[PATCH for-stable-v4.4 37/53] ARM: spectre-v1: use get_user() for __get_user()

From: Ard Biesheuvel <ardb@kernel.org>
To: linux-arm-kernel@lists.infradead.org
Cc: Mark Rutland <mark.rutland@arm.com>,
	Florian Fainelli <f.fainelli@gmail.com>,
	Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	Tony Lindgren <tony@atomide.com>,
	Catalin Marinas <catalin.marinas@arm.com>,
	Russell King <rmk+kernel@armlinux.org.uk>,
	"David A . Long" <dave.long@linaro.org>,
	Marc Zyngier <maz@kernel.org>, Will Deacon <will@kernel.org>,
	Ard Biesheuvel <ardb@kernel.org>
Subject: [PATCH for-stable-v4.4 37/53] ARM: spectre-v1: use get_user() for __get_user()
Date: Tue,  5 Nov 2019 21:58:30 +0100	[thread overview]
Message-ID: <20191105205846.1394-38-ardb@kernel.org> (raw)
In-Reply-To: <20191105205846.1394-1-ardb@kernel.org>

From: Russell King <rmk+kernel@armlinux.org.uk>

Commit b1cd0a14806321721aae45f5446ed83a3647c914 upstream.

Fixing __get_user() for spectre variant 1 is not sane: we would have to
add address space bounds checking in order to validate that the location
should be accessed, and then zero the address if found to be invalid.

Since __get_user() is supposed to avoid the bounds check, and this is
exactly what get_user() does, there's no point having two different
implementations that are doing the same thing.  So, when the Spectre
workarounds are required, make __get_user() an alias of get_user().

Acked-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: David A. Long <dave.long@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
---
 arch/arm/include/asm/uaccess.h | 17 +++++++++++------
 1 file changed, 11 insertions(+), 6 deletions(-)

diff --git a/arch/arm/include/asm/uaccess.h b/arch/arm/include/asm/uaccess.h
index 99005567fb92..fd33021da6f6 100644
--- a/arch/arm/include/asm/uaccess.h
+++ b/arch/arm/include/asm/uaccess.h
@@ -288,6 +288,16 @@ static inline void set_fs(mm_segment_t fs)
 #define user_addr_max() \
 	(segment_eq(get_fs(), KERNEL_DS) ? ~0UL : get_fs())
 
+#ifdef CONFIG_CPU_SPECTRE
+/*
+ * When mitigating Spectre variant 1, it is not worth fixing the non-
+ * verifying accessors, because we need to add verification of the
+ * address space there.  Force these to use the standard get_user()
+ * version instead.
+ */
+#define __get_user(x, ptr) get_user(x, ptr)
+#else
+
 /*
  * The "__xxx" versions of the user access functions do not verify the
  * address space - it must have been done previously with a separate
@@ -304,12 +314,6 @@ static inline void set_fs(mm_segment_t fs)
 	__gu_err;							\
 })
 
-#define __get_user_error(x, ptr, err)					\
-({									\
-	__get_user_err((x), (ptr), err);				\
-	(void) 0;							\
-})
-
 #define __get_user_err(x, ptr, err)					\
 do {									\
 	unsigned long __gu_addr = (unsigned long)(ptr);			\
@@ -369,6 +373,7 @@ do {									\
 
 #define __get_user_asm_word(x, addr, err)			\
 	__get_user_asm(x, addr, err, ldr)
+#endif
 
 
 #define __put_user_switch(x, ptr, __err, __fn)				\
-- 
2.17.1


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
