qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Peter Maydell <peter.maydell@linaro.org>
To: qemu-devel@nongnu.org
Subject: [PULL 06/23] target/arm: Reorg ARMCPRegInfo type field bits
Date: Thu,  5 May 2022 10:11:30 +0100	[thread overview]
Message-ID: <20220505091147.2657652-7-peter.maydell@linaro.org> (raw)
In-Reply-To: <20220505091147.2657652-1-peter.maydell@linaro.org>

From: Richard Henderson <richard.henderson@linaro.org>

Instead of defining ARM_CP_FLAG_MASK to remove flags,
define ARM_CP_SPECIAL_MASK to isolate special cases.
Sort the specials to the low bits. Use an enum.

Split the large comment block so as to document each
value separately.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 20220501055028.646596-6-richard.henderson@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
---
 target/arm/cpregs.h        | 130 +++++++++++++++++++++++--------------
 target/arm/cpu.c           |   4 +-
 target/arm/helper.c        |   4 +-
 target/arm/translate-a64.c |   6 +-
 target/arm/translate.c     |   6 +-
 5 files changed, 92 insertions(+), 58 deletions(-)

diff --git a/target/arm/cpregs.h b/target/arm/cpregs.h
index a5231504d58..ff3817decbd 100644
--- a/target/arm/cpregs.h
+++ b/target/arm/cpregs.h
@@ -22,57 +22,87 @@
 #define TARGET_ARM_CPREGS_H
 
 /*
- * ARMCPRegInfo type field bits. If the SPECIAL bit is set this is a
- * special-behaviour cp reg and bits [11..8] indicate what behaviour
- * it has. Otherwise it is a simple cp reg, where CONST indicates that
- * TCG can assume the value to be constant (ie load at translate time)
- * and 64BIT indicates a 64 bit wide coprocessor register. SUPPRESS_TB_END
- * indicates that the TB should not be ended after a write to this register
- * (the default is that the TB ends after cp writes). OVERRIDE permits
- * a register definition to override a previous definition for the
- * same (cp, is64, crn, crm, opc1, opc2) tuple: either the new or the
- * old must have the OVERRIDE bit set.
- * ALIAS indicates that this register is an alias view of some underlying
- * state which is also visible via another register, and that the other
- * register is handling migration and reset; registers marked ALIAS will not be
- * migrated but may have their state set by syncing of register state from KVM.
- * NO_RAW indicates that this register has no underlying state and does not
- * support raw access for state saving/loading; it will not be used for either
- * migration or KVM state synchronization. (Typically this is for "registers"
- * which are actually used as instructions for cache maintenance and so on.)
- * IO indicates that this register does I/O and therefore its accesses
- * need to be marked with gen_io_start() and also end the TB. In particular,
- * registers which implement clocks or timers require this.
- * RAISES_EXC is for when the read or write hook might raise an exception;
- * the generated code will synchronize the CPU state before calling the hook
- * so that it is safe for the hook to call raise_exception().
- * NEWEL is for writes to registers that might change the exception
- * level - typically on older ARM chips. For those cases we need to
- * re-read the new el when recomputing the translation flags.
+ * ARMCPRegInfo type field bits:
  */
-#define ARM_CP_SPECIAL           0x0001
-#define ARM_CP_CONST             0x0002
-#define ARM_CP_64BIT             0x0004
-#define ARM_CP_SUPPRESS_TB_END   0x0008
-#define ARM_CP_OVERRIDE          0x0010
-#define ARM_CP_ALIAS             0x0020
-#define ARM_CP_IO                0x0040
-#define ARM_CP_NO_RAW            0x0080
-#define ARM_CP_NOP               (ARM_CP_SPECIAL | 0x0100)
-#define ARM_CP_WFI               (ARM_CP_SPECIAL | 0x0200)
-#define ARM_CP_NZCV              (ARM_CP_SPECIAL | 0x0300)
-#define ARM_CP_CURRENTEL         (ARM_CP_SPECIAL | 0x0400)
-#define ARM_CP_DC_ZVA            (ARM_CP_SPECIAL | 0x0500)
-#define ARM_CP_DC_GVA            (ARM_CP_SPECIAL | 0x0600)
-#define ARM_CP_DC_GZVA           (ARM_CP_SPECIAL | 0x0700)
-#define ARM_LAST_SPECIAL         ARM_CP_DC_GZVA
-#define ARM_CP_FPU               0x1000
-#define ARM_CP_SVE               0x2000
-#define ARM_CP_NO_GDB            0x4000
-#define ARM_CP_RAISES_EXC        0x8000
-#define ARM_CP_NEWEL             0x10000
-/* Mask of only the flag bits in a type field */
-#define ARM_CP_FLAG_MASK         0x1f0ff
+enum {
+    /*
+     * Register must be handled specially during translation.
+     * The method is one of the values below:
+     */
+    ARM_CP_SPECIAL_MASK          = 0x000f,
+    /* Special: no change to PE state: writes ignored, reads ignored. */
+    ARM_CP_NOP                   = 0x0001,
+    /* Special: sysreg is WFI, for v5 and v6. */
+    ARM_CP_WFI                   = 0x0002,
+    /* Special: sysreg is NZCV. */
+    ARM_CP_NZCV                  = 0x0003,
+    /* Special: sysreg is CURRENTEL. */
+    ARM_CP_CURRENTEL             = 0x0004,
+    /* Special: sysreg is DC ZVA or similar. */
+    ARM_CP_DC_ZVA                = 0x0005,
+    ARM_CP_DC_GVA                = 0x0006,
+    ARM_CP_DC_GZVA               = 0x0007,
+
+    /* Flag: reads produce resetvalue; writes ignored. */
+    ARM_CP_CONST                 = 1 << 4,
+    /* Flag: For ARM_CP_STATE_AA32, sysreg is 64-bit. */
+    ARM_CP_64BIT                 = 1 << 5,
+    /*
+     * Flag: TB should not be ended after a write to this register
+     * (the default is that the TB ends after cp writes).
+     */
+    ARM_CP_SUPPRESS_TB_END       = 1 << 6,
+    /*
+     * Flag: Permit a register definition to override a previous definition
+     * for the same (cp, is64, crn, crm, opc1, opc2) tuple: either the new
+     * or the old must have the ARM_CP_OVERRIDE bit set.
+     */
+    ARM_CP_OVERRIDE              = 1 << 7,
+    /*
+     * Flag: Register is an alias view of some underlying state which is also
+     * visible via another register, and that the other register is handling
+     * migration and reset; registers marked ARM_CP_ALIAS will not be migrated
+     * but may have their state set by syncing of register state from KVM.
+     */
+    ARM_CP_ALIAS                 = 1 << 8,
+    /*
+     * Flag: Register does I/O and therefore its accesses need to be marked
+     * with gen_io_start() and also end the TB. In particular, registers which
+     * implement clocks or timers require this.
+     */
+    ARM_CP_IO                    = 1 << 9,
+    /*
+     * Flag: Register has no underlying state and does not support raw access
+     * for state saving/loading; it will not be used for either migration or
+     * KVM state synchronization. Typically this is for "registers" which are
+     * actually used as instructions for cache maintenance and so on.
+     */
+    ARM_CP_NO_RAW                = 1 << 10,
+    /*
+     * Flag: The read or write hook might raise an exception; the generated
+     * code will synchronize the CPU state before calling the hook so that it
+     * is safe for the hook to call raise_exception().
+     */
+    ARM_CP_RAISES_EXC            = 1 << 11,
+    /*
+     * Flag: Writes to the sysreg might change the exception level - typically
+     * on older ARM chips. For those cases we need to re-read the new el when
+     * recomputing the translation flags.
+     */
+    ARM_CP_NEWEL                 = 1 << 12,
+    /*
+     * Flag: Access check for this sysreg is identical to accessing FPU state
+     * from an instruction: use translation fp_access_check().
+     */
+    ARM_CP_FPU                   = 1 << 13,
+    /*
+     * Flag: Access check for this sysreg is identical to accessing SVE state
+     * from an instruction: use translation sve_access_check().
+     */
+    ARM_CP_SVE                   = 1 << 14,
+    /* Flag: Do not expose in gdb sysreg xml. */
+    ARM_CP_NO_GDB                = 1 << 15,
+};
 
 /*
  * Valid values for ARMCPRegInfo state field, indicating which of
diff --git a/target/arm/cpu.c b/target/arm/cpu.c
index 18212eb6eef..a7cd692010c 100644
--- a/target/arm/cpu.c
+++ b/target/arm/cpu.c
@@ -117,7 +117,7 @@ static void cp_reg_reset(gpointer key, gpointer value, gpointer opaque)
     ARMCPRegInfo *ri = value;
     ARMCPU *cpu = opaque;
 
-    if (ri->type & (ARM_CP_SPECIAL | ARM_CP_ALIAS)) {
+    if (ri->type & (ARM_CP_SPECIAL_MASK | ARM_CP_ALIAS)) {
         return;
     }
 
@@ -153,7 +153,7 @@ static void cp_reg_check_reset(gpointer key, gpointer value,  gpointer opaque)
     ARMCPU *cpu = opaque;
     uint64_t oldvalue, newvalue;
 
-    if (ri->type & (ARM_CP_SPECIAL | ARM_CP_ALIAS | ARM_CP_NO_RAW)) {
+    if (ri->type & (ARM_CP_SPECIAL_MASK | ARM_CP_ALIAS | ARM_CP_NO_RAW)) {
         return;
     }
 
diff --git a/target/arm/helper.c b/target/arm/helper.c
index ca6ba9bd820..f84377babe1 100644
--- a/target/arm/helper.c
+++ b/target/arm/helper.c
@@ -8600,7 +8600,7 @@ static void add_cpreg_to_hashtable(ARMCPU *cpu, const ARMCPRegInfo *r,
      * multiple times. Special registers (ie NOP/WFI) are
      * never migratable and not even raw-accessible.
      */
-    if ((r->type & ARM_CP_SPECIAL)) {
+    if (r->type & ARM_CP_SPECIAL_MASK) {
         r2->type |= ARM_CP_NO_RAW;
     }
     if (((r->crm == CP_ANY) && crm != 0) ||
@@ -8750,7 +8750,7 @@ void define_one_arm_cp_reg_with_opaque(ARMCPU *cpu,
     /* Check that the register definition has enough info to handle
      * reads and writes if they are permitted.
      */
-    if (!(r->type & (ARM_CP_SPECIAL|ARM_CP_CONST))) {
+    if (!(r->type & (ARM_CP_SPECIAL_MASK | ARM_CP_CONST))) {
         if (r->access & PL3_R) {
             assert((r->fieldoffset ||
                    (r->bank_fieldoffsets[0] && r->bank_fieldoffsets[1])) ||
diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c
index 348a638c5cb..a82f5d5984b 100644
--- a/target/arm/translate-a64.c
+++ b/target/arm/translate-a64.c
@@ -1833,7 +1833,9 @@ static void handle_sys(DisasContext *s, uint32_t insn, bool isread,
     }
 
     /* Handle special cases first */
-    switch (ri->type & ~(ARM_CP_FLAG_MASK & ~ARM_CP_SPECIAL)) {
+    switch (ri->type & ARM_CP_SPECIAL_MASK) {
+    case 0:
+        break;
     case ARM_CP_NOP:
         return;
     case ARM_CP_NZCV:
@@ -1908,7 +1910,7 @@ static void handle_sys(DisasContext *s, uint32_t insn, bool isread,
         }
         return;
     default:
-        break;
+        g_assert_not_reached();
     }
     if ((ri->type & ARM_CP_FPU) && !fp_access_check(s)) {
         return;
diff --git a/target/arm/translate.c b/target/arm/translate.c
index fc7917cdf44..050c237b076 100644
--- a/target/arm/translate.c
+++ b/target/arm/translate.c
@@ -4744,7 +4744,9 @@ static void do_coproc_insn(DisasContext *s, int cpnum, int is64,
         }
 
         /* Handle special cases first */
-        switch (ri->type & ~(ARM_CP_FLAG_MASK & ~ARM_CP_SPECIAL)) {
+        switch (ri->type & ARM_CP_SPECIAL_MASK) {
+        case 0:
+            break;
         case ARM_CP_NOP:
             return;
         case ARM_CP_WFI:
@@ -4756,7 +4758,7 @@ static void do_coproc_insn(DisasContext *s, int cpnum, int is64,
             s->base.is_jmp = DISAS_WFI;
             return;
         default:
-            break;
+            g_assert_not_reached();
         }
 
         if ((tb_cflags(s->base.tb) & CF_USE_ICOUNT) && (ri->type & ARM_CP_IO)) {
-- 
2.25.1



  parent reply	other threads:[~2022-05-05 10:09 UTC|newest]

Thread overview: 25+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-05-05  9:11 [PULL 00/23] target-arm queue Peter Maydell
2022-05-05  9:11 ` [PULL 01/23] target/arm: Enable SCTLR_EL1.BT0 for aarch64-linux-user Peter Maydell
2022-05-05  9:11 ` [PULL 02/23] target/arm: Split out cpregs.h Peter Maydell
2022-05-05  9:11 ` [PULL 03/23] target/arm: Reorg CPAccessResult and access_check_cp_reg Peter Maydell
2022-05-05  9:11 ` [PULL 04/23] target/arm: Replace sentinels with ARRAY_SIZE in cpregs.h Peter Maydell
2022-05-05  9:11 ` [PULL 05/23] target/arm: Make some more cpreg data static const Peter Maydell
2022-05-05  9:11 ` Peter Maydell [this message]
2022-05-05  9:11 ` [PULL 07/23] target/arm: Avoid bare abort() or assert(0) Peter Maydell
2022-05-05  9:11 ` [PULL 08/23] target/arm: Change cpreg access permissions to enum Peter Maydell
2022-05-05  9:11 ` [PULL 09/23] target/arm: Name CPState type Peter Maydell
2022-05-05  9:11 ` [PULL 10/23] target/arm: Name CPSecureState type Peter Maydell
2022-05-05  9:11 ` [PULL 11/23] target/arm: Drop always-true test in define_arm_vh_e2h_redirects_aliases Peter Maydell
2022-05-05  9:11 ` [PULL 12/23] target/arm: Store cpregs key in the hash table directly Peter Maydell
2022-05-05  9:11 ` [PULL 13/23] target/arm: Merge allocation of the cpreg and its name Peter Maydell
2022-05-05  9:11 ` [PULL 14/23] target/arm: Hoist computation of key in add_cpreg_to_hashtable Peter Maydell
2022-05-05  9:11 ` [PULL 15/23] target/arm: Consolidate cpreg updates " Peter Maydell
2022-05-05  9:11 ` [PULL 16/23] target/arm: Use bool for is64 and ns " Peter Maydell
2022-05-05  9:11 ` [PULL 17/23] target/arm: Hoist isbanked computation " Peter Maydell
2022-05-05  9:11 ` [PULL 18/23] target/arm: Perform override check early " Peter Maydell
2022-05-05  9:11 ` [PULL 19/23] target/arm: Reformat comments " Peter Maydell
2022-05-05  9:11 ` [PULL 20/23] target/arm: Remove HOST_BIG_ENDIAN ifdef " Peter Maydell
2022-05-05  9:11 ` [PULL 21/23] target/arm: Add isar predicates for FEAT_Debugv8p2 Peter Maydell
2022-05-05  9:11 ` [PULL 22/23] target/arm: Add isar_feature_{aa64,any}_ras Peter Maydell
2022-05-05  9:11 ` [PULL 23/23] target/arm: read access to performance counters from EL0 Peter Maydell
2022-05-05 17:56 ` [PULL 00/23] target-arm queue Richard Henderson

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=20220505091147.2657652-7-peter.maydell@linaro.org \
    --to=peter.maydell@linaro.org \
    --cc=qemu-devel@nongnu.org \
    /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).