Hi Ming,
sorry for the delay.
On 2023/11/08 10:39, Ming Wang wrote:
 @@ -512,6 +527,10 @@ get_current_configuration(struct
supported_gdb_version *sp)
   		    (target_data.target != MIPS64))
   			arch_mismatch(sp);
   
 +		if ((target_data.initial_gdb_target == LOONGARCH64) &&
 +			(target_data.target != LOONGARCH64))
 +			arch_mismatch(sp);
 +
 
Please add the "if (target_data.target == X86_64)" block same as
RISCV64 and etc. otherwise implied (sticky) target does not work:
$ make target=LOONGARCH64
...
$ make     # without target
The initial build in this source tree was for the LOONGARCH64 architecture.
In order to build a crash binary for the X86_64 architecture:
  1. remove the gdb-10.2 subdirectory
  2. perform a "make clean"
  3. retry the build
make: *** [Makefile:253: all] Error 1
$
 @@ -2227,6 +2233,8 @@ struct offset_table {                    /*
stash of commonly-used offsets */
   	long module_memory_size;
   	long irq_data_irq;
   	long zspage_huge;
 +	long task_struct_thread_reg01;
 +	long task_struct_thread_reg03;
 
These are not used in this patch, could you move to 05/10?
 diff --git a/gdb-10.2.patch b/gdb-10.2.patch
 index d81030d..c28fc38 100644
 
Could you add "gdb-10.2/bfd/elf-bfd.h" and "gdb-10.2/gdb/objfiles.h"
to the list of the restored files at the beginning of gdb-10.2.patch?
Please see the comment there for details.  Now they each have two fixes.
Thanks,
Kazu