Hi Will,
Thanks for reviewing.
On Wed, Jul 31, 2019 at 05:08:37PM +0100, Will Deacon wrote:
> On Tue, Jul 16, 2019 at 07:13:00PM +0800, Leo Yan wrote:
> > This patch implement regs_set_return_value() and
> > override_function_with_return() to support function error injection
> > for arm64.
> >
> > In the exception flow, arm64's general register x30 contains the value
> > for the link register; so we can just update pt_regs::pc with it rather
> > than redirecting execution to a dummy function that returns.
> >
> > This patch is heavily inspired by the commit 7cd01b08d35f ("powerpc:
> > Add support for function error injection").
> >
> > Signed-off-by: Leo Yan <leo.yan@xxxxxxxxxx>
> > ---
> > arch/arm64/Kconfig | 1 +
> > arch/arm64/include/asm/error-injection.h | 13 +++++++++++++
> > arch/arm64/include/asm/ptrace.h | 5 +++++
> > arch/arm64/lib/Makefile | 2 ++
> > arch/arm64/lib/error-inject.c | 19 +++++++++++++++++++
> > 5 files changed, 40 insertions(+)
> > create mode 100644 arch/arm64/include/asm/error-injection.h
> > create mode 100644 arch/arm64/lib/error-inject.c
> >
> > diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
> > index 697ea0510729..a6d9e622977d 100644
> > --- a/arch/arm64/Kconfig
> > +++ b/arch/arm64/Kconfig
> > @@ -142,6 +142,7 @@ config ARM64
> > select HAVE_EFFICIENT_UNALIGNED_ACCESS
> > select HAVE_FTRACE_MCOUNT_RECORD
> > select HAVE_FUNCTION_TRACER
> > + select HAVE_FUNCTION_ERROR_INJECTION
> > select HAVE_FUNCTION_GRAPH_TRACER
> > select HAVE_GCC_PLUGINS
> > select HAVE_HW_BREAKPOINT if PERF_EVENTS
> > diff --git a/arch/arm64/include/asm/error-injection.h b/arch/arm64/include/asm/error-injection.h
> > new file mode 100644
> > index 000000000000..da057e8ed224
> > --- /dev/null
> > +++ b/arch/arm64/include/asm/error-injection.h
> > @@ -0,0 +1,13 @@
> > +/* SPDX-License-Identifier: GPL-2.0+ */
> > +
> > +#ifndef __ASM_ERROR_INJECTION_H_
> > +#define __ASM_ERROR_INJECTION_H_
> > +
> > +#include <linux/compiler.h>
> > +#include <linux/linkage.h>
> > +#include <asm/ptrace.h>
> > +#include <asm-generic/error-injection.h>
> > +
> > +void override_function_with_return(struct pt_regs *regs);
> > +
> > +#endif /* __ASM_ERROR_INJECTION_H_ */
>
> Why isn't this prototype in the asm-generic header? Seems weird to have to
> duplicate it for each architecture.
Yeah. When I spin for new version patches, will try to refactor in
the asm-generic header.
> > diff --git a/arch/arm64/include/asm/ptrace.h b/arch/arm64/include/asm/ptrace.h
> > index dad858b6adc6..3aafbbe218a2 100644
> > --- a/arch/arm64/include/asm/ptrace.h
> > +++ b/arch/arm64/include/asm/ptrace.h
> > @@ -294,6 +294,11 @@ static inline unsigned long regs_return_value(struct pt_regs *regs)
> > return regs->regs[0];
> > }
> >
> > +static inline void regs_set_return_value(struct pt_regs *regs, unsigned long rc)
> > +{
> > + regs->regs[0] = rc;
> > +}
> > +
> > /**
> > * regs_get_kernel_argument() - get Nth function argument in kernel
> > * @regs: pt_regs of that context
> > diff --git a/arch/arm64/lib/Makefile b/arch/arm64/lib/Makefile
> > index 33c2a4abda04..f182ccb0438e 100644
> > --- a/arch/arm64/lib/Makefile
> > +++ b/arch/arm64/lib/Makefile
> > @@ -33,3 +33,5 @@ UBSAN_SANITIZE_atomic_ll_sc.o := n
> > lib-$(CONFIG_ARCH_HAS_UACCESS_FLUSHCACHE) += uaccess_flushcache.o
> >
> > obj-$(CONFIG_CRC32) += crc32.o
> > +
> > +obj-$(CONFIG_FUNCTION_ERROR_INJECTION) += error-inject.o
> > diff --git a/arch/arm64/lib/error-inject.c b/arch/arm64/lib/error-inject.c
> > new file mode 100644
> > index 000000000000..35661c2de4b0
> > --- /dev/null
> > +++ b/arch/arm64/lib/error-inject.c
> > @@ -0,0 +1,19 @@
> > +// SPDX-License-Identifier: GPL-2.0
> > +
> > +#include <linux/error-injection.h>
> > +#include <linux/kprobes.h>
> > +
> > +void override_function_with_return(struct pt_regs *regs)
> > +{
> > + /*
> > + * 'regs' represents the state on entry of a predefined function in
> > + * the kernel/module and which is captured on a kprobe.
> > + *
> > + * 'regs->regs[30]' contains the the link register for the probed
>
> extra "the"
Will fix.
> > + * function and assign it to 'regs->pc', so when kprobe returns
> > + * back from exception it will override the end of probed function
> > + * and drirectly return to the predefined function's caller.
>
> directly
Will fix.
> > + */
> > + regs->pc = regs->regs[30];
>
> I suppose we could be all fancy and do:
>
> instruction_pointer_set(regs, procedure_link_pointer(regs));
>
> How about that?
Ah, good point. Will change to use the common APIs.
Thanks,
Leo Yan
