Re: [Crash-utility] using crash for ARM

----- Original Message ----- I do not have an ARM machine to test with, but it is my understanding that if you attempt to build on an ARM host, it should just work by entering "make" alone, i.e., just like any of the other architectures. In other words, the --host and --target values should default to the actual host. Although you haven't stated it specifically, it sounds like you are trying to build an ARM binary on an x86 or x86_64 host? The problem with that is that there is currently no capability of changing the host compiler that is used to build the top-level crash source files. The crash sources themselves are always compiled using the host's default compiler. Passing --host and --target args to the gdb build would only address the embedded gdb module. You could try modifying GDB_TARGET_ARM_ON_X86 or GDB_TARGET_ARM_ON_X86_64 (whichever is appropriate) in configure.c with whatever you want. As they are now, both #define's presume that the host is either x86 or x86_64, and therefore --host is not passed: #define GDB_TARGET_ARM_ON_X86 "GDB_CONF_FLAGS=--target=arm-elf-linux" #define GDB_TARGET_ARM_ON_X86_64 "GDB_CONF_FLAGS=--target=arm-elf-linux CFLAGS=-m32" For the crash sources you would have to set things up such that "cc" points to your cross-compiler. You could try changing the Makefile to assign CC= to your cross-compiler, and then change all of the "cc" commands to ${CC}. The ARM developers on this mailing list may have some input on this -- I've added a few of them to the cc list. Dave