x86/tsc: Use fallback for random_get_entropy() instead of zero
authorJason A. Donenfeld <Jason@zx2c4.com>
Fri, 8 Apr 2022 16:03:13 +0000 (18:03 +0200)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Mon, 30 May 2022 07:29:14 +0000 (09:29 +0200)
commite017497815f3e676421a2afb0abe4ae03dbaad2a
tree970534e001e96dcba466f98dbb17a47bb1086e6d
parent38fbfa404da6118f9a86b7f3681eab09cfbfcb35
x86/tsc: Use fallback for random_get_entropy() instead of zero

commit 3bd4abc07a267e6a8b33d7f8717136e18f921c53 upstream.

In the event that random_get_entropy() can't access a cycle counter or
similar, falling back to returning 0 is suboptimal. Instead, fallback
to calling random_get_entropy_fallback(), which isn't extremely high
precision or guaranteed to be entropic, but is certainly better than
returning zero all the time.

If CONFIG_X86_TSC=n, then it's possible for the kernel to run on systems
without RDTSC, such as 486 and certain 586, so the fallback code is only
required for that case.

As well, fix up both the new function and the get_cycles() function from
which it was derived to use cpu_feature_enabled() rather than
boot_cpu_has(), and use !IS_ENABLED() instead of #ifndef.

Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Borislav Petkov <bp@alien8.de>
Cc: x86@kernel.org
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
arch/x86/include/asm/timex.h
arch/x86/include/asm/tsc.h