perf/core: Force USER_DS when recording user stack data

Perf can record user stack data in response to a synchronous request, such
as a tracepoint firing. If this happens under set_fs(KERNEL_DS), then we
end up reading user stack data using __copy_from_user_inatomic() under
set_fs(KERNEL_DS). I think this conflicts with the intention of using
set_fs(KERNEL_DS). And it is explicitly forbidden by hardware on ARM64
when both CONFIG_ARM64_UAO and CONFIG_ARM64_PAN are used.

So fix this by forcing USER_DS when recording user stack data.

Signed-off-by: Yabin Cui <[e-mail]>
Acked-by: Peter Zijlstra (Intel) <[e-mail]>
Cc: <[e-mail]>
Cc: Alexander Shishkin <[e-mail]>
Cc: Arnaldo Carvalho de Melo <[e-mail]>
Cc: Jiri Olsa <[e-mail]>
Cc: Linus Torvalds <[e-mail]>
Cc: Namhyung Kim <[e-mail]>
Cc: Peter Zijlstra <[e-mail]>
Cc: Thomas Gleixner <[e-mail]>
Fixes: 88b0193d9418 ("perf/callchain: Force USER_DS when invoking perf_callchain_user()")
Link: http://lkml.kernel.org/r/[e-mail]
Signed-off-by: Ingo Molnar <[e-mail]>

kernel/events/core.c

···
 		unsigned long sp;
 		unsigned int rem;
 		u64 dyn_size;
+		mm_segment_t fs;
 
 		/*
 		 * We dump:
···
 
 		/* Data. */
 		sp = perf_user_stack_pointer(regs);
+		fs = get_fs();
+		set_fs(USER_DS);
 		rem = __output_copy_user(handle, (void *) sp, dump_size);
+		set_fs(fs);
 		dyn_size = dump_size - rem;
 
 		perf_output_skip(handle, rem);