From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:34204) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1ebIBW-0005z9-Gc for qemu-devel@nongnu.org; Mon, 15 Jan 2018 22:45:15 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1ebIBV-00078s-Ju for qemu-devel@nongnu.org; Mon, 15 Jan 2018 22:45:14 -0500 Received: from mail-pl0-x241.google.com ([2607:f8b0:400e:c01::241]:39996) by eggs.gnu.org with esmtps (TLS1.0:RSA_AES_128_CBC_SHA1:16) (Exim 4.71) (envelope-from ) id 1ebIBV-00078J-Cb for qemu-devel@nongnu.org; Mon, 15 Jan 2018 22:45:13 -0500 Received: by mail-pl0-x241.google.com with SMTP id 62so5141614pld.7 for ; Mon, 15 Jan 2018 19:45:13 -0800 (PST) From: Richard Henderson Date: Mon, 15 Jan 2018 19:45:03 -0800 Message-Id: <20180116034503.31926-3-richard.henderson@linaro.org> In-Reply-To: <20180116034503.31926-1-richard.henderson@linaro.org> References: <20180116034503.31926-1-richard.henderson@linaro.org> Subject: [Qemu-devel] [PATCH v2 2/4] tcg/arm: Support tlb offsets larger than 64k List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel@nongnu.org Cc: peter.maydell@linaro.org AArch64 with SVE has an offset of 80k to the 8th TLB. Signed-off-by: Richard Henderson --- tcg/arm/tcg-target.inc.c | 30 +++++++++++++++++------------- 1 file changed, 17 insertions(+), 13 deletions(-) diff --git a/tcg/arm/tcg-target.inc.c b/tcg/arm/tcg-target.inc.c index 0ff283d84f..8f5d4f208d 100644 --- a/tcg/arm/tcg-target.inc.c +++ b/tcg/arm/tcg-target.inc.c @@ -1246,12 +1246,6 @@ static TCGReg tcg_out_arg_reg64(TCGContext *s, TCGReg argreg, /* We're expecting to use an 8-bit immediate and to mask. */ QEMU_BUILD_BUG_ON(CPU_TLB_BITS > 8); -/* We're expecting to use an 8-bit immediate add + 8-bit ldrd offset. - Using the offset of the second entry in the last tlb table ensures - that we can index all of the elements of the first entry. */ -QEMU_BUILD_BUG_ON(offsetof(CPUArchState, tlb_table[NB_MMU_MODES - 1][1]) - > 0xffff); - /* Load and compare a TLB entry, leaving the flags set. Returns the register containing the addend of the tlb entry. Clobbers R0, R1, R2, TMP. */ @@ -1264,6 +1258,7 @@ static TCGReg tcg_out_tlb_read(TCGContext *s, TCGReg addrlo, TCGReg addrhi, ? offsetof(CPUArchState, tlb_table[mem_index][0].addr_read) : offsetof(CPUArchState, tlb_table[mem_index][0].addr_write)); int add_off = offsetof(CPUArchState, tlb_table[mem_index][0].addend); + int mask_off; unsigned s_bits = opc & MO_SIZE; unsigned a_bits = get_alignment_bits(opc); @@ -1295,16 +1290,25 @@ static TCGReg tcg_out_tlb_read(TCGContext *s, TCGReg addrlo, TCGReg addrhi, 0, addrlo, SHIFT_IMM_LSR(TARGET_PAGE_BITS)); } - /* We checked that the offset is contained within 16 bits above. */ - if (add_off > 0xfff - || (use_armv6_instructions && TARGET_LONG_BITS == 64 - && cmp_off > 0xff)) { + /* Add portions of the offset until the memory access is in range. + * If we plan on using ldrd, reduce to an 8-bit offset; otherwise + * we can use a 12-bit offset. */ + if (use_armv6_instructions && TARGET_LONG_BITS == 64) { + mask_off = 0xff; + } else { + mask_off = 0xfff; + } + while (cmp_off > mask_off) { + int shift = ctz32(cmp_off & ~mask_off) & ~1; + int rot = ((32 - shift) << 7) & 0xf00; + int addend = cmp_off & (0xff << shift); tcg_out_dat_imm(s, COND_AL, ARITH_ADD, TCG_REG_R2, base, - (24 << 7) | (cmp_off >> 8)); + rot | ((cmp_off >> shift) & 0xff)); base = TCG_REG_R2; - add_off -= cmp_off & 0xff00; - cmp_off &= 0xff; + add_off -= addend; + cmp_off -= addend; } + if (!use_armv7_instructions) { tcg_out_dat_imm(s, COND_AL, ARITH_AND, TCG_REG_R0, TCG_REG_TMP, CPU_TLB_SIZE - 1); -- 2.14.3