[PULL 00/32] Accelerator patches for 2026-07-06

[PULL 00/32] Accelerator patches for 2026-07-06

[Philippe Mathieu-Daudé]

The following changes since commit 4ee536fac748b70e6f3d8568ddd20cfbaa9cf7bf:

  Merge tag 'firmware-20260704-pull-request' of https://gitlab.com/kraxel/qemu into staging (2026-07-05 08:42:47 +0200)

are available in the Git repository at:

  https://github.com/philmd/qemu.git tags/accel-20260706

for you to fetch changes up to c0df53752c86e5f51e12e7b99501b6f4d2e52c01:

  cpu: Only check SSTEP_ENABLE flag in cpu_single_stepping() (2026-07-06 15:42:18 +0200)

Ignoring following checkpatch.pl warnings:

  WARNING: Saw acceptable license 'LGPL-2.1-or-later' ... Please explain the license
  WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?
    accel/tcg/system-page-protection.h
  WARNING: line over 80 characters
  #71: FILE: gdbstub/gdbstub.c:75:
  +    gdbserver_state.accel_config.sstep_flags = accel_supported_gdbstub_sstep_flags();
  WARNING: line over 80 characters
  #582: FILE: target/i386/kvm/kvm.c:6167:
  +static int find_hw_breakpoint(target_ulong addr, int len, GdbBreakpointType type)

----------------------------------------------------------------
Accelerators patches queue

- Various cleanups around debugging APIs
- Correctly check singlestep flag enabled in CPUState
- Fix possible memory corruption with MSHV (CID 1660876)
----------------------------------------------------------------

Magnus Kulke (1):
  accel/mshv: Fix pointer to proc feature bitfield

Philippe Mathieu-Daudé (30):
  cpu: Constify CPUState::cc (cached CPUClass pointer)
  target/i386: Remove duplicate tlb_flush() call in cpu_post_load()
  accel/tcg: Restrict tlb_protect/unprotect_code() to TCG
  accel/hvf: Remove left-over comment
  accel/mshv: Replace @dirty field by generic CPUState::vcpu_dirty field
  gdbstub: Add trace event for STEP packet handler
  gdbstub: Only return E22 when reverse GDB is not supported
  accel/whpx: Implement missing
    AccelClass::gdbstub_supported_sstep_flags
  accel/kvm: Always define AccelOpsClass::supports_guest_debug
  accel/kvm: Simplify kvm_init() w.r.t. TARGET_KVM_HAVE_GUEST_DEBUG
  accel/kvm: Hold have_guest_debug in KVMState
  gdbstub: Reduce gdb_supports_guest_debug() scope
  gdbstub: Move supported_sstep_flags in AccelGdbConfig structure
  accel: Have each implementation return their AccelGdbConfig
  gdbstub: Make default replay_mode value explicit in stubs
  accel: Hold @can_reverse information in AccelGdbConfig
  accel: Remove AccelOpsClass::supports_guest_debug
  cpu: Move BREAKPOINT definitions to 'exec/breakpoint.h'
  accel: Remove unnecessary 'inline' qualifier in remove_all_breakpoints
  gdbstub/user: Directly call gdb_breakpoint_remove_all() in user mode
  gdbstub: Reduce @type variable scope
  gdbstub: Introduce GdbBreakpointType enumerator
  accel: Use GdbBreakpointType enum
  target/arm: Inline check_watchpoints() in arm_debug_check_watchpoint()
  target/ppc: Ensure TCG is used in ppc_update_daw()
  accel/tcg: Improve docstrings around TCGCPUOps::*watchpoint* handlers
  cpu: Better name cpu_single_step() trace event
  cpu: Introduce cpu_single_stepping() helper
  cpu: Rename CPUState @singlestep_enabled -> @singlestep_flags
  cpu: Only check SSTEP_ENABLE flag in cpu_single_stepping()

Richard Henderson (1):
  cpu: Move cpu_breakpoint_test out of line

 docs/system/gdb.rst                |  2 +-
 accel/kvm/kvm-cpus.h               | 12 ++++---
 accel/tcg/system-page-protection.h | 17 ++++++++++
 gdbstub/internals.h                | 12 ++++---
 include/accel/accel-cpu-ops.h      | 10 +++---
 include/accel/accel-ops.h          |  5 ++-
 include/accel/tcg/cpu-ops.h        |  5 ++-
 include/exec/breakpoint.h          | 14 ++++++++
 include/exec/cputlb.h              |  8 ++---
 include/gdbstub/enums.h            | 12 ++++---
 include/hw/core/cpu.h              | 49 +++++++++-------------------
 include/qemu/accel.h               | 13 +++++---
 include/system/hvf_int.h           | 14 ++++----
 include/system/kvm.h               |  9 ++++--
 include/system/mshv_int.h          |  1 -
 include/system/whpx-all.h          |  3 --
 target/arm/internals.h             |  5 +--
 accel/accel-common.c               | 15 +++------
 accel/hvf/hvf-accel-ops.c          | 23 +++++++------
 accel/hvf/hvf-all.c                |  8 ++---
 accel/kvm/kvm-accel-ops.c          |  7 ++--
 accel/kvm/kvm-all.c                | 52 ++++++++----------------------
 accel/mshv/mshv-all.c              | 22 ++++++-------
 accel/tcg/cpu-exec-common.c        |  2 +-
 accel/tcg/cpu-exec.c               | 12 +++----
 accel/tcg/cputlb.c                 |  3 +-
 accel/tcg/tb-maint.c               |  1 +
 accel/tcg/tcg-accel-ops-rr.c       |  2 +-
 accel/tcg/tcg-accel-ops.c          | 22 ++++++-------
 accel/tcg/tcg-all.c                | 30 ++++++++---------
 accel/tcg/translate-all.c          |  3 +-
 accel/whpx/whpx-accel-ops.c        |  6 ----
 cpu-common.c                       | 15 +++++++++
 cpu-target.c                       | 10 +++---
 gdbstub/gdbstub.c                  | 18 ++++++-----
 gdbstub/system.c                   | 37 ++++++++-------------
 gdbstub/user.c                     | 20 +++---------
 linux-user/riscv/cpu_loop.c        |  2 +-
 linux-user/s390x/cpu_loop.c        |  2 +-
 stubs/replay-mode.c                |  2 +-
 system/cpus.c                      |  2 +-
 target/arm/hvf/hvf.c               | 27 +++++++---------
 target/arm/hyp_gdbstub.c           |  8 ++---
 target/arm/kvm.c                   | 14 ++++----
 target/arm/tcg/debug.c             | 40 ++++++++++-------------
 target/arm/whpx/whpx-all.c         |  5 ---
 target/i386/hvf/hvf.c              | 13 +++-----
 target/i386/kvm/kvm.c              | 14 ++++----
 target/i386/machine.c              |  1 -
 target/i386/mshv/mshv-cpu.c        |  6 ++--
 target/i386/whpx/whpx-all.c        | 13 +++-----
 target/loongarch/kvm/kvm.c         | 10 +++---
 target/microblaze/translate.c      |  2 +-
 target/ppc/cpu.c                   |  2 ++
 target/ppc/kvm.c                   | 14 ++++----
 target/ppc/translate.c             | 16 ++++-----
 target/riscv/kvm/kvm-cpu.c         |  8 +++--
 target/s390x/kvm/kvm.c             | 10 +++---
 trace-events                       |  2 +-
 59 files changed, 338 insertions(+), 374 deletions(-)
 create mode 100644 accel/tcg/system-page-protection.h

-- 
2.53.0

[PULL 01/32] cpu: Constify CPUState::cc (cached CPUClass pointer)

[Philippe Mathieu-Daudé]

Various CPUState can share the same CPUClass parent,
and must not update its fields. Protect the CPUClass
by marking the CPUState pointer const.

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Daniel Henrique Barboza <daniel.barboza@oss.qualcomm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Chao Liu <chao.liu.zevorn@gmail.com>
Message-ID: <20260705215729.62196-2-philmd@oss.qualcomm.com>
---
 include/hw/core/cpu.h     | 2 +-
 accel/tcg/cpu-exec.c      | 2 +-
 accel/tcg/translate-all.c | 3 +--
 gdbstub/gdbstub.c         | 2 +-
 4 files changed, 4 insertions(+), 5 deletions(-)

diff --git a/include/hw/core/cpu.h b/include/hw/core/cpu.h
index 59d601465b4..60a23d6f38b 100644
--- a/include/hw/core/cpu.h
+++ b/include/hw/core/cpu.h
@@ -480,7 +480,7 @@ struct CPUState {
     /*< private >*/
     DeviceState parent_obj;
     /* cache to avoid expensive CPU_GET_CLASS */
-    CPUClass *cc;
+    const CPUClass *cc;
     /*< public >*/
 
     int nr_threads;
diff --git a/accel/tcg/cpu-exec.c b/accel/tcg/cpu-exec.c
index 9c754b03656..d7f91dce9db 100644
--- a/accel/tcg/cpu-exec.c
+++ b/accel/tcg/cpu-exec.c
@@ -460,7 +460,7 @@ cpu_tb_exec(CPUState *cpu, TranslationBlock *itb, int *tb_exit)
          * counter hit zero); we must restore the guest PC to the address
          * of the start of the TB.
          */
-        CPUClass *cc = cpu->cc;
+        const CPUClass *cc = cpu->cc;
         const TCGCPUOps *tcg_ops = cc->tcg_ops;
 
         if (tcg_ops->synchronize_from_tb) {
diff --git a/accel/tcg/translate-all.c b/accel/tcg/translate-all.c
index 3f1a3a28438..328d96aba30 100644
--- a/accel/tcg/translate-all.c
+++ b/accel/tcg/translate-all.c
@@ -575,7 +575,7 @@ void tb_check_watchpoint(CPUState *cpu, uintptr_t retaddr)
 void cpu_io_recompile(CPUState *cpu, uintptr_t retaddr)
 {
     TranslationBlock *tb;
-    CPUClass *cc;
+    const CPUClass *cc = cpu->cc;
     uint32_t n;
 
     tb = tcg_tb_lookup(retaddr);
@@ -591,7 +591,6 @@ void cpu_io_recompile(CPUState *cpu, uintptr_t retaddr)
      * to account for the re-execution of the branch.
      */
     n = 1;
-    cc = cpu->cc;
     if (cc->tcg_ops->io_recompile_replay_branch &&
         cc->tcg_ops->io_recompile_replay_branch(cpu, tb)) {
         cpu->neg.icount_decr.u16.low++;
diff --git a/gdbstub/gdbstub.c b/gdbstub/gdbstub.c
index c3c944e965b..26366acb674 100644
--- a/gdbstub/gdbstub.c
+++ b/gdbstub/gdbstub.c
@@ -578,7 +578,7 @@ static void gdb_register_feature(CPUState *cpu, int base_reg,
 
 static const char *gdb_get_core_xml_file(CPUState *cpu)
 {
-    CPUClass *cc = cpu->cc;
+    const CPUClass *cc = cpu->cc;
 
     /*
      * The CPU class can provide the XML filename via a method,
-- 
2.53.0

[PULL 02/32] target/i386: Remove duplicate tlb_flush() call in cpu_post_load()

[Philippe Mathieu-Daudé]

Common vCPU cpu_common_post_load() handler calls tlb_flush() since
commit 9656f324d25 ("Move interrupt_request and user_mode_only to
common cpu state..."), no need to call it twice.

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Chao Liu <chao.liu.zevorn@gmail.com>
Message-ID: <20260705215729.62196-3-philmd@oss.qualcomm.com>
---
 target/i386/machine.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/target/i386/machine.c b/target/i386/machine.c
index 1f5c517e562..df0e0c178e4 100644
--- a/target/i386/machine.c
+++ b/target/i386/machine.c
@@ -401,7 +401,6 @@ static int cpu_post_load(void *opaque, int version_id)
         env->dr[7] = dr7 & ~(DR7_GLOBAL_BP_MASK | DR7_LOCAL_BP_MASK);
         cpu_x86_update_dr7(env, dr7);
     }
-    tlb_flush(cs);
     return 0;
 }
 
-- 
2.53.0

[PULL 03/32] accel/tcg: Restrict tlb_protect/unprotect_code() to TCG

[Philippe Mathieu-Daudé]

Both tlb_protect() and tlb_unprotect_code() are only used
within accel/tcg/. Avoid exposing them to the whole code base,
declare them in a new "system-page-protection.h" local header
(maintaining the previous LGPL-2.1-or-later license).

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Daniel Henrique Barboza <daniel.barboza@oss.qualcomm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Chao Liu <chao.liu.zevorn@gmail.com>
Message-ID: <20260705215729.62196-4-philmd@oss.qualcomm.com>
---
 accel/tcg/system-page-protection.h | 17 +++++++++++++++++
 include/exec/cputlb.h              |  8 ++------
 accel/tcg/cputlb.c                 |  3 ++-
 accel/tcg/tb-maint.c               |  1 +
 4 files changed, 22 insertions(+), 7 deletions(-)
 create mode 100644 accel/tcg/system-page-protection.h

diff --git a/accel/tcg/system-page-protection.h b/accel/tcg/system-page-protection.h
new file mode 100644
index 00000000000..52896c6d24a
--- /dev/null
+++ b/accel/tcg/system-page-protection.h
@@ -0,0 +1,17 @@
+/*
+ * QEMU page protection (system emulation)
+ *
+ *  Copyright (c) 2003 Fabrice Bellard
+ *
+ * SPDX-License-Identifier: LGPL-2.1-or-later
+ */
+#ifndef SYSTEM_PAGE_PROTECTION_H
+#define SYSTEM_PAGE_PROTECTION_H
+
+#include "system/ram_addr.h"
+
+void tlb_protect_code(ram_addr_t ram_addr);
+void tlb_unprotect_code(ram_addr_t ram_addr);
+
+#endif
+
diff --git a/include/exec/cputlb.h b/include/exec/cputlb.h
index 3a9603a6965..f7f7020382d 100644
--- a/include/exec/cputlb.h
+++ b/include/exec/cputlb.h
@@ -24,14 +24,10 @@
 #include "exec/hwaddr.h"
 #include "exec/memattrs.h"
 #include "exec/vaddr.h"
-#include "system/ram_addr.h"
-
-#if defined(CONFIG_TCG) && !defined(CONFIG_USER_ONLY)
-void tlb_protect_code(ram_addr_t ram_addr);
-void tlb_unprotect_code(ram_addr_t ram_addr);
-#endif
 
 #ifndef CONFIG_USER_ONLY
+#include "system/ram_addr.h"
+
 void tlb_reset_dirty(CPUState *cpu, uintptr_t start, uintptr_t length);
 void tlb_reset_dirty_range_all(ram_addr_t start, ram_addr_t length);
 #endif
diff --git a/accel/tcg/cputlb.c b/accel/tcg/cputlb.c
index 6e66f3c6e93..7f7c208ba12 100644
--- a/accel/tcg/cputlb.c
+++ b/accel/tcg/cputlb.c
@@ -1,5 +1,5 @@
 /*
- *  Common CPU TLB handling
+ *  Common CPU TLB handling (system emulation)
  *
  *  Copyright (c) 2003 Fabrice Bellard
  *
@@ -47,6 +47,7 @@
 #include "tb-hash.h"
 #include "tlb-bounds.h"
 #include "internal-common.h"
+#include "system-page-protection.h"
 #ifdef CONFIG_PLUGIN
 #include "qemu/plugin-memory.h"
 #endif
diff --git a/accel/tcg/tb-maint.c b/accel/tcg/tb-maint.c
index c33dbf65e7f..4313864d2c5 100644
--- a/accel/tcg/tb-maint.c
+++ b/accel/tcg/tb-maint.c
@@ -38,6 +38,7 @@
 #include "user/page-protection.h"
 #define runstate_is_running()  true
 #else
+#include "system-page-protection.h"
 #include "system/runstate.h"
 #endif
 #include "trace.h"
-- 
2.53.0

[PULL 04/32] accel/hvf: Remove left-over comment

[Philippe Mathieu-Daudé]

Missed in commit 6bf331164c5 ("accel/hvf: Drop hvf_slot and
hvf_find_overlap_slot").

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Daniel Henrique Barboza <daniel.barboza@oss.qualcomm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20260705215729.62196-5-philmd@oss.qualcomm.com>
---
 accel/hvf/hvf-accel-ops.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/accel/hvf/hvf-accel-ops.c b/accel/hvf/hvf-accel-ops.c
index b74a5779c3d..dd2b45b7cb5 100644
--- a/accel/hvf/hvf-accel-ops.c
+++ b/accel/hvf/hvf-accel-ops.c
@@ -62,8 +62,6 @@
 
 HVFState *hvf_state;
 
-/* Memory slots */
-
 static void do_hvf_cpu_synchronize_state(CPUState *cpu, run_on_cpu_data arg)
 {
     if (!cpu->vcpu_dirty) {
-- 
2.53.0

[PULL 05/32] accel/mshv: Replace @dirty field by generic CPUState::vcpu_dirty field

[Philippe Mathieu-Daudé]

No need for accel-specific @dirty field when we have
a generic one in CPUState. (Other accelerators already
did that in commits 6f13a0ada01..36ab216b81d).

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Tested-by: Magnus Kulke <magnuskulke@linux.microsoft.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20260705215729.62196-6-philmd@oss.qualcomm.com>
---
 include/system/mshv_int.h   |  1 -
 accel/mshv/mshv-all.c       | 18 +++++++++---------
 target/i386/mshv/mshv-cpu.c |  6 +++---
 3 files changed, 12 insertions(+), 13 deletions(-)

diff --git a/include/system/mshv_int.h b/include/system/mshv_int.h
index ba3073888ac..b91c4d661a3 100644
--- a/include/system/mshv_int.h
+++ b/include/system/mshv_int.h
@@ -43,7 +43,6 @@ typedef struct MshvHvCallArgs {
 
 struct AccelCPUState {
     int cpufd;
-    bool dirty;
     MshvHvCallArgs hvcall_args;
 };
 
diff --git a/accel/mshv/mshv-all.c b/accel/mshv/mshv-all.c
index 9452504ac28..1959baf6672 100644
--- a/accel/mshv/mshv-all.c
+++ b/accel/mshv/mshv-all.c
@@ -531,7 +531,7 @@ static int mshv_init_vcpu(CPUState *cpu)
     }
 
     mshv_arch_init_vcpu(cpu);
-    cpu->accel->dirty = true;
+    cpu->vcpu_dirty = true;
 
     return 0;
 }
@@ -617,7 +617,7 @@ static int mshv_cpu_exec(CPUState *cpu)
     cpu_exec_start(cpu);
 
     do {
-        if (cpu->accel->dirty) {
+        if (cpu->vcpu_dirty) {
             ret = mshv_arch_store_vcpu_state(cpu);
             if (ret) {
                 error_report("Failed to put registers after init: %s",
@@ -625,7 +625,7 @@ static int mshv_cpu_exec(CPUState *cpu)
                 ret = -1;
                 break;
             }
-            cpu->accel->dirty = false;
+            cpu->vcpu_dirty = false;
         }
 
         ret = mshv_run_vcpu(mshv_state->vm, cpu, &mshv_msg, &exit_reason);
@@ -746,7 +746,7 @@ static void do_mshv_cpu_synchronize_post_init(CPUState *cpu,
         abort();
     }
 
-    cpu->accel->dirty = false;
+    cpu->vcpu_dirty = false;
 }
 
 static void mshv_cpu_synchronize_post_init(CPUState *cpu)
@@ -763,13 +763,13 @@ static void mshv_cpu_synchronize_post_reset(CPUState *cpu)
         cpu_dump_state(cpu, stderr, CPU_DUMP_CODE);
         vm_stop(RUN_STATE_INTERNAL_ERROR);
     }
-    cpu->accel->dirty = false;
+    cpu->vcpu_dirty = false;
 }
 
 static void do_mshv_cpu_synchronize_pre_loadvm(CPUState *cpu,
                                                run_on_cpu_data arg)
 {
-    cpu->accel->dirty = true;
+    cpu->vcpu_dirty = true;
 }
 
 static void mshv_cpu_synchronize_pre_loadvm(CPUState *cpu)
@@ -779,7 +779,7 @@ static void mshv_cpu_synchronize_pre_loadvm(CPUState *cpu)
 
 static void do_mshv_cpu_synchronize(CPUState *cpu, run_on_cpu_data arg)
 {
-    if (!cpu->accel->dirty) {
+    if (!cpu->vcpu_dirty) {
         int ret = mshv_arch_load_vcpu_state(cpu);
         if (ret < 0) {
             error_report("Failed to load registers for vcpu %d",
@@ -789,13 +789,13 @@ static void do_mshv_cpu_synchronize(CPUState *cpu, run_on_cpu_data arg)
             vm_stop(RUN_STATE_INTERNAL_ERROR);
         }
 
-        cpu->accel->dirty = true;
+        cpu->vcpu_dirty = true;
     }
 }
 
 static void mshv_cpu_synchronize(CPUState *cpu)
 {
-    if (!cpu->accel->dirty) {
+    if (!cpu->vcpu_dirty) {
         run_on_cpu(cpu, do_mshv_cpu_synchronize, RUN_ON_CPU_NULL);
     }
 }
diff --git a/target/i386/mshv/mshv-cpu.c b/target/i386/mshv/mshv-cpu.c
index 126ca40b481..1c433c408c8 100644
--- a/target/i386/mshv/mshv-cpu.c
+++ b/target/i386/mshv/mshv-cpu.c
@@ -1692,7 +1692,7 @@ static int pio_write(uint64_t port, const uint8_t *data, uintptr_t size,
     return ret;
 }
 
-static int handle_pio_non_str(const CPUState *cpu,
+static int handle_pio_non_str(CPUState *cpu,
                               hv_x64_io_port_intercept_message *info)
 {
     size_t len = info->access_info.access_size;
@@ -1748,7 +1748,7 @@ static int handle_pio_non_str(const CPUState *cpu,
         return -1;
     }
 
-    cpu->accel->dirty = false;
+    cpu->vcpu_dirty = false;
 
     return 0;
 }
@@ -1909,7 +1909,7 @@ static int handle_pio_str(CPUState *cpu, hv_x64_io_port_intercept_message *info)
         return -1;
     }
 
-    cpu->accel->dirty = false;
+    cpu->vcpu_dirty = false;
 
     return 0;
 }
-- 
2.53.0

[PULL 06/32] accel/mshv: Fix pointer to proc feature bitfield

[Philippe Mathieu-Daudé]

From: Magnus Kulke <magnuskulke@linux.microsoft.com>

Processor features are stored in a union containing two "banks":

union hv_partition_processor_features {
    uint64_t as_uint[2];
    struct {
        uint64_t sse3_support:1;
        ...
    }
}

get_proc_features() to retrieve the 2nd bank was passing a pointer that
steps over the whole union (+16B) instead of picking the 2nd bank _in_
the union. This manifests in mismatching feature bits for the 2nd bank
and possibly other side-effects caused by writing beyond the union.

We need to step over the first bank (+8B) by using as_uint64[0/1] to
correct this behaviour.

Resolves: Coverity CID 1660876
Fixes: 2f6da91e8a ("accel/mshv: store partition proc features")
Signed-off-by: Magnus Kulke <magnuskulke@linux.microsoft.com>
Reviewed-by: Doru Blânzeanu <dblanzeanu@linux.microsoft.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-ID: <20260701130335.418156-1-magnuskulke@linux.microsoft.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
---
 accel/mshv/mshv-all.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/accel/mshv/mshv-all.c b/accel/mshv/mshv-all.c
index 1959baf6672..72721d0f0de 100644
--- a/accel/mshv/mshv-all.c
+++ b/accel/mshv/mshv-all.c
@@ -167,7 +167,7 @@ static int get_proc_features(int vm_fd,
 
     ret = get_partition_property(vm_fd,
                                  HV_PARTITION_PROPERTY_PROCESSOR_FEATURES0,
-                                 features[0].as_uint64);
+                                 &features->as_uint64[0]);
     if (ret < 0) {
         error_report("Failed to get processor features bank 0");
         return -1;
@@ -175,7 +175,7 @@ static int get_proc_features(int vm_fd,
 
     ret = get_partition_property(vm_fd,
                                  HV_PARTITION_PROPERTY_PROCESSOR_FEATURES1,
-                                 features[1].as_uint64);
+                                 &features->as_uint64[1]);
     if (ret < 0) {
         error_report("Failed to get processor features bank 1");
         return -1;
-- 
2.53.0

[PULL 07/32] gdbstub: Add trace event for STEP packet handler

[Philippe Mathieu-Daudé]

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Daniel Henrique Barboza <daniel.barboza@oss.qualcomm.com>
Acked-by: Alex Bennée <alex.bennee@linaro.org>
Message-ID: <20260705215729.62196-7-philmd@oss.qualcomm.com>
---
 gdbstub/gdbstub.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/gdbstub/gdbstub.c b/gdbstub/gdbstub.c
index 26366acb674..246e7746300 100644
--- a/gdbstub/gdbstub.c
+++ b/gdbstub/gdbstub.c
@@ -1369,6 +1369,7 @@ static void handle_step(GArray *params, void *user_ctx)
         gdb_set_cpu_pc(gdb_get_cmd_param(params, 0)->val_ull);
     }
 
+    trace_gdbstub_op_stepping(gdbserver_state.c_cpu->cpu_index);
     cpu_single_step(gdbserver_state.c_cpu, gdbserver_state.sstep_flags);
     gdb_continue();
 }
-- 
2.53.0

[PULL 08/32] gdbstub: Only return E22 when reverse GDB is not supported

[Philippe Mathieu-Daudé]

No need to keep processing the arguments when we know
reverse debugging is not available.

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Daniel Henrique Barboza <daniel.barboza@oss.qualcomm.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-ID: <20260705215729.62196-8-philmd@oss.qualcomm.com>
---
 gdbstub/gdbstub.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/gdbstub/gdbstub.c b/gdbstub/gdbstub.c
index 246e7746300..f54415a9db7 100644
--- a/gdbstub/gdbstub.c
+++ b/gdbstub/gdbstub.c
@@ -1378,6 +1378,7 @@ static void handle_backward(GArray *params, void *user_ctx)
 {
     if (!gdb_can_reverse()) {
         gdb_put_packet("E22");
+        return;
     }
     if (params->len == 1) {
         switch (gdb_get_cmd_param(params, 0)->opcode) {
-- 
2.53.0

[PULL 09/32] accel/whpx: Implement missing AccelClass::gdbstub_supported_sstep_flags

[Philippe Mathieu-Daudé]

Correct gdbstub support requires some gdbstub_supported_sstep_flags.
Apparently missed in commit d7482ffe975 ("whpx: Added support for
breakpoints and stepping"), even with the recent 19b48084f71 ("whpx:
i386: re-enable guest debug support") fixes.

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Daniel Henrique Barboza <daniel.barboza@oss.qualcomm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20260705215729.62196-9-philmd@oss.qualcomm.com>
---
 include/system/whpx-all.h   | 1 +
 accel/whpx/whpx-common.c    | 1 +
 target/arm/whpx/whpx-all.c  | 5 +++++
 target/i386/whpx/whpx-all.c | 5 +++++
 4 files changed, 12 insertions(+)

diff --git a/include/system/whpx-all.h b/include/system/whpx-all.h
index 4022571ffff..0e2d333e919 100644
--- a/include/system/whpx-all.h
+++ b/include/system/whpx-all.h
@@ -23,6 +23,7 @@ void whpx_arch_destroy_vcpu(CPUState *cpu);
 void whpx_arch_accel_class_init(ObjectClass *oc);
 
 /* called by whpx-accel-ops */
+int whpx_arch_gdbstub_sstep_flags(AccelState *as);
 bool whpx_arch_supports_guest_debug(void);
 
 #endif
diff --git a/accel/whpx/whpx-common.c b/accel/whpx/whpx-common.c
index 247e12db812..e3f93efc61f 100644
--- a/accel/whpx/whpx-common.c
+++ b/accel/whpx/whpx-common.c
@@ -527,6 +527,7 @@ static void whpx_accel_class_init(ObjectClass *oc, const void *data)
     ac->init_machine = whpx_accel_init;
     ac->pre_resume_vm = whpx_pre_resume_vm;
     ac->allowed = &whpx_allowed;
+    ac->gdbstub_supported_sstep_flags = whpx_arch_gdbstub_sstep_flags;
 
     object_class_property_add(oc, "kernel-irqchip", "on|off|split",
         NULL, whpx_set_kernel_irqchip,
diff --git a/target/arm/whpx/whpx-all.c b/target/arm/whpx/whpx-all.c
index 3079c6293c8..5208d7e1dfe 100644
--- a/target/arm/whpx/whpx-all.c
+++ b/target/arm/whpx/whpx-all.c
@@ -295,6 +295,11 @@ void whpx_translate_cpu_breakpoints(
     /* Breakpoints aren’t supported on this platform */
 }
 
+int whpx_arch_gdbstub_sstep_flags(AccelState *as)
+{
+    return 0;
+}
+
 bool whpx_arch_supports_guest_debug(void) 
 {
     return false;
diff --git a/target/i386/whpx/whpx-all.c b/target/i386/whpx/whpx-all.c
index e626acef2f0..f6e1636a4e0 100644
--- a/target/i386/whpx/whpx-all.c
+++ b/target/i386/whpx/whpx-all.c
@@ -1853,6 +1853,11 @@ void whpx_apply_breakpoints(
     }
 }
 
+int whpx_arch_gdbstub_sstep_flags(AccelState *as)
+{
+    return SSTEP_ENABLE;
+}
+
 bool whpx_arch_supports_guest_debug(void) 
 {
     return true;
-- 
2.53.0

[PULL 10/32] accel/kvm: Always define AccelOpsClass::supports_guest_debug

[Philippe Mathieu-Daudé]

Whether TARGET_KVM_HAVE_GUEST_DEBUG is defined or not,
kvm_supports_guest_debug() still exists and can be called.

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Daniel Henrique Barboza <daniel.barboza@oss.qualcomm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20260705215729.62196-10-philmd@oss.qualcomm.com>
---
 accel/kvm/kvm-accel-ops.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/accel/kvm/kvm-accel-ops.c b/accel/kvm/kvm-accel-ops.c
index 6d9140e549f..d7967bc922f 100644
--- a/accel/kvm/kvm-accel-ops.c
+++ b/accel/kvm/kvm-accel-ops.c
@@ -105,9 +105,9 @@ static void kvm_accel_ops_class_init(ObjectClass *oc, const void *data)
     ops->synchronize_pre_loadvm = kvm_cpu_synchronize_pre_loadvm;
     ops->handle_interrupt = generic_handle_interrupt;
 
+    ops->supports_guest_debug = kvm_supports_guest_debug;
 #ifdef TARGET_KVM_HAVE_GUEST_DEBUG
     ops->update_guest_debug = kvm_update_guest_debug_ops;
-    ops->supports_guest_debug = kvm_supports_guest_debug;
     ops->insert_breakpoint = kvm_insert_breakpoint;
     ops->remove_breakpoint = kvm_remove_breakpoint;
     ops->remove_all_breakpoints = kvm_remove_all_breakpoints;
-- 
2.53.0

[PULL 11/32] accel/kvm: Simplify kvm_init() w.r.t. TARGET_KVM_HAVE_GUEST_DEBUG

[Philippe Mathieu-Daudé]

Simplify #ifdef'ry using TARGET_KVM_HAVE_GUEST_DEBUG in kvm_init().
No need to zero-initialize kvm_sstep_flags again.

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20260705215729.62196-11-philmd@oss.qualcomm.com>
---
 accel/kvm/kvm-all.c | 5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)

diff --git a/accel/kvm/kvm-all.c b/accel/kvm/kvm-all.c
index 5d55cb45cf7..304467704ee 100644
--- a/accel/kvm/kvm-all.c
+++ b/accel/kvm/kvm-all.c
@@ -3040,21 +3040,18 @@ static int kvm_init(AccelState *as, MachineState *ms)
 #ifdef TARGET_KVM_HAVE_GUEST_DEBUG
     kvm_has_guest_debug =
         (kvm_check_extension(s, KVM_CAP_SET_GUEST_DEBUG) > 0);
-#endif
 
-    kvm_sstep_flags = 0;
     if (kvm_has_guest_debug) {
         kvm_sstep_flags = SSTEP_ENABLE;
 
-#if defined TARGET_KVM_HAVE_GUEST_DEBUG
         int guest_debug_flags =
             kvm_check_extension(s, KVM_CAP_SET_GUEST_DEBUG2);
 
         if (guest_debug_flags & KVM_GUESTDBG_BLOCKIRQ) {
             kvm_sstep_flags |= SSTEP_NOIRQ;
         }
-#endif
     }
+#endif
 
     kvm_state = s;
 
-- 
2.53.0

[PULL 12/32] accel/kvm: Hold have_guest_debug in KVMState

[Philippe Mathieu-Daudé]

Prefer to store per-accelerator variables in the per-accelerator
state, rather than as static variables. This is a good practice
to allow concurrent accelerators in the future.

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Daniel Henrique Barboza <daniel.barboza@oss.qualcomm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20260705215729.62196-12-philmd@oss.qualcomm.com>
---
 include/system/kvm_int.h | 1 +
 accel/kvm/kvm-all.c      | 9 +++++----
 2 files changed, 6 insertions(+), 4 deletions(-)

diff --git a/include/system/kvm_int.h b/include/system/kvm_int.h
index 0876aac938d..53bb0fcf180 100644
--- a/include/system/kvm_int.h
+++ b/include/system/kvm_int.h
@@ -118,6 +118,7 @@ struct KVMState
 #endif
     int max_nested_state_len;
     int kvm_shadow_mem;
+    bool have_guest_debug;
     bool kernel_irqchip_allowed;
     bool kernel_irqchip_required;
     OnOffAuto kernel_irqchip_split;
diff --git a/accel/kvm/kvm-all.c b/accel/kvm/kvm-all.c
index 304467704ee..b6125125f82 100644
--- a/accel/kvm/kvm-all.c
+++ b/accel/kvm/kvm-all.c
@@ -104,7 +104,6 @@ bool kvm_readonly_mem_allowed;
 bool kvm_vm_attributes_allowed;
 bool kvm_msi_use_devid;
 bool kvm_pre_fault_memory_supported;
-static bool kvm_has_guest_debug;
 static int kvm_sstep_flags;
 static bool kvm_immediate_exit;
 static uint64_t kvm_supported_memory_attributes;
@@ -3038,10 +3037,10 @@ static int kvm_init(AccelState *as, MachineState *ms)
         (kvm_check_extension(s, KVM_CAP_VM_ATTRIBUTES) > 0);
 
 #ifdef TARGET_KVM_HAVE_GUEST_DEBUG
-    kvm_has_guest_debug =
+    s->have_guest_debug =
         (kvm_check_extension(s, KVM_CAP_SET_GUEST_DEBUG) > 0);
 
-    if (kvm_has_guest_debug) {
+    if (s->have_guest_debug) {
         kvm_sstep_flags = SSTEP_ENABLE;
 
         int guest_debug_flags =
@@ -3833,8 +3832,10 @@ int kvm_update_guest_debug(CPUState *cpu, unsigned long reinject_trap)
 
 bool kvm_supports_guest_debug(void)
 {
+    KVMState *s = KVM_STATE(as);
+
     /* probed during kvm_init() */
-    return kvm_has_guest_debug;
+    return s->have_guest_debug;
 }
 
 int kvm_insert_breakpoint(CPUState *cpu, int type, vaddr addr, vaddr len)
-- 
2.53.0

[PULL 13/32] gdbstub: Reduce gdb_supports_guest_debug() scope

[Philippe Mathieu-Daudé]

gdb_supports_guest_debug() is only required for system emulation,
reduce its scope by making it private; remove the user emulation
variant.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Message-ID: <20260705215729.62196-13-philmd@oss.qualcomm.com>
---
 gdbstub/internals.h | 1 -
 gdbstub/system.c    | 4 +++-
 gdbstub/user.c      | 6 ------
 3 files changed, 3 insertions(+), 8 deletions(-)

diff --git a/gdbstub/internals.h b/gdbstub/internals.h
index 9b25bf58b8e..60ddc146661 100644
--- a/gdbstub/internals.h
+++ b/gdbstub/internals.h
@@ -217,7 +217,6 @@ void gdb_syscall_handling(const char *syscall_packet);
  * Break/Watch point support - there is an implementation for system
  * and user mode.
  */
-bool gdb_supports_guest_debug(void);
 int gdb_breakpoint_insert(CPUState *cs, int type, vaddr addr, vaddr len);
 int gdb_breakpoint_remove(CPUState *cs, int type, vaddr addr, vaddr len);
 void gdb_breakpoint_remove_all(CPUState *cs);
diff --git a/gdbstub/system.c b/gdbstub/system.c
index 2063b63b2f8..1179c82cb86 100644
--- a/gdbstub/system.c
+++ b/gdbstub/system.c
@@ -331,6 +331,8 @@ static void create_processes(GDBState *s)
     gdb_create_default_process(s);
 }
 
+static bool gdb_supports_guest_debug(void);
+
 bool gdbserver_start(const char *device, Error **errp)
 {
     Chardev *chr = NULL;
@@ -627,7 +629,7 @@ int gdb_signal_to_target(int sig)
  * Break/Watch point helpers
  */
 
-bool gdb_supports_guest_debug(void)
+static bool gdb_supports_guest_debug(void)
 {
     const AccelOpsClass *ops = cpus_get_accel();
     if (ops->supports_guest_debug) {
diff --git a/gdbstub/user.c b/gdbstub/user.c
index 97eb13e796a..f7bc4e63dfd 100644
--- a/gdbstub/user.c
+++ b/gdbstub/user.c
@@ -796,12 +796,6 @@ bool gdb_can_reverse(void)
  * Break/Watch point helpers
  */
 
-bool gdb_supports_guest_debug(void)
-{
-    /* user-mode == TCG == supported */
-    return true;
-}
-
 int gdb_breakpoint_insert(CPUState *cs, int type, vaddr addr, vaddr len)
 {
     CPUState *cpu;
-- 
2.53.0

[PULL 14/32] gdbstub: Move supported_sstep_flags in AccelGdbConfig structure

[Philippe Mathieu-Daudé]

supported_sstep_flags are per-accelerators. Move them
to a new AccelGdbConfig structure, still in GDBState.

Suggested-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20260705215729.62196-14-philmd@oss.qualcomm.com>
---
 gdbstub/internals.h  |  3 ++-
 include/qemu/accel.h |  9 +++++++++
 gdbstub/gdbstub.c    | 10 +++++-----
 3 files changed, 16 insertions(+), 6 deletions(-)

diff --git a/gdbstub/internals.h b/gdbstub/internals.h
index 60ddc146661..0444dee2246 100644
--- a/gdbstub/internals.h
+++ b/gdbstub/internals.h
@@ -9,6 +9,7 @@
 #ifndef GDBSTUB_INTERNALS_H
 #define GDBSTUB_INTERNALS_H
 
+#include "qemu/accel.h"
 #include "exec/cpu-common.h"
 
 /*
@@ -83,8 +84,8 @@ typedef struct GDBState {
     int process_num;
     GString *str_buf;
     GByteArray *mem_buf;
+    AccelGdbConfig accel_config;
     int sstep_flags;
-    int supported_sstep_flags;
     /*
      * Whether we are allowed to send a stop reply packet at this moment.
      * Must be set off after sending the stop reply itself.
diff --git a/include/qemu/accel.h b/include/qemu/accel.h
index d3638c7bfda..052b4c59313 100644
--- a/include/qemu/accel.h
+++ b/include/qemu/accel.h
@@ -73,6 +73,15 @@ bool accel_cpu_common_realize(CPUState *cpu, Error **errp);
  */
 void accel_cpu_common_unrealize(CPUState *cpu);
 
+/**
+ * struct AccelGdbConfig - gdbstub configuration for an accelerator.
+ *
+ * @sstep_flags: Set SSTEP_* flags that accelerator supports for guest debug.
+ */
+typedef struct AccelGdbConfig {
+    unsigned sstep_flags;
+} AccelGdbConfig;
+
 /**
  * accel_supported_gdbstub_sstep_flags:
  *
diff --git a/gdbstub/gdbstub.c b/gdbstub/gdbstub.c
index f54415a9db7..06b712562c5 100644
--- a/gdbstub/gdbstub.c
+++ b/gdbstub/gdbstub.c
@@ -72,9 +72,9 @@ void gdb_init_gdbserver_state(void)
      * By default try to use no IRQs and no timers while single
      * stepping so as to make single stepping like a typical ICE HW step.
      */
-    gdbserver_state.supported_sstep_flags = accel_supported_gdbstub_sstep_flags();
+    gdbserver_state.accel_config.sstep_flags = accel_supported_gdbstub_sstep_flags();
     gdbserver_state.sstep_flags = SSTEP_ENABLE | SSTEP_NOIRQ | SSTEP_NOTIMER;
-    gdbserver_state.sstep_flags &= gdbserver_state.supported_sstep_flags;
+    gdbserver_state.sstep_flags &= gdbserver_state.accel_config.sstep_flags;
 }
 
 /* writes 2*len+1 bytes in buf */
@@ -1537,12 +1537,12 @@ static void handle_query_qemu_sstepbits(GArray *params, void *user_ctx)
 {
     g_string_printf(gdbserver_state.str_buf, "ENABLE=%x", SSTEP_ENABLE);
 
-    if (gdbserver_state.supported_sstep_flags & SSTEP_NOIRQ) {
+    if (gdbserver_state.accel_config.sstep_flags & SSTEP_NOIRQ) {
         g_string_append_printf(gdbserver_state.str_buf, ",NOIRQ=%x",
                                SSTEP_NOIRQ);
     }
 
-    if (gdbserver_state.supported_sstep_flags & SSTEP_NOTIMER) {
+    if (gdbserver_state.accel_config.sstep_flags & SSTEP_NOTIMER) {
         g_string_append_printf(gdbserver_state.str_buf, ",NOTIMER=%x",
                                SSTEP_NOTIMER);
     }
@@ -1560,7 +1560,7 @@ static void handle_set_qemu_sstep(GArray *params, void *user_ctx)
 
     new_sstep_flags = gdb_get_cmd_param(params, 0)->val_ul;
 
-    if (new_sstep_flags  & ~gdbserver_state.supported_sstep_flags) {
+    if (new_sstep_flags & ~gdbserver_state.accel_config.sstep_flags) {
         gdb_put_packet("E22");
         return;
     }
-- 
2.53.0

[PULL 15/32] accel: Have each implementation return their AccelGdbConfig

[Philippe Mathieu-Daudé]

Hold the per-accelerator AccelGdbConfig in AccelState, set its
single @sstep_flags field in AccelClass::init_machine handlers.

Remove the AccelClass::gdbstub_supported_sstep_flags() getter
and inline the single accel_supported_gdbstub_sstep_flags() call
in gdb_init_gdbserver_state().

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20260705215729.62196-15-philmd@oss.qualcomm.com>
---
 include/accel/accel-ops.h   |  5 ++---
 include/qemu/accel.h        |  8 --------
 include/system/whpx-all.h   |  1 -
 accel/accel-common.c        | 10 ----------
 accel/hvf/hvf-all.c         |  8 ++------
 accel/kvm/kvm-all.c         | 17 ++---------------
 accel/tcg/tcg-all.c         | 27 +++++++++++----------------
 accel/whpx/whpx-common.c    |  1 -
 gdbstub/gdbstub.c           |  2 +-
 target/arm/whpx/whpx-all.c  |  5 -----
 target/i386/whpx/whpx-all.c |  7 ++-----
 11 files changed, 20 insertions(+), 71 deletions(-)

diff --git a/include/accel/accel-ops.h b/include/accel/accel-ops.h
index f46492e3fe1..9f3f329c448 100644
--- a/include/accel/accel-ops.h
+++ b/include/accel/accel-ops.h
@@ -13,6 +13,8 @@
 
 struct AccelState {
     Object parent_obj;
+
+    AccelGdbConfig gdbstub;
 };
 
 struct AccelClass {
@@ -36,9 +38,6 @@ struct AccelClass {
     bool (*has_memory)(AccelState *accel, AddressSpace *as,
                        hwaddr start_addr, hwaddr size);
 
-    /* gdbstub related hooks */
-    int (*gdbstub_supported_sstep_flags)(AccelState *as);
-
     bool *allowed;
     /*
      * Array of global properties that would be applied when specific
diff --git a/include/qemu/accel.h b/include/qemu/accel.h
index 052b4c59313..3b01d55741b 100644
--- a/include/qemu/accel.h
+++ b/include/qemu/accel.h
@@ -82,12 +82,4 @@ typedef struct AccelGdbConfig {
     unsigned sstep_flags;
 } AccelGdbConfig;
 
-/**
- * accel_supported_gdbstub_sstep_flags:
- *
- * Returns the supported single step modes for the configured
- * accelerator.
- */
-int accel_supported_gdbstub_sstep_flags(void);
-
 #endif /* QEMU_ACCEL_H */
diff --git a/include/system/whpx-all.h b/include/system/whpx-all.h
index 0e2d333e919..4022571ffff 100644
--- a/include/system/whpx-all.h
+++ b/include/system/whpx-all.h
@@ -23,7 +23,6 @@ void whpx_arch_destroy_vcpu(CPUState *cpu);
 void whpx_arch_accel_class_init(ObjectClass *oc);
 
 /* called by whpx-accel-ops */
-int whpx_arch_gdbstub_sstep_flags(AccelState *as);
 bool whpx_arch_supports_guest_debug(void);
 
 #endif
diff --git a/accel/accel-common.c b/accel/accel-common.c
index 62590a7d9a6..f31d68a9b78 100644
--- a/accel/accel-common.c
+++ b/accel/accel-common.c
@@ -113,16 +113,6 @@ void accel_cpu_common_unrealize(CPUState *cpu)
     }
 }
 
-int accel_supported_gdbstub_sstep_flags(void)
-{
-    AccelState *accel = current_accel();
-    AccelClass *acc = ACCEL_GET_CLASS(accel);
-    if (acc->gdbstub_supported_sstep_flags) {
-        return acc->gdbstub_supported_sstep_flags(accel);
-    }
-    return 0;
-}
-
 static const TypeInfo accel_types[] = {
     {
         .name           = TYPE_ACCEL,
diff --git a/accel/hvf/hvf-all.c b/accel/hvf/hvf-all.c
index 21b9b71a6df..8bd9154a50b 100644
--- a/accel/hvf/hvf-all.c
+++ b/accel/hvf/hvf-all.c
@@ -220,6 +220,8 @@ static int hvf_accel_init(AccelState *as, MachineState *ms)
     }
     assert_hvf_ok(ret);
 
+    as->gdbstub.sstep_flags = SSTEP_ENABLE | SSTEP_NOIRQ;
+
     QTAILQ_INIT(&s->hvf_sw_breakpoints);
 
     hvf_state = s;
@@ -228,11 +230,6 @@ static int hvf_accel_init(AccelState *as, MachineState *ms)
     return hvf_arch_init();
 }
 
-static int hvf_gdbstub_sstep_flags(AccelState *as)
-{
-    return SSTEP_ENABLE | SSTEP_NOIRQ;
-}
-
 static void hvf_set_kernel_irqchip(Object *obj, Visitor *v,
                                    const char *name, void *opaque,
                                    Error **errp)
@@ -278,7 +275,6 @@ static void hvf_accel_class_init(ObjectClass *oc, const void *data)
     ac->name = "HVF";
     ac->init_machine = hvf_accel_init;
     ac->allowed = &hvf_allowed;
-    ac->gdbstub_supported_sstep_flags = hvf_gdbstub_sstep_flags;
     hvf_kernel_irqchip_override = false;
     hvf_kernel_irqchip = false;
     object_class_property_add(oc, "kernel-irqchip", "on|off|split",
diff --git a/accel/kvm/kvm-all.c b/accel/kvm/kvm-all.c
index b6125125f82..ceb0b33abdd 100644
--- a/accel/kvm/kvm-all.c
+++ b/accel/kvm/kvm-all.c
@@ -104,7 +104,6 @@ bool kvm_readonly_mem_allowed;
 bool kvm_vm_attributes_allowed;
 bool kvm_msi_use_devid;
 bool kvm_pre_fault_memory_supported;
-static int kvm_sstep_flags;
 static bool kvm_immediate_exit;
 static uint64_t kvm_supported_memory_attributes;
 static bool kvm_guest_memfd_supported;
@@ -3041,13 +3040,13 @@ static int kvm_init(AccelState *as, MachineState *ms)
         (kvm_check_extension(s, KVM_CAP_SET_GUEST_DEBUG) > 0);
 
     if (s->have_guest_debug) {
-        kvm_sstep_flags = SSTEP_ENABLE;
+        as->gdbstub.sstep_flags = SSTEP_ENABLE;
 
         int guest_debug_flags =
             kvm_check_extension(s, KVM_CAP_SET_GUEST_DEBUG2);
 
         if (guest_debug_flags & KVM_GUESTDBG_BLOCKIRQ) {
-            kvm_sstep_flags |= SSTEP_NOIRQ;
+            as->gdbstub.sstep_flags |= SSTEP_NOIRQ;
         }
     }
 #endif
@@ -4279,17 +4278,6 @@ static void kvm_accel_instance_init(Object *obj)
     s->honor_guest_pat = ON_OFF_AUTO_OFF;
 }
 
-/**
- * kvm_gdbstub_sstep_flags():
- *
- * Returns: SSTEP_* flags that KVM supports for guest debug. The
- * support is probed during kvm_init()
- */
-static int kvm_gdbstub_sstep_flags(AccelState *as)
-{
-    return kvm_sstep_flags;
-}
-
 static void kvm_accel_class_init(ObjectClass *oc, const void *data)
 {
     AccelClass *ac = ACCEL_CLASS(oc);
@@ -4298,7 +4286,6 @@ static void kvm_accel_class_init(ObjectClass *oc, const void *data)
     ac->rebuild_guest = kvm_reset_vmfd;
     ac->has_memory = kvm_accel_has_memory;
     ac->allowed = &kvm_allowed;
-    ac->gdbstub_supported_sstep_flags = kvm_gdbstub_sstep_flags;
 
     object_class_property_add(oc, "kernel-irqchip", "on|off|split",
         NULL, kvm_set_kernel_irqchip,
diff --git a/accel/tcg/tcg-all.c b/accel/tcg/tcg-all.c
index 8eb4a6b89e8..d30b53cd3b7 100644
--- a/accel/tcg/tcg-all.c
+++ b/accel/tcg/tcg-all.c
@@ -149,6 +149,17 @@ static int tcg_init_machine(AccelState *as, MachineState *ms)
 
     tcg_allowed = true;
 
+    as->gdbstub.sstep_flags = SSTEP_ENABLE;
+    if (replay_mode == REPLAY_MODE_NONE) {
+        /*
+         * In replay mode all events will come from the log and can't be
+         * suppressed otherwise we would break determinism. However as those
+         * events are tied to the number of executed instructions we won't see
+         * them occurring every time we single step.
+         */
+        as->gdbstub.sstep_flags |= SSTEP_NOIRQ | SSTEP_NOTIMER;
+    }
+
     page_init();
     tb_htable_init();
     tcg_init(s->tb_size * MiB, s->splitwx_enabled, max_threads);
@@ -242,21 +253,6 @@ static void tcg_set_one_insn_per_tb(Object *obj, bool value, Error **errp)
     qatomic_set(&one_insn_per_tb, value);
 }
 
-static int tcg_gdbstub_supported_sstep_flags(AccelState *as)
-{
-    /*
-     * In replay mode all events will come from the log and can't be
-     * suppressed otherwise we would break determinism. However as those
-     * events are tied to the number of executed instructions we won't see
-     * them occurring every time we single step.
-     */
-    if (replay_mode != REPLAY_MODE_NONE) {
-        return SSTEP_ENABLE;
-    } else {
-        return SSTEP_ENABLE | SSTEP_NOIRQ | SSTEP_NOTIMER;
-    }
-}
-
 static void tcg_accel_class_init(ObjectClass *oc, const void *data)
 {
     AccelClass *ac = ACCEL_CLASS(oc);
@@ -266,7 +262,6 @@ static void tcg_accel_class_init(ObjectClass *oc, const void *data)
     ac->cpu_common_unrealize = tcg_exec_unrealizefn;
     ac->get_stats = tcg_get_stats;
     ac->allowed = &tcg_allowed;
-    ac->gdbstub_supported_sstep_flags = tcg_gdbstub_supported_sstep_flags;
 
     object_class_property_add_str(oc, "thread",
                                   tcg_get_thread,
diff --git a/accel/whpx/whpx-common.c b/accel/whpx/whpx-common.c
index e3f93efc61f..247e12db812 100644
--- a/accel/whpx/whpx-common.c
+++ b/accel/whpx/whpx-common.c
@@ -527,7 +527,6 @@ static void whpx_accel_class_init(ObjectClass *oc, const void *data)
     ac->init_machine = whpx_accel_init;
     ac->pre_resume_vm = whpx_pre_resume_vm;
     ac->allowed = &whpx_allowed;
-    ac->gdbstub_supported_sstep_flags = whpx_arch_gdbstub_sstep_flags;
 
     object_class_property_add(oc, "kernel-irqchip", "on|off|split",
         NULL, whpx_set_kernel_irqchip,
diff --git a/gdbstub/gdbstub.c b/gdbstub/gdbstub.c
index 06b712562c5..6c4fa065638 100644
--- a/gdbstub/gdbstub.c
+++ b/gdbstub/gdbstub.c
@@ -72,7 +72,7 @@ void gdb_init_gdbserver_state(void)
      * By default try to use no IRQs and no timers while single
      * stepping so as to make single stepping like a typical ICE HW step.
      */
-    gdbserver_state.accel_config.sstep_flags = accel_supported_gdbstub_sstep_flags();
+    gdbserver_state.accel_config = current_accel()->gdbstub;
     gdbserver_state.sstep_flags = SSTEP_ENABLE | SSTEP_NOIRQ | SSTEP_NOTIMER;
     gdbserver_state.sstep_flags &= gdbserver_state.accel_config.sstep_flags;
 }
diff --git a/target/arm/whpx/whpx-all.c b/target/arm/whpx/whpx-all.c
index 5208d7e1dfe..3079c6293c8 100644
--- a/target/arm/whpx/whpx-all.c
+++ b/target/arm/whpx/whpx-all.c
@@ -295,11 +295,6 @@ void whpx_translate_cpu_breakpoints(
     /* Breakpoints aren’t supported on this platform */
 }
 
-int whpx_arch_gdbstub_sstep_flags(AccelState *as)
-{
-    return 0;
-}
-
 bool whpx_arch_supports_guest_debug(void) 
 {
     return false;
diff --git a/target/i386/whpx/whpx-all.c b/target/i386/whpx/whpx-all.c
index f6e1636a4e0..467ac6bed45 100644
--- a/target/i386/whpx/whpx-all.c
+++ b/target/i386/whpx/whpx-all.c
@@ -1853,11 +1853,6 @@ void whpx_apply_breakpoints(
     }
 }
 
-int whpx_arch_gdbstub_sstep_flags(AccelState *as)
-{
-    return SSTEP_ENABLE;
-}
-
 bool whpx_arch_supports_guest_debug(void) 
 {
     return true;
@@ -3348,6 +3343,8 @@ int whpx_accel_init(AccelState *as, MachineState *ms)
     whpx_memory_init();
     whpx_init_emu();
 
+    as->gdbstub.sstep_flags = SSTEP_ENABLE;
+
     return 0;
 
 error:
-- 
2.53.0

[PULL 16/32] gdbstub: Make default replay_mode value explicit in stubs

[Philippe Mathieu-Daudé]

Make explicit @replay_mode is stubbed as REPLAY_MODE_NONE.

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20260705215729.62196-16-philmd@oss.qualcomm.com>
---
 stubs/replay-mode.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/stubs/replay-mode.c b/stubs/replay-mode.c
index 439d97e4a87..5bc8f7d6589 100644
--- a/stubs/replay-mode.c
+++ b/stubs/replay-mode.c
@@ -1,4 +1,4 @@
 #include "qemu/osdep.h"
 #include "system/replay.h"
 
-ReplayMode replay_mode;
+ReplayMode replay_mode = REPLAY_MODE_NONE;
-- 
2.53.0

[PULL 17/32] accel: Hold @can_reverse information in AccelGdbConfig

[Philippe Mathieu-Daudé]

Hold @can_reverse in AccelGdbConfig, set it when initializing
AccelState in AccelClass::init_machine handlers (only TCG sets
it). Remove gdb_can_reverse() as now unused.

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Manos Pitsidianakis <manos.pitsidianakis@linaro.org>
Message-ID: <20260705215729.62196-17-philmd@oss.qualcomm.com>
---
 gdbstub/internals.h  | 1 -
 include/qemu/accel.h | 2 ++
 accel/tcg/tcg-all.c  | 3 +++
 gdbstub/gdbstub.c    | 4 ++--
 gdbstub/system.c     | 5 -----
 gdbstub/user.c       | 6 ------
 6 files changed, 7 insertions(+), 14 deletions(-)

diff --git a/gdbstub/internals.h b/gdbstub/internals.h
index 0444dee2246..0b74ea4000f 100644
--- a/gdbstub/internals.h
+++ b/gdbstub/internals.h
@@ -154,7 +154,6 @@ CPUState *gdb_first_attached_cpu(void);
 void gdb_append_thread_id(CPUState *cpu, GString *buf);
 int gdb_get_cpu_index(CPUState *cpu);
 unsigned int gdb_get_max_cpus(void); /* both */
-bool gdb_can_reverse(void); /* system emulation, stub for user */
 int gdb_target_sigtrap(void); /* user */
 
 void gdb_create_default_process(GDBState *s);
diff --git a/include/qemu/accel.h b/include/qemu/accel.h
index 3b01d55741b..8e0c3dfe088 100644
--- a/include/qemu/accel.h
+++ b/include/qemu/accel.h
@@ -77,9 +77,11 @@ void accel_cpu_common_unrealize(CPUState *cpu);
  * struct AccelGdbConfig - gdbstub configuration for an accelerator.
  *
  * @sstep_flags: Set SSTEP_* flags that accelerator supports for guest debug.
+ * @can_reverse: Whether reverse mode is supported.
  */
 typedef struct AccelGdbConfig {
     unsigned sstep_flags;
+    bool can_reverse;
 } AccelGdbConfig;
 
 #endif /* QEMU_ACCEL_H */
diff --git a/accel/tcg/tcg-all.c b/accel/tcg/tcg-all.c
index d30b53cd3b7..7186c10cf02 100644
--- a/accel/tcg/tcg-all.c
+++ b/accel/tcg/tcg-all.c
@@ -159,6 +159,9 @@ static int tcg_init_machine(AccelState *as, MachineState *ms)
          */
         as->gdbstub.sstep_flags |= SSTEP_NOIRQ | SSTEP_NOTIMER;
     }
+    if (replay_mode == REPLAY_MODE_PLAY) {
+        as->gdbstub.can_reverse = true;
+    }
 
     page_init();
     tb_htable_init();
diff --git a/gdbstub/gdbstub.c b/gdbstub/gdbstub.c
index 6c4fa065638..0a328b0dd48 100644
--- a/gdbstub/gdbstub.c
+++ b/gdbstub/gdbstub.c
@@ -1376,7 +1376,7 @@ static void handle_step(GArray *params, void *user_ctx)
 
 static void handle_backward(GArray *params, void *user_ctx)
 {
-    if (!gdb_can_reverse()) {
+    if (!gdbserver_state.accel_config.can_reverse) {
         gdb_put_packet("E22");
         return;
     }
@@ -1684,7 +1684,7 @@ static void handle_query_supported(GArray *params, void *user_ctx)
         g_string_append(gdbserver_state.str_buf, ";qXfer:features:read+");
     }
 
-    if (gdb_can_reverse()) {
+    if (gdbserver_state.accel_config.can_reverse) {
         g_string_append(gdbserver_state.str_buf,
             ";ReverseStep+;ReverseContinue+");
     }
diff --git a/gdbstub/system.c b/gdbstub/system.c
index 1179c82cb86..2e8c457097b 100644
--- a/gdbstub/system.c
+++ b/gdbstub/system.c
@@ -479,11 +479,6 @@ unsigned int gdb_get_max_cpus(void)
     return ms->smp.max_cpus;
 }
 
-bool gdb_can_reverse(void)
-{
-    return replay_mode == REPLAY_MODE_PLAY;
-}
-
 /*
  * Softmmu specific command helpers
  */
diff --git a/gdbstub/user.c b/gdbstub/user.c
index f7bc4e63dfd..299388b9f71 100644
--- a/gdbstub/user.c
+++ b/gdbstub/user.c
@@ -786,12 +786,6 @@ unsigned int gdb_get_max_cpus(void)
     return max_cpus;
 }
 
-/* replay not supported for user-mode */
-bool gdb_can_reverse(void)
-{
-    return false;
-}
-
 /*
  * Break/Watch point helpers
  */
-- 
2.53.0

[PULL 18/32] accel: Remove AccelOpsClass::supports_guest_debug

[Philippe Mathieu-Daudé]

Now accelerators hold the 'guest debug supported' information
in their state, accessible by the common code. No need to call
a per-accelerator handler, simply check for the SSTEP_ENABLE
in AccelGdbConfig::sstep_flags.

Remove all AccelOpsClass::supports_guest_debug implementations,
inline gdb_supports_guest_debug() and remove the now unnecessary
KVMState::have_guest_debug field.

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20260705215729.62196-18-philmd@oss.qualcomm.com>
---
 docs/system/gdb.rst           |  2 +-
 accel/kvm/kvm-cpus.h          |  1 -
 include/accel/accel-cpu-ops.h |  1 -
 include/qemu/accel.h          |  2 ++
 include/system/hvf_int.h      |  5 -----
 include/system/kvm_int.h      |  1 -
 include/system/whpx-all.h     |  3 ---
 accel/accel-common.c          |  5 +++++
 accel/hvf/hvf-accel-ops.c     |  1 -
 accel/kvm/kvm-accel-ops.c     |  1 -
 accel/kvm/kvm-all.c           | 13 +------------
 accel/tcg/tcg-accel-ops.c     |  6 ------
 accel/whpx/whpx-accel-ops.c   |  6 ------
 gdbstub/system.c              | 13 +------------
 target/arm/hvf/hvf.c          |  5 -----
 target/arm/whpx/whpx-all.c    |  5 -----
 target/i386/hvf/hvf.c         |  5 -----
 target/i386/whpx/whpx-all.c   |  5 -----
 18 files changed, 10 insertions(+), 70 deletions(-)

diff --git a/docs/system/gdb.rst b/docs/system/gdb.rst
index d50470b135e..b6d8f8e77da 100644
--- a/docs/system/gdb.rst
+++ b/docs/system/gdb.rst
@@ -54,7 +54,7 @@ While GDB can always fall back to inserting breakpoints into memory
 accelerator. For TCG system emulation we advertise an infinite number
 of hardware assisted breakpoints and watchpoints. For other
 accelerators it will depend on if support has been added (see
-supports_guest_debug and related hooks in AccelOpsClass).
+the AccelGdbConfig structure).
 
 As TCG cannot track all memory accesses in user-mode there is no
 support for watchpoints.
diff --git a/accel/kvm/kvm-cpus.h b/accel/kvm/kvm-cpus.h
index 688511151c8..3185659562d 100644
--- a/accel/kvm/kvm-cpus.h
+++ b/accel/kvm/kvm-cpus.h
@@ -16,7 +16,6 @@ void kvm_destroy_vcpu(CPUState *cpu);
 void kvm_cpu_synchronize_post_reset(CPUState *cpu);
 void kvm_cpu_synchronize_post_init(CPUState *cpu);
 void kvm_cpu_synchronize_pre_loadvm(CPUState *cpu);
-bool kvm_supports_guest_debug(void);
 int kvm_insert_breakpoint(CPUState *cpu, int type, vaddr addr, vaddr len);
 int kvm_remove_breakpoint(CPUState *cpu, int type, vaddr addr, vaddr len);
 void kvm_remove_all_breakpoints(CPUState *cpu);
diff --git a/include/accel/accel-cpu-ops.h b/include/accel/accel-cpu-ops.h
index 9c07a903ea0..b23a0606c7e 100644
--- a/include/accel/accel-cpu-ops.h
+++ b/include/accel/accel-cpu-ops.h
@@ -84,7 +84,6 @@ struct AccelOpsClass {
     int64_t (*get_elapsed_ticks)(void);
 
     /* gdbstub hooks */
-    bool (*supports_guest_debug)(void);
     int (*update_guest_debug)(CPUState *cpu);
     int (*insert_breakpoint)(CPUState *cpu, int type, vaddr addr, vaddr len);
     int (*remove_breakpoint)(CPUState *cpu, int type, vaddr addr, vaddr len);
diff --git a/include/qemu/accel.h b/include/qemu/accel.h
index 8e0c3dfe088..6f65f1cb89a 100644
--- a/include/qemu/accel.h
+++ b/include/qemu/accel.h
@@ -84,4 +84,6 @@ typedef struct AccelGdbConfig {
     bool can_reverse;
 } AccelGdbConfig;
 
+bool accel_supports_guest_debug(AccelState *accel);
+
 #endif /* QEMU_ACCEL_H */
diff --git a/include/system/hvf_int.h b/include/system/hvf_int.h
index 58fb865ebae..d64f4942556 100644
--- a/include/system/hvf_int.h
+++ b/include/system/hvf_int.h
@@ -104,11 +104,6 @@ void hvf_arch_remove_all_hw_breakpoints(void);
  */
 int hvf_update_guest_debug(CPUState *cpu);
 
-/*
- * Return whether the guest supports debugging.
- */
-bool hvf_arch_supports_guest_debug(void);
-
 bool hvf_arch_cpu_realize(CPUState *cpu, Error **errp);
 uint32_t hvf_arch_get_default_ipa_bit_size(void);
 uint32_t hvf_arch_get_max_ipa_bit_size(void);
diff --git a/include/system/kvm_int.h b/include/system/kvm_int.h
index 53bb0fcf180..0876aac938d 100644
--- a/include/system/kvm_int.h
+++ b/include/system/kvm_int.h
@@ -118,7 +118,6 @@ struct KVMState
 #endif
     int max_nested_state_len;
     int kvm_shadow_mem;
-    bool have_guest_debug;
     bool kernel_irqchip_allowed;
     bool kernel_irqchip_required;
     OnOffAuto kernel_irqchip_split;
diff --git a/include/system/whpx-all.h b/include/system/whpx-all.h
index 4022571ffff..d018e67e4f7 100644
--- a/include/system/whpx-all.h
+++ b/include/system/whpx-all.h
@@ -22,7 +22,4 @@ void whpx_translate_cpu_breakpoints(
 void whpx_arch_destroy_vcpu(CPUState *cpu);
 void whpx_arch_accel_class_init(ObjectClass *oc);
 
-/* called by whpx-accel-ops */
-bool whpx_arch_supports_guest_debug(void);
-
 #endif
diff --git a/accel/accel-common.c b/accel/accel-common.c
index f31d68a9b78..00a400243f0 100644
--- a/accel/accel-common.c
+++ b/accel/accel-common.c
@@ -70,6 +70,11 @@ void accel_init_interfaces(AccelClass *ac)
     accel_init_cpu_interfaces(ac);
 }
 
+bool accel_supports_guest_debug(AccelState *accel)
+{
+    return accel->gdbstub.sstep_flags & SSTEP_ENABLE;
+}
+
 void accel_cpu_instance_init(CPUState *cpu)
 {
     if (cpu->cc->accel_cpu && cpu->cc->accel_cpu->cpu_instance_init) {
diff --git a/accel/hvf/hvf-accel-ops.c b/accel/hvf/hvf-accel-ops.c
index dd2b45b7cb5..ecc0cd1b558 100644
--- a/accel/hvf/hvf-accel-ops.c
+++ b/accel/hvf/hvf-accel-ops.c
@@ -369,7 +369,6 @@ static void hvf_accel_ops_class_init(ObjectClass *oc, const void *data)
     ops->remove_breakpoint = hvf_remove_breakpoint;
     ops->remove_all_breakpoints = hvf_remove_all_breakpoints;
     ops->update_guest_debug = hvf_update_guest_debug;
-    ops->supports_guest_debug = hvf_arch_supports_guest_debug;
 
     ops->get_vcpu_stats = hvf_get_vcpu_stats;
 };
diff --git a/accel/kvm/kvm-accel-ops.c b/accel/kvm/kvm-accel-ops.c
index d7967bc922f..45330327909 100644
--- a/accel/kvm/kvm-accel-ops.c
+++ b/accel/kvm/kvm-accel-ops.c
@@ -105,7 +105,6 @@ static void kvm_accel_ops_class_init(ObjectClass *oc, const void *data)
     ops->synchronize_pre_loadvm = kvm_cpu_synchronize_pre_loadvm;
     ops->handle_interrupt = generic_handle_interrupt;
 
-    ops->supports_guest_debug = kvm_supports_guest_debug;
 #ifdef TARGET_KVM_HAVE_GUEST_DEBUG
     ops->update_guest_debug = kvm_update_guest_debug_ops;
     ops->insert_breakpoint = kvm_insert_breakpoint;
diff --git a/accel/kvm/kvm-all.c b/accel/kvm/kvm-all.c
index ceb0b33abdd..b03488ebac3 100644
--- a/accel/kvm/kvm-all.c
+++ b/accel/kvm/kvm-all.c
@@ -3036,10 +3036,7 @@ static int kvm_init(AccelState *as, MachineState *ms)
         (kvm_check_extension(s, KVM_CAP_VM_ATTRIBUTES) > 0);
 
 #ifdef TARGET_KVM_HAVE_GUEST_DEBUG
-    s->have_guest_debug =
-        (kvm_check_extension(s, KVM_CAP_SET_GUEST_DEBUG) > 0);
-
-    if (s->have_guest_debug) {
+    if (kvm_check_extension(s, KVM_CAP_SET_GUEST_DEBUG) > 0) {
         as->gdbstub.sstep_flags = SSTEP_ENABLE;
 
         int guest_debug_flags =
@@ -3829,14 +3826,6 @@ int kvm_update_guest_debug(CPUState *cpu, unsigned long reinject_trap)
     return data.err;
 }
 
-bool kvm_supports_guest_debug(void)
-{
-    KVMState *s = KVM_STATE(as);
-
-    /* probed during kvm_init() */
-    return s->have_guest_debug;
-}
-
 int kvm_insert_breakpoint(CPUState *cpu, int type, vaddr addr, vaddr len)
 {
     struct kvm_sw_breakpoint *bp;
diff --git a/accel/tcg/tcg-accel-ops.c b/accel/tcg/tcg-accel-ops.c
index c179cd4adeb..927af6c1db0 100644
--- a/accel/tcg/tcg-accel-ops.c
+++ b/accel/tcg/tcg-accel-ops.c
@@ -109,11 +109,6 @@ void tcg_handle_interrupt(CPUState *cpu, int mask)
     }
 }
 
-static bool tcg_supports_guest_debug(void)
-{
-    return true;
-}
-
 /* Translate GDB watchpoint type to a flags value for cpu_watchpoint_* */
 static inline int xlat_gdb_type(CPUState *cpu, int gdbtype)
 {
@@ -221,7 +216,6 @@ static void tcg_accel_ops_init(AccelClass *ac)
     }
 
     ops->cpu_reset_hold = tcg_cpu_reset_hold;
-    ops->supports_guest_debug = tcg_supports_guest_debug;
     ops->insert_breakpoint = tcg_insert_breakpoint;
     ops->remove_breakpoint = tcg_remove_breakpoint;
     ops->remove_all_breakpoints = tcg_remove_all_breakpoints;
diff --git a/accel/whpx/whpx-accel-ops.c b/accel/whpx/whpx-accel-ops.c
index b8f41544cbe..ca5a119521a 100644
--- a/accel/whpx/whpx-accel-ops.c
+++ b/accel/whpx/whpx-accel-ops.c
@@ -82,11 +82,6 @@ static bool whpx_vcpu_thread_is_idle(CPUState *cpu)
     return !whpx_irqchip_in_kernel();
 }
 
-static bool whpx_supports_guest_debug(void)
-{
-    return whpx_arch_supports_guest_debug();
-}
-
 
 static void whpx_accel_ops_class_init(ObjectClass *oc, const void *data)
 {
@@ -96,7 +91,6 @@ static void whpx_accel_ops_class_init(ObjectClass *oc, const void *data)
     ops->kick_vcpu_thread = whpx_kick_vcpu_thread;
     ops->cpu_thread_is_idle = whpx_vcpu_thread_is_idle;
     ops->handle_interrupt = generic_handle_interrupt;
-    ops->supports_guest_debug = whpx_supports_guest_debug;
 
     ops->synchronize_post_reset = whpx_cpu_synchronize_post_reset;
     ops->synchronize_post_init = whpx_cpu_synchronize_post_init;
diff --git a/gdbstub/system.c b/gdbstub/system.c
index 2e8c457097b..bf7aa1ac234 100644
--- a/gdbstub/system.c
+++ b/gdbstub/system.c
@@ -331,8 +331,6 @@ static void create_processes(GDBState *s)
     gdb_create_default_process(s);
 }
 
-static bool gdb_supports_guest_debug(void);
-
 bool gdbserver_start(const char *device, Error **errp)
 {
     Chardev *chr = NULL;
@@ -345,7 +343,7 @@ bool gdbserver_start(const char *device, Error **errp)
         return false;
     }
 
-    if (!gdb_supports_guest_debug()) {
+    if (!accel_supports_guest_debug(current_accel())) {
         error_setg(errp, "gdbstub: current accelerator doesn't "
                    "support guest debugging");
         return false;
@@ -624,15 +622,6 @@ int gdb_signal_to_target(int sig)
  * Break/Watch point helpers
  */
 
-static bool gdb_supports_guest_debug(void)
-{
-    const AccelOpsClass *ops = cpus_get_accel();
-    if (ops->supports_guest_debug) {
-        return ops->supports_guest_debug();
-    }
-    return false;
-}
-
 int gdb_breakpoint_insert(CPUState *cs, int type, vaddr addr, vaddr len)
 {
     const AccelOpsClass *ops = cpus_get_accel();
diff --git a/target/arm/hvf/hvf.c b/target/arm/hvf/hvf.c
index 8b902c68829..e16457c3cfb 100644
--- a/target/arm/hvf/hvf.c
+++ b/target/arm/hvf/hvf.c
@@ -2900,8 +2900,3 @@ void hvf_arch_update_guest_debug(CPUState *cpu)
 
     hvf_arch_set_traps(cpu);
 }
-
-bool hvf_arch_supports_guest_debug(void)
-{
-    return true;
-}
diff --git a/target/arm/whpx/whpx-all.c b/target/arm/whpx/whpx-all.c
index 3079c6293c8..00a5de8cdc1 100644
--- a/target/arm/whpx/whpx-all.c
+++ b/target/arm/whpx/whpx-all.c
@@ -295,11 +295,6 @@ void whpx_translate_cpu_breakpoints(
     /* Breakpoints aren’t supported on this platform */
 }
 
-bool whpx_arch_supports_guest_debug(void) 
-{
-    return false;
-}
-
 void whpx_arch_destroy_vcpu(CPUState *cpu)
 {
     /* currently empty on Arm */
diff --git a/target/i386/hvf/hvf.c b/target/i386/hvf/hvf.c
index 2d1a943b96c..19ad64a1d93 100644
--- a/target/i386/hvf/hvf.c
+++ b/target/i386/hvf/hvf.c
@@ -1066,8 +1066,3 @@ void hvf_arch_remove_all_hw_breakpoints(void)
 void hvf_arch_update_guest_debug(CPUState *cpu)
 {
 }
-
-bool hvf_arch_supports_guest_debug(void)
-{
-    return false;
-}
diff --git a/target/i386/whpx/whpx-all.c b/target/i386/whpx/whpx-all.c
index 467ac6bed45..11b5d1fc607 100644
--- a/target/i386/whpx/whpx-all.c
+++ b/target/i386/whpx/whpx-all.c
@@ -1853,11 +1853,6 @@ void whpx_apply_breakpoints(
     }
 }
 
-bool whpx_arch_supports_guest_debug(void) 
-{
-    return true;
-}
-
 void whpx_arch_destroy_vcpu(CPUState *cpu)
 {
     X86CPU *x86cpu = X86_CPU(cpu);
-- 
2.53.0

[PULL 19/32] cpu: Move cpu_breakpoint_test out of line

[Philippe Mathieu-Daudé]

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

Move the function to cpu-common.c, with the other
breakpoint functions.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Message-ID: <20260705215729.62196-19-philmd@oss.qualcomm.com>
---
 include/hw/core/cpu.h | 16 +---------------
 cpu-common.c          | 15 +++++++++++++++
 2 files changed, 16 insertions(+), 15 deletions(-)

diff --git a/include/hw/core/cpu.h b/include/hw/core/cpu.h
index 60a23d6f38b..31753f331fb 100644
--- a/include/hw/core/cpu.h
+++ b/include/hw/core/cpu.h
@@ -1157,21 +1157,7 @@ int cpu_breakpoint_insert(CPUState *cpu, vaddr pc, int flags,
 int cpu_breakpoint_remove(CPUState *cpu, vaddr pc, int flags);
 void cpu_breakpoint_remove_by_ref(CPUState *cpu, CPUBreakpoint *breakpoint);
 void cpu_breakpoint_remove_all(CPUState *cpu, int mask);
-
-/* Return true if PC matches an installed breakpoint.  */
-static inline bool cpu_breakpoint_test(CPUState *cpu, vaddr pc, int mask)
-{
-    CPUBreakpoint *bp;
-
-    if (unlikely(!QTAILQ_EMPTY(&cpu->breakpoints))) {
-        QTAILQ_FOREACH(bp, &cpu->breakpoints, entry) {
-            if (bp->pc == pc && (bp->flags & mask)) {
-                return true;
-            }
-        }
-    }
-    return false;
-}
+bool cpu_breakpoint_test(CPUState *cpu, vaddr pc, int mask);
 
 /**
  * cpu_get_address_space:
diff --git a/cpu-common.c b/cpu-common.c
index 988d057d844..adb76b3a786 100644
--- a/cpu-common.c
+++ b/cpu-common.c
@@ -388,6 +388,21 @@ void process_queued_cpu_work(CPUState *cpu)
     qemu_cond_broadcast(&qemu_work_cond);
 }
 
+/* Return true if PC matches an installed breakpoint.  */
+bool cpu_breakpoint_test(CPUState *cpu, vaddr pc, int mask)
+{
+    CPUBreakpoint *bp;
+
+    if (unlikely(!QTAILQ_EMPTY(&cpu->breakpoints))) {
+        QTAILQ_FOREACH(bp, &cpu->breakpoints, entry) {
+            if (bp->pc == pc && (bp->flags & mask)) {
+                return true;
+            }
+        }
+    }
+    return false;
+}
+
 /* Add a breakpoint.  */
 int cpu_breakpoint_insert(CPUState *cpu, vaddr pc, int flags,
                           CPUBreakpoint **breakpoint)
-- 
2.53.0

[PULL 20/32] cpu: Move BREAKPOINT definitions to 'exec/breakpoint.h'

[Philippe Mathieu-Daudé]

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Daniel Henrique Barboza <daniel.barboza@oss.qualcomm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20260705215729.62196-20-philmd@oss.qualcomm.com>
---
 include/exec/breakpoint.h | 14 ++++++++++++++
 include/hw/core/cpu.h     | 14 --------------
 2 files changed, 14 insertions(+), 14 deletions(-)

diff --git a/include/exec/breakpoint.h b/include/exec/breakpoint.h
index 95f0482e6d0..cc6fedd09fe 100644
--- a/include/exec/breakpoint.h
+++ b/include/exec/breakpoint.h
@@ -12,6 +12,20 @@
 #include "exec/vaddr.h"
 #include "exec/memattrs.h"
 
+/* Breakpoint/watchpoint flags */
+#define BP_MEM_READ           0x01
+#define BP_MEM_WRITE          0x02
+#define BP_MEM_ACCESS         (BP_MEM_READ | BP_MEM_WRITE)
+#define BP_STOP_BEFORE_ACCESS 0x04
+/* 0x08 currently unused */
+#define BP_GDB                0x10
+#define BP_CPU                0x20
+#define BP_ANY                (BP_GDB | BP_CPU)
+#define BP_HIT_SHIFT          6
+#define BP_WATCHPOINT_HIT_READ  (BP_MEM_READ << BP_HIT_SHIFT)
+#define BP_WATCHPOINT_HIT_WRITE (BP_MEM_WRITE << BP_HIT_SHIFT)
+#define BP_WATCHPOINT_HIT       (BP_MEM_ACCESS << BP_HIT_SHIFT)
+
 typedef struct CPUBreakpoint {
     vaddr pc;
     int flags; /* BP_* */
diff --git a/include/hw/core/cpu.h b/include/hw/core/cpu.h
index 31753f331fb..e5b68bced3d 100644
--- a/include/hw/core/cpu.h
+++ b/include/hw/core/cpu.h
@@ -1138,20 +1138,6 @@ void qemu_init_vcpu(CPUState *cpu);
  */
 void cpu_single_step(CPUState *cpu, int enabled);
 
-/* Breakpoint/watchpoint flags */
-#define BP_MEM_READ           0x01
-#define BP_MEM_WRITE          0x02
-#define BP_MEM_ACCESS         (BP_MEM_READ | BP_MEM_WRITE)
-#define BP_STOP_BEFORE_ACCESS 0x04
-/* 0x08 currently unused */
-#define BP_GDB                0x10
-#define BP_CPU                0x20
-#define BP_ANY                (BP_GDB | BP_CPU)
-#define BP_HIT_SHIFT          6
-#define BP_WATCHPOINT_HIT_READ  (BP_MEM_READ << BP_HIT_SHIFT)
-#define BP_WATCHPOINT_HIT_WRITE (BP_MEM_WRITE << BP_HIT_SHIFT)
-#define BP_WATCHPOINT_HIT       (BP_MEM_ACCESS << BP_HIT_SHIFT)
-
 int cpu_breakpoint_insert(CPUState *cpu, vaddr pc, int flags,
                           CPUBreakpoint **breakpoint);
 int cpu_breakpoint_remove(CPUState *cpu, vaddr pc, int flags);
-- 
2.53.0

[PULL 21/32] accel: Remove unnecessary 'inline' qualifier in remove_all_breakpoints

[Philippe Mathieu-Daudé]

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Daniel Henrique Barboza <daniel.barboza@oss.qualcomm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20260705215729.62196-22-philmd@oss.qualcomm.com>
---
 accel/tcg/tcg-accel-ops.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/accel/tcg/tcg-accel-ops.c b/accel/tcg/tcg-accel-ops.c
index 927af6c1db0..1043fc61874 100644
--- a/accel/tcg/tcg-accel-ops.c
+++ b/accel/tcg/tcg-accel-ops.c
@@ -188,7 +188,7 @@ static int tcg_remove_breakpoint(CPUState *cs, int type, vaddr addr, vaddr len)
     }
 }
 
-static inline void tcg_remove_all_breakpoints(CPUState *cpu)
+static void tcg_remove_all_breakpoints(CPUState *cpu)
 {
     cpu_breakpoint_remove_all(cpu, BP_GDB);
     cpu_watchpoint_remove_all(cpu, BP_GDB);
-- 
2.53.0

[PULL 22/32] gdbstub/user: Directly call gdb_breakpoint_remove_all() in user mode

[Philippe Mathieu-Daudé]

No need to deref external methods with unused argument to
end up calling a method defined in the same unit file, call
it directly.

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Daniel Henrique Barboza <daniel.barboza@oss.qualcomm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20260705215729.62196-23-philmd@oss.qualcomm.com>
---
 gdbstub/user.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/gdbstub/user.c b/gdbstub/user.c
index 299388b9f71..90d6ee49fb5 100644
--- a/gdbstub/user.c
+++ b/gdbstub/user.c
@@ -557,7 +557,7 @@ static void disable_gdbstub(CPUState *thread_cpu)
     close(gdbserver_user_state.fd);
     gdbserver_user_state.fd = -1;
     CPU_FOREACH(cpu) {
-        cpu_breakpoint_remove_all(cpu, BP_GDB);
+        gdb_breakpoint_remove_all(cpu);
         /* no cpu_watchpoint_remove_all for user-mode */
         cpu_single_step(cpu, 0);
     }
-- 
2.53.0

[PULL 23/32] gdbstub: Reduce @type variable scope

[Philippe Mathieu-Daudé]

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Daniel Henrique Barboza <daniel.barboza@oss.qualcomm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20260705215729.62196-24-philmd@oss.qualcomm.com>
---
 gdbstub/system.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/gdbstub/system.c b/gdbstub/system.c
index bf7aa1ac234..9c4bd8ebcba 100644
--- a/gdbstub/system.c
+++ b/gdbstub/system.c
@@ -125,7 +125,6 @@ static void gdb_vm_state_change(void *opaque, bool running, RunState state)
     CPUState *cpu = gdbserver_state.c_cpu;
     g_autoptr(GString) buf = g_string_new(NULL);
     g_autoptr(GString) tid = g_string_new(NULL);
-    const char *type;
     int ret;
 
     if (running || gdbserver_state.state == RS_INACTIVE) {
@@ -151,6 +150,8 @@ static void gdb_vm_state_change(void *opaque, bool running, RunState state)
     switch (state) {
     case RUN_STATE_DEBUG:
         if (cpu->watchpoint_hit) {
+            const char *type;
+
             switch (cpu->watchpoint_hit->flags & BP_MEM_ACCESS) {
             case BP_MEM_READ:
                 type = "r";
-- 
2.53.0

[PULL 24/32] gdbstub: Introduce GdbBreakpointType enumerator

[Philippe Mathieu-Daudé]

Introduce the GdbBreakpointType enumerator to better follow
code related to GDB protocol handling.

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Daniel Henrique Barboza <daniel.barboza@oss.qualcomm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20260705215729.62196-25-philmd@oss.qualcomm.com>
---
 gdbstub/internals.h     |  7 +++++--
 include/gdbstub/enums.h | 12 +++++++-----
 gdbstub/system.c        |  6 ++++--
 gdbstub/user.c          |  6 ++++--
 4 files changed, 20 insertions(+), 11 deletions(-)

diff --git a/gdbstub/internals.h b/gdbstub/internals.h
index 0b74ea4000f..33f68672d1f 100644
--- a/gdbstub/internals.h
+++ b/gdbstub/internals.h
@@ -11,6 +11,7 @@
 
 #include "qemu/accel.h"
 #include "exec/cpu-common.h"
+#include "gdbstub/enums.h"
 
 /*
  * Most "large" transfers (e.g. memory reads, feature XML
@@ -217,8 +218,10 @@ void gdb_syscall_handling(const char *syscall_packet);
  * Break/Watch point support - there is an implementation for system
  * and user mode.
  */
-int gdb_breakpoint_insert(CPUState *cs, int type, vaddr addr, vaddr len);
-int gdb_breakpoint_remove(CPUState *cs, int type, vaddr addr, vaddr len);
+int gdb_breakpoint_insert(CPUState *cs, GdbBreakpointType type,
+                          vaddr addr, vaddr len);
+int gdb_breakpoint_remove(CPUState *cs, GdbBreakpointType type,
+                          vaddr addr, vaddr len);
 void gdb_breakpoint_remove_all(CPUState *cs);
 
 /**
diff --git a/include/gdbstub/enums.h b/include/gdbstub/enums.h
index c4d54a1d083..e4e2043d92c 100644
--- a/include/gdbstub/enums.h
+++ b/include/gdbstub/enums.h
@@ -12,10 +12,12 @@
 #define DEFAULT_GDBSTUB_PORT "1234"
 
 /* GDB breakpoint/watchpoint types */
-#define GDB_BREAKPOINT_SW        0
-#define GDB_BREAKPOINT_HW        1
-#define GDB_WATCHPOINT_WRITE     2
-#define GDB_WATCHPOINT_READ      3
-#define GDB_WATCHPOINT_ACCESS    4
+typedef enum GdbBreakpointType {
+    GDB_BREAKPOINT_SW       = 0,
+    GDB_BREAKPOINT_HW       = 1,
+    GDB_WATCHPOINT_WRITE    = 2,
+    GDB_WATCHPOINT_READ     = 3,
+    GDB_WATCHPOINT_ACCESS   = 4,
+} GdbBreakpointType;
 
 #endif /* GDBSTUB_ENUMS_H */
diff --git a/gdbstub/system.c b/gdbstub/system.c
index 9c4bd8ebcba..6e2dbc823e3 100644
--- a/gdbstub/system.c
+++ b/gdbstub/system.c
@@ -623,7 +623,8 @@ int gdb_signal_to_target(int sig)
  * Break/Watch point helpers
  */
 
-int gdb_breakpoint_insert(CPUState *cs, int type, vaddr addr, vaddr len)
+int gdb_breakpoint_insert(CPUState *cs, GdbBreakpointType type,
+                          vaddr addr, vaddr len)
 {
     const AccelOpsClass *ops = cpus_get_accel();
     if (ops->insert_breakpoint) {
@@ -632,7 +633,8 @@ int gdb_breakpoint_insert(CPUState *cs, int type, vaddr addr, vaddr len)
     return -ENOSYS;
 }
 
-int gdb_breakpoint_remove(CPUState *cs, int type, vaddr addr, vaddr len)
+int gdb_breakpoint_remove(CPUState *cs, GdbBreakpointType type,
+                          vaddr addr, vaddr len)
 {
     const AccelOpsClass *ops = cpus_get_accel();
     if (ops->remove_breakpoint) {
diff --git a/gdbstub/user.c b/gdbstub/user.c
index 90d6ee49fb5..9e6f9a6f376 100644
--- a/gdbstub/user.c
+++ b/gdbstub/user.c
@@ -790,7 +790,8 @@ unsigned int gdb_get_max_cpus(void)
  * Break/Watch point helpers
  */
 
-int gdb_breakpoint_insert(CPUState *cs, int type, vaddr addr, vaddr len)
+int gdb_breakpoint_insert(CPUState *cs, GdbBreakpointType type,
+                          vaddr addr, vaddr len)
 {
     CPUState *cpu;
     int err = 0;
@@ -811,7 +812,8 @@ int gdb_breakpoint_insert(CPUState *cs, int type, vaddr addr, vaddr len)
     }
 }
 
-int gdb_breakpoint_remove(CPUState *cs, int type, vaddr addr, vaddr len)
+int gdb_breakpoint_remove(CPUState *cs, GdbBreakpointType type,
+                          vaddr addr, vaddr len)
 {
     CPUState *cpu;
     int err = 0;
-- 
2.53.0

[PULL 25/32] accel: Use GdbBreakpointType enum

[Philippe Mathieu-Daudé]

Include '_gdbstub_' in the AccelOpsClass handlers to emphasize
we are handling gdbstub-related requests.

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Daniel Henrique Barboza <daniel.barboza@oss.qualcomm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20260705215729.62196-26-philmd@oss.qualcomm.com>
---
 accel/kvm/kvm-cpus.h          | 11 ++++++++---
 include/accel/accel-cpu-ops.h |  9 ++++++---
 include/system/hvf_int.h      |  9 ++++++---
 include/system/kvm.h          |  9 ++++++---
 target/arm/internals.h        |  5 +++--
 accel/hvf/hvf-accel-ops.c     | 20 +++++++++++---------
 accel/kvm/kvm-accel-ops.c     |  6 +++---
 accel/kvm/kvm-all.c           | 14 ++++++++------
 accel/tcg/tcg-accel-ops.c     | 16 +++++++++-------
 gdbstub/system.c              | 12 ++++++------
 target/arm/hvf/hvf.c          | 12 +++++++-----
 target/arm/hyp_gdbstub.c      |  8 ++++----
 target/arm/kvm.c              | 12 +++++++-----
 target/i386/hvf/hvf.c         |  8 +++++---
 target/i386/kvm/kvm.c         | 12 +++++++-----
 target/loongarch/kvm/kvm.c    |  8 +++++---
 target/ppc/kvm.c              | 12 +++++++-----
 target/riscv/kvm/kvm-cpu.c    |  8 +++++---
 target/s390x/kvm/kvm.c        |  8 +++++---
 19 files changed, 118 insertions(+), 81 deletions(-)

diff --git a/accel/kvm/kvm-cpus.h b/accel/kvm/kvm-cpus.h
index 3185659562d..bc2dd82e68e 100644
--- a/accel/kvm/kvm-cpus.h
+++ b/accel/kvm/kvm-cpus.h
@@ -10,13 +10,18 @@
 #ifndef KVM_CPUS_H
 #define KVM_CPUS_H
 
+#include "gdbstub/enums.h"
+
 int kvm_init_vcpu(CPUState *cpu, Error **errp);
 int kvm_cpu_exec(CPUState *cpu);
 void kvm_destroy_vcpu(CPUState *cpu);
 void kvm_cpu_synchronize_post_reset(CPUState *cpu);
 void kvm_cpu_synchronize_post_init(CPUState *cpu);
 void kvm_cpu_synchronize_pre_loadvm(CPUState *cpu);
-int kvm_insert_breakpoint(CPUState *cpu, int type, vaddr addr, vaddr len);
-int kvm_remove_breakpoint(CPUState *cpu, int type, vaddr addr, vaddr len);
-void kvm_remove_all_breakpoints(CPUState *cpu);
+int kvm_insert_gdbstub_breakpoint(CPUState *cpu, GdbBreakpointType type,
+                                  vaddr addr, vaddr len);
+int kvm_remove_gdbstub_breakpoint(CPUState *cpu, GdbBreakpointType type,
+                                  vaddr addr, vaddr len);
+void kvm_remove_all_gdbstub_breakpoints(CPUState *cpu);
+
 #endif /* KVM_CPUS_H */
diff --git a/include/accel/accel-cpu-ops.h b/include/accel/accel-cpu-ops.h
index b23a0606c7e..f0c7ee7542c 100644
--- a/include/accel/accel-cpu-ops.h
+++ b/include/accel/accel-cpu-ops.h
@@ -13,6 +13,7 @@
 #include "qemu/accel.h"
 #include "exec/vaddr.h"
 #include "qom/object.h"
+#include "gdbstub/enums.h"
 
 #define ACCEL_OPS_SUFFIX "-ops"
 #define TYPE_ACCEL_OPS "accel" ACCEL_OPS_SUFFIX
@@ -85,9 +86,11 @@ struct AccelOpsClass {
 
     /* gdbstub hooks */
     int (*update_guest_debug)(CPUState *cpu);
-    int (*insert_breakpoint)(CPUState *cpu, int type, vaddr addr, vaddr len);
-    int (*remove_breakpoint)(CPUState *cpu, int type, vaddr addr, vaddr len);
-    void (*remove_all_breakpoints)(CPUState *cpu);
+    int (*insert_gdbstub_breakpoint)(CPUState *cpu, GdbBreakpointType type,
+                                     vaddr addr, vaddr len);
+    int (*remove_gdbstub_breakpoint)(CPUState *cpu, GdbBreakpointType type,
+                                     vaddr addr, vaddr len);
+    void (*remove_all_gdbstub_breakpoints)(CPUState *cpu);
 };
 
 void generic_handle_interrupt(CPUState *cpu, int mask);
diff --git a/include/system/hvf_int.h b/include/system/hvf_int.h
index d64f4942556..a01691ce172 100644
--- a/include/system/hvf_int.h
+++ b/include/system/hvf_int.h
@@ -13,6 +13,7 @@
 
 #include "qemu/queue.h"
 #include "exec/vaddr.h"
+#include "gdbstub/enums.h"
 #include "qom/object.h"
 #include "accel/accel-ops.h"
 
@@ -91,9 +92,11 @@ int hvf_sw_breakpoints_active(CPUState *cpu);
 
 int hvf_arch_insert_sw_breakpoint(CPUState *cpu, struct hvf_sw_breakpoint *bp);
 int hvf_arch_remove_sw_breakpoint(CPUState *cpu, struct hvf_sw_breakpoint *bp);
-int hvf_arch_insert_hw_breakpoint(vaddr addr, vaddr len, int type);
-int hvf_arch_remove_hw_breakpoint(vaddr addr, vaddr len, int type);
-void hvf_arch_remove_all_hw_breakpoints(void);
+int hvf_arch_insert_gdbstub_hw_breakpoint(vaddr addr, vaddr len,
+                                          GdbBreakpointType type);
+int hvf_arch_remove_gdbstub_hw_breakpoint(vaddr addr, vaddr len,
+                                          GdbBreakpointType type);
+void hvf_arch_remove_all_gdbstub_hw_breakpoints(void);
 
 /*
  * hvf_update_guest_debug:
diff --git a/include/system/kvm.h b/include/system/kvm.h
index cdd1856ac5f..714b8c7b011 100644
--- a/include/system/kvm.h
+++ b/include/system/kvm.h
@@ -17,6 +17,7 @@
 #define QEMU_KVM_H
 
 #include "exec/memattrs.h"
+#include "gdbstub/enums.h"
 #include "qemu/accel.h"
 #include "accel/accel-route.h"
 #include "qom/object.h"
@@ -412,9 +413,11 @@ int kvm_arch_insert_sw_breakpoint(CPUState *cpu,
                                   struct kvm_sw_breakpoint *bp);
 int kvm_arch_remove_sw_breakpoint(CPUState *cpu,
                                   struct kvm_sw_breakpoint *bp);
-int kvm_arch_insert_hw_breakpoint(vaddr addr, vaddr len, int type);
-int kvm_arch_remove_hw_breakpoint(vaddr addr, vaddr len, int type);
-void kvm_arch_remove_all_hw_breakpoints(void);
+int kvm_arch_insert_gdbstub_hw_breakpoint(vaddr addr, vaddr len,
+                                          GdbBreakpointType type);
+int kvm_arch_remove_gdbstub_hw_breakpoint(vaddr addr, vaddr len,
+                                          GdbBreakpointType type);
+void kvm_arch_remove_all_gdbstub_hw_breakpoints(void);
 
 void kvm_arch_update_guest_debug(CPUState *cpu, struct kvm_guest_debug *dbg);
 
diff --git a/target/arm/internals.h b/target/arm/internals.h
index fcce3804f34..067c3f2b8bb 100644
--- a/target/arm/internals.h
+++ b/target/arm/internals.h
@@ -29,6 +29,7 @@
 #include "exec/vaddr.h"
 #include "exec/breakpoint.h"
 #include "exec/memop.h"
+#include "gdbstub/enums.h"
 #ifdef CONFIG_TCG
 #include "accel/tcg/tb-cpu-state.h"
 #include "tcg/tcg-gvec-desc.h"
@@ -1968,8 +1969,8 @@ int delete_hw_breakpoint(vaddr pc);
 
 bool check_watchpoint_in_range(int i, vaddr addr);
 CPUWatchpoint *find_hw_watchpoint(CPUState *cpu, vaddr addr);
-int insert_hw_watchpoint(vaddr addr, vaddr len, int type);
-int delete_hw_watchpoint(vaddr addr, vaddr len, int type);
+int insert_gdbstub_hw_watchpoint(vaddr addr, vaddr len, GdbBreakpointType type);
+int delete_gdbstub_hw_watchpoint(vaddr addr, vaddr len, GdbBreakpointType type);
 
 /* Return the current value of the system counter in ticks */
 uint64_t gt_get_countervalue(CPUARMState *env);
diff --git a/accel/hvf/hvf-accel-ops.c b/accel/hvf/hvf-accel-ops.c
index ecc0cd1b558..d2276d8513e 100644
--- a/accel/hvf/hvf-accel-ops.c
+++ b/accel/hvf/hvf-accel-ops.c
@@ -233,7 +233,8 @@ int hvf_update_guest_debug(CPUState *cpu)
     return 0;
 }
 
-static int hvf_insert_breakpoint(CPUState *cpu, int type, vaddr addr, vaddr len)
+static int hvf_insert_gdbstub_breakpoint(CPUState *cpu, GdbBreakpointType type,
+                                         vaddr addr, vaddr len)
 {
     struct hvf_sw_breakpoint *bp;
     int err;
@@ -256,7 +257,7 @@ static int hvf_insert_breakpoint(CPUState *cpu, int type, vaddr addr, vaddr len)
 
         QTAILQ_INSERT_HEAD(&hvf_state->hvf_sw_breakpoints, bp, entry);
     } else {
-        err = hvf_arch_insert_hw_breakpoint(addr, len, type);
+        err = hvf_arch_insert_gdbstub_hw_breakpoint(addr, len, type);
         if (err) {
             return err;
         }
@@ -271,7 +272,8 @@ static int hvf_insert_breakpoint(CPUState *cpu, int type, vaddr addr, vaddr len)
     return 0;
 }
 
-static int hvf_remove_breakpoint(CPUState *cpu, int type, vaddr addr, vaddr len)
+static int hvf_remove_gdbstub_breakpoint(CPUState *cpu, GdbBreakpointType type,
+                                         vaddr addr, vaddr len)
 {
     struct hvf_sw_breakpoint *bp;
     int err;
@@ -295,7 +297,7 @@ static int hvf_remove_breakpoint(CPUState *cpu, int type, vaddr addr, vaddr len)
         QTAILQ_REMOVE(&hvf_state->hvf_sw_breakpoints, bp, entry);
         g_free(bp);
     } else {
-        err = hvf_arch_remove_hw_breakpoint(addr, len, type);
+        err = hvf_arch_remove_gdbstub_hw_breakpoint(addr, len, type);
         if (err) {
             return err;
         }
@@ -310,7 +312,7 @@ static int hvf_remove_breakpoint(CPUState *cpu, int type, vaddr addr, vaddr len)
     return 0;
 }
 
-static void hvf_remove_all_breakpoints(CPUState *cpu)
+static void hvf_remove_all_gdbstub_breakpoints(CPUState *cpu)
 {
     struct hvf_sw_breakpoint *bp, *next;
     CPUState *tmpcpu;
@@ -328,7 +330,7 @@ static void hvf_remove_all_breakpoints(CPUState *cpu)
         QTAILQ_REMOVE(&hvf_state->hvf_sw_breakpoints, bp, entry);
         g_free(bp);
     }
-    hvf_arch_remove_all_hw_breakpoints();
+    hvf_arch_remove_all_gdbstub_hw_breakpoints();
 
     CPU_FOREACH(cpu) {
         hvf_update_guest_debug(cpu);
@@ -365,9 +367,9 @@ static void hvf_accel_ops_class_init(ObjectClass *oc, const void *data)
     ops->synchronize_state = hvf_cpu_synchronize_state;
     ops->synchronize_pre_loadvm = hvf_cpu_synchronize_pre_loadvm;
 
-    ops->insert_breakpoint = hvf_insert_breakpoint;
-    ops->remove_breakpoint = hvf_remove_breakpoint;
-    ops->remove_all_breakpoints = hvf_remove_all_breakpoints;
+    ops->insert_gdbstub_breakpoint = hvf_insert_gdbstub_breakpoint;
+    ops->remove_gdbstub_breakpoint = hvf_remove_gdbstub_breakpoint;
+    ops->remove_all_gdbstub_breakpoints = hvf_remove_all_gdbstub_breakpoints;
     ops->update_guest_debug = hvf_update_guest_debug;
 
     ops->get_vcpu_stats = hvf_get_vcpu_stats;
diff --git a/accel/kvm/kvm-accel-ops.c b/accel/kvm/kvm-accel-ops.c
index 45330327909..c8e7aa38709 100644
--- a/accel/kvm/kvm-accel-ops.c
+++ b/accel/kvm/kvm-accel-ops.c
@@ -107,9 +107,9 @@ static void kvm_accel_ops_class_init(ObjectClass *oc, const void *data)
 
 #ifdef TARGET_KVM_HAVE_GUEST_DEBUG
     ops->update_guest_debug = kvm_update_guest_debug_ops;
-    ops->insert_breakpoint = kvm_insert_breakpoint;
-    ops->remove_breakpoint = kvm_remove_breakpoint;
-    ops->remove_all_breakpoints = kvm_remove_all_breakpoints;
+    ops->insert_gdbstub_breakpoint = kvm_insert_gdbstub_breakpoint;
+    ops->remove_gdbstub_breakpoint = kvm_remove_gdbstub_breakpoint;
+    ops->remove_all_gdbstub_breakpoints = kvm_remove_all_gdbstub_breakpoints;
 #endif
 }
 
diff --git a/accel/kvm/kvm-all.c b/accel/kvm/kvm-all.c
index b03488ebac3..feffca0d82d 100644
--- a/accel/kvm/kvm-all.c
+++ b/accel/kvm/kvm-all.c
@@ -3826,7 +3826,8 @@ int kvm_update_guest_debug(CPUState *cpu, unsigned long reinject_trap)
     return data.err;
 }
 
-int kvm_insert_breakpoint(CPUState *cpu, int type, vaddr addr, vaddr len)
+int kvm_insert_gdbstub_breakpoint(CPUState *cpu, GdbBreakpointType type,
+                                  vaddr addr, vaddr len)
 {
     struct kvm_sw_breakpoint *bp;
     int err;
@@ -3849,7 +3850,7 @@ int kvm_insert_breakpoint(CPUState *cpu, int type, vaddr addr, vaddr len)
 
         QTAILQ_INSERT_HEAD(&cpu->kvm_state->kvm_sw_breakpoints, bp, entry);
     } else {
-        err = kvm_arch_insert_hw_breakpoint(addr, len, type);
+        err = kvm_arch_insert_gdbstub_hw_breakpoint(addr, len, type);
         if (err) {
             return err;
         }
@@ -3864,7 +3865,8 @@ int kvm_insert_breakpoint(CPUState *cpu, int type, vaddr addr, vaddr len)
     return 0;
 }
 
-int kvm_remove_breakpoint(CPUState *cpu, int type, vaddr addr, vaddr len)
+int kvm_remove_gdbstub_breakpoint(CPUState *cpu, GdbBreakpointType type,
+                                  vaddr addr, vaddr len)
 {
     struct kvm_sw_breakpoint *bp;
     int err;
@@ -3888,7 +3890,7 @@ int kvm_remove_breakpoint(CPUState *cpu, int type, vaddr addr, vaddr len)
         QTAILQ_REMOVE(&cpu->kvm_state->kvm_sw_breakpoints, bp, entry);
         g_free(bp);
     } else {
-        err = kvm_arch_remove_hw_breakpoint(addr, len, type);
+        err = kvm_arch_remove_gdbstub_hw_breakpoint(addr, len, type);
         if (err) {
             return err;
         }
@@ -3903,7 +3905,7 @@ int kvm_remove_breakpoint(CPUState *cpu, int type, vaddr addr, vaddr len)
     return 0;
 }
 
-void kvm_remove_all_breakpoints(CPUState *cpu)
+void kvm_remove_all_gdbstub_breakpoints(CPUState *cpu)
 {
     struct kvm_sw_breakpoint *bp, *next;
     KVMState *s = cpu->kvm_state;
@@ -3921,7 +3923,7 @@ void kvm_remove_all_breakpoints(CPUState *cpu)
         QTAILQ_REMOVE(&s->kvm_sw_breakpoints, bp, entry);
         g_free(bp);
     }
-    kvm_arch_remove_all_hw_breakpoints();
+    kvm_arch_remove_all_gdbstub_hw_breakpoints();
 
     CPU_FOREACH(cpu) {
         kvm_update_guest_debug(cpu, 0);
diff --git a/accel/tcg/tcg-accel-ops.c b/accel/tcg/tcg-accel-ops.c
index 1043fc61874..560fe2554ba 100644
--- a/accel/tcg/tcg-accel-ops.c
+++ b/accel/tcg/tcg-accel-ops.c
@@ -110,7 +110,7 @@ void tcg_handle_interrupt(CPUState *cpu, int mask)
 }
 
 /* Translate GDB watchpoint type to a flags value for cpu_watchpoint_* */
-static inline int xlat_gdb_type(CPUState *cpu, int gdbtype)
+static inline int xlat_gdb_type(CPUState *cpu, GdbBreakpointType gdbtype)
 {
     static const int xlat[] = {
         [GDB_WATCHPOINT_WRITE]  = BP_GDB | BP_MEM_WRITE,
@@ -126,7 +126,8 @@ static inline int xlat_gdb_type(CPUState *cpu, int gdbtype)
     return cputype;
 }
 
-static int tcg_insert_breakpoint(CPUState *cs, int type, vaddr addr, vaddr len)
+static int tcg_insert_gdbstub_breakpoint(CPUState *cs, GdbBreakpointType type,
+                                         vaddr addr, vaddr len)
 {
     CPUState *cpu;
     int err = 0;
@@ -157,7 +158,8 @@ static int tcg_insert_breakpoint(CPUState *cs, int type, vaddr addr, vaddr len)
     }
 }
 
-static int tcg_remove_breakpoint(CPUState *cs, int type, vaddr addr, vaddr len)
+static int tcg_remove_gdbstub_breakpoint(CPUState *cs, GdbBreakpointType type,
+                                         vaddr addr, vaddr len)
 {
     CPUState *cpu;
     int err = 0;
@@ -188,7 +190,7 @@ static int tcg_remove_breakpoint(CPUState *cs, int type, vaddr addr, vaddr len)
     }
 }
 
-static void tcg_remove_all_breakpoints(CPUState *cpu)
+static void tcg_remove_all_gdbstub_breakpoints(CPUState *cpu)
 {
     cpu_breakpoint_remove_all(cpu, BP_GDB);
     cpu_watchpoint_remove_all(cpu, BP_GDB);
@@ -216,9 +218,9 @@ static void tcg_accel_ops_init(AccelClass *ac)
     }
 
     ops->cpu_reset_hold = tcg_cpu_reset_hold;
-    ops->insert_breakpoint = tcg_insert_breakpoint;
-    ops->remove_breakpoint = tcg_remove_breakpoint;
-    ops->remove_all_breakpoints = tcg_remove_all_breakpoints;
+    ops->insert_gdbstub_breakpoint = tcg_insert_gdbstub_breakpoint;
+    ops->remove_gdbstub_breakpoint = tcg_remove_gdbstub_breakpoint;
+    ops->remove_all_gdbstub_breakpoints = tcg_remove_all_gdbstub_breakpoints;
 }
 
 static void tcg_accel_ops_class_init(ObjectClass *oc, const void *data)
diff --git a/gdbstub/system.c b/gdbstub/system.c
index 6e2dbc823e3..3098fd55740 100644
--- a/gdbstub/system.c
+++ b/gdbstub/system.c
@@ -627,8 +627,8 @@ int gdb_breakpoint_insert(CPUState *cs, GdbBreakpointType type,
                           vaddr addr, vaddr len)
 {
     const AccelOpsClass *ops = cpus_get_accel();
-    if (ops->insert_breakpoint) {
-        return ops->insert_breakpoint(cs, type, addr, len);
+    if (ops->insert_gdbstub_breakpoint) {
+        return ops->insert_gdbstub_breakpoint(cs, type, addr, len);
     }
     return -ENOSYS;
 }
@@ -637,8 +637,8 @@ int gdb_breakpoint_remove(CPUState *cs, GdbBreakpointType type,
                           vaddr addr, vaddr len)
 {
     const AccelOpsClass *ops = cpus_get_accel();
-    if (ops->remove_breakpoint) {
-        return ops->remove_breakpoint(cs, type, addr, len);
+    if (ops->remove_gdbstub_breakpoint) {
+        return ops->remove_gdbstub_breakpoint(cs, type, addr, len);
     }
     return -ENOSYS;
 }
@@ -646,8 +646,8 @@ int gdb_breakpoint_remove(CPUState *cs, GdbBreakpointType type,
 void gdb_breakpoint_remove_all(CPUState *cs)
 {
     const AccelOpsClass *ops = cpus_get_accel();
-    if (ops->remove_all_breakpoints) {
-        ops->remove_all_breakpoints(cs);
+    if (ops->remove_all_gdbstub_breakpoints) {
+        ops->remove_all_gdbstub_breakpoints(cs);
     }
 }
 
diff --git a/target/arm/hvf/hvf.c b/target/arm/hvf/hvf.c
index e16457c3cfb..d5b1966a867 100644
--- a/target/arm/hvf/hvf.c
+++ b/target/arm/hvf/hvf.c
@@ -2726,7 +2726,8 @@ int hvf_arch_remove_sw_breakpoint(CPUState *cpu, struct hvf_sw_breakpoint *bp)
     return 0;
 }
 
-int hvf_arch_insert_hw_breakpoint(vaddr addr, vaddr len, int type)
+int hvf_arch_insert_gdbstub_hw_breakpoint(vaddr addr, vaddr len,
+                                          GdbBreakpointType type)
 {
     switch (type) {
     case GDB_BREAKPOINT_HW:
@@ -2734,13 +2735,14 @@ int hvf_arch_insert_hw_breakpoint(vaddr addr, vaddr len, int type)
     case GDB_WATCHPOINT_READ:
     case GDB_WATCHPOINT_WRITE:
     case GDB_WATCHPOINT_ACCESS:
-        return insert_hw_watchpoint(addr, len, type);
+        return insert_gdbstub_hw_watchpoint(addr, len, type);
     default:
         return -ENOSYS;
     }
 }
 
-int hvf_arch_remove_hw_breakpoint(vaddr addr, vaddr len, int type)
+int hvf_arch_remove_gdbstub_hw_breakpoint(vaddr addr, vaddr len,
+                                          GdbBreakpointType type)
 {
     switch (type) {
     case GDB_BREAKPOINT_HW:
@@ -2748,13 +2750,13 @@ int hvf_arch_remove_hw_breakpoint(vaddr addr, vaddr len, int type)
     case GDB_WATCHPOINT_READ:
     case GDB_WATCHPOINT_WRITE:
     case GDB_WATCHPOINT_ACCESS:
-        return delete_hw_watchpoint(addr, len, type);
+        return delete_gdbstub_hw_watchpoint(addr, len, type);
     default:
         return -ENOSYS;
     }
 }
 
-void hvf_arch_remove_all_hw_breakpoints(void)
+void hvf_arch_remove_all_gdbstub_hw_breakpoints(void)
 {
     if (cur_hw_wps > 0) {
         g_array_remove_range(hw_watchpoints, 0, cur_hw_wps);
diff --git a/target/arm/hyp_gdbstub.c b/target/arm/hyp_gdbstub.c
index bb5969720ce..dda6946ddaa 100644
--- a/target/arm/hyp_gdbstub.c
+++ b/target/arm/hyp_gdbstub.c
@@ -94,7 +94,7 @@ int delete_hw_breakpoint(vaddr pc)
 }
 
 /**
- * insert_hw_watchpoint()
+ * insert_gdbstub_hw_watchpoint()
  * @addr: address of watch point
  * @len: size of area
  * @type: type of watch point
@@ -125,7 +125,7 @@ int delete_hw_breakpoint(vaddr pc)
  * need to ensure you mask the address as required and set BAS=0xff
  */
 
-int insert_hw_watchpoint(vaddr addr, vaddr len, int type)
+int insert_gdbstub_hw_watchpoint(vaddr addr, vaddr len, GdbBreakpointType type)
 {
     HWWatchpoint wp = {
         .wcr = R_DBGWCR_E_MASK, /* E=1, enable */
@@ -208,13 +208,13 @@ bool check_watchpoint_in_range(int i, vaddr addr)
 }
 
 /**
- * delete_hw_watchpoint()
+ * delete_gdbstub_hw_watchpoint()
  * @addr: address of breakpoint
  *
  * Delete a breakpoint and shuffle any above down
  */
 
-int delete_hw_watchpoint(vaddr addr, vaddr len, int type)
+int delete_gdbstub_hw_watchpoint(vaddr addr, vaddr len, GdbBreakpointType type)
 {
     int i;
     for (i = 0; i < cur_hw_wps; i++) {
diff --git a/target/arm/kvm.c b/target/arm/kvm.c
index a54ef51ec2a..01f42f42538 100644
--- a/target/arm/kvm.c
+++ b/target/arm/kvm.c
@@ -1784,7 +1784,8 @@ void kvm_arch_accel_class_init(ObjectClass *oc)
         "Eager Page Split chunk size for hugepages. (default: 0, disabled)");
 }
 
-int kvm_arch_insert_hw_breakpoint(vaddr addr, vaddr len, int type)
+int kvm_arch_insert_gdbstub_hw_breakpoint(vaddr addr, vaddr len,
+                                          GdbBreakpointType type)
 {
     switch (type) {
     case GDB_BREAKPOINT_HW:
@@ -1793,13 +1794,14 @@ int kvm_arch_insert_hw_breakpoint(vaddr addr, vaddr len, int type)
     case GDB_WATCHPOINT_READ:
     case GDB_WATCHPOINT_WRITE:
     case GDB_WATCHPOINT_ACCESS:
-        return insert_hw_watchpoint(addr, len, type);
+        return insert_gdbstub_hw_watchpoint(addr, len, type);
     default:
         return -ENOSYS;
     }
 }
 
-int kvm_arch_remove_hw_breakpoint(vaddr addr, vaddr len, int type)
+int kvm_arch_remove_gdbstub_hw_breakpoint(vaddr addr, vaddr len,
+                                          GdbBreakpointType type)
 {
     switch (type) {
     case GDB_BREAKPOINT_HW:
@@ -1807,13 +1809,13 @@ int kvm_arch_remove_hw_breakpoint(vaddr addr, vaddr len, int type)
     case GDB_WATCHPOINT_READ:
     case GDB_WATCHPOINT_WRITE:
     case GDB_WATCHPOINT_ACCESS:
-        return delete_hw_watchpoint(addr, len, type);
+        return delete_gdbstub_hw_watchpoint(addr, len, type);
     default:
         return -ENOSYS;
     }
 }
 
-void kvm_arch_remove_all_hw_breakpoints(void)
+void kvm_arch_remove_all_gdbstub_hw_breakpoints(void)
 {
     if (cur_hw_wps > 0) {
         g_array_remove_range(hw_watchpoints, 0, cur_hw_wps);
diff --git a/target/i386/hvf/hvf.c b/target/i386/hvf/hvf.c
index 19ad64a1d93..150598418e2 100644
--- a/target/i386/hvf/hvf.c
+++ b/target/i386/hvf/hvf.c
@@ -1049,17 +1049,19 @@ int hvf_arch_remove_sw_breakpoint(CPUState *cpu, struct hvf_sw_breakpoint *bp)
     return -ENOSYS;
 }
 
-int hvf_arch_insert_hw_breakpoint(vaddr addr, vaddr len, int type)
+int hvf_arch_insert_gdbstub_hw_breakpoint(vaddr addr, vaddr len,
+                                          GdbBreakpointType type)
 {
     return -ENOSYS;
 }
 
-int hvf_arch_remove_hw_breakpoint(vaddr addr, vaddr len, int type)
+int hvf_arch_remove_gdbstub_hw_breakpoint(vaddr addr, vaddr len,
+                                          GdbBreakpointType type)
 {
     return -ENOSYS;
 }
 
-void hvf_arch_remove_all_hw_breakpoints(void)
+void hvf_arch_remove_all_gdbstub_hw_breakpoints(void)
 {
 }
 
diff --git a/target/i386/kvm/kvm.c b/target/i386/kvm/kvm.c
index 64cc421abe6..1e09155e92c 100644
--- a/target/i386/kvm/kvm.c
+++ b/target/i386/kvm/kvm.c
@@ -6159,12 +6159,12 @@ int kvm_arch_remove_sw_breakpoint(CPUState *cs, struct kvm_sw_breakpoint *bp)
 static struct {
     target_ulong addr;
     int len;
-    int type;
+    GdbBreakpointType type;
 } hw_breakpoint[4];
 
 static int nb_hw_breakpoint;
 
-static int find_hw_breakpoint(target_ulong addr, int len, int type)
+static int find_hw_breakpoint(target_ulong addr, int len, GdbBreakpointType type)
 {
     int n;
 
@@ -6177,7 +6177,8 @@ static int find_hw_breakpoint(target_ulong addr, int len, int type)
     return -1;
 }
 
-int kvm_arch_insert_hw_breakpoint(vaddr addr, vaddr len, int type)
+int kvm_arch_insert_gdbstub_hw_breakpoint(vaddr addr, vaddr len,
+                                          GdbBreakpointType type)
 {
     switch (type) {
     case GDB_BREAKPOINT_HW:
@@ -6217,7 +6218,8 @@ int kvm_arch_insert_hw_breakpoint(vaddr addr, vaddr len, int type)
     return 0;
 }
 
-int kvm_arch_remove_hw_breakpoint(vaddr addr, vaddr len, int type)
+int kvm_arch_remove_gdbstub_hw_breakpoint(vaddr addr, vaddr len,
+                                          GdbBreakpointType type)
 {
     int n;
 
@@ -6231,7 +6233,7 @@ int kvm_arch_remove_hw_breakpoint(vaddr addr, vaddr len, int type)
     return 0;
 }
 
-void kvm_arch_remove_all_hw_breakpoints(void)
+void kvm_arch_remove_all_gdbstub_hw_breakpoints(void)
 {
     nb_hw_breakpoint = 0;
 }
diff --git a/target/loongarch/kvm/kvm.c b/target/loongarch/kvm/kvm.c
index d6539c12acd..beda6965f8c 100644
--- a/target/loongarch/kvm/kvm.c
+++ b/target/loongarch/kvm/kvm.c
@@ -1418,17 +1418,19 @@ int kvm_arch_remove_sw_breakpoint(CPUState *cs, struct kvm_sw_breakpoint *bp)
     return 0;
 }
 
-int kvm_arch_insert_hw_breakpoint(vaddr addr, vaddr len, int type)
+int kvm_arch_insert_gdbstub_hw_breakpoint(vaddr addr, vaddr len,
+                                          GdbBreakpointType type)
 {
     return -ENOSYS;
 }
 
-int kvm_arch_remove_hw_breakpoint(vaddr addr, vaddr len, int type)
+int kvm_arch_remove_gdbstub_hw_breakpoint(vaddr addr, vaddr len,
+                                          GdbBreakpointType type)
 {
     return -ENOSYS;
 }
 
-void kvm_arch_remove_all_hw_breakpoints(void)
+void kvm_arch_remove_all_gdbstub_hw_breakpoints(void)
 {
 }
 
diff --git a/target/ppc/kvm.c b/target/ppc/kvm.c
index b94c2997a07..93ec0cf6a72 100644
--- a/target/ppc/kvm.c
+++ b/target/ppc/kvm.c
@@ -455,7 +455,7 @@ unsigned long kvm_arch_vcpu_id(CPUState *cpu)
 
 static struct HWBreakpoint {
     target_ulong addr;
-    int type;
+    GdbBreakpointType type;
 } hw_debug_points[MAX_HW_BKPTS];
 
 static CPUWatchpoint hw_watchpoint;
@@ -1412,7 +1412,7 @@ int kvm_arch_remove_sw_breakpoint(CPUState *cs, struct kvm_sw_breakpoint *bp)
     return 0;
 }
 
-static int find_hw_breakpoint(target_ulong addr, int type)
+static int find_hw_breakpoint(target_ulong addr, GdbBreakpointType type)
 {
     int n;
 
@@ -1454,7 +1454,8 @@ static int find_hw_watchpoint(target_ulong addr, int *flag)
     return -1;
 }
 
-int kvm_arch_insert_hw_breakpoint(vaddr addr, vaddr len, int type)
+int kvm_arch_insert_gdbstub_hw_breakpoint(vaddr addr, vaddr len,
+                                          GdbBreakpointType type)
 {
     const unsigned breakpoint_index = nb_hw_breakpoint + nb_hw_watchpoint;
     if (breakpoint_index >= ARRAY_SIZE(hw_debug_points)) {
@@ -1498,7 +1499,8 @@ int kvm_arch_insert_hw_breakpoint(vaddr addr, vaddr len, int type)
     return 0;
 }
 
-int kvm_arch_remove_hw_breakpoint(vaddr addr, vaddr len, int type)
+int kvm_arch_remove_gdbstub_hw_breakpoint(vaddr addr, vaddr len,
+                                          GdbBreakpointType type)
 {
     int n;
 
@@ -1526,7 +1528,7 @@ int kvm_arch_remove_hw_breakpoint(vaddr addr, vaddr len, int type)
     return 0;
 }
 
-void kvm_arch_remove_all_hw_breakpoints(void)
+void kvm_arch_remove_all_gdbstub_hw_breakpoints(void)
 {
     nb_hw_breakpoint = nb_hw_watchpoint = 0;
 }
diff --git a/target/riscv/kvm/kvm-cpu.c b/target/riscv/kvm/kvm-cpu.c
index 39d48a9db4d..2a5440d5846 100644
--- a/target/riscv/kvm/kvm-cpu.c
+++ b/target/riscv/kvm/kvm-cpu.c
@@ -2215,19 +2215,21 @@ int kvm_arch_remove_sw_breakpoint(CPUState *cs, struct kvm_sw_breakpoint *bp)
     return 0;
 }
 
-int kvm_arch_insert_hw_breakpoint(vaddr addr, vaddr len, int type)
+int kvm_arch_insert_gdbstub_hw_breakpoint(vaddr addr, vaddr len,
+                                          GdbBreakpointType type)
 {
     /* TODO; To be implemented later. */
     return -EINVAL;
 }
 
-int kvm_arch_remove_hw_breakpoint(vaddr addr, vaddr len, int type)
+int kvm_arch_remove_gdbstub_hw_breakpoint(vaddr addr, vaddr len,
+                                          GdbBreakpointType type)
 {
     /* TODO; To be implemented later. */
     return -EINVAL;
 }
 
-void kvm_arch_remove_all_hw_breakpoints(void)
+void kvm_arch_remove_all_gdbstub_hw_breakpoints(void)
 {
     /* TODO; To be implemented later. */
 }
diff --git a/target/s390x/kvm/kvm.c b/target/s390x/kvm/kvm.c
index fdef8f9e8ac..195e39df032 100644
--- a/target/s390x/kvm/kvm.c
+++ b/target/s390x/kvm/kvm.c
@@ -907,7 +907,8 @@ static int insert_hw_breakpoint(vaddr addr, int len, int type)
     return 0;
 }
 
-int kvm_arch_insert_hw_breakpoint(vaddr addr, vaddr len, int type)
+int kvm_arch_insert_gdbstub_hw_breakpoint(vaddr addr, vaddr len,
+                                          GdbBreakpointType type)
 {
     switch (type) {
     case GDB_BREAKPOINT_HW:
@@ -925,7 +926,8 @@ int kvm_arch_insert_hw_breakpoint(vaddr addr, vaddr len, int type)
     return insert_hw_breakpoint(addr, len, type);
 }
 
-int kvm_arch_remove_hw_breakpoint(vaddr addr, vaddr len, int type)
+int kvm_arch_remove_gdbstub_hw_breakpoint(vaddr addr, vaddr len,
+                                          GdbBreakpointType type)
 {
     int size;
     struct kvm_hw_breakpoint *bp = find_hw_breakpoint(addr, len, type);
@@ -954,7 +956,7 @@ int kvm_arch_remove_hw_breakpoint(vaddr addr, vaddr len, int type)
     return 0;
 }
 
-void kvm_arch_remove_all_hw_breakpoints(void)
+void kvm_arch_remove_all_gdbstub_hw_breakpoints(void)
 {
     nb_hw_breakpoints = 0;
     g_free(hw_breakpoints);
-- 
2.53.0

[PULL 26/32] target/arm: Inline check_watchpoints() in arm_debug_check_watchpoint()

[Philippe Mathieu-Daudé]

check_watchpoints() is called once, by arm_debug_check_watchpoint(),
which doesn't do more than this call. Merge both. No logical change
intended.

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20260705215729.62196-27-philmd@oss.qualcomm.com>
---
 target/arm/tcg/debug.c | 40 +++++++++++++++++-----------------------
 1 file changed, 17 insertions(+), 23 deletions(-)

diff --git a/target/arm/tcg/debug.c b/target/arm/tcg/debug.c
index 07a52643e71..528d2889c3a 100644
--- a/target/arm/tcg/debug.c
+++ b/target/arm/tcg/debug.c
@@ -351,28 +351,6 @@ static bool bp_wp_matches(ARMCPU *cpu, int n, bool is_wp)
     return true;
 }
 
-static bool check_watchpoints(ARMCPU *cpu)
-{
-    CPUARMState *env = &cpu->env;
-    int n;
-
-    /*
-     * If watchpoints are disabled globally or we can't take debug
-     * exceptions here then watchpoint firings are ignored.
-     */
-    if (extract32(env->cp15.mdscr_el1, 15, 1) == 0
-        || !arm_generate_debug_exceptions(env)) {
-        return false;
-    }
-
-    for (n = 0; n < ARRAY_SIZE(env->cpu_watchpoint); n++) {
-        if (bp_wp_matches(cpu, n, true)) {
-            return true;
-        }
-    }
-    return false;
-}
-
 bool arm_debug_check_breakpoint(CPUState *cs)
 {
     ARMCPU *cpu = ARM_CPU(cs);
@@ -426,8 +404,24 @@ bool arm_debug_check_watchpoint(CPUState *cs, CPUWatchpoint *wp)
      * is also an architectural watchpoint match.
      */
     ARMCPU *cpu = ARM_CPU(cs);
+    CPUARMState *env = &cpu->env;
+    int n;
 
-    return check_watchpoints(cpu);
+    /*
+     * If watchpoints are disabled globally or we can't take debug
+     * exceptions here then watchpoint firings are ignored.
+     */
+    if (extract32(env->cp15.mdscr_el1, 15, 1) == 0
+        || !arm_generate_debug_exceptions(env)) {
+        return false;
+    }
+
+    for (n = 0; n < ARRAY_SIZE(env->cpu_watchpoint); n++) {
+        if (bp_wp_matches(cpu, n, true)) {
+            return true;
+        }
+    }
+    return false;
 }
 
 /*
-- 
2.53.0

[PULL 27/32] target/ppc: Ensure TCG is used in ppc_update_daw()

[Philippe Mathieu-Daudé]

Per commit d5ee641cfc5 ("target/ppc: Implement watchpoint debug
facility for v2.07S"), only TCG is implemented:

    ISA v2.07S introduced the watchpoint facility based on the DAWR0
    and DAWRX0 SPRs. Implement this in TCG.
                     ^^^^^^^^^^^^^^^^^^^^^

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Daniel Henrique Barboza <daniel.barboza@oss.qualcomm.com>
Message-ID: <20260705215729.62196-28-philmd@oss.qualcomm.com>
---
 target/ppc/cpu.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/target/ppc/cpu.c b/target/ppc/cpu.c
index 41edb18643d..47467ee1635 100644
--- a/target/ppc/cpu.c
+++ b/target/ppc/cpu.c
@@ -147,6 +147,8 @@ void ppc_update_daw(CPUPPCState *env, int rid)
     vaddr len;
     int flags;
 
+    assert(tcg_enabled());
+
     if (env->dawr_watchpoint[rid]) {
         cpu_watchpoint_remove_by_ref(cs, env->dawr_watchpoint[rid]);
         env->dawr_watchpoint[rid] = NULL;
-- 
2.53.0

[PULL 28/32] accel/tcg: Improve docstrings around TCGCPUOps::*watchpoint* handlers

[Philippe Mathieu-Daudé]

Commit d5ee641cfc5 ("target/ppc: Implement watchpoint debug facility
for v2.07S") also implemented TCGCPUOps::debug_check_watchpoint for
PPC: make the comment generic.

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Daniel Henrique Barboza <daniel.barboza@oss.qualcomm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20260705215729.62196-29-philmd@oss.qualcomm.com>
---
 include/accel/tcg/cpu-ops.h | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/include/accel/tcg/cpu-ops.h b/include/accel/tcg/cpu-ops.h
index 710da12b828..3ff6e6810e0 100644
--- a/include/accel/tcg/cpu-ops.h
+++ b/include/accel/tcg/cpu-ops.h
@@ -250,14 +250,13 @@ struct TCGCPUOps {
                                            int mmu_idx, uintptr_t retaddr);
 
     /**
-     * @adjust_watchpoint_address: hack for cpu_check_watchpoint used by ARM
+     * @adjust_watchpoint_address: hack for cpu_check_watchpoint (used by ARM)
      */
     vaddr (*adjust_watchpoint_address)(CPUState *cpu, vaddr addr, int len);
 
     /**
      * @debug_check_watchpoint: return true if the architectural
-     * watchpoint whose address has matched should really fire, used by ARM
-     * and RISC-V
+     * watchpoint whose address has matched should really fire.
      */
     bool (*debug_check_watchpoint)(CPUState *cpu, CPUWatchpoint *wp);
 
-- 
2.53.0

[PULL 29/32] cpu: Better name cpu_single_step() trace event

[Philippe Mathieu-Daudé]

cpu_single_step() is not related to breakpoints. Rename the
trace event.

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Daniel Henrique Barboza <daniel.barboza@oss.qualcomm.com>
Message-ID: <20260705215729.62196-30-philmd@oss.qualcomm.com>
---
 cpu-target.c | 4 ++--
 trace-events | 2 +-
 2 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/cpu-target.c b/cpu-target.c
index f030e2c642e..019906b32eb 100644
--- a/cpu-target.c
+++ b/cpu-target.c
@@ -31,6 +31,8 @@
 void cpu_single_step(CPUState *cpu, int enabled)
 {
     if (cpu->singlestep_enabled != enabled) {
+        trace_cpu_change_singlestep_flags(cpu->cpu_index,
+                                          cpu->singlestep_enabled, enabled);
         cpu->singlestep_enabled = enabled;
 
 #if !defined(CONFIG_USER_ONLY)
@@ -39,8 +41,6 @@ void cpu_single_step(CPUState *cpu, int enabled)
             ops->update_guest_debug(cpu);
         }
 #endif
-
-        trace_breakpoint_singlestep(cpu->cpu_index, enabled);
     }
 }
 
diff --git a/trace-events b/trace-events
index faeba6242fa..ea1d2b11a22 100644
--- a/trace-events
+++ b/trace-events
@@ -28,7 +28,7 @@
 # cpu.c
 breakpoint_insert(int cpu_index, uint64_t pc, int flags) "cpu=%d pc=0x%" PRIx64 " flags=0x%x"
 breakpoint_remove(int cpu_index, uint64_t pc, int flags) "cpu=%d pc=0x%" PRIx64 " flags=0x%x"
-breakpoint_singlestep(int cpu_index, int enabled) "cpu=%d enable=%d"
+cpu_change_singlestep_flags(int cpu_index, int old_flags, int new_flags) "cpu=%d flags=0x%x -> 0x%x"
 cpu_exec_start(int cpu_index) "cpu=%d"
 cpu_exec_end(int cpu_index) "cpu=%d"
 
-- 
2.53.0

[PULL 30/32] cpu: Introduce cpu_single_stepping() helper

[Philippe Mathieu-Daudé]

Access CPUState::@singlestep_enabled field with a helper.

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20260705215729.62196-31-philmd@oss.qualcomm.com>
---
 include/hw/core/cpu.h         | 11 +++++++++++
 accel/kvm/kvm-all.c           |  2 +-
 accel/tcg/cpu-exec-common.c   |  2 +-
 accel/tcg/cpu-exec.c          |  8 ++++----
 linux-user/riscv/cpu_loop.c   |  2 +-
 linux-user/s390x/cpu_loop.c   |  2 +-
 system/cpus.c                 |  2 +-
 target/arm/hvf/hvf.c          |  8 ++++----
 target/arm/kvm.c              |  2 +-
 target/i386/kvm/kvm.c         |  2 +-
 target/i386/whpx/whpx-all.c   |  6 +++---
 target/loongarch/kvm/kvm.c    |  2 +-
 target/microblaze/translate.c |  2 +-
 target/ppc/kvm.c              |  2 +-
 target/s390x/kvm/kvm.c        |  2 +-
 15 files changed, 33 insertions(+), 22 deletions(-)

diff --git a/include/hw/core/cpu.h b/include/hw/core/cpu.h
index e5b68bced3d..00f6c5d5de7 100644
--- a/include/hw/core/cpu.h
+++ b/include/hw/core/cpu.h
@@ -1138,6 +1138,17 @@ void qemu_init_vcpu(CPUState *cpu);
  */
 void cpu_single_step(CPUState *cpu, int enabled);
 
+/**
+ * cpu_single_stepping:
+ * @cpu: The vCPU to check
+ *
+ * Returns whether the vCPU has single-stepping enabled.
+ */
+static inline bool cpu_single_stepping(const CPUState *cpu)
+{
+    return cpu->singlestep_enabled;
+}
+
 int cpu_breakpoint_insert(CPUState *cpu, vaddr pc, int flags,
                           CPUBreakpoint **breakpoint);
 int cpu_breakpoint_remove(CPUState *cpu, vaddr pc, int flags);
diff --git a/accel/kvm/kvm-all.c b/accel/kvm/kvm-all.c
index feffca0d82d..963a4edd262 100644
--- a/accel/kvm/kvm-all.c
+++ b/accel/kvm/kvm-all.c
@@ -3812,7 +3812,7 @@ int kvm_update_guest_debug(CPUState *cpu, unsigned long reinject_trap)
 
     data.dbg.control = reinject_trap;
 
-    if (cpu->singlestep_enabled) {
+    if (cpu_single_stepping(cpu)) {
         data.dbg.control |= KVM_GUESTDBG_ENABLE | KVM_GUESTDBG_SINGLESTEP;
 
         if (cpu->singlestep_enabled & SSTEP_NOIRQ) {
diff --git a/accel/tcg/cpu-exec-common.c b/accel/tcg/cpu-exec-common.c
index e48ea31373f..44e84344f3b 100644
--- a/accel/tcg/cpu-exec-common.c
+++ b/accel/tcg/cpu-exec-common.c
@@ -47,7 +47,7 @@ uint32_t curr_cflags(CPUState *cpu)
      * For singlestep and -d nochain, suppress goto_tb so that
      * we can log -d cpu,exec after every TB.
      */
-    if (unlikely(cpu->singlestep_enabled)) {
+    if (unlikely(cpu_single_stepping(cpu))) {
         cflags |= CF_NO_GOTO_TB | CF_NO_GOTO_PTR | CF_SINGLE_STEP | 1;
     } else if (qatomic_read(&one_insn_per_tb)) {
         cflags |= CF_NO_GOTO_TB | 1;
diff --git a/accel/tcg/cpu-exec.c b/accel/tcg/cpu-exec.c
index d7f91dce9db..0386ac49551 100644
--- a/accel/tcg/cpu-exec.c
+++ b/accel/tcg/cpu-exec.c
@@ -308,7 +308,7 @@ static bool check_for_breakpoints_slow(CPUState *cpu, vaddr pc,
      * so that one could (gdb) singlestep into the guest kernel's
      * architectural breakpoint handler.
      */
-    if (cpu->singlestep_enabled) {
+    if (cpu_single_stepping(cpu)) {
         return false;
     }
 
@@ -485,7 +485,7 @@ cpu_tb_exec(CPUState *cpu, TranslationBlock *itb, int *tb_exit)
      * raise a debug exception.  Single-step with another exception
      * is handled in cpu_handle_exception.
      */
-    if (unlikely(cpu->singlestep_enabled) && cpu->exception_index == -1) {
+    if (unlikely(cpu_single_stepping(cpu)) && cpu->exception_index == -1) {
         cpu->exception_index = EXCP_DEBUG;
         cpu_loop_exit(cpu);
     }
@@ -732,7 +732,7 @@ static inline bool cpu_handle_exception(CPUState *cpu, int *ret)
         bql_unlock();
         cpu->exception_index = -1;
 
-        if (unlikely(cpu->singlestep_enabled)) {
+        if (unlikely(cpu_single_stepping(cpu))) {
             /*
              * After processing the exception, ensure an EXCP_DEBUG is
              * raised when single-stepping so that GDB doesn't miss the
@@ -849,7 +849,7 @@ static inline bool cpu_handle_interrupt(CPUState *cpu,
                  * raised when single-stepping so that GDB doesn't miss the
                  * next instruction.
                  */
-                if (unlikely(cpu->singlestep_enabled)) {
+                if (unlikely(cpu_single_stepping(cpu))) {
                     cpu->exception_index = EXCP_DEBUG;
                     bql_unlock();
                     return true;
diff --git a/linux-user/riscv/cpu_loop.c b/linux-user/riscv/cpu_loop.c
index eecc8d15178..6f27a06da49 100644
--- a/linux-user/riscv/cpu_loop.c
+++ b/linux-user/riscv/cpu_loop.c
@@ -68,7 +68,7 @@ void cpu_loop(CPURISCVState *env)
             } else if (ret != -QEMU_ESIGRETURN && ret != -QEMU_ESETPC) {
                 env->gpr[xA0] = ret;
             }
-            if (cs->singlestep_enabled) {
+            if (cpu_single_stepping(cs)) {
                 goto gdbstep;
             }
             break;
diff --git a/linux-user/s390x/cpu_loop.c b/linux-user/s390x/cpu_loop.c
index 67d2a803fbc..25f19f725e7 100644
--- a/linux-user/s390x/cpu_loop.c
+++ b/linux-user/s390x/cpu_loop.c
@@ -87,7 +87,7 @@ void cpu_loop(CPUS390XState *env)
                 env->regs[2] = ret;
             }
 
-            if (unlikely(cs->singlestep_enabled)) {
+            if (unlikely(cpu_single_stepping(cs))) {
                 /*
                  * cpu_tb_exec() did not raise EXCP_DEBUG, because it has seen
                  * that EXCP_SVC was already pending.
diff --git a/system/cpus.c b/system/cpus.c
index b31c825b464..97e5a5edee2 100644
--- a/system/cpus.c
+++ b/system/cpus.c
@@ -344,7 +344,7 @@ bool cpu_can_run(CPUState *cpu)
 void cpu_handle_guest_debug(CPUState *cpu)
 {
     if (replay_running_debug()) {
-        if (!cpu->singlestep_enabled) {
+        if (!cpu_single_stepping(cpu)) {
             /*
              * Report about the breakpoint and
              * make a single step to skip it
diff --git a/target/arm/hvf/hvf.c b/target/arm/hvf/hvf.c
index d5b1966a867..640ef665593 100644
--- a/target/arm/hvf/hvf.c
+++ b/target/arm/hvf/hvf.c
@@ -2340,7 +2340,7 @@ static int hvf_handle_exception(CPUState *cpu, hv_vcpu_exit_exception_t *excp)
     case EC_SOFTWARESTEP: {
         ret = EXCP_DEBUG;
 
-        if (!cpu->singlestep_enabled) {
+        if (!cpu_single_stepping(cpu)) {
             error_report("EC_SOFTWARESTEP but single-stepping not enabled");
         }
         break;
@@ -2549,7 +2549,7 @@ static int hvf_handle_exception(CPUState *cpu, hv_vcpu_exit_exception_t *excp)
         assert_hvf_ok(r);
 
         /* Handle single-stepping over instructions which trigger a VM exit */
-        if (cpu->singlestep_enabled) {
+        if (cpu_single_stepping(cpu)) {
             ret = EXCP_DEBUG;
         }
     }
@@ -2868,7 +2868,7 @@ void hvf_arch_update_guest_debug(CPUState *cpu)
     CPUARMState *env = &arm_cpu->env;
 
     /* Check whether guest debugging is enabled */
-    cpu->accel->guest_debug_enabled = cpu->singlestep_enabled ||
+    cpu->accel->guest_debug_enabled = cpu_single_stepping(cpu) ||
                                     hvf_sw_breakpoints_active(cpu) ||
                                     hvf_arm_hw_debug_active(cpu);
 
@@ -2882,7 +2882,7 @@ void hvf_arch_update_guest_debug(CPUState *cpu)
     cpu_synchronize_state(cpu);
 
     /* Enable/disable single-stepping */
-    if (cpu->singlestep_enabled) {
+    if (cpu_single_stepping(cpu)) {
         env->cp15.mdscr_el1 =
             deposit64(env->cp15.mdscr_el1, MDSCR_EL1_SS_SHIFT, 1, 1);
         pstate_write(env, pstate_read(env) | PSTATE_SS);
diff --git a/target/arm/kvm.c b/target/arm/kvm.c
index 01f42f42538..d40a6a98591 100644
--- a/target/arm/kvm.c
+++ b/target/arm/kvm.c
@@ -1492,7 +1492,7 @@ static bool kvm_arm_handle_debug(ARMCPU *cpu,
 
     switch (hsr_ec) {
     case EC_SOFTWARESTEP:
-        if (cs->singlestep_enabled) {
+        if (cpu_single_stepping(cs)) {
             return true;
         } else {
             /*
diff --git a/target/i386/kvm/kvm.c b/target/i386/kvm/kvm.c
index 1e09155e92c..4272b6770c7 100644
--- a/target/i386/kvm/kvm.c
+++ b/target/i386/kvm/kvm.c
@@ -6250,7 +6250,7 @@ static int kvm_handle_debug(X86CPU *cpu,
 
     if (arch_info->exception == EXCP01_DB) {
         if (arch_info->dr6 & DR6_BS) {
-            if (cs->singlestep_enabled) {
+            if (cpu_single_stepping(cs)) {
                 ret = EXCP_DEBUG;
             }
         } else {
diff --git a/target/i386/whpx/whpx-all.c b/target/i386/whpx/whpx-all.c
index 11b5d1fc607..634d5428219 100644
--- a/target/i386/whpx/whpx-all.c
+++ b/target/i386/whpx/whpx-all.c
@@ -2266,7 +2266,7 @@ int whpx_vcpu_run(CPUState *cpu)
             }
         }
 
-        if (exclusive_step_mode != WHPX_STEP_NONE || cpu->singlestep_enabled) {
+        if (exclusive_step_mode != WHPX_STEP_NONE || cpu_single_stepping(cpu)) {
             whpx_vcpu_configure_single_stepping(cpu, true, NULL);
         }
 
@@ -2283,7 +2283,7 @@ int whpx_vcpu_run(CPUState *cpu)
             break;
         }
 
-        if (exclusive_step_mode != WHPX_STEP_NONE || cpu->singlestep_enabled) {
+        if (exclusive_step_mode != WHPX_STEP_NONE || cpu_single_stepping(cpu)) {
             whpx_vcpu_configure_single_stepping(cpu,
                 false,
                 &vcpu->exit_ctx.VpContext.Rflags);
@@ -2648,7 +2648,7 @@ int whpx_vcpu_run(CPUState *cpu)
                 cpu->exception_index = EXCP_DEBUG;
             } else if ((vcpu->exit_ctx.VpException.ExceptionType ==
                         WHvX64ExceptionTypeDebugTrapOrFault) &&
-                       !cpu->singlestep_enabled) {
+                       !cpu_single_stepping(cpu)) {
                 /*
                  * Just finished stepping over a breakpoint, but the
                  * gdb does not expect us to do single-stepping.
diff --git a/target/loongarch/kvm/kvm.c b/target/loongarch/kvm/kvm.c
index beda6965f8c..48cdb78a4ca 100644
--- a/target/loongarch/kvm/kvm.c
+++ b/target/loongarch/kvm/kvm.c
@@ -1440,7 +1440,7 @@ static bool kvm_loongarch_handle_debug(CPUState *cs, struct kvm_run *run)
     CPULoongArchState *env = &cpu->env;
 
     kvm_cpu_synchronize_state(cs);
-    if (cs->singlestep_enabled) {
+    if (cpu_single_stepping(cs)) {
         return true;
     }
 
diff --git a/target/microblaze/translate.c b/target/microblaze/translate.c
index a93f14d6e09..8b219afb5dd 100644
--- a/target/microblaze/translate.c
+++ b/target/microblaze/translate.c
@@ -1772,7 +1772,7 @@ static void mb_tr_tb_stop(DisasContextBase *dcb, CPUState *cs)
     }
 
     /* Finish DISAS_EXIT_* */
-    if (unlikely(cs->singlestep_enabled)) {
+    if (unlikely(cpu_single_stepping(cs))) {
         gen_raise_exception(dc, EXCP_DEBUG);
     } else {
         tcg_gen_exit_tb(NULL, 0);
diff --git a/target/ppc/kvm.c b/target/ppc/kvm.c
index 93ec0cf6a72..19d41c8e91a 100644
--- a/target/ppc/kvm.c
+++ b/target/ppc/kvm.c
@@ -1615,7 +1615,7 @@ static int kvm_handle_debug(PowerPCCPU *cpu, struct kvm_run *run)
     CPUPPCState *env = &cpu->env;
     struct kvm_debug_exit_arch *arch_info = &run->debug.arch;
 
-    if (cs->singlestep_enabled) {
+    if (cpu_single_stepping(cs)) {
         return kvm_handle_singlestep();
     }
 
diff --git a/target/s390x/kvm/kvm.c b/target/s390x/kvm/kvm.c
index 195e39df032..19ddf59e002 100644
--- a/target/s390x/kvm/kvm.c
+++ b/target/s390x/kvm/kvm.c
@@ -1874,7 +1874,7 @@ static int kvm_arch_handle_debug_exit(S390CPU *cpu)
         }
         break;
     case KVM_SINGLESTEP:
-        if (cs->singlestep_enabled) {
+        if (cpu_single_stepping(cs)) {
             ret = EXCP_DEBUG;
         }
         break;
-- 
2.53.0

[PULL 31/32] cpu: Rename CPUState @singlestep_enabled -> @singlestep_flags

[Philippe Mathieu-Daudé]

CPUState::singlestep_enabled contains multiple flags since
commit 60897d369f1 ("Debugger single step without interrupts").
Use an unsigned type and rename the field to avoid mistakes.

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20260705215729.62196-32-philmd@oss.qualcomm.com>
---
 include/hw/core/cpu.h        | 10 +++++-----
 accel/kvm/kvm-all.c          |  2 +-
 accel/tcg/cpu-exec.c         |  2 +-
 accel/tcg/tcg-accel-ops-rr.c |  2 +-
 cpu-target.c                 |  8 ++++----
 target/arm/hvf/hvf.c         |  2 +-
 target/ppc/translate.c       | 16 ++++++++--------
 7 files changed, 21 insertions(+), 21 deletions(-)

diff --git a/include/hw/core/cpu.h b/include/hw/core/cpu.h
index 00f6c5d5de7..db4865bd63d 100644
--- a/include/hw/core/cpu.h
+++ b/include/hw/core/cpu.h
@@ -440,7 +440,7 @@ struct qemu_work_item;
  * @stopped: Indicates the CPU has been artificially stopped.
  * @unplug: Indicates a pending CPU unplug request.
  * @crash_occurred: Indicates the OS reported a crash (panic) for this CPU
- * @singlestep_enabled: Flags for single-stepping.
+ * @singlestep_flags: Flags for single-stepping.
  * @icount_extra: Instructions until next timer event.
  * @cpu_ases: Pointer to array of CPUAddressSpaces (which define the
  *            AddressSpaces this CPU has)
@@ -505,7 +505,7 @@ struct CPUState {
     int exclusive_context_count;
     uint32_t cflags_next_tb;
     uint32_t interrupt_request;
-    int singlestep_enabled;
+    unsigned singlestep_flags;
     int64_t icount_budget;
     int64_t icount_extra;
     uint64_t random_seed;
@@ -1132,11 +1132,11 @@ void qemu_init_vcpu(CPUState *cpu);
 /**
  * cpu_single_step:
  * @cpu: CPU to the flags for.
- * @enabled: Flags to enable.
+ * @flags: Flags to enable.
  *
  * Enables or disables single-stepping for @cpu.
  */
-void cpu_single_step(CPUState *cpu, int enabled);
+void cpu_single_step(CPUState *cpu, unsigned flags);
 
 /**
  * cpu_single_stepping:
@@ -1146,7 +1146,7 @@ void cpu_single_step(CPUState *cpu, int enabled);
  */
 static inline bool cpu_single_stepping(const CPUState *cpu)
 {
-    return cpu->singlestep_enabled;
+    return cpu->singlestep_flags;
 }
 
 int cpu_breakpoint_insert(CPUState *cpu, vaddr pc, int flags,
diff --git a/accel/kvm/kvm-all.c b/accel/kvm/kvm-all.c
index 963a4edd262..83cbd120a84 100644
--- a/accel/kvm/kvm-all.c
+++ b/accel/kvm/kvm-all.c
@@ -3815,7 +3815,7 @@ int kvm_update_guest_debug(CPUState *cpu, unsigned long reinject_trap)
     if (cpu_single_stepping(cpu)) {
         data.dbg.control |= KVM_GUESTDBG_ENABLE | KVM_GUESTDBG_SINGLESTEP;
 
-        if (cpu->singlestep_enabled & SSTEP_NOIRQ) {
+        if (cpu->singlestep_flags & SSTEP_NOIRQ) {
             data.dbg.control |= KVM_GUESTDBG_BLOCKIRQ;
         }
     }
diff --git a/accel/tcg/cpu-exec.c b/accel/tcg/cpu-exec.c
index 0386ac49551..257211235db 100644
--- a/accel/tcg/cpu-exec.c
+++ b/accel/tcg/cpu-exec.c
@@ -828,7 +828,7 @@ static inline bool cpu_handle_interrupt(CPUState *cpu,
                 return true;
             }
 
-            if (unlikely(cpu->singlestep_enabled & SSTEP_NOIRQ)) {
+            if (unlikely(cpu->singlestep_flags & SSTEP_NOIRQ)) {
                 /* Mask out external interrupts for this step. */
                 interrupt_request &= ~CPU_INTERRUPT_SSTEP_MASK;
             }
diff --git a/accel/tcg/tcg-accel-ops-rr.c b/accel/tcg/tcg-accel-ops-rr.c
index 5b132d3d5d8..cdaa3e11808 100644
--- a/accel/tcg/tcg-accel-ops-rr.c
+++ b/accel/tcg/tcg-accel-ops-rr.c
@@ -274,7 +274,7 @@ static void *rr_cpu_thread_fn(void *arg)
             current_cpu = cpu;
 
             qemu_clock_enable(QEMU_CLOCK_VIRTUAL,
-                              (cpu->singlestep_enabled & SSTEP_NOTIMER) == 0);
+                              (cpu->singlestep_flags & SSTEP_NOTIMER) == 0);
 
             if (cpu_can_run(cpu)) {
                 int r;
diff --git a/cpu-target.c b/cpu-target.c
index 019906b32eb..4783845c9bf 100644
--- a/cpu-target.c
+++ b/cpu-target.c
@@ -28,12 +28,12 @@
 
 /* enable or disable single step mode. EXCP_DEBUG is returned by the
    CPU loop after each instruction */
-void cpu_single_step(CPUState *cpu, int enabled)
+void cpu_single_step(CPUState *cpu, unsigned flags)
 {
-    if (cpu->singlestep_enabled != enabled) {
+    if (cpu->singlestep_flags != flags) {
         trace_cpu_change_singlestep_flags(cpu->cpu_index,
-                                          cpu->singlestep_enabled, enabled);
-        cpu->singlestep_enabled = enabled;
+                                          cpu->singlestep_flags, flags);
+        cpu->singlestep_flags = flags;
 
 #if !defined(CONFIG_USER_ONLY)
         const AccelOpsClass *ops = cpus_get_accel();
diff --git a/target/arm/hvf/hvf.c b/target/arm/hvf/hvf.c
index 640ef665593..f5dd7e8e029 100644
--- a/target/arm/hvf/hvf.c
+++ b/target/arm/hvf/hvf.c
@@ -2603,7 +2603,7 @@ int hvf_arch_vcpu_exec(CPUState *cpu)
     flush_cpu_state(cpu);
 
     do {
-        if (!(cpu->singlestep_enabled & SSTEP_NOIRQ) &&
+        if (!(cpu->singlestep_flags & SSTEP_NOIRQ) &&
             hvf_inject_interrupts(cpu)) {
             return EXCP_INTERRUPT;
         }
diff --git a/target/ppc/translate.c b/target/ppc/translate.c
index 3f6d326cef3..06ed2adf105 100644
--- a/target/ppc/translate.c
+++ b/target/ppc/translate.c
@@ -198,7 +198,7 @@ struct DisasContext {
     bool pmu_insn_cnt;
     bool bhrb_enable;
     ppc_spr_t *spr_cb; /* Needed to check rights for mfspr/mtspr */
-    int singlestep_enabled;
+    int singlestep_flags;
     uint32_t flags;
     uint64_t insns_flags;
     uint64_t insns_flags2;
@@ -367,7 +367,7 @@ static void gen_debug_exception(DisasContext *ctx, bool rfi_type)
 #if !defined(CONFIG_USER_ONLY)
     if (ctx->flags & POWERPC_FLAG_DE) {
         target_ulong dbsr = 0;
-        if (ctx->singlestep_enabled & CPU_SINGLE_STEP) {
+        if (ctx->singlestep_flags & CPU_SINGLE_STEP) {
             dbsr = DBCR0_ICMP;
         } else {
             /* Must have been branch */
@@ -3645,7 +3645,7 @@ static void pmu_count_insns(DisasContext *ctx)
 
 static inline bool use_goto_tb(DisasContext *ctx, target_ulong dest)
 {
-    if (unlikely(ctx->singlestep_enabled)) {
+    if (unlikely(ctx->singlestep_flags)) {
         return false;
     }
     return translator_use_goto_tb(&ctx->base, dest);
@@ -3653,7 +3653,7 @@ static inline bool use_goto_tb(DisasContext *ctx, target_ulong dest)
 
 static void gen_lookup_and_goto_ptr(DisasContext *ctx)
 {
-    if (unlikely(ctx->singlestep_enabled)) {
+    if (unlikely(ctx->singlestep_flags)) {
         gen_debug_exception(ctx, false);
     } else {
         /*
@@ -6559,13 +6559,13 @@ static void ppc_tr_init_disas_context(DisasContextBase *dcbase, CPUState *cs)
     ctx->pmu_insn_cnt = (hflags >> HFLAGS_INSN_CNT) & 1;
     ctx->bhrb_enable = (hflags >> HFLAGS_BHRB_ENABLE) & 1;
 
-    ctx->singlestep_enabled = 0;
+    ctx->singlestep_flags = 0;
     if ((hflags >> HFLAGS_SE) & 1) {
-        ctx->singlestep_enabled |= CPU_SINGLE_STEP;
+        ctx->singlestep_flags |= CPU_SINGLE_STEP;
         ctx->base.max_insns = 1;
     }
     if ((hflags >> HFLAGS_BE) & 1) {
-        ctx->singlestep_enabled |= CPU_BRANCH_STEP;
+        ctx->singlestep_flags |= CPU_BRANCH_STEP;
     }
 }
 
@@ -6641,7 +6641,7 @@ static void ppc_tr_tb_stop(DisasContextBase *dcbase, CPUState *cs)
     }
 
     /* Honor single stepping. */
-    if (unlikely(ctx->singlestep_enabled & CPU_SINGLE_STEP)) {
+    if (unlikely(ctx->singlestep_flags & CPU_SINGLE_STEP)) {
         bool rfi_type = false;
 
         switch (is_jmp) {
-- 
2.53.0

[PULL 32/32] cpu: Only check SSTEP_ENABLE flag in cpu_single_stepping()

[Philippe Mathieu-Daudé]

Only the SSTEP_ENABLE bitmask means single-step is enabled.

Fixes: 60897d369f1 ("Debugger single step without interrupts")
Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
Reviewed-by: Daniel Henrique Barboza <daniel.barboza@oss.qualcomm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20260705215729.62196-33-philmd@oss.qualcomm.com>
---
 include/hw/core/cpu.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/include/hw/core/cpu.h b/include/hw/core/cpu.h
index db4865bd63d..b54035fb13b 100644
--- a/include/hw/core/cpu.h
+++ b/include/hw/core/cpu.h
@@ -1146,7 +1146,7 @@ void cpu_single_step(CPUState *cpu, unsigned flags);
  */
 static inline bool cpu_single_stepping(const CPUState *cpu)
 {
-    return cpu->singlestep_flags;
+    return cpu->singlestep_flags & SSTEP_ENABLE;
 }
 
 int cpu_breakpoint_insert(CPUState *cpu, vaddr pc, int flags,
-- 
2.53.0

Re: [PULL 00/32] Accelerator patches for 2026-07-06

[Stefan Hajnoczi]

[-- Attachment #1: Type: text/plain, Size: 116 bytes --]

Applied, thanks.

Please update the changelog at https://wiki.qemu.org/ChangeLog/11.1 for any user-visible changes.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]