Kernel+LibC: Allow clock_gettime() to run without syscalls

This patch adds a vDSO-like mechanism for exposing the current time as
an array of per-clock-source timestamps.

LibC's clock_gettime() calls sys$map_time_page() to map the kernel's
"time page" into the process address space (at a random address, ofc.)
This is only done on first call, and from then on the timestamps are
fetched from the time page.

This first patch only adds support for CLOCK_REALTIME, but eventually
we should be able to support all clock sources this way and get rid of
sys$clock_gettime() in the kernel entirely. :^)

Accesses are synchronized using two atomic integers that are incremented
at the start and finish of the kernel's time page update cycle.

Kernel/API/Syscall.h

@@ -118,6 +118,7 @@ enum class NeedsBigProcessLock {
     S(listen, NeedsBigProcessLock::Yes)                     \
     S(lseek, NeedsBigProcessLock::Yes)                      \
     S(madvise, NeedsBigProcessLock::Yes)                    \
+    S(map_time_page, NeedsBigProcessLock::Yes)              \
     S(mkdir, NeedsBigProcessLock::Yes)                      \
     S(mknod, NeedsBigProcessLock::Yes)                      \
     S(mmap, NeedsBigProcessLock::Yes)                       \

Kernel/API/TimePage.h

@@ -0,0 +1,30 @@
+/*
+ * Copyright (c) 2021, Andreas Kling <kling@serenityos.org>
+ *
+ * SPDX-License-Identifier: BSD-2-Clause
+ */
+
+#pragma once
+
+#include <AK/Types.h>
+
+#ifdef KERNEL
+#    include <Kernel/UnixTypes.h>
+#else
+#    include <time.h>
+#endif
+
+namespace Kernel {
+
+inline bool time_page_supports(clockid_t clock_id)
+{
+    return clock_id == CLOCK_REALTIME;
+}
+
+struct TimePage {
+    volatile u32 update1;
+    struct timespec clocks[CLOCK_ID_COUNT];
+    volatile u32 update2;
+};
+
+}