Skip to contentSkip to conversation

target/arm: Allow CONFIG_TCG=n builds

PATCH Superseded Fabiano Rosas 3 participants 13 messages Mar 13, 2023 – Apr 11, 2023 Arm/Amlogic Meson SoC support

v9 of 11 Superseded Latest is v11 →

lore
Share & tools
lore
#0 lore raw

v9 cover target/arm: Allow CONFIG_TCG=n builds

Changes since v8:

- reverted back to keeping -cpu max code for TCG in tcg/cpu64.c;

- pauth: used tcg_enabled instead of CONFIG_TCG in gdbstub.c. However we still need to keep the ifdef around the function definition in gdbstub64.c;

- moved all the regression and test fixes after the patches that enable the --disable-tcg flag;

- tests with --enable-xen --disable-tcg:

Tests now explicitly skip instead of passing.

We cannot do the check at build time because we build the tests only once for all the QEMU binaries and the native binary will have support for KVM, thus breaking our !CONFIG_TCG && !CONFIG_KVM verification.

I also don't think we can have a wrapper because which accelerators are used is a test-specific information. The test should be able to specify which accelerators it needs. In other words there will always be hardcoded "tcg" and "kvm" strings somewhere.

CI run: https://gitlab.com/farosas/qemu/-/pipelines/804459781

v8:
https://lore.kernel.org/r/[email protected]

v7 resend:
https://lore.kernel.org/r/[email protected]

v7:
https://lore.kernel.org/r/[email protected]

v6:
https://lore.kernel.org/r/[email protected]

v5 resend:
https://lore.kernel.org/r/[email protected]

v5:
https://lore.kernel.org/r/[email protected]

v4:
https://lore.kernel.org/r/[email protected]

v3:
https://lore.kernel.org/r/[email protected]

v2:
https://lore.kernel.org/r/[email protected]

v1:
https://lore.kernel.org/r/[email protected]

Claudio Fontana (1):
target/arm: move cpu_tcg to tcg/cpu32.c

Fabiano Rosas (8):
target/arm: Move cortex sysregs into a separate file target/arm: Move 64-bit TCG CPUs into tcg/ target/arm: gdbstub: Guard M-profile code with CONFIG_TCG target/arm: gdbstub: Guard pauth code with CONFIG_TCG tests/qtest: Fix tests when no KVM or TCG are present tests/avocado: Pass parameters to migration test arm/Kconfig: Always select SEMIHOSTING when TCG is present arm/Kconfig: Do not build TCG-only boards on a KVM-only build

Philippe Mathieu-Daudé (1):
gitlab-ci: Check building KVM-only aarch64 target

 .gitlab-ci.d/crossbuilds.yml                  |  11 +
 .../custom-runners/ubuntu-22.04-aarch64.yml   |   4 -
 configs/devices/aarch64-softmmu/default.mak   |   4 -
 configs/devices/arm-softmmu/default.mak       |  39 --
 hw/arm/Kconfig                                |  43 +-
 hw/arm/virt.c                                 |   6 +-
 target/arm/Kconfig                            |   7 +
 target/arm/cortex-regs.c                      |  69 ++
 target/arm/cpregs.h                           |   6 +
 target/arm/cpu64.c                            | 633 +----------------
 target/arm/gdbstub.c                          |   8 +-
 target/arm/gdbstub64.c                        |   2 +
 target/arm/internals.h                        |  10 +-
 target/arm/meson.build                        |   2 +-
 target/arm/{cpu_tcg.c => tcg/cpu32.c}         |  72 +-
 target/arm/tcg/cpu64.c                        | 655 ++++++++++++++++++
 target/arm/tcg/meson.build                    |   2 +
 tests/avocado/migration.py                    |  83 ++-
 tests/qtest/arm-cpu-features.c                |  12 +-
 tests/qtest/bios-tables-test.c                |  11 +-
 tests/qtest/boot-serial-test.c                |   5 +
 tests/qtest/migration-test.c                  |  10 +-
 tests/qtest/pxe-test.c                        |   8 +-
 tests/qtest/vmgenid-test.c                    |   9 +-
 24 files changed, 949 insertions(+), 762 deletions(-)
 create mode 100644 target/arm/cortex-regs.c
 rename target/arm/{cpu_tcg.c => tcg/cpu32.c} (93%)
 create mode 100644 target/arm/tcg/cpu64.c
Signature · 1 line
2.35.3
#1 lore raw

PATCH v9 1/10 target/arm: Move cortex sysregs into a separate file

The file cpu_tcg.c is about to be moved into the tcg/ directory, so move the register definitions into a new file.

Also move the function declaration to the more appropriate cpregs.h.

---
 target/arm/cortex-regs.c | 69 ++++++++++++++++++++++++++++++++++++++++
 target/arm/cpregs.h      |  6 ++++
 target/arm/cpu64.c       |  1 +
 target/arm/cpu_tcg.c     | 59 ----------------------------------
 target/arm/internals.h   |  6 ----
 target/arm/meson.build   |  1 +
 6 files changed, 77 insertions(+), 65 deletions(-)
 create mode 100644 target/arm/cortex-regs.c
target/arm/cortex-regs.c new file +69 -0
new file mode 100644
index 0000000000..17708480e7
@@ -0,0 +1,69 @@
+/*
+ * ARM Cortex-A registers
+ *
+ * This code is licensed under the GNU GPL v2 or later.
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+#include "qemu/osdep.h"
+#include "cpu.h"
+#include "cpregs.h"
+
+
+static uint64_t l2ctlr_read(CPUARMState *env, const ARMCPRegInfo *ri)
+{
+    ARMCPU *cpu = env_archcpu(env);
+
+    /* Number of cores is in [25:24]; otherwise we RAZ */
+    return (cpu->core_count - 1) << 24;
+}
+
+static const ARMCPRegInfo cortex_a72_a57_a53_cp_reginfo[] = {
+    { .name = "L2CTLR_EL1", .state = ARM_CP_STATE_AA64,
+      .opc0 = 3, .opc1 = 1, .crn = 11, .crm = 0, .opc2 = 2,
+      .access = PL1_RW, .readfn = l2ctlr_read,
+      .writefn = arm_cp_write_ignore },
+    { .name = "L2CTLR",
+      .cp = 15, .opc1 = 1, .crn = 9, .crm = 0, .opc2 = 2,
+      .access = PL1_RW, .readfn = l2ctlr_read,
+      .writefn = arm_cp_write_ignore },
+    { .name = "L2ECTLR_EL1", .state = ARM_CP_STATE_AA64,
+      .opc0 = 3, .opc1 = 1, .crn = 11, .crm = 0, .opc2 = 3,
+      .access = PL1_RW, .type = ARM_CP_CONST, .resetvalue = 0 },
+    { .name = "L2ECTLR",
+      .cp = 15, .opc1 = 1, .crn = 9, .crm = 0, .opc2 = 3,
+      .access = PL1_RW, .type = ARM_CP_CONST, .resetvalue = 0 },
+    { .name = "L2ACTLR", .state = ARM_CP_STATE_BOTH,
+      .opc0 = 3, .opc1 = 1, .crn = 15, .crm = 0, .opc2 = 0,
+      .access = PL1_RW, .type = ARM_CP_CONST, .resetvalue = 0 },
+    { .name = "CPUACTLR_EL1", .state = ARM_CP_STATE_AA64,
+      .opc0 = 3, .opc1 = 1, .crn = 15, .crm = 2, .opc2 = 0,
+      .access = PL1_RW, .type = ARM_CP_CONST, .resetvalue = 0 },
+    { .name = "CPUACTLR",
+      .cp = 15, .opc1 = 0, .crm = 15,
+      .access = PL1_RW, .type = ARM_CP_CONST | ARM_CP_64BIT, .resetvalue = 0 },
+    { .name = "CPUECTLR_EL1", .state = ARM_CP_STATE_AA64,
+      .opc0 = 3, .opc1 = 1, .crn = 15, .crm = 2, .opc2 = 1,
+      .access = PL1_RW, .type = ARM_CP_CONST, .resetvalue = 0 },
+    { .name = "CPUECTLR",
+      .cp = 15, .opc1 = 1, .crm = 15,
+      .access = PL1_RW, .type = ARM_CP_CONST | ARM_CP_64BIT, .resetvalue = 0 },
+    { .name = "CPUMERRSR_EL1", .state = ARM_CP_STATE_AA64,
+      .opc0 = 3, .opc1 = 1, .crn = 15, .crm = 2, .opc2 = 2,
+      .access = PL1_RW, .type = ARM_CP_CONST, .resetvalue = 0 },
+    { .name = "CPUMERRSR",
+      .cp = 15, .opc1 = 2, .crm = 15,
+      .access = PL1_RW, .type = ARM_CP_CONST | ARM_CP_64BIT, .resetvalue = 0 },
+    { .name = "L2MERRSR_EL1", .state = ARM_CP_STATE_AA64,
+      .opc0 = 3, .opc1 = 1, .crn = 15, .crm = 2, .opc2 = 3,
+      .access = PL1_RW, .type = ARM_CP_CONST, .resetvalue = 0 },
+    { .name = "L2MERRSR",
+      .cp = 15, .opc1 = 3, .crm = 15,
+      .access = PL1_RW, .type = ARM_CP_CONST | ARM_CP_64BIT, .resetvalue = 0 },
+};
+
+void define_cortex_a72_a57_a53_cp_reginfo(ARMCPU *cpu)
+{
+    define_arm_cp_regs(cpu, cortex_a72_a57_a53_cp_reginfo);
+}
target/arm/cpregs.h +6 -0
index 1ee64e99de..b04d344a9f 100644
@@ -1071,4 +1071,10 @@ static inline bool arm_cpreg_in_idspace(const ARMCPRegInfo *ri)
                                       ri->crn, ri->crm);
 }
 
+#ifdef CONFIG_USER_ONLY
+static inline void define_cortex_a72_a57_a53_cp_reginfo(ARMCPU *cpu) { }
+#else
+void define_cortex_a72_a57_a53_cp_reginfo(ARMCPU *cpu);
+#endif
+
 #endif /* TARGET_ARM_CPREGS_H */
target/arm/cpu64.c +1 -0
index 4066950da1..9f193927d8 100644
@@ -29,6 +29,7 @@
 #include "qapi/visitor.h"
 #include "hw/qdev-properties.h"
 #include "internals.h"
+#include "cpregs.h"
 
 static void aarch64_a35_initfn(Object *obj)
 {
target/arm/cpu_tcg.c +0 -59

65 lines not on this page · raw message

target/arm/internals.h +0 -6
index b1ef05963f..46a1d06e15 100644
@@ -1370,12 +1370,6 @@ uint32_t arm_v7m_mrs_control(CPUARMState *env, uint32_t secure);
 uint32_t *arm_v7m_get_sp_ptr(CPUARMState *env, bool secure,
                              bool threadmode, bool spsel);
 
-#ifdef CONFIG_USER_ONLY
-static inline void define_cortex_a72_a57_a53_cp_reginfo(ARMCPU *cpu) { }
-#else
-void define_cortex_a72_a57_a53_cp_reginfo(ARMCPU *cpu);
-#endif
-
 bool el_is_in_host(CPUARMState *env, int el);
 
 void aa32_max_features(ARMCPU *cpu);
target/arm/meson.build +1 -0
index 6226098ad5..3469926295 100644
@@ -21,6 +21,7 @@ arm_softmmu_ss.add(files(
   'arch_dump.c',
   'arm-powerctl.c',
   'arm-qmp-cmds.c',
+  'cortex-regs.c',
   'machine.c',
   'ptw.c',
 ))
Signature · 1 line
2.35.3
#2 lore raw

PATCH v9 2/10 target/arm: Move 64-bit TCG CPUs into tcg/

In reply to Fabiano Rosas

Move the 64-bit CPUs that are TCG-only:
- cortex-a35
- cortex-a55
- cortex-a72
- cortex-a76
- a64fx
- neoverse-n1

Keep the CPUs that can be used with KVM:
- cortex-a57
- cortex-a53
- max
- host

For the special case "max" CPU, there's a nuance that while KVM/HVF use the "host" model instead, we still cannot move all of the TCG code into the tcg directory because the qtests might reach the !kvm && !hvf branch. Keep the cortex_a57_initfn() call to cover that scenario.

---
 hw/arm/virt.c              |   6 +-
 target/arm/cpu64.c         | 632 +----------------------------------
 target/arm/internals.h     |   4 +
 target/arm/tcg/cpu64.c     | 655 +++++++++++++++++++++++++++++++++++++
 target/arm/tcg/meson.build |   1 +
 5 files changed, 676 insertions(+), 622 deletions(-)
 create mode 100644 target/arm/tcg/cpu64.c
hw/arm/virt.c +4 -2
index ac626b3bef..999c1ada79 100644
@@ -206,14 +206,16 @@ static const int a15irqmap[] = {
 static const char *valid_cpus[] = {
     ARM_CPU_TYPE_NAME("cortex-a7"),
     ARM_CPU_TYPE_NAME("cortex-a15"),
+#ifdef CONFIG_TCG
     ARM_CPU_TYPE_NAME("cortex-a35"),
-    ARM_CPU_TYPE_NAME("cortex-a53"),
     ARM_CPU_TYPE_NAME("cortex-a55"),
-    ARM_CPU_TYPE_NAME("cortex-a57"),
     ARM_CPU_TYPE_NAME("cortex-a72"),
     ARM_CPU_TYPE_NAME("cortex-a76"),
     ARM_CPU_TYPE_NAME("a64fx"),
     ARM_CPU_TYPE_NAME("neoverse-n1"),
+#endif
+    ARM_CPU_TYPE_NAME("cortex-a53"),
+    ARM_CPU_TYPE_NAME("cortex-a57"),
     ARM_CPU_TYPE_NAME("host"),
     ARM_CPU_TYPE_NAME("max"),
 };
target/arm/cpu64.c +12 -620

700 lines not on this page · raw message

target/arm/internals.h +4 -0
index 46a1d06e15..be07a87b2c 100644
@@ -1355,6 +1355,10 @@ void arm_cpu_sve_finalize(ARMCPU *cpu, Error **errp);
 void arm_cpu_sme_finalize(ARMCPU *cpu, Error **errp);
 void arm_cpu_pauth_finalize(ARMCPU *cpu, Error **errp);
 void arm_cpu_lpa2_finalize(ARMCPU *cpu, Error **errp);
+void aarch64_max_tcg_initfn(Object *obj);
+void aarch64_add_pauth_properties(Object *obj);
+void aarch64_add_sve_properties(Object *obj);
+void aarch64_add_sme_properties(Object *obj);
 #endif
 
 /* Read the CONTROL register as the MRS instruction would. */
target/arm/tcg/cpu64.c new file +655 -0

655 lines not on this page · raw message

target/arm/tcg/meson.build +1 -0
index d27e76af6c..128f782816 100644
@@ -35,6 +35,7 @@ arm_ss.add(files(
 ))
 
 arm_ss.add(when: 'TARGET_AARCH64', if_true: files(
+  'cpu64.c',
   'translate-a64.c',
   'translate-sve.c',
   'translate-sme.c',
Signature · 1 line
2.35.3
#3 lore raw

Re: [PATCH v9 02/10] target/arm: Move 64-bit TCG CPUs into tcg/

Quoted 35 lines · Fabiano Rosas

Move the 64-bit CPUs that are TCG-only:
- cortex-a35
- cortex-a55
- cortex-a72
- cortex-a76
- a64fx
- neoverse-n1

Keep the CPUs that can be used with KVM:
- cortex-a57
- cortex-a53
- max
- host

For the special case "max" CPU, there's a nuance that while KVM/HVF use the "host" model instead, we still cannot move all of the TCG code into the tcg directory because the qtests might reach the !kvm && !hvf branch. Keep the cortex_a57_initfn() call to cover that scenario.

target/arm/cpu64.c +0 -15
-static void cpu_max_set_sve_max_vq(Object *obj, Visitor *v, const char *name,
-                                   void *opaque, Error **errp)
-{
-    ARMCPU *cpu = ARM_CPU(obj);
-    uint32_t max_vq;
-
-    if (!visit_type_uint32(v, name, &max_vq, errp)) {
-        return;
-    }
-
-    if (kvm_enabled() && !kvm_arm_sve_supported()) {
-        error_setg(errp, "cannot set sve-max-vq");
-        error_append_hint(errp, "SVE not supported by KVM on this host\n");
-        return;
-    }

target/arm/cpu64.c void arm_cpu_sve_finalize(ARMCPU *cpu, Error **errp) · line 352

Quoted 10 lines
target/arm/cpu64.c +0 -10
-
-    if (max_vq == 0 || max_vq > ARM_MAX_VQ) {
-        error_setg(errp, "unsupported SVE vector length");
-        error_append_hint(errp, "Valid sve-max-vq in range [1-%d]\n",
-                          ARM_MAX_VQ);
-        return;
-    }
-
-    cpu->sve_max_vq = max_vq;
-}

The deleted code has 3 if() conditions...

Quoted 9 lines
target/arm/tcg/cpu64.c new file +9 -0
+static void cpu_max_set_sve_max_vq(Object *obj, Visitor *v, const char *name,
+                                   void *opaque, Error **errp)
+{
+    ARMCPU *cpu = ARM_CPU(obj);
+    uint32_t max_vq;
+
+    if (!visit_type_uint32(v, name, &max_vq, errp)) {
+        return;
+    }

target/arm/tcg/cpu64.c line 143

Quoted 10 lines
target/arm/tcg/cpu64.c new file +10 -0
+
+    if (max_vq == 0 || max_vq > ARM_MAX_VQ) {
+        error_setg(errp, "unsupported SVE vector length");
+        error_append_hint(errp, "Valid sve-max-vq in range [1-%d]\n",
+                          ARM_MAX_VQ);
+        return;
+    }
+
+    cpu->sve_max_vq = max_vq;
+}

...but the version of the function moved into the new file has only 2. What's going on here?

Anyway, this fails to apply to current head of git. I had a go at trying to sort it out but gave up when I found this place where what looks like it's basically just a "move code around" patch is actually also making code changes.

Please can you rebase your series and rearrange it so you aren't making code changes in the middle of +700-700 patches?

thanks
-- PMM

#4 lore raw

PATCH v9 3/10 target/arm: move cpu_tcg to tcg/cpu32.c

In reply to Fabiano Rosas

From: Claudio Fontana <[email protected]>

move the module containing cpu models definitions for 32bit TCG-only CPUs to tcg/ and rename it for clarity.

---
 hw/arm/virt.c                         |  2 +-
 target/arm/meson.build                |  1 -
 target/arm/{cpu_tcg.c => tcg/cpu32.c} | 13 +++----------
 target/arm/tcg/cpu64.c                |  2 +-
 target/arm/tcg/meson.build            |  1 +
 tests/qtest/arm-cpu-features.c        | 12 +++++++++---
 6 files changed, 15 insertions(+), 16 deletions(-)
 rename target/arm/{cpu_tcg.c => tcg/cpu32.c} (99%)
hw/arm/virt.c +1 -1
index 999c1ada79..b661b8d91b 100644
@@ -204,9 +204,9 @@ static const int a15irqmap[] = {
 };
 
 static const char *valid_cpus[] = {
+#ifdef CONFIG_TCG
     ARM_CPU_TYPE_NAME("cortex-a7"),
     ARM_CPU_TYPE_NAME("cortex-a15"),
-#ifdef CONFIG_TCG
     ARM_CPU_TYPE_NAME("cortex-a35"),
     ARM_CPU_TYPE_NAME("cortex-a55"),
     ARM_CPU_TYPE_NAME("cortex-a72"),
target/arm/meson.build +0 -1
index 3469926295..359a649eaf 100644
@@ -5,7 +5,6 @@ arm_ss.add(files(
   'gdbstub.c',
   'helper.c',
   'vfp_helper.c',
-  'cpu_tcg.c',
 ))
 arm_ss.add(zlib)
 
target/arm/tcg/cpu32.c renamed from target/arm/cpu_tcg.c +3 -10
similarity index 99%
rename from target/arm/cpu_tcg.c
rename to target/arm/tcg/cpu32.c
index 6ce728134f..a03c819b41 100644
@@ -1,5 +1,5 @@
 /*
- * QEMU ARM TCG CPUs.
+ * QEMU ARM TCG-only CPUs.
  *
  * Copyright (c) 2012 SUSE LINUX Products GmbH
  *
@@ -10,9 +10,7 @@
 
 #include "qemu/osdep.h"
 #include "cpu.h"
-#ifdef CONFIG_TCG
 #include "hw/core/tcg-cpu-ops.h"
-#endif /* CONFIG_TCG */
 #include "internals.h"
 #include "target/arm/idau.h"
 #if !defined(CONFIG_USER_ONLY)
@@ -96,7 +94,7 @@ void aa32_max_features(ARMCPU *cpu)
 /* CPU models. These are not needed for the AArch64 linux-user build. */
 #if !defined(CONFIG_USER_ONLY) || !defined(TARGET_AARCH64)
 
-#if !defined(CONFIG_USER_ONLY) && defined(CONFIG_TCG)
+#if !defined(CONFIG_USER_ONLY)
 static bool arm_v7m_cpu_exec_interrupt(CPUState *cs, int interrupt_request)
 {
     CPUClass *cc = CPU_GET_CLASS(cs);
@@ -120,7 +118,7 @@ static bool arm_v7m_cpu_exec_interrupt(CPUState *cs, int interrupt_request)
     }
     return ret;
 }
-#endif /* !CONFIG_USER_ONLY && CONFIG_TCG */
+#endif /* !CONFIG_USER_ONLY */
 
 static void arm926_initfn(Object *obj)
 {
@@ -1016,7 +1014,6 @@ static void pxa270c5_initfn(Object *obj)
     cpu->reset_sctlr = 0x00000078;
 }
 
-#ifdef CONFIG_TCG
 static const struct TCGCPUOps arm_v7m_tcg_ops = {
     .initialize = arm_translate_init,
     .synchronize_from_tb = arm_cpu_synchronize_from_tb,
@@ -1037,7 +1034,6 @@ static const struct TCGCPUOps arm_v7m_tcg_ops = {
     .debug_check_breakpoint = arm_debug_check_breakpoint,
 #endif /* !CONFIG_USER_ONLY */
 };
-#endif /* CONFIG_TCG */
 
 static void arm_v7m_class_init(ObjectClass *oc, void *data)
 {
@@ -1045,10 +1041,7 @@ static void arm_v7m_class_init(ObjectClass *oc, void *data)
     CPUClass *cc = CPU_CLASS(oc);
 
     acc->info = data;
-#ifdef CONFIG_TCG
     cc->tcg_ops = &arm_v7m_tcg_ops;
-#endif /* CONFIG_TCG */
-
     cc->gdb_core_xml_file = "arm-m-profile.xml";
 }
 
target/arm/tcg/cpu64.c +1 -1
index 89d54ec0f6..d0feb3eb52 100644
@@ -457,7 +457,7 @@ static void aarch64_neoverse_n1_initfn(Object *obj)
 
 /*
  * -cpu max: a CPU with as many features enabled as our emulation supports.
- * The version of '-cpu max' for qemu-system-arm is defined in cpu_tcg.c;
+ * The version of '-cpu max' for qemu-system-arm is defined in cpu32.c;
  * this only needs to handle 64 bits.
  */
 void aarch64_max_tcg_initfn(Object *obj)
target/arm/tcg/meson.build +1 -0
index 128f782816..4d99f6dacb 100644
@@ -18,6 +18,7 @@ gen = [
 arm_ss.add(gen)
 
 arm_ss.add(files(
+  'cpu32.c',
   'translate.c',
   'translate-m-nocp.c',
   'translate-mve.c',
tests/qtest/arm-cpu-features.c +9 -3
index 1cb08138ad..1555b0bab8 100644
@@ -506,9 +506,15 @@ static void test_query_cpu_model_expansion_kvm(const void *data)
         QDict *resp;
         char *error;
 
-        assert_error(qts, "cortex-a15",
-            "We cannot guarantee the CPU type 'cortex-a15' works "
-            "with KVM on this host", NULL);
+        if (qtest_has_accel("tcg")) {
+            assert_error(qts, "cortex-a15",
+                         "We cannot guarantee the CPU type 'cortex-a15' works "
+                         "with KVM on this host", NULL);
+        } else {
+            assert_error(qts, "cortex-a15",
+                         "The CPU type 'cortex-a15' is not a "
+                         "recognized ARM CPU type", NULL);
+        }
 
         assert_has_feature_enabled(qts, "host", "aarch64");
 
Signature · 1 line
2.35.3
#5 lore raw

PATCH v9 4/10 target/arm: gdbstub: Guard M-profile code with CONFIG_TCG

In reply to Fabiano Rosas

This code is only relevant when TCG is present in the build. We're about to enable the build with --disable-tcg, so avoid the error:

libqemu-aarch64-softmmu.fa.p/target_arm_gdbstub.c.o: in function m_sysreg_ptr': ../target/arm/gdbstub.c:356: undefined reference to arm_v7m_get_sp_ptr'

---
 target/arm/gdbstub.c | 4 ++++
 1 file changed, 4 insertions(+)
target/arm/gdbstub.c +4 -0
index 78105b8078..36a1d7dbc5 100644
@@ -323,6 +323,7 @@ static int arm_gen_dynamic_sysreg_xml(CPUState *cs, int base_reg)
     return cpu->dyn_sysreg_xml.num;
 }
 
+#ifdef CONFIG_TCG
 typedef enum {
     M_SYSREG_MSP,
     M_SYSREG_PSP,
@@ -480,6 +481,7 @@ static int arm_gen_dynamic_m_secextreg_xml(CPUState *cs, int orig_base_reg)
     return cpu->dyn_m_secextreg_xml.num;
 }
 #endif
+#endif /* CONFIG_TCG */
 
 const char *arm_gdb_get_dynamic_xml(CPUState *cs, const char *xmlname)
 {
@@ -554,6 +556,7 @@ void arm_cpu_register_gdb_regs_for_features(ARMCPU *cpu)
                              arm_gen_dynamic_sysreg_xml(cs, cs->gdb_num_regs),
                              "system-registers.xml", 0);
 
+#ifdef CONFIG_TCG
     if (arm_feature(env, ARM_FEATURE_M)) {
         gdb_register_coprocessor(cs,
             arm_gdb_get_m_systemreg, arm_gdb_set_m_systemreg,
@@ -568,4 +571,5 @@ void arm_cpu_register_gdb_regs_for_features(ARMCPU *cpu)
         }
 #endif
     }
+#endif
 }
Signature · 1 line
2.35.3
#6 lore raw

PATCH v9 5/10 target/arm: gdbstub: Guard pauth code with CONFIG_TCG

In reply to Fabiano Rosas

We currently don't have the reading of pauth regs implemented for KVM so wrap the pauth registration with CONFIG_TCG.

This avoids the build error when using --disable-tcg:

libqemu-aarch64-softmmu.fa.p/target_arm_gdbstub64.c.o: in function aarch64_gdb_get_pauth_reg': ../target/arm/gdbstub64.c:233: undefined reference to pauth_ptr_mask'

---
 target/arm/gdbstub.c   | 4 +++-
 target/arm/gdbstub64.c | 2 ++
 2 files changed, 5 insertions(+), 1 deletion(-)
target/arm/gdbstub.c +3 -1
index 36a1d7dbc5..3ea50bcaaf 100644
@@ -23,6 +23,7 @@
 #include "gdbstub/helpers.h"
 #include "internals.h"
 #include "cpregs.h"
+#include "sysemu/tcg.h"
 
 typedef struct RegisterSysregXmlParam {
     CPUState *cs;
@@ -522,7 +523,8 @@ void arm_cpu_register_gdb_regs_for_features(ARMCPU *cpu)
                                      aarch64_gdb_set_fpu_reg,
                                      34, "aarch64-fpu.xml", 0);
         }
-        if (isar_feature_aa64_pauth(&cpu->isar)) {
+
+        if (tcg_enabled() && isar_feature_aa64_pauth(&cpu->isar)) {
             gdb_register_coprocessor(cs, aarch64_gdb_get_pauth_reg,
                                      aarch64_gdb_set_pauth_reg,
                                      4, "aarch64-pauth.xml", 0);
target/arm/gdbstub64.c +2 -0
index ec1e07f139..104a43e471 100644
@@ -210,6 +210,7 @@ int aarch64_gdb_set_sve_reg(CPUARMState *env, uint8_t *buf, int reg)
     return 0;
 }
 
+#ifdef CONFIG_TCG
 int aarch64_gdb_get_pauth_reg(CPUARMState *env, GByteArray *buf, int reg)
 {
     switch (reg) {
@@ -243,6 +244,7 @@ int aarch64_gdb_set_pauth_reg(CPUARMState *env, uint8_t *buf, int reg)
     /* All pseudo registers are read-only. */
     return 0;
 }
+#endif
 
 static void output_vector_union_type(GString *s, int reg_width,
                                      const char *name)
Signature · 1 line
2.35.3
Reviewed-by on v9 5/10
#7 lore raw

Re: [PATCH v9 05/10] target/arm: gdbstub: Guard pauth code with CONFIG_TCG

Quoted 9 lines · Fabiano Rosas

We currently don't have the reading of pauth regs implemented for KVM so wrap the pauth registration with CONFIG_TCG.

This avoids the build error when using --disable-tcg:

libqemu-aarch64-softmmu.fa.p/target_arm_gdbstub64.c.o: in function aarch64_gdb_get_pauth_reg': ../target/arm/gdbstub64.c:233: undefined reference to pauth_ptr_mask'

Quoted 5 lines
---
  target/arm/gdbstub.c   | 4 +++-
  target/arm/gdbstub64.c | 2 ++
  2 files changed, 5 insertions(+), 1 deletion(-)

r~

#8 lore raw

PATCH v9 6/10 tests/qtest: Fix tests when no KVM or TCG are present

In reply to Fabiano Rosas

It is possible to have a build with both TCG and KVM disabled due to Xen requiring the i386 and x86_64 binaries to be present in an aarch64 host.

If we build with --disable-tcg on the aarch64 host, we will end-up with a QEMU binary (x86) that does not support TCG nor KVM.

Skip tests that crash or hang in the above scenario. Do not include any test cases if TCG and KVM are missing.

Make sure that calls to qtest_has_accel are placed after g_test_init in similar fashion to commit ae4b01b349 ("tests: Ensure TAP version is printed before other messages") to avoid TAP parsing errors.

---
 tests/qtest/bios-tables-test.c | 11 +++++++++--
 tests/qtest/boot-serial-test.c |  5 +++++
 tests/qtest/migration-test.c   | 10 +++++++++-
 tests/qtest/pxe-test.c         |  8 +++++++-
 tests/qtest/vmgenid-test.c     |  9 +++++++--
 5 files changed, 37 insertions(+), 6 deletions(-)
tests/qtest/bios-tables-test.c +9 -2
index 76d5100911..2315adc70c 100644
@@ -2082,8 +2082,7 @@ static void test_acpi_virt_oem_fields(void)
 int main(int argc, char *argv[])
 {
     const char *arch = qtest_get_arch();
-    const bool has_kvm = qtest_has_accel("kvm");
-    const bool has_tcg = qtest_has_accel("tcg");
+    bool has_kvm, has_tcg;
     char *v_env = getenv("V");
     int ret;
 
@@ -2093,6 +2092,14 @@ int main(int argc, char *argv[])
 
     g_test_init(&argc, &argv, NULL);
 
+    has_kvm = qtest_has_accel("kvm");
+    has_tcg = qtest_has_accel("tcg");
+
+    if (!has_tcg && !has_kvm) {
+        g_test_skip("No accelerator available");
+        return 0;
+    }
+
     if (strcmp(arch, "i386") == 0 || strcmp(arch, "x86_64") == 0) {
         ret = boot_sector_init(disk);
         if (ret) {
tests/qtest/boot-serial-test.c +5 -0
index 3aef3a97a9..28bab3d5e4 100644
@@ -287,6 +287,11 @@ int main(int argc, char *argv[])
 
     g_test_init(&argc, &argv, NULL);
 
+    if (!qtest_has_accel("tcg") && !qtest_has_accel("kvm")) {
+        g_test_skip("No accelerator available");
+        return 0;
+    }
+
     for (i = 0; tests[i].arch != NULL; i++) {
         if (g_str_equal(arch, tests[i].arch) &&
             qtest_has_machine(tests[i].machine)) {
tests/qtest/migration-test.c +9 -1
index d4ab3934ed..5e4d562f0b 100644
@@ -2459,7 +2459,7 @@ static bool kvm_dirty_ring_supported(void)
 
 int main(int argc, char **argv)
 {
-    const bool has_kvm = qtest_has_accel("kvm");
+    bool has_kvm, has_tcg;
     const bool has_uffd = ufd_version_check();
     const char *arch = qtest_get_arch();
     g_autoptr(GError) err = NULL;
@@ -2467,6 +2467,14 @@ int main(int argc, char **argv)
 
     g_test_init(&argc, &argv, NULL);
 
+    has_kvm = qtest_has_accel("kvm");
+    has_tcg = qtest_has_accel("tcg");
+
+    if (!has_tcg && !has_kvm) {
+        g_test_skip("No accelerator available");
+        return 0;
+    }
+
     /*
      * On ppc64, the test only works with kvm-hv, but not with kvm-pr and TCG
      * is touchy due to race conditions on dirty bits (especially on PPC for
tests/qtest/pxe-test.c +7 -1
index 62b6eef464..5813adcf56 100644
@@ -131,11 +131,17 @@ int main(int argc, char *argv[])
     int ret;
     const char *arch = qtest_get_arch();
 
+    g_test_init(&argc, &argv, NULL);
+
+    if (!qtest_has_accel("tcg") && !qtest_has_accel("kvm")) {
+        g_test_skip("No accelerator available");
+        return 0;
+    }
+
     ret = boot_sector_init(disk);
     if(ret)
         return ret;
 
-    g_test_init(&argc, &argv, NULL);
 
     if (strcmp(arch, "i386") == 0 || strcmp(arch, "x86_64") == 0) {
         test_batch(x86_tests, false);
tests/qtest/vmgenid-test.c +7 -2
index efba76e716..a597cde707 100644
@@ -165,13 +165,18 @@ int main(int argc, char **argv)
 {
     int ret;
 
+    g_test_init(&argc, &argv, NULL);
+
+    if (!qtest_has_accel("tcg") && !qtest_has_accel("kvm")) {
+        g_test_skip("No accelerator available");
+        return 0;
+    }
+
     ret = boot_sector_init(disk);
     if (ret) {
         return ret;
     }
 
-    g_test_init(&argc, &argv, NULL);
-
     qtest_add_func("/vmgenid/vmgenid/set-guid",
                    vmgenid_set_guid_test);
     qtest_add_func("/vmgenid/vmgenid/set-guid-auto",
Signature · 1 line
2.35.3
#9 lore raw

PATCH v9 7/10 tests/avocado: Pass parameters to migration test

In reply to Fabiano Rosas

The migration tests are currently broken for an aarch64 host because the tests pass no 'machine' and 'cpu' options on the QEMU command line.

Add a separate class to each architecture so that we can specify 'machine' and 'cpu' options instead of relying on defaults.

Add a skip decorator to keep the current behavior of only running migration tests when the qemu target matches the host architecture.

---
 tests/avocado/migration.py | 83 +++++++++++++++++++++++++++++++++++---
 1 file changed, 78 insertions(+), 5 deletions(-)
tests/avocado/migration.py +78 -5
index 4b25680c50..8b2ec0e3c4 100644
@@ -11,6 +11,8 @@
 
 
 import tempfile
+import os
+
 from avocado_qemu import QemuSystemTest
 from avocado import skipUnless
 
@@ -19,7 +21,7 @@
 from avocado.utils.path import find_command
 
 
-class Migration(QemuSystemTest):
+class MigrationTest(QemuSystemTest):
     """
     :avocado: tags=migration
     """
@@ -62,20 +64,91 @@ def _get_free_port(self):
             self.cancel('Failed to find a free port')
         return port
 
-
-    def test_migration_with_tcp_localhost(self):
+    def migration_with_tcp_localhost(self):
         dest_uri = 'tcp:localhost:%u' % self._get_free_port()
         self.do_migrate(dest_uri)
 
-    def test_migration_with_unix(self):
+    def migration_with_unix(self):
         with tempfile.TemporaryDirectory(prefix='socket_') as socket_path:
             dest_uri = 'unix:%s/qemu-test.sock' % socket_path
             self.do_migrate(dest_uri)
 
     @skipUnless(find_command('nc', default=False), "'nc' command not found")
-    def test_migration_with_exec(self):
+    def migration_with_exec(self):
         """The test works for both netcat-traditional and netcat-openbsd packages."""
         free_port = self._get_free_port()
         dest_uri = 'exec:nc -l localhost %u' % free_port
         src_uri = 'exec:nc localhost %u' % free_port
         self.do_migrate(dest_uri, src_uri)
+
+
+@skipUnless('aarch64' in os.uname()[4], "host != target")
+class Aarch64(MigrationTest):
+    """
+    :avocado: tags=arch:aarch64
+    :avocado: tags=machine:virt
+    :avocado: tags=cpu:max
+    """
+
+    def test_migration_with_tcp_localhost(self):
+        self.migration_with_tcp_localhost()
+
+    def test_migration_with_unix(self):
+        self.migration_with_unix()
+
+    def test_migration_with_exec(self):
+        self.migration_with_exec()
+
+
+@skipUnless('x86_64' in os.uname()[4], "host != target")
+class X86_64(MigrationTest):
+    """
+    :avocado: tags=arch:x86_64
+    :avocado: tags=machine:pc
+    :avocado: tags=cpu:qemu64
+    """
+
+    def test_migration_with_tcp_localhost(self):
+        self.migration_with_tcp_localhost()
+
+    def test_migration_with_unix(self):
+        self.migration_with_unix()
+
+    def test_migration_with_exec(self):
+        self.migration_with_exec()
+
+
+@skipUnless('ppc64le' in os.uname()[4], "host != target")
+class PPC64(MigrationTest):
+    """
+    :avocado: tags=arch:ppc64
+    :avocado: tags=machine:pseries
+    :avocado: tags=cpu:power9_v2.0
+    """
+
+    def test_migration_with_tcp_localhost(self):
+        self.migration_with_tcp_localhost()
+
+    def test_migration_with_unix(self):
+        self.migration_with_unix()
+
+    def test_migration_with_exec(self):
+        self.migration_with_exec()
+
+
+@skipUnless('s390x' in os.uname()[4], "host != target")
+class S390X(MigrationTest):
+    """
+    :avocado: tags=arch:s390x
+    :avocado: tags=machine:s390-ccw-virtio
+    :avocado: tags=cpu:qemu
+    """
+
+    def test_migration_with_tcp_localhost(self):
+        self.migration_with_tcp_localhost()
+
+    def test_migration_with_unix(self):
+        self.migration_with_unix()
+
+    def test_migration_with_exec(self):
+        self.migration_with_exec()
Signature · 1 line
2.35.3
#10 lore raw

PATCH v9 8/10 arm/Kconfig: Always select SEMIHOSTING when TCG is present

In reply to Fabiano Rosas

We are about to enable the build without TCG, so CONFIG_SEMIHOSTING and CONFIG_ARM_COMPATIBLE_SEMIHOSTING cannot be unconditionally set in default.mak anymore. So reflect the change in a Kconfig.

Instead of using semihosting/Kconfig, use a target-specific file, so that the change doesn't affect other architectures which might implement semihosting in a way compatible with KVM.

The selection from ARM_v7M needs to be removed to avoid a cycle during parsing.

---
 configs/devices/arm-softmmu/default.mak | 2 --
 hw/arm/Kconfig                          | 1 -
 target/arm/Kconfig                      | 7 +++++++
 3 files changed, 7 insertions(+), 3 deletions(-)
configs/devices/arm-softmmu/default.mak +0 -2
index 1b49a7830c..cb3e5aea65 100644
@@ -40,6 +40,4 @@ CONFIG_MICROBIT=y
 CONFIG_FSL_IMX25=y
 CONFIG_FSL_IMX7=y
 CONFIG_FSL_IMX6UL=y
-CONFIG_SEMIHOSTING=y
-CONFIG_ARM_COMPATIBLE_SEMIHOSTING=y
 CONFIG_ALLWINNER_H3=y
hw/arm/Kconfig +0 -1
index b5aed4aff5..c0b213f42d 100644
@@ -316,7 +316,6 @@ config ARM_V7M
     # currently v7M must be included in a TCG build due to translate.c
     default y if TCG && (ARM || AARCH64)
     select PTIMER
-    select ARM_COMPATIBLE_SEMIHOSTING
 
 config ALLWINNER_A10
     bool
target/arm/Kconfig +7 -0
index 3f3394a22b..39f05b6420 100644
@@ -4,3 +4,10 @@ config ARM
 config AARCH64
     bool
     select ARM
+
+# This config exists just so we can make SEMIHOSTING default when TCG
+# is selected without also changing it for other architectures.
+config ARM_SEMIHOSTING
+    bool
+    default y if TCG && ARM
+    select ARM_COMPATIBLE_SEMIHOSTING
Signature · 1 line
2.35.3
#11 lore raw

PATCH v9 9/10 arm/Kconfig: Do not build TCG-only boards on a KVM-only build

In reply to Fabiano Rosas

Move all the CONFIG_FOO=y from default.mak into "default y if TCG" statements in Kconfig. That way they won't be selected when CONFIG_TCG=n.

I'm leaving CONFIG_ARM_VIRT in default.mak because it allows us to keep the two default.mak files not empty and keep aarch64-default.mak including arm-default.mak. That way we don't surprise anyone that's used to altering these files.

With this change we can start building with --disable-tcg.

---
 configs/devices/aarch64-softmmu/default.mak |  4 --
 configs/devices/arm-softmmu/default.mak     | 37 ------------------
 hw/arm/Kconfig                              | 42 ++++++++++++++++++++-
 3 files changed, 41 insertions(+), 42 deletions(-)
configs/devices/aarch64-softmmu/default.mak +0 -4
index cf43ac8da1..70e05a197d 100644
@@ -2,7 +2,3 @@
 
 # We support all the 32 bit boards so need all their config
 include ../arm-softmmu/default.mak
-
-CONFIG_XLNX_ZYNQMP_ARM=y
-CONFIG_XLNX_VERSAL=y
-CONFIG_SBSA_REF=y
configs/devices/arm-softmmu/default.mak +0 -37
index cb3e5aea65..647fbce88d 100644
@@ -4,40 +4,3 @@
 # CONFIG_TEST_DEVICES=n
 
 CONFIG_ARM_VIRT=y
-CONFIG_CUBIEBOARD=y
-CONFIG_EXYNOS4=y
-CONFIG_HIGHBANK=y
-CONFIG_INTEGRATOR=y
-CONFIG_FSL_IMX31=y
-CONFIG_MUSICPAL=y
-CONFIG_MUSCA=y
-CONFIG_CHEETAH=y
-CONFIG_SX1=y
-CONFIG_NSERIES=y
-CONFIG_STELLARIS=y
-CONFIG_STM32VLDISCOVERY=y
-CONFIG_REALVIEW=y
-CONFIG_VERSATILE=y
-CONFIG_VEXPRESS=y
-CONFIG_ZYNQ=y
-CONFIG_MAINSTONE=y
-CONFIG_GUMSTIX=y
-CONFIG_SPITZ=y
-CONFIG_TOSA=y
-CONFIG_Z2=y
-CONFIG_NPCM7XX=y
-CONFIG_COLLIE=y
-CONFIG_ASPEED_SOC=y
-CONFIG_NETDUINO2=y
-CONFIG_NETDUINOPLUS2=y
-CONFIG_OLIMEX_STM32_H405=y
-CONFIG_MPS2=y
-CONFIG_RASPI=y
-CONFIG_DIGIC=y
-CONFIG_SABRELITE=y
-CONFIG_EMCRAFT_SF2=y
-CONFIG_MICROBIT=y
-CONFIG_FSL_IMX25=y
-CONFIG_FSL_IMX7=y
-CONFIG_FSL_IMX6UL=y
-CONFIG_ALLWINNER_H3=y
hw/arm/Kconfig +41 -1

275 lines not on this page · raw message

Signature · 1 line
2.35.3
#12 lore raw

PATCH v9 10/10 gitlab-ci: Check building KVM-only aarch64 target

In reply to Fabiano Rosas

From: Philippe Mathieu-Daudé <[email protected]>

Add a manual new job to cross-build the aarch64 target with only the KVM accelerator enabled (in particular, no TCG).

Re-enable running the similar job on the project Aarch64 custom runner.

---
 .gitlab-ci.d/crossbuilds.yml                         | 11 +++++++++++
 .gitlab-ci.d/custom-runners/ubuntu-22.04-aarch64.yml |  4 ----
 2 files changed, 11 insertions(+), 4 deletions(-)
.gitlab-ci.d/crossbuilds.yml +11 -0
index 61b8ac86ee..da787ea9bf 100644
@@ -196,3 +196,14 @@ cross-arm64-xen-only:
     IMAGE: debian-arm64-cross
     ACCEL: xen
     EXTRA_CONFIGURE_OPTS: --disable-tcg --disable-kvm
+
+# Similar job is run by qemu-project's custom runner by default
+cross-arm64-kvm-only:
+  extends: .cross_accel_build_job
+  needs:
+    job: arm64-debian-cross-container
+  variables:
+    QEMU_JOB_OPTIONAL: 1
+    IMAGE: debian-arm64-cross
+    ACCEL: kvm
+    EXTRA_CONFIGURE_OPTS: --disable-tcg --disable-xen --without-default-devices
.gitlab-ci.d/custom-runners/ubuntu-22.04-aarch64.yml +0 -4
index 13e14a0f87..c61be46b82 100644
@@ -115,11 +115,7 @@ ubuntu-22.04-aarch64-notcg:
  - aarch64
  rules:
  - if: '$CI_PROJECT_NAMESPACE == "qemu-project" && $CI_COMMIT_BRANCH =~ /^staging/'
-   when: manual
-   allow_failure: true
  - if: "$AARCH64_RUNNER_AVAILABLE"
-   when: manual
-   allow_failure: true
  script:
  - mkdir build
  - cd build
Signature · 1 line
2.35.3