Re: [PATCH v2 4/9] kgdb: Delay "kgdbwait" to dbg_late_init() by default

[Daniel Thompson]

On Tue, Apr 21, 2020 at 02:14:42PM -0700, Douglas Anderson wrote:
> Using kgdb requires at least some level of architecture-level
> initialization.  If nothing else, it relies on the architecture to
> pass breakpoints / crashes onto kgdb.
> 
> On some architectures this all works super early, specifically it
> starts working at some point in time before Linux parses
> early_params's.  On other architectures it doesn't.  A survey of a few
> platforms:
> 
> a) x86: Presumably it all works early since "ekgdboc" is documented to
>    work here.
> b) arm64: Catching crashes works; with a simple patch breakpoints can
>    also be made to work.
> c) arm: Nothing in kgdb works until
>    paging_init() -> devicemaps_init() -> early_trap_init()
> 
> Let's be conservative and, by default, process "kgdbwait" (which tells
> the kernel to drop into the debugger ASAP at boot) a bit later at
> dbg_late_init() time.  If an architecture has tested it and wants to
> re-enable super early debugging, they can implement the weak function
> kgdb_arch_can_debug_early() to return true.  We'll do this for x86 to
> start.  It should be noted that dbg_late_init() is still called quite
> early in the system.
> 
> Note that this patch doesn't affect when kgdb runs its init.  If kgdb
> is set to initialize early it will still initialize when parsing
> early_params's.  This patch _only_ inhibits the initial breakpoint
> from "kgdbwait".  This means:
> 
> * Without any extra patches arm64 platforms will at least catch
>   crashes after kgdb inits.
> * arm platforms will catch crashes (and could handle a hardcoded
>   kgdb_breakpoint()) any time after early_trap_init() runs, even
>   before dbg_late_init().
> 
> Signed-off-by: Douglas Anderson <dianders@xxxxxxxxxxxx>
> Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
> Cc: Ingo Molnar <mingo@xxxxxxxxxx>
> Cc: Borislav Petkov <bp@xxxxxxxxx>

Overall this looks good but there is a small quibble below...


> diff --git a/include/linux/kgdb.h b/include/linux/kgdb.h
> index b072aeb1fd78..7371517aeacc 100644
> --- a/include/linux/kgdb.h
> +++ b/include/linux/kgdb.h
> @@ -226,6 +226,28 @@ extern int kgdb_arch_remove_breakpoint(struct kgdb_bkpt *bpt);
>   */
>  extern void kgdb_arch_late(void);
>  
> +/**
> + *	kgdb_arch_can_debug_early - Check if OK to break before dbg_late_init()
> + *
> + *	If an architecture can definitely handle entering the debugger when
> + *	early_param's are parsed then it can override this function to return
> + *	true.  Otherwise if "kgdbwait" is passed on the kernel command line it
> + *	won't actually be processed until dbg_late_init() just after the call
> + *	to kgdb_arch_late() is made.
> + *
> + *	NOTE: Even if this returns false we will still try to register kgdb to
> + *	handle breakpoints and crashes when early_params's are parsed, we just
> + *	won't act on the "kgdbwait" parameter until dbg_late_init().  If you
> + *	get a crash and try to drop into kgdb somewhere between these two
> + *	places you might or might not end up being able to use kgdb depending
> + *	on exactly how far along the architecture has initted.
> + *
> + *	ALSO: dbg_late_init() is actually still fairly early in the system
> + *	boot process.
> + *
> + *	Return: true if platform can handle kgdb early.
> + */
> +extern bool kgdb_arch_can_debug_early(void);

Does this need to be a function? It looks like all implementations are
either return true or return false (e.g. CONFIG_ARCH_HAVE_EARLY_DEBUG
would do the same thing).


Daniel.