and-add-required-realtime-library.patchNeeded for uClibc
posix_spawnp().
Then it shouldn't be needed, if it is only for uClibc-ng.
best regards
Waldemar
cover toolchain-external: remove support for glibc before 2.17
Hi Waldemar,
Quoted 7 lines · Waldemar Brodkorb
Baruch Siach wrote,
Quoted 3 lines
package/vlc/0008-fix-fallback-code-and-add-required-realtime-library.patchNeeded for uClibc
posix_spawnp().Then it shouldn't be needed, if it is only for uClibc-ng.
It's your patch (commit d3af7772d07aa5). Feel free to update/remove it.
baruch
Signature · 3 lines
http://baruch.siach.name/blog/ ~. .~ Tk Open Systems
=}------------------------------------------------ooO--U--Ooo------------{=
- baruch at tkos.co.il - tel: +972.2.679.5364, http://www.tkos.co.il -