qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Alvise Rigo <a.rigo@virtualopensystems.com>
To: qemu-devel@nongnu.org
Cc: mttcg@listserver.greensocs.com, claudio.fontana@huawei.com,
	cota@braap.org, jani.kokkonen@huawei.com,
	tech@virtualopensystems.com, alex.bennee@linaro.org,
	rth@twiddle.net
Subject: [Qemu-devel] [RFC v2 7/7] target-i386: translate: implement qemu_ldlink and qemu_stcond ops
Date: Mon, 15 Jun 2015 13:51:28 +0200	[thread overview]
Message-ID: <1434369088-15076-8-git-send-email-a.rigo@virtualopensystems.com> (raw)
In-Reply-To: <1434369088-15076-1-git-send-email-a.rigo@virtualopensystems.com>

Implement strex and ldrex instruction relying on TCG's qemu_ldlink and
qemu_stcond.  For the time being only 32bit configurations are supported.

Suggested-by: Jani Kokkonen <jani.kokkonen@huawei.com>
Suggested-by: Claudio Fontana <claudio.fontana@huawei.com>
Signed-off-by: Alvise Rigo <a.rigo@virtualopensystems.com>
---
 tcg/i386/tcg-target.c | 136 ++++++++++++++++++++++++++++++++++++++++++--------
 1 file changed, 114 insertions(+), 22 deletions(-)

diff --git a/tcg/i386/tcg-target.c b/tcg/i386/tcg-target.c
index ff4d9cf..1ffc88a 100644
--- a/tcg/i386/tcg-target.c
+++ b/tcg/i386/tcg-target.c
@@ -1137,6 +1137,17 @@ static void * const qemu_ld_helpers[16] = {
     [MO_BEQ]  = helper_be_ldq_mmu,
 };
 
+/* LoadLink helpers, only unsigned. Use the macro below to access them. */
+static void * const qemu_ldex_helpers[16] = {
+    [MO_LEUL] = helper_le_ldlinkul_mmu,
+};
+
+#define LDEX_HELPER(mem_op)                                             \
+({                                                                      \
+    assert(mem_op & MO_EXCL);                                           \
+    qemu_ldex_helpers[((int)mem_op - MO_EXCL)];                         \
+})
+
 /* helper signature: helper_ret_st_mmu(CPUState *env, target_ulong addr,
  *                                     uintxx_t val, int mmu_idx, uintptr_t ra)
  */
@@ -1150,6 +1161,17 @@ static void * const qemu_st_helpers[16] = {
     [MO_BEQ]  = helper_be_stq_mmu,
 };
 
+/* StoreConditional helpers. Use the macro below to access them. */
+static void * const qemu_stex_helpers[16] = {
+    [MO_LEUL] = helper_le_stcondl_mmu,
+};
+
+#define STEX_HELPER(mem_op)                                             \
+({                                                                      \
+    assert(mem_op & MO_EXCL);                                           \
+    qemu_stex_helpers[(int)mem_op - MO_EXCL];                           \
+})
+
 /* Perform the TLB load and compare.
 
    Inputs:
@@ -1245,6 +1267,7 @@ static inline void tcg_out_tlb_load(TCGContext *s, TCGReg addrlo, TCGReg addrhi,
  * for a load or store, so that we can later generate the correct helper code
  */
 static void add_qemu_ldst_label(TCGContext *s, bool is_ld, TCGMemOpIdx oi,
+                                TCGReg llsc_success,
                                 TCGReg datalo, TCGReg datahi,
                                 TCGReg addrlo, TCGReg addrhi,
                                 tcg_insn_unit *raddr,
@@ -1253,6 +1276,7 @@ static void add_qemu_ldst_label(TCGContext *s, bool is_ld, TCGMemOpIdx oi,
     TCGLabelQemuLdst *label = new_ldst_label(s);
 
     label->is_ld = is_ld;
+    label->llsc_success = llsc_success;
     label->oi = oi;
     label->datalo_reg = datalo;
     label->datahi_reg = datahi;
@@ -1307,7 +1331,11 @@ static void tcg_out_qemu_ld_slow_path(TCGContext *s, TCGLabelQemuLdst *l)
                      (uintptr_t)l->raddr);
     }
 
-    tcg_out_call(s, qemu_ld_helpers[opc & (MO_BSWAP | MO_SIZE)]);
+    if (opc & MO_EXCL) {
+        tcg_out_call(s, LDEX_HELPER(opc));
+    } else {
+        tcg_out_call(s, qemu_ld_helpers[opc & (MO_BSWAP | MO_SIZE)]);
+    }
 
     data_reg = l->datalo_reg;
     switch (opc & MO_SSIZE) {
@@ -1411,9 +1439,16 @@ static void tcg_out_qemu_st_slow_path(TCGContext *s, TCGLabelQemuLdst *l)
         }
     }
 
-    /* "Tail call" to the helper, with the return address back inline.  */
-    tcg_out_push(s, retaddr);
-    tcg_out_jmp(s, qemu_st_helpers[opc & (MO_BSWAP | MO_SIZE)]);
+    if (opc & MO_EXCL) {
+        tcg_out_call(s, STEX_HELPER(opc));
+        /* Save the output of the StoreConditional */
+        tcg_out_mov(s, TCG_TYPE_I32, l->llsc_success, TCG_REG_EAX);
+        tcg_out_jmp(s, l->raddr);
+    } else {
+        /* "Tail call" to the helper, with the return address back inline.  */
+        tcg_out_push(s, retaddr);
+        tcg_out_jmp(s, qemu_st_helpers[opc & (MO_BSWAP | MO_SIZE)]);
+    }
 }
 #elif defined(__x86_64__) && defined(__linux__)
 # include <asm/prctl.h>
@@ -1526,7 +1561,8 @@ static void tcg_out_qemu_ld_direct(TCGContext *s, TCGReg datalo, TCGReg datahi,
 /* XXX: qemu_ld and qemu_st could be modified to clobber only EDX and
    EAX. It will be useful once fixed registers globals are less
    common. */
-static void tcg_out_qemu_ld(TCGContext *s, const TCGArg *args, bool is64)
+static void tcg_out_qemu_ld(TCGContext *s, const TCGArg *args, bool is64,
+                            bool isLoadLink)
 {
     TCGReg datalo, datahi, addrlo;
     TCGReg addrhi __attribute__((unused));
@@ -1549,14 +1585,34 @@ static void tcg_out_qemu_ld(TCGContext *s, const TCGArg *args, bool is64)
     mem_index = get_mmuidx(oi);
     s_bits = opc & MO_SIZE;
 
-    tcg_out_tlb_load(s, addrlo, addrhi, mem_index, s_bits,
-                     label_ptr, offsetof(CPUTLBEntry, addr_read));
+    if (isLoadLink) {
+        TCGType t = ((TCG_TARGET_REG_BITS == 64) && (TARGET_LONG_BITS == 64)) ?
+                                                   TCG_TYPE_I64 : TCG_TYPE_I32;
+        /* The JMP address will be patched afterwards,
+         * in tcg_out_qemu_ld_slow_path (two times when
+         * TARGET_LONG_BITS > TCG_TARGET_REG_BITS). */
+        tcg_out_mov(s, t, TCG_REG_L1, addrlo);
+
+        if (TARGET_LONG_BITS > TCG_TARGET_REG_BITS) {
+            /* Store the second part of the address. */
+            tcg_out_mov(s, t, TCG_REG_L0, addrhi);
+            /* We add 4 to include the jmp that follows. */
+            label_ptr[1] = s->code_ptr + 4;
+        }
 
-    /* TLB Hit.  */
-    tcg_out_qemu_ld_direct(s, datalo, datahi, TCG_REG_L1, 0, 0, opc);
+        tcg_out_opc(s, OPC_JMP_long, 0, 0, 0);
+        label_ptr[0] = s->code_ptr;
+        s->code_ptr += 4;
+    } else {
+        tcg_out_tlb_load(s, addrlo, addrhi, mem_index, s_bits,
+                         label_ptr, offsetof(CPUTLBEntry, addr_read));
+
+        /* TLB Hit.  */
+        tcg_out_qemu_ld_direct(s, datalo, datahi, TCG_REG_L1, 0, 0, opc);
+    }
 
     /* Record the current context of a load into ldst label */
-    add_qemu_ldst_label(s, true, oi, datalo, datahi, addrlo, addrhi,
+    add_qemu_ldst_label(s, true, oi, 0, datalo, datahi, addrlo, addrhi,
                         s->code_ptr, label_ptr);
 #else
     {
@@ -1659,9 +1715,10 @@ static void tcg_out_qemu_st_direct(TCGContext *s, TCGReg datalo, TCGReg datahi,
     }
 }
 
-static void tcg_out_qemu_st(TCGContext *s, const TCGArg *args, bool is64)
+static void tcg_out_qemu_st(TCGContext *s, const TCGArg *args, bool is64,
+                            bool isStoreCond)
 {
-    TCGReg datalo, datahi, addrlo;
+    TCGReg datalo, datahi, addrlo, llsc_success;
     TCGReg addrhi __attribute__((unused));
     TCGMemOpIdx oi;
     TCGMemOp opc;
@@ -1671,6 +1728,9 @@ static void tcg_out_qemu_st(TCGContext *s, const TCGArg *args, bool is64)
     tcg_insn_unit *label_ptr[2];
 #endif
 
+    /* The stcond variant has one more param */
+    llsc_success = (isStoreCond ? *args++ : 0);
+
     datalo = *args++;
     datahi = (TCG_TARGET_REG_BITS == 32 && is64 ? *args++ : 0);
     addrlo = *args++;
@@ -1682,15 +1742,35 @@ static void tcg_out_qemu_st(TCGContext *s, const TCGArg *args, bool is64)
     mem_index = get_mmuidx(oi);
     s_bits = opc & MO_SIZE;
 
-    tcg_out_tlb_load(s, addrlo, addrhi, mem_index, s_bits,
-                     label_ptr, offsetof(CPUTLBEntry, addr_write));
+    if (isStoreCond) {
+        TCGType t = ((TCG_TARGET_REG_BITS == 64) && (TARGET_LONG_BITS == 64)) ?
+                                                   TCG_TYPE_I64 : TCG_TYPE_I32;
+        /* The JMP address will be filled afterwards,
+         * in tcg_out_qemu_ld_slow_path (two times when
+         * TARGET_LONG_BITS > TCG_TARGET_REG_BITS). */
+        tcg_out_mov(s, t, TCG_REG_L1, addrlo);
+
+        if (TARGET_LONG_BITS > TCG_TARGET_REG_BITS) {
+            /* Store the second part of the address. */
+            tcg_out_mov(s, t, TCG_REG_L0, addrhi);
+            /* We add 4 to include the jmp that follows. */
+            label_ptr[1] = s->code_ptr + 4;
+        }
 
-    /* TLB Hit.  */
-    tcg_out_qemu_st_direct(s, datalo, datahi, TCG_REG_L1, 0, 0, opc);
+        tcg_out_opc(s, OPC_JMP_long, 0, 0, 0);
+        label_ptr[0] = s->code_ptr;
+        s->code_ptr += 4;
+    } else {
+        tcg_out_tlb_load(s, addrlo, addrhi, mem_index, s_bits,
+                         label_ptr, offsetof(CPUTLBEntry, addr_write));
+
+        /* TLB Hit.  */
+        tcg_out_qemu_st_direct(s, datalo, datahi, TCG_REG_L1, 0, 0, opc);
+    }
 
     /* Record the current context of a store into ldst label */
-    add_qemu_ldst_label(s, false, oi, datalo, datahi, addrlo, addrhi,
-                        s->code_ptr, label_ptr);
+    add_qemu_ldst_label(s, false, oi, llsc_success, datalo, datahi, addrlo,
+                        addrhi, s->code_ptr, label_ptr);
 #else
     {
         int32_t offset = GUEST_BASE;
@@ -1951,16 +2031,22 @@ static inline void tcg_out_op(TCGContext *s, TCGOpcode opc,
         break;
 
     case INDEX_op_qemu_ld_i32:
-        tcg_out_qemu_ld(s, args, 0);
+        tcg_out_qemu_ld(s, args, 0, 0);
+        break;
+    case INDEX_op_qemu_ldlink_i32:
+        tcg_out_qemu_ld(s, args, 0, 1);
         break;
     case INDEX_op_qemu_ld_i64:
-        tcg_out_qemu_ld(s, args, 1);
+        tcg_out_qemu_ld(s, args, 1, 0);
         break;
     case INDEX_op_qemu_st_i32:
-        tcg_out_qemu_st(s, args, 0);
+        tcg_out_qemu_st(s, args, 0, 0);
+        break;
+    case INDEX_op_qemu_stcond_i32:
+        tcg_out_qemu_st(s, args, 0, 1);
         break;
     case INDEX_op_qemu_st_i64:
-        tcg_out_qemu_st(s, args, 1);
+        tcg_out_qemu_st(s, args, 1, 0);
         break;
 
     OP_32_64(mulu2):
@@ -2182,17 +2268,23 @@ static const TCGTargetOpDef x86_op_defs[] = {
 
 #if TCG_TARGET_REG_BITS == 64
     { INDEX_op_qemu_ld_i32, { "r", "L" } },
+    { INDEX_op_qemu_ldlink_i32, { "r", "L" } },
     { INDEX_op_qemu_st_i32, { "L", "L" } },
+    { INDEX_op_qemu_stcond_i32, { "r", "L", "L" } },
     { INDEX_op_qemu_ld_i64, { "r", "L" } },
     { INDEX_op_qemu_st_i64, { "L", "L" } },
 #elif TARGET_LONG_BITS <= TCG_TARGET_REG_BITS
     { INDEX_op_qemu_ld_i32, { "r", "L" } },
+    { INDEX_op_qemu_ldlink_i32, { "r", "L" } },
     { INDEX_op_qemu_st_i32, { "L", "L" } },
+    { INDEX_op_qemu_stcond_i32, { "r", "L", "L" } },
     { INDEX_op_qemu_ld_i64, { "r", "r", "L" } },
     { INDEX_op_qemu_st_i64, { "L", "L", "L" } },
 #else
     { INDEX_op_qemu_ld_i32, { "r", "L", "L" } },
+    { INDEX_op_qemu_ldlink_i32, { "r", "L", "L" } },
     { INDEX_op_qemu_st_i32, { "L", "L", "L" } },
+    { INDEX_op_qemu_stcond_i32, { "r", "L", "L", "L" } },
     { INDEX_op_qemu_ld_i64, { "r", "r", "L", "L" } },
     { INDEX_op_qemu_st_i64, { "L", "L", "L", "L" } },
 #endif
-- 
2.4.3

      parent reply	other threads:[~2015-06-15 11:50 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-06-15 11:51 [Qemu-devel] [RFC v2 0/7] Slow-path for atomic instruction translation Alvise Rigo
2015-06-15 11:51 ` [Qemu-devel] [RFC v2 1/7] bitmap: Add bitmap_one_extend operation Alvise Rigo
2015-06-15 11:51 ` [Qemu-devel] [RFC v2 2/7] exec: Add new exclusive bitmap to ram_list Alvise Rigo
2015-06-15 11:51 ` [Qemu-devel] [RFC v2 3/7] Add new TLB_EXCL flag Alvise Rigo
2015-06-15 11:51 ` [Qemu-devel] [RFC v2 4/7] softmmu: Add helpers for a new slow-path Alvise Rigo
2015-06-15 11:51 ` [Qemu-devel] [RFC v2 5/7] tcg-op: create new TCG qemu_ldlink and qemu_stcond instructions Alvise Rigo
2015-06-15 11:51 ` [Qemu-devel] [RFC v2 6/7] target-arm: translate: implement qemu_ldlink and qemu_stcond ops Alvise Rigo
2015-06-15 11:51 ` Alvise Rigo [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1434369088-15076-8-git-send-email-a.rigo@virtualopensystems.com \
    --to=a.rigo@virtualopensystems.com \
    --cc=alex.bennee@linaro.org \
    --cc=claudio.fontana@huawei.com \
    --cc=cota@braap.org \
    --cc=jani.kokkonen@huawei.com \
    --cc=mttcg@listserver.greensocs.com \
    --cc=qemu-devel@nongnu.org \
    --cc=rth@twiddle.net \
    --cc=tech@virtualopensystems.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).