On Tue, Mar 16, 2010 at 10:40, <[email protected]> wrote:
> Added: trunk/arch/blackfin/include/asm/pseudo_instructions.h (0 => 8470)
>
> @@ -0,0 +1 @@
> +extern bool execute_pseudodbg_assert(struct pt_regs *fp, unsigned int
> opcode);
needs a comment header, and include some linux/ headers for "bool" and
"struct pt_regs"
> Added: trunk/arch/blackfin/kernel/pseudodgba.c (0 => 8470)
>
> @@ -0,0 +1,58 @@
> +/* The fake debug assert instructions
> + *
> + * Copyright 2004-2009 Analog Devices Inc.
should it be "2010" instead ?
> +#include <linux/ptrace.h>
this thing needs at least linux/types.h and linux/kernel.h i think ...
> Modified: trunk/arch/blackfin/kernel/traps.c (8469 => 8470)
>
> @@ -14,6 +14,9 @@
> #include <linux/irq.h>
> #include <asm/trace.h>
> #include <asm/fixed_code.h>
> +#ifdef CONFIG_BFIN_PSEUDO_DBGA
> +#include <asm/pseudo_instructions.h>
> +#endif
i dont think the #ifdef is needed
> +#ifdef CONFIG_BFIN_PSEUDO_DBGA
> + /*
> + * Support for the fake instructions, if the instruction fails,
> + * then just execute a illegal opcode failure (like normal).
> + * Don't support these instructions inside the kernel
> + */
> + if (!kernel_mode_regs(fp) && get_instruction(&opcode, (unsigned
> short
> *)fp->pc)) {
> + if (execute_pseudodbg_assert(fp, opcode)) {
> + fp->pc += 4;
> + goto traps_done;
> + }
> + }
> +#endif
perhaps the pc update should be in the execute_pseudodbg_assert() instead ?
-mike
_______________________________________________
Linux-kernel-commits mailing list
[email protected]
https://blackfin.uclinux.org/mailman/listinfo/linux-kernel-commits
