[SPARC64]: Fix pfn_pte() build failure.

The "%uhi" needs to be "%%uhi" because we want a real
"%" character in the assembler here, instead of an
assembler variable expansion.

Aparently older GCCs were more liberal and interpreted
this %-letter as a literal "%" for whatever reason.

Based upon a build failure report from Meelis Roos.

Signed-off-by: David S. Miller <davem@davemloft.net>

diff --git a/include/asm-sparc64/pgtable.h b/include/asm-sparc64/pgtable.h
index 1ba19eb34ce32120b7a71366ba9e7a401597c91b..ebfe395cfb87dab31b3c79a2b545d13709f8f368 100644 (file)
--- a/include/asm-sparc64/pgtable.h
+++ b/include/asm-sparc64/pgtable.h
@@ -234,7 +234,7 @@ static inline pte_t pfn_pte(unsigned long pfn, pgprot_t prot)
        sz_bits = 0UL;
        if (_PAGE_SZBITS_4U != 0UL || _PAGE_SZBITS_4V != 0UL) {
                __asm__ __volatile__(
-               "\n661: sethi           %uhi(%1), %0\n"
+               "\n661: sethi           %%uhi(%1), %0\n"
                "       sllx            %0, 32, %0\n"
                "       .section        .sun4v_2insn_patch, \"ax\"\n"
                "       .word           661b\n"