| Message ID | 1399419149-26685-3-git-send-email-larry.bassel@linaro.org |
|---|---|
| State | New |
| Headers | show |
On Wed, May 07, 2014 at 12:32:29AM +0100, Larry Bassel wrote: > Make calls to ct_user_enter when the kernel is exited > and ct_user_exit when the kernel is entered (in el0_da, > el0_ia, el0_svc, el0_irq). Why only these entry points? I can reschedule after any exception from EL0, so I'd expect all exceptions from userspace to need annotating, no? > These macros expand to function calls which will only work > properly if el0_sync and related code has been rearranged > (in a previous patch of this series). > > In order to avoid saving registers, the slow syscall path > is forced (as x86 does). ... and if you decide to handle undef exceptions, I think you'll need the register saving too, in case the kernel needs to perform emulation. Will -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
On 07 May 14 11:17, Will Deacon wrote: > On Wed, May 07, 2014 at 12:32:29AM +0100, Larry Bassel wrote: > > Make calls to ct_user_enter when the kernel is exited > > and ct_user_exit when the kernel is entered (in el0_da, > > el0_ia, el0_svc, el0_irq). > > Why only these entry points? I can reschedule after any exception from EL0, > so I'd expect all exceptions from userspace to need annotating, no? > > > These macros expand to function calls which will only work > > properly if el0_sync and related code has been rearranged > > (in a previous patch of this series). > > > > In order to avoid saving registers, the slow syscall path > > is forced (as x86 does). > > ... and if you decide to handle undef exceptions, I think you'll need > the register saving too, in case the kernel needs to perform emulation. These are excellent points, I will rework the patch and submit v3. Thanks for the feedback. > > Will Larry -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
On Thu, May 08, 2014 at 12:49:04AM +0100, Kevin Hilman wrote: > Hi Will, Hello Kevin, > Will Deacon <will.deacon@arm.com> writes: > > On Wed, May 07, 2014 at 12:32:29AM +0100, Larry Bassel wrote: > >> Make calls to ct_user_enter when the kernel is exited > >> and ct_user_exit when the kernel is entered (in el0_da, > >> el0_ia, el0_svc, el0_irq). > > > > Why only these entry points? I can reschedule after any exception from EL0, > > so I'd expect all exceptions from userspace to need annotating, no? > > In my initial approach to this, you might recall (though it was over a > year ago now) was to just instrument kernel_enter rather than sprinkle > the instrumentaion in cl0_*. However, your concern at the time was that > since it was before debugging was enabled it would complicate debugging > these paths. > > Any chance you have any other suggestion on how we might do this in > kernel_entry rather than sprinkling them all over cl0_*? or is the > sprinkling the only good way to handle this. Unfortunately, different exceptions do subtly different things before invoking the main handler. For example: - Stashing the far - Enabling IRQs - Enabling debug - All the stuff on the syscall path so putting the logic in kernel_entry isn't really do-able, unfortunately. Will -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig index e6e4d37..152d92b 100644 --- a/arch/arm64/Kconfig +++ b/arch/arm64/Kconfig @@ -55,6 +55,7 @@ config ARM64 select RTC_LIB select SPARSE_IRQ select SYSCTL_EXCEPTION_TRACE + select HAVE_CONTEXT_TRACKING help ARM 64-bit (AArch64) Linux support. diff --git a/arch/arm64/include/asm/thread_info.h b/arch/arm64/include/asm/thread_info.h index 720e70b..301ea6a 100644 --- a/arch/arm64/include/asm/thread_info.h +++ b/arch/arm64/include/asm/thread_info.h @@ -108,6 +108,7 @@ static inline struct thread_info *current_thread_info(void) #define TIF_SINGLESTEP 21 #define TIF_32BIT 22 /* 32bit process */ #define TIF_SWITCH_MM 23 /* deferred switch_mm */ +#define TIF_NOHZ 24 #define _TIF_SIGPENDING (1 << TIF_SIGPENDING) #define _TIF_NEED_RESCHED (1 << TIF_NEED_RESCHED) diff --git a/arch/arm64/kernel/entry.S b/arch/arm64/kernel/entry.S index d920d7f..5fe447c 100644 --- a/arch/arm64/kernel/entry.S +++ b/arch/arm64/kernel/entry.S @@ -30,6 +30,22 @@ #include <asm/unistd32.h> /* + * Context tracking subsystem. Used to instrument transitions + * between user and kernel mode. + */ + .macro ct_user_exit +#ifdef CONFIG_CONTEXT_TRACKING + bl context_tracking_user_exit +#endif + .endm + + .macro ct_user_enter +#ifdef CONFIG_CONTEXT_TRACKING + bl context_tracking_user_enter +#endif + .endm + +/* * Bad Abort numbers *----------------- */ @@ -88,6 +104,7 @@ .macro kernel_exit, el, ret = 0 ldp x21, x22, [sp, #S_PC] // load ELR, SPSR .if \el == 0 + ct_user_enter ldr x23, [sp, #S_SP] // load return stack pointer .endif .if \ret @@ -427,6 +444,7 @@ el0_da: enable_dbg // enable interrupts before calling the main handler enable_irq + ct_user_exit mov x0, x26 bic x0, x0, #(0xff << 56) mov x1, x25 @@ -443,6 +461,7 @@ el0_ia: enable_dbg // enable interrupts before calling the main handler enable_irq + ct_user_exit mov x0, x26 orr x1, x25, #1 << 24 // use reserved ISS bit for instruction aborts mov x2, sp @@ -511,6 +530,7 @@ el0_irq_naked: bl trace_hardirqs_off #endif + ct_user_exit irq_handler get_thread_info tsk @@ -633,10 +653,12 @@ el0_svc_naked: // compat entry point isb enable_dbg enable_irq + ct_user_exit get_thread_info tsk ldr x16, [tsk, #TI_FLAGS] // check for syscall tracing tbnz x16, #TIF_SYSCALL_TRACE, __sys_trace // are we tracing syscalls? + tbnz x16, #TIF_NOHZ, __sys_trace adr lr, ret_fast_syscall // return address cmp scno, sc_nr // check upper syscall limit b.hs ni_sys