[PATCH] arm64: kgdb: handle read-only text / modules

[AKASHI Takahiro]

On Tue, Sep 20, 2016 at 11:33:34AM +0100, Mark Rutland wrote:
> On Tue, Sep 20, 2016 at 07:03:21PM +0900, AKASHI Takahiro wrote:
> > Handle read-only cases (CONFIG_DEBUG_RODATA/CONFIG_DEBUG_SET_MODULE_RONX)
> > by using aarch64_insn_write() instead of probe_kernel_write().
> > See how this works:
> >     commit 2f896d586610 ("arm64: use fixmap for text patching")
> > 
> > Signed-off-by: AKASHI Takahiro <takahiro.akashi at linaro.org>
> > Cc: Catalin Marinas <catalin.marinas at arm.com>
> > Cc: Will Deacon <will.deacon at arm.com>
> > Cc: Jason Wessel <jason.wessel at windriver.com>
> > Cc: <stable at vger.kernel.org> # 4.0-
> 
> We had SET_MODULE_RONX in v3.17, and we had KGDB before that, so we need
> something for v3.17+.

Right, but 3.18+ :)
Unfortunately, the patch ("arm64: use fixmap for text patching")
is merged only in v4.0 or later. So it is also a pre-requisite.

> > ---
> >  arch/arm64/kernel/kgdb.c | 20 ++++++++++++++++++++
> >  1 file changed, 20 insertions(+)
> > 
> > diff --git a/arch/arm64/kernel/kgdb.c b/arch/arm64/kernel/kgdb.c
> > index 6732a27..133cfe3 100644
> > --- a/arch/arm64/kernel/kgdb.c
> > +++ b/arch/arm64/kernel/kgdb.c
> > @@ -382,3 +382,23 @@ struct kgdb_arch arch_kgdb_ops = {
> >  		KGDB_DYN_BRK_INS_BYTE(3),
> >  	}
> >  };
> > +int kgdb_arch_set_breakpoint(struct kgdb_bkpt *bpt)
> > +{
> > +	int err;
> > +
> > +	BUILD_BUG_ON(AARCH64_INSN_SIZE != BREAK_INSTR_SIZE);
> > +
> > +	err = aarch64_insn_read((void *)bpt->bpt_addr, (u32 *)bpt->saved_instr);
> > +	if (err)
> > +		return err;
> > +
> > +	return aarch64_insn_write((void *)bpt->bpt_addr,
> > +			(u32)AARCH64_BREAK_KGDB_DYN_DBG);
> > +}
> 
> This changes the endianness of saved_instr (on BE), but it looks like
> that's handed as an opaque token by the core code anyway, so that should
> be fine.
> 
> This also renders arch_kgdb_ops.gdb_bpt_instr unused. Can/should we get
> rid of that?

Yes, we can. But arch_kgdb_ops is still needed for compiling anyway.

> > +int kgdb_arch_remove_breakpoint(struct kgdb_bkpt *bpt)
> > +{
> > +	return aarch64_insn_write((void *)bpt->bpt_addr,
> > +			*(u32 *)bpt->saved_instr);
> > +}
> 
> We also need a few additional includes:
> 
> <asm/debug-monitors.h> # for BREAK_INSTR_SIZE, AARCH64_BREAK_KGDB_DYN_DBG
> <asm/insn.h> # for AARCH64_INSN_SIZE, insn_{read,write}
> <linux/bug.h> # for BUILD_BUG_ON()

Added.

> I take it that we're protected against nesting within
> aarch64_insn_write(), so that we can't deadlock on patch_lock?
> 
> Other than that, this looks good to me.

Thanks,
-Takahiro AKASHI

> Thanks,
> Mark.