Re: Fixes for uaccess.h with gcc >= 4.0.1

[Richard Sandiford <richard@codesourcery.com>]

Ralf Baechle <ralf@linux-mips.org> writes:
> I'd appreciate if somebody with gcc 4.0.1 could test this kernel patch
> below.

Sorry in advance if this is a dup, but...

This patch caused a miscompilation of the restore_gp_regs() block
in restore_sigcontext().  This was in a 32-bit kernel compiled with
GCC CVS head.

restore_gp_regs() copies 64-bit user fields into 32-bit variables,
and in this combination, the new __get_user_asm_ll32() clobbers too
many registers.  It says:

/*
 * Get a long long 64 using 32 bit registers.
 */
#define __get_user_asm_ll32(val, addr)                                  \
{                                                                       \
        __asm__ __volatile__(                                           \
        "1:     lw      %1, (%3)                                \n"     \
        "2:     lw      %D1, 4(%3)                              \n"     \
        "       move    %0, $0                                  \n"     \
        "3:     .section        .fixup,\"ax\"                   \n"     \
        "4:     li      %0, %4                                  \n"     \
        "       move    %1, $0                                  \n"     \
        "       move    %D1, $0                                 \n"     \
        "       j       3b                                      \n"     \
        "       .previous                                       \n"     \
        "       .section        __ex_table,\"a\"                \n"     \
        "       " __UA_ADDR "   1b, 4b                          \n"     \
        "       " __UA_ADDR "   2b, 4b                          \n"     \
        "       .previous                                       \n"     \
        : "=r" (__gu_err), "=&r" (val)                                  \
        : "0" (0), "r" (addr), "i" (-EFAULT));                          \
}

and this requires val (%1) to be a 64-bit value.  In the case I saw,
gcc was using $3 for the 32-bit val, and wasn't expecting $4 to be
clobbered.

FWIW, the patch below fixes it for me.

Richard


diff --git a/include/asm-mips/uaccess.h b/include/asm-mips/uaccess.h
index 91d813a..1208cae 100644
--- a/include/asm-mips/uaccess.h
+++ b/include/asm-mips/uaccess.h
@@ -266,6 +266,7 @@ do {                                                        
                \
  */
 #define __get_user_asm_ll32(val, addr)                                 \
 {                                                                      \
+        unsigned long long __gu_tmp;                                   \
        __asm__ __volatile__(                                           \
        "1:     lw      %1, (%3)                                \n"     \
        "2:     lw      %D1, 4(%3)                              \n"     \
@@ -280,8 +281,9 @@ do {                                                        
                \
        "       " __UA_ADDR "   1b, 4b                          \n"     \
        "       " __UA_ADDR "   2b, 4b                          \n"     \
        "       .previous                                       \n"     \
-       : "=r" (__gu_err), "=&r" (val)                                  \
+       : "=r" (__gu_err), "=&r" (__gu_tmp)                             \
        : "0" (0), "r" (addr), "i" (-EFAULT));                          \
+       (val) = __gu_tmp;                                               \
 }
 
 /*