Merge branch 'gdb-10' of https://github.com/cpackham/crosstool-ng

Signed-off-by: Chris Packham <judge.packham@gmail.com>

config/debug/gdb.in

@@ -40,3 +40,9 @@ config GDB_HAS_IPA_LIB
 config GDB_CONST_GET_THREAD_AREA
     bool
     default y if GDB_older_than_7_12
+
+# As of GDB 10.x gdbserver is now at the toplevel of the distributed
+# tarball (instead of a subdirectory of gdb)
+config GDB_GDBSERVER_TOPLEVEL
+    def_bool y
+    depends on GDB_10_2_or_later

packages/gdb/10.2/0000-musl_fix.patch

@@ -0,0 +1,33 @@
+---
+ gdb/linux-nat.c |    5 +++++
+ gdb/stopcode.h  |    4 ++++
+ 2 files changed, 9 insertions(+)
+
+--- a/gdb/linux-nat.c
++++ b/gdb/linux-nat.c
+@@ -17,6 +17,7 @@
+    You should have received a copy of the GNU General Public License
+    along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
+ 
++#include "stopcode.h"
+ #include "defs.h"
+ #include "inferior.h"
+ #include "infrun.h"
+@@ -70,6 +71,10 @@
+ #include "gdbsupport/gdb-sigmask.h"
+ #include "debug.h"
+ 
++#ifndef __SIGRTMIN
++#define __SIGRTMIN SIGRTMIN
++#endif
++
+ /* This comment documents high-level logic of this file.
+ 
+ Waiting for events in sync mode
+--- a/gdb/stopcode.h
++++ b/gdb/stopcode.h
+@@ -0,0 +1,4 @@
++#ifndef W_STOPCODE
++#define W_STOPCODE(sig) ((sig) << 8 | 0x7f)
++#endif
++

packages/gdb/10.2/0001-uclibc-no-gettimeofday-clobber.patch

@@ -0,0 +1,35 @@
+Improve gnulib in gdb's guess work, gettimeofday() works in uClibcm promise.
+
+This patch helps building x86_64-unknown-linux-uclibc toolchains, the final
+gdb-native step otherwise fails when linking the libinproctrace.so
+
+Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
+Signed-off-by: Alexey Neyman <stilor@att.net>
+
+---
+ gnulib/configure                 |    2 ++
+ gnulib/import/m4/gettimeofday.m4 |    2 ++
+ 2 files changed, 4 insertions(+)
+
+--- a/gnulib/configure
++++ b/gnulib/configure
+@@ -20745,6 +20745,8 @@
+       case "$host_os" in
+                        # Guess all is fine on glibc systems.
+         *-gnu* | gnu*) gl_cv_func_gettimeofday_clobber="guessing no" ;;
++                       # Guess all is fine on uclibc systems.
++        *-uclibc*)     gl_cv_func_gettimeofday_clobber="guessing no" ;;
+                        # Guess all is fine on musl systems.
+         *-musl*)       gl_cv_func_gettimeofday_clobber="guessing no" ;;
+                        # Guess no on native Windows.
+--- a/gnulib/import/m4/gettimeofday.m4
++++ b/gnulib/import/m4/gettimeofday.m4
+@@ -105,6 +105,8 @@
+       case "$host_os" in
+                        # Guess all is fine on glibc systems.
+         *-gnu* | gnu*) gl_cv_func_gettimeofday_clobber="guessing no" ;;
++                       # Guess all is fine on uclibc systems.
++        *-uclibc*)     gl_cv_func_gettimeofday_clobber="guessing no" ;;
+                        # Guess all is fine on musl systems.
+         *-musl*)       gl_cv_func_gettimeofday_clobber="guessing no" ;;
+                        # Guess no on native Windows.

packages/gdb/10.2/0002-xtensa-make-sure-ar_base-is-initialized.patch

@@ -0,0 +1,30 @@
+From 208ea73d38c9c16cf983b6419f58050dbadcb6a9 Mon Sep 17 00:00:00 2001
+From: Max Filippov <jcmvbkbc@gmail.com>
+Date: Sun, 7 Jun 2015 22:43:49 +0300
+Subject: [PATCH 2/2] xtensa: make sure ar_base is initialized
+
+ar_base is uninitialized for cores w/o windowed registers as their
+regmap doesn't have register 0x0100.
+Check that ar_base is initialized and if not initialize it with a0_base.
+
+gdb/
+	* xtensa-tdep.c (xtensa_derive_tdep): Make sure ar_base is
+	initialized.
+
+Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
+---
+ gdb/xtensa-tdep.c |    3 +++
+ 1 file changed, 3 insertions(+)
+
+--- a/gdb/xtensa-tdep.c
++++ b/gdb/xtensa-tdep.c
+@@ -3137,6 +3137,9 @@
+   if (tdep->num_regs == 0)
+     tdep->num_regs = tdep->num_nopriv_regs;
+ 
++  if (tdep->ar_base == -1)
++    tdep->ar_base = tdep->a0_base;
++
+   /* Number of pseudo registers.  */
+   tdep->num_pseudo_regs = n - tdep->num_regs;
+ 

packages/gdb/10.2/0003-WIP-end-of-prologue-detection-hack.patch

@@ -0,0 +1,26 @@
+From 7f8eacbb468575fb67db7fd1155a3aedaa91911b Mon Sep 17 00:00:00 2001
+From: Max Filippov <jcmvbkbc@gmail.com>
+Date: Sun, 7 Jun 2015 23:15:39 +0300
+Subject: [PATCH] WIP: *end of prologue* detection hack
+
+see
+  http://www.esp8266.com/viewtopic.php?p=18461#p18461
+  http://www.esp8266.com/viewtopic.php?p=19026#p19026
+  http://www.esp8266.com/viewtopic.php?p=19683#p19683
+
+Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
+---
+ gdb/xtensa-tdep.c |    2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+--- a/gdb/xtensa-tdep.c
++++ b/gdb/xtensa-tdep.c
+@@ -2385,7 +2385,7 @@
+   /* Find out, if we have an information about the prologue from DWARF.  */
+   prologue_sal = find_pc_line (start, 0);
+   if (prologue_sal.line != 0) /* Found debug info.  */
+-    body_pc = prologue_sal.end;
++    body_pc = prologue_sal.end + 40;
+ 
+   /* If we are going to analyze the prologue in general without knowing about
+      the current PC, make the best assumption for the end of the prologue.  */

packages/gdb/10.2/0004-allow-android.patch

@@ -0,0 +1,45 @@
+---
+ gdb/gdbserver/configure    |   11 -----------
+ gdb/gdbserver/configure.ac |   11 -----------
+ 2 files changed, 22 deletions(-)
+
+--- a/gdbserver/configure
++++ b/gdbserver/configure
+@@ -10069,17 +10069,6 @@
+ 
+ 
+ case "${target}" in
+-  *-android*)
+-    # Starting with NDK version 9, <elf.h> actually includes definitions
+-    # of Elf32_auxv_t and Elf64_auxv_t. But sadly, <elf.h> includes
+-    # <sys/exec_elf.h> which defines some of the ELF types incorrectly,
+-    # leading to conflicts with the defintions from <linux/elf.h>.
+-    # This makes it impossible for us to include both <elf.h> and
+-    # <linux/elf.h>, which means that, in practice, we do not have
+-    # access to Elf32_auxv_t and Elf64_auxv_t on this platform.
+-    # Therefore, do not try to auto-detect availability, as it would
+-    # get it wrong on this platform.
+-    ;;
+   *)
+     ac_fn_c_check_type "$LINENO" "Elf32_auxv_t" "ac_cv_type_Elf32_auxv_t" "#include <elf.h>
+ 
+--- a/gdbserver/configure.ac
++++ b/gdbserver/configure.ac
+@@ -155,17 +155,6 @@
+ ])
+ 
+ case "${target}" in
+-  *-android*)
+-    # Starting with NDK version 9, <elf.h> actually includes definitions
+-    # of Elf32_auxv_t and Elf64_auxv_t. But sadly, <elf.h> includes
+-    # <sys/exec_elf.h> which defines some of the ELF types incorrectly,
+-    # leading to conflicts with the defintions from <linux/elf.h>.
+-    # This makes it impossible for us to include both <elf.h> and
+-    # <linux/elf.h>, which means that, in practice, we do not have
+-    # access to Elf32_auxv_t and Elf64_auxv_t on this platform.
+-    # Therefore, do not try to auto-detect availability, as it would
+-    # get it wrong on this platform.
+-    ;;
+   *)
+     AC_CHECK_TYPES([Elf32_auxv_t, Elf64_auxv_t], [], [],
+     #include <elf.h>

packages/gdb/10.2/0005-xtensa-Fix-compilation-of-gdbserver.patch

@@ -0,0 +1,35 @@
+From 6f110500ed04148e3734f5b6a4fc573477c8398a Mon Sep 17 00:00:00 2001
+From: Alexey Brodkin <abrodkin@synopsys.com>
+Date: Tue, 24 Aug 2021 12:29:18 -0700
+Subject: [PATCH] xtensa: Fix compilation of gdbserver
+
+Otherwise a lot of errors like the following happen:
+--------------------------->8---------------------------------
+Error:     .../gdb/xtensa-linux-nat.c:66:17: error: 'gdbarch_pc_regnum' was not declared in this scope
+Error:     .../gdb/xtensa-linux-nat.c:68:17: error: 'gdbarch_ps_regnum' was not declared in this scope
+Error:     .../gdb/xtensa-linux-nat.c:71:38: error: no matching function for call to 'gdbarch_tdep::gdbarch_tdep(gdbarch*&)'
+...
+--------------------------->8---------------------------------
+
+Initially caught by https://github.com/crosstool-ng/crosstool-ng/pull/1573/checks?check_run_id=3361561567
+
+Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
+---
+ gdb/xtensa-linux-nat.c | 1 +
+ 1 file changed, 1 insertion(+)
+
+diff --git a/gdb/xtensa-linux-nat.c b/gdb/xtensa-linux-nat.c
+index b2b3db182bc..03f3ba7b50a 100644
+--- a/gdb/xtensa-linux-nat.c
++++ b/gdb/xtensa-linux-nat.c
+@@ -34,6 +34,7 @@
+ #include "nat/gdb_ptrace.h"
+ #include <asm/ptrace.h>
+
++#include "gdbarch.h"
+ #include "gregset.h"
+ #include "xtensa-tdep.h"
+
+--
+2.16.2
+

packages/gdb/10.2/chksum

@@ -0,0 +1,8 @@
+md5 gdb-10.2.tar.xz c044b7146903ec51c9d2337a29aee93b
+sha1 gdb-10.2.tar.xz 1056e2743a825ecce46ec9eec37f0b357831012b
+sha256 gdb-10.2.tar.xz aaa1223d534c9b700a8bec952d9748ee1977513f178727e1bee520ee000b4f29
+sha512 gdb-10.2.tar.xz 3653762ac008e065c37cd641653184c9ff7ce51ee2222ade1122bec9d6cc64dffd4fb74888ef11ac1942064a08910e96b7865112ad37f4602eb0a16bed074caa
+md5 gdb-10.2.tar.gz 7aeb896762924ae9a2ec59525088bada
+sha1 gdb-10.2.tar.gz 6bf5ee7877a4740835745ed97ce525a00bb2232c
+sha256 gdb-10.2.tar.gz b33ad58d687487a821ec8d878daab0f716be60d0936f2e3ac5cf08419ce70350
+sha512 gdb-10.2.tar.gz aa89caf47c1c84366020377d47e7c51ddbc48e5b7686f244e38797c8eb88411cf57fcdc37eb669961efb41ceeac4181747f429625fd1acce7712cb9a1fea9c41

packages/gdb/package.desc

@@ -2,5 +2,5 @@ origin='GNU'
 repository='git git://sourceware.org/git/binutils-gdb.git'
 mirrors='$(CT_Mirrors GNU gdb) $(CT_Mirrors sourceware gdb/releases)'
 relevantpattern='*.*|[a-z.]'
-milestones='7.0 7.2 7.12 8.0 8.3'
+milestones='7.0 7.2 7.12 8.0 8.3 10.2'
 archive_formats='.tar.xz .tar.gz'

scripts/build/debug/300-gdb.sh

@@ -50,6 +50,25 @@ do_debug_gdb_build()
             cross_extra_config+=("--enable-build-warnings=,-Wno-format-nonliteral,-Wno-format-security")
         fi
 
+        # Target libexpat resides in sysroot and does not have
+        # any dependencies, so just passing '-lexpat' to gcc is enough.
+        #
+        # By default gdb configure looks for expat in '$prefix/lib'
+        # directory. In our case '$prefix/lib' resolves to '/usr/lib'
+        # where libexpat for build platform lives, which is
+        # unacceptable for cross-compiling.
+        #
+        # To prevent this '--without-libexpat-prefix' flag must be passed.
+        # Thus configure falls back to '-lexpat', which is exactly what we want.
+        #
+        # NOTE: DO NOT USE --with-libexpat-prefix (until GDB configure is smarter)!!!
+        # It conflicts with a static build: GDB's configure script will find the shared
+        # version of expat and will attempt to link that, despite the -static flag.
+        # The link will fail, and configure will abort with "expat missing or unusable"
+        # message.
+        extra_config+=("--with-expat")
+        extra_config+=("--without-libexpat-prefix")
+
         do_gdb_backend \
             buildtype=cross \
             host="${CT_HOST}" \
@@ -88,7 +107,7 @@ do_debug_gdb_build()
         CT_EndStep
     fi
 
-    if [ "${CT_GDB_NATIVE}" = "y" -o "${CT_GDB_GDBSERVER}" = "y" ]; then
+    if [ "${CT_GDB_NATIVE}" = "y" ]; then
         local -a native_extra_config
         local subdir
 
@@ -102,17 +121,78 @@ do_debug_gdb_build()
             native_extra_config+=("--with-curses")
         fi
 
-        # Build a native gdbserver if needed. If building only
-        # gdbserver, configure in the subdirectory.
-        # Newer versions enable it automatically for a native target by default.
-        if [ "${CT_GDB_GDBSERVER}" != "y" ]; then
-            native_extra_config+=("--disable-gdbserver")
+        if [ "${CT_GDB_NATIVE_BUILD_IPA_LIB}" = "y" ]; then
+            native_extra_config+=("--enable-inprocess-agent")
         else
-            native_extra_config+=("--enable-gdbserver")
-            if [ "${CT_GDB_NATIVE}" != "y" ]; then
+            native_extra_config+=("--disable-inprocess-agent")
+        fi
+
+        export ac_cv_func_strncmp_works=yes
+
+        # TBD do we need all these? Eg why do we disable TUI if we build curses for target?
+        native_extra_config+=(
+            --without-uiout
+            --disable-tui
+            --disable-gdbtk
+            --without-x
+            --disable-sim
+            --without-included-gettext
+            --without-develop
+            --sysconfdir=/etc
+            --localstatedir=/var
+        )
+
+        # Target libexpat resides in sysroot and does not have
+        # any dependencies, so just passing '-lexpat' to gcc is enough.
+        #
+        # By default gdb configure looks for expat in '$prefix/lib'
+        # directory. In our case '$prefix/lib' resolves to '/usr/lib'
+        # where libexpat for build platform lives, which is
+        # unacceptable for cross-compiling.
+        #
+        # To prevent this '--without-libexpat-prefix' flag must be passed.
+        # Thus configure falls back to '-lexpat', which is exactly what we want.
+        #
+        # NOTE: DO NOT USE --with-libexpat-prefix (until GDB configure is smarter)!!!
+        # It conflicts with a static build: GDB's configure script will find the shared
+        # version of expat and will attempt to link that, despite the -static flag.
+        # The link will fail, and configure will abort with "expat missing or unusable"
+        # message.
+        extra_config+=("--with-expat")
+        extra_config+=("--without-libexpat-prefix")
+
+        do_gdb_backend \
+            buildtype=native \
+            subdir=${subdir} \
+            host="${CT_TARGET}" \
+            cflags="${CT_ALL_TARGET_CFLAGS}" \
+            ldflags="${CT_ALL_TARGET_LDFLAGS}" \
+            static="${CT_GDB_NATIVE_STATIC}" \
+            static_libstdc="${CT_GDB_NATIVE_STATIC_LIBSTDC}" \
+            prefix=/usr \
+            destdir="${CT_DEBUGROOT_DIR}" \
+            "${native_extra_config[@]}"
+
+        unset ac_cv_func_strncmp_works
+
+        CT_Popd
+        CT_EndStep # native gdb build
+    fi
+
+    if [ "${CT_GDB_GDBSERVER}" = "y" ]; then
+        local -a native_extra_config
+        local subdir
+
+        if [ "${CT_GDB_GDBSERVER_TOPLEVEL}" != "y" ]; then
             subdir=gdb/gdbserver/
         fi
-        fi
+
+        CT_DoStep INFO "Installing gdb server"
+        CT_mkdir_pushd "${CT_BUILD_DIR}/build-gdb-server"
+
+        native_extra_config+=("--program-prefix=")
+        native_extra_config+=("--enable-gdbserver")
+
         if [ "${CT_GDB_NATIVE_BUILD_IPA_LIB}" = "y" ]; then
             native_extra_config+=("--enable-inprocess-agent")
         else
@@ -149,7 +229,7 @@ do_debug_gdb_build()
         unset ac_cv_func_strncmp_works
 
         CT_Popd
-        CT_EndStep # native gdb build
+        CT_EndStep # gdb server build
     fi
 }
 
@@ -208,25 +288,6 @@ do_gdb_backend()
         extra_config+=("--disable-nls")
     fi
 
-    # Target libexpat resides in sysroot and does not have
-    # any dependencies, so just passing '-lexpat' to gcc is enough.
-    #
-    # By default gdb configure looks for expat in '$prefix/lib'
-    # directory. In our case '$prefix/lib' resolves to '/usr/lib'
-    # where libexpat for build platform lives, which is
-    # unacceptable for cross-compiling.
-    #
-    # To prevent this '--without-libexpat-prefix' flag must be passed.
-    # Thus configure falls back to '-lexpat', which is exactly what we want.
-    #
-    # NOTE: DO NOT USE --with-libexpat-prefix (until GDB configure is smarter)!!!
-    # It conflicts with a static build: GDB's configure script will find the shared
-    # version of expat and will attempt to link that, despite the -static flag.
-    # The link will fail, and configure will abort with "expat missing or unusable"
-    # message.
-    extra_config+=("--with-expat")
-    extra_config+=("--without-libexpat-prefix")
-
     if [ "${static}" = "y" ]; then
         cflags+=" -static"
         ldflags+=" -static"