Re: [LEDE-DEV] [PATCH v2] gdb: bump to 8.0.1
On 2017-09-27 19:41, Stijn Tintel wrote: > On 27-09-17 10:31, Felix Fietkau wrote: >> On 2017-09-26 17:10, Stijn Tintel wrote: >>> On 25-09-17 20:18, Felix Fietkau wrote: On 2017-09-25 15:36, Stijn Tintel wrote: > Since version 8.0, gdbserver seems to have a hard dependency on > libstdc++, even with --disable-libstdcxx. I think it's worth adding -static-libstdc++ to LDFLAGS. With a bit of luck, gdbserver does not need much of libstdc++. >>> That seems to work. Sizes on x86/64: >>> gdbserver 151K >>> libstdcpp 335K >>> >>> With TARGET_LDFLAGS+=-static-libstdc++: >>> gdbserver 320K >> Please add -Wl,--gc-sections as well. It seems to bring the package size >> down to 146K on MIPS. > Have this queued in my staging tree. I'll push it when the buildbot is > upgraded: https://git.lede-project.org/9eb0ccfe > > Size of gdbserver on x86/64: 171K Looks good to me. - Felix ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] [PATCH v2] gdb: bump to 8.0.1
On 27-09-17 10:31, Felix Fietkau wrote: > On 2017-09-26 17:10, Stijn Tintel wrote: >> On 25-09-17 20:18, Felix Fietkau wrote: >>> On 2017-09-25 15:36, Stijn Tintel wrote: Since version 8.0, gdbserver seems to have a hard dependency on libstdc++, even with --disable-libstdcxx. >>> I think it's worth adding -static-libstdc++ to LDFLAGS. With a bit of >>> luck, gdbserver does not need much of libstdc++. >> That seems to work. Sizes on x86/64: >> gdbserver 151K >> libstdcpp 335K >> >> With TARGET_LDFLAGS+=-static-libstdc++: >> gdbserver 320K > Please add -Wl,--gc-sections as well. It seems to bring the package size > down to 146K on MIPS. Have this queued in my staging tree. I'll push it when the buildbot is upgraded: https://git.lede-project.org/9eb0ccfe Size of gdbserver on x86/64: 171K Thanks, Stijn ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] [PATCH v2] gdb: bump to 8.0.1
On 2017-09-26 17:10, Stijn Tintel wrote: > On 25-09-17 20:18, Felix Fietkau wrote: >> On 2017-09-25 15:36, Stijn Tintel wrote: >>> Since version 8.0, gdbserver seems to have a hard dependency on >>> libstdc++, even with --disable-libstdcxx. >>> >>> Fixes CVE-2017-9778. >>> >>> Signed-off-by: Stijn Tintel >>> --- >>> V2: add hard dependency on libstdc++ >>> >>> Also in my staging tree: https://git.lede-project.org/a96fc0d9 >> I think it's worth adding -static-libstdc++ to LDFLAGS. With a bit of >> luck, gdbserver does not need much of libstdc++. > That seems to work. Sizes on x86/64: > gdbserver 151K > libstdcpp 335K > > With TARGET_LDFLAGS+=-static-libstdc++: > gdbserver 320K > > But as gdb 8.0.1 doesn't compile on the buildbot master due to it still > running gcc 4.7, the buildbot will need to be upgraded first. Please add -Wl,--gc-sections as well. It seems to bring the package size down to 146K on MIPS. - Felix ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] [PATCH v2] gdb: bump to 8.0.1
On 25-09-17 20:18, Felix Fietkau wrote: > On 2017-09-25 15:36, Stijn Tintel wrote: >> Since version 8.0, gdbserver seems to have a hard dependency on >> libstdc++, even with --disable-libstdcxx. >> >> Fixes CVE-2017-9778. >> >> Signed-off-by: Stijn Tintel >> --- >> V2: add hard dependency on libstdc++ >> >> Also in my staging tree: https://git.lede-project.org/a96fc0d9 > I think it's worth adding -static-libstdc++ to LDFLAGS. With a bit of > luck, gdbserver does not need much of libstdc++. That seems to work. Sizes on x86/64: gdbserver 151K libstdcpp 335K With TARGET_LDFLAGS+=-static-libstdc++: gdbserver 320K But as gdb 8.0.1 doesn't compile on the buildbot master due to it still running gcc 4.7, the buildbot will need to be upgraded first. Stijn ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] [PATCH v2] gdb: bump to 8.0.1
On 2017-09-25 15:36, Stijn Tintel wrote: > Since version 8.0, gdbserver seems to have a hard dependency on > libstdc++, even with --disable-libstdcxx. > > Fixes CVE-2017-9778. > > Signed-off-by: Stijn Tintel > --- > V2: add hard dependency on libstdc++ > > Also in my staging tree: https://git.lede-project.org/a96fc0d9 I think it's worth adding -static-libstdc++ to LDFLAGS. With a bit of luck, gdbserver does not need much of libstdc++. - Felix ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
[LEDE-DEV] [PATCH v2] gdb: bump to 8.0.1
Since version 8.0, gdbserver seems to have a hard dependency on libstdc++, even with --disable-libstdcxx. Fixes CVE-2017-9778. Signed-off-by: Stijn Tintel --- V2: add hard dependency on libstdc++ Also in my staging tree: https://git.lede-project.org/a96fc0d9 package/devel/gdb/Makefile | 7 --- package/devel/gdb/patches/110-shared_libgcc.patch| 12 ++-- package/devel/gdb/patches/120-sigprocmask-invalid-call.patch | 5 - 3 files changed, 10 insertions(+), 14 deletions(-) diff --git a/package/devel/gdb/Makefile b/package/devel/gdb/Makefile index 23039910ce..e0c0207802 100644 --- a/package/devel/gdb/Makefile +++ b/package/devel/gdb/Makefile @@ -8,12 +8,12 @@ include $(TOPDIR)/rules.mk PKG_NAME:=gdb -PKG_VERSION:=7.12.1 -PKG_RELEASE:=3 +PKG_VERSION:=8.0.1 +PKG_RELEASE:=1 PKG_SOURCE:=$(PKG_NAME)-$(PKG_VERSION).tar.xz PKG_SOURCE_URL:=@GNU/gdb -PKG_HASH:=4607680b973d3ec92c30ad029f1b7dbde3876869e6b3a117d8a7e90081113186 +PKG_HASH:=3dbd5f93e36ba2815ad0efab030dcd0c7b211d7b353a40a53f4c02d7d56295e3 PKG_BUILD_PARALLEL:=1 PKG_INSTALL:=1 @@ -43,6 +43,7 @@ endef define Package/gdbserver $(call Package/gdb/Default) TITLE:=Remote server for GNU Debugger + DEPENDS+=+libstdcpp endef define Package/gdbserver/description diff --git a/package/devel/gdb/patches/110-shared_libgcc.patch b/package/devel/gdb/patches/110-shared_libgcc.patch index e44616f4a2..929a5ae977 100644 --- a/package/devel/gdb/patches/110-shared_libgcc.patch +++ b/package/devel/gdb/patches/110-shared_libgcc.patch @@ -1,6 +1,6 @@ --- a/configure.ac +++ b/configure.ac -@@ -1406,13 +1406,13 @@ if test -z "$LD"; then +@@ -1300,13 +1300,13 @@ if test -z "$LD"; then fi fi @@ -17,7 +17,7 @@ AC_LANG_PUSH(C++) AC_LINK_IFELSE([ #if (__GNUC__ < 4) || (__GNUC__ == 4 && __GNUC_MINOR__ < 5) -@@ -1740,7 +1740,7 @@ AC_ARG_WITH(stage1-ldflags, +@@ -1632,7 +1632,7 @@ AC_ARG_WITH(stage1-ldflags, # if supported. But if the user explicitly specified the libraries to use, # trust that they are doing what they want. if test "$stage1_libs" = "" -a "$have_static_libs" = yes; then @@ -26,7 +26,7 @@ fi]) AC_SUBST(stage1_ldflags) -@@ -1769,7 +1769,7 @@ AC_ARG_WITH(boot-ldflags, +@@ -1661,7 +1661,7 @@ AC_ARG_WITH(boot-ldflags, # statically. But if the user explicitly specified the libraries to # use, trust that they are doing what they want. if test "$poststage1_libs" = ""; then @@ -37,7 +37,7 @@ --- a/configure +++ b/configure -@@ -5109,14 +5109,14 @@ if test -z "$LD"; then +@@ -5005,14 +5005,14 @@ if test -z "$LD"; then fi fi @@ -56,7 +56,7 @@ ac_ext=cpp ac_cpp='$CXXCPP $CPPFLAGS' ac_compile='$CXX -c $CXXFLAGS $CPPFLAGS conftest.$ac_ext >&5' -@@ -5902,7 +5902,7 @@ else +@@ -5795,7 +5795,7 @@ else # if supported. But if the user explicitly specified the libraries to use, # trust that they are doing what they want. if test "$stage1_libs" = "" -a "$have_static_libs" = yes; then @@ -65,7 +65,7 @@ fi fi -@@ -5938,7 +5938,7 @@ else +@@ -5831,7 +5831,7 @@ else # statically. But if the user explicitly specified the libraries to # use, trust that they are doing what they want. if test "$poststage1_libs" = ""; then diff --git a/package/devel/gdb/patches/120-sigprocmask-invalid-call.patch b/package/devel/gdb/patches/120-sigprocmask-invalid-call.patch index 7090db170a..8afa8ccf81 100644 --- a/package/devel/gdb/patches/120-sigprocmask-invalid-call.patch +++ b/package/devel/gdb/patches/120-sigprocmask-invalid-call.patch @@ -25,8 +25,6 @@ gdb/ChangeLog: gdb/common/signals-state-save-restore.c | 2 +- 2 files changed, 6 insertions(+), 1 deletion(-) -diff --git a/gdb/common/signals-state-save-restore.c b/gdb/common/signals-state-save-restore.c -index d11a9ae..734335c 100644 --- a/gdb/common/signals-state-save-restore.c +++ b/gdb/common/signals-state-save-restore.c @@ -41,7 +41,7 @@ save_original_signals_state (void) @@ -38,6 +36,3 @@ index d11a9ae..734335c 100644 if (res == -1) perror_with_name (("sigprocmask")); --- -2.6.4 - -- 2.13.5 ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev