Skip to contentSkip to conversation

hw/core: Handle cpu_model_from_type() returning NULL value

PATCH In review Philippe Mathieu-Daudé 2 participants 5 messages Jan 11, 2024 – Jan 12, 2024

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 hw/core: Handle cpu_model_from_type() returning NULL value

Per cpu_model_from_type() docstring (added in commit 445946f4dd):

* Returns: CPU model name or NULL if the CPU class doesn't exist

We must check the return value in order to avoid surprises, i.e.:

 $ qemu-system-arm -machine virt -cpu cortex-a9
  qemu-system-arm: Invalid CPU model: cortex-a9
  The valid models are: cortex-a7, cortex-a15, (null), (null), (null), (null), (null), (null), (null), (null), (null), (null), (null), max

Add assertions when the call can not fail (because the CPU type must be registered).

Fixes 5422d2a8fa ("machine: Print CPU model name instead of CPU type")
---
 cpu-target.c          | 1 +
 hw/core/machine.c     | 5 +++++
 target/ppc/cpu_init.c | 1 +
 3 files changed, 7 insertions(+)
cpu-target.c +1 -0
index 5eecd7ea2d..b0f6deb13b 100644
@@ -291,6 +291,7 @@ static void cpu_list_entry(gpointer data, gpointer user_data)
     const char *typename = object_class_get_name(OBJECT_CLASS(data));
     g_autofree char *model = cpu_model_from_type(typename);
 
+    assert(model);
     if (cc->deprecation_note) {
         qemu_printf("  %s (deprecated)\n", model);
     } else {
hw/core/machine.c +5 -0
index fc239101f9..730ec10328 100644
@@ -1422,16 +1422,21 @@ static bool is_cpu_type_supported(const MachineState *machine, Error **errp)
         /* The user specified CPU type isn't valid */
         if (!mc->valid_cpu_types[i]) {
             g_autofree char *requested = cpu_model_from_type(machine->cpu_type);
+            assert(requested);
             error_setg(errp, "Invalid CPU model: %s", requested);
             if (!mc->valid_cpu_types[1]) {
                 g_autofree char *model = cpu_model_from_type(
                                                  mc->valid_cpu_types[0]);
+                assert(model);
                 error_append_hint(errp, "The only valid type is: %s\n", model);
             } else {
                 error_append_hint(errp, "The valid models are: ");
                 for (i = 0; mc->valid_cpu_types[i]; i++) {
                     g_autofree char *model = cpu_model_from_type(
                                                  mc->valid_cpu_types[i]);
+                    if (!model) {
+                        continue;
+                    }
                     error_append_hint(errp, "%s%s",
                                       model,
                                       mc->valid_cpu_types[i + 1] ? ", " : "");
target/ppc/cpu_init.c +1 -0
index 344196a8ce..58f0c1e30e 100644
@@ -7037,6 +7037,7 @@ static void ppc_cpu_list_entry(gpointer data, gpointer user_data)
     }
 
     name = cpu_model_from_type(typename);
+    assert(name);
     qemu_printf("PowerPC %-16s PVR %08x\n", name, pcc->pvr);
     for (i = 0; ppc_cpu_aliases[i].alias != NULL; i++) {
         PowerPCCPUAlias *alias = &ppc_cpu_aliases[i];
Signature · 1 line
2.41.0
#1 lore raw
Quoted 7 lines · Philippe Mathieu-Daudé

Per cpu_model_from_type() docstring (added in commit 445946f4dd):

* Returns: CPU model name or NULL if the CPU class doesn't exist

We must check the return value in order to avoid surprises, i.e.:

  $ qemu-system-arm -machine virt -cpu cortex-a9

Doh I missed one space before the '$' character when pasting.

Quoted 14 lines

qemu-system-arm: Invalid CPU model: cortex-a9
The valid models are: cortex-a7, cortex-a15, (null), (null), (null), (null), (null), (null), (null), (null), (null), (null), (null), max

Add assertions when the call can not fail (because the CPU type must be registered).

Fixes 5422d2a8fa ("machine: Print CPU model name instead of CPU type")
---
  cpu-target.c          | 1 +
  hw/core/machine.c     | 5 +++++
  target/ppc/cpu_init.c | 1 +
  3 files changed, 7 insertions(+)
#2 lore raw
In reply to Philippe Mathieu-Daudé

Hi Phil,

Quoted 47 lines · Philippe Mathieu-Daudé

Per cpu_model_from_type() docstring (added in commit 445946f4dd):

* Returns: CPU model name or NULL if the CPU class doesn't exist

We must check the return value in order to avoid surprises, i.e.:

  $ qemu-system-arm -machine virt -cpu cortex-a9
   qemu-system-arm: Invalid CPU model: cortex-a9
   The valid models are: cortex-a7, cortex-a15, (null), (null), (null), (null), (null), (null), (null), (null), (null), (null), (null), max

Add assertions when the call can not fail (because the CPU type must be registered).

Fixes 5422d2a8fa ("machine: Print CPU model name instead of CPU type")
---
  cpu-target.c          | 1 +
  hw/core/machine.c     | 5 +++++
  target/ppc/cpu_init.c | 1 +
  3 files changed, 7 insertions(+)
cpu-target.c +1 -0
index 5eecd7ea2d..b0f6deb13b 100644
@@ -291,6 +291,7 @@ static void cpu_list_entry(gpointer data, gpointer user_data)
      const char *typename = object_class_get_name(OBJECT_CLASS(data));
      g_autofree char *model = cpu_model_from_type(typename);
 
+    assert(model);
      if (cc->deprecation_note) {
          qemu_printf("  %s (deprecated)\n", model);
      } else {
hw/core/machine.c +1 -0
index fc239101f9..730ec10328 100644
@@ -1422,16 +1422,21 @@ static bool is_cpu_type_supported(const MachineState *machine, Error **errp)
          /* The user specified CPU type isn't valid */
          if (!mc->valid_cpu_types[i]) {
              g_autofree char *requested = cpu_model_from_type(machine->cpu_type);
+            assert(requested);
              error_setg(errp, "Invalid CPU model: %s", requested);
              if (!mc->valid_cpu_types[1]) {
                  g_autofree char *model = cpu_model_from_type(
                                                   mc->valid_cpu_types[0]);

hw/core/machine.c static bool is_cpu_type_supported(const MachineState *machine, Error **errp) · line 1439

Quoted 10 lines
hw/core/machine.c +4 -0
+                assert(model);
                  error_append_hint(errp, "The only valid type is: %s\n", model);
              } else {
                  error_append_hint(errp, "The valid models are: ");
                  for (i = 0; mc->valid_cpu_types[i]; i++) {
                      g_autofree char *model = cpu_model_from_type(
                                                   mc->valid_cpu_types[i]);
+                    if (!model) {
+                        continue;
+                    }

Shall we assert(model) for this case, to be consistent with other cases? :)

hw/core/machine.c static bool is_cpu_type_supported(const MachineState *machine, Error **errp) · line 1442

Quoted 3 lines
hw/core/machine.c +0 -0
                      error_append_hint(errp, "%s%s",
                                        model,
                                        mc->valid_cpu_types[i + 1] ? ", " : "");

Otherwise, the separator here need to be adjusted because it's uncertain that mc->valid_cpu_types[i+1] ... mc->valid_cpu_types[END] are valid.

target/ppc/cpu_init.c static void ppc_cpu_list_entry(gpointer data, gpointer user_data) · line 7043

Quoted 12 lines
target/ppc/cpu_init.c +1 -0
index 344196a8ce..58f0c1e30e 100644
@@ -7037,6 +7037,7 @@ static void ppc_cpu_list_entry(gpointer data, gpointer user_data)
      }
 
      name = cpu_model_from_type(typename);
+    assert(name);
      qemu_printf("PowerPC %-16s PVR %08x\n", name, pcc->pvr);
      for (i = 0; ppc_cpu_aliases[i].alias != NULL; i++) {
          PowerPCCPUAlias *alias = &ppc_cpu_aliases[i];

Thanks,
Gavin

#3 lore raw

Hi Gavin,

Quoted 59 lines · Gavin Shan

Hi Phil,

Quoted 56 lines · Philippe Mathieu-Daudé

Per cpu_model_from_type() docstring (added in commit 445946f4dd):

   * Returns: CPU model name or NULL if the CPU class doesn't exist

We must check the return value in order to avoid surprises, i.e.:

  $ qemu-system-arm -machine virt -cpu cortex-a9
   qemu-system-arm: Invalid CPU model: cortex-a9
   The valid models are: cortex-a7, cortex-a15, (null), (null), (null), (null), (null), (null), (null), (null), (null), (null), (null), max

Add assertions when the call can not fail (because the CPU type must be registered).

Fixes 5422d2a8fa ("machine: Print CPU model name instead of CPU type")

---
  cpu-target.c          | 1 +
  hw/core/machine.c     | 5 +++++
  target/ppc/cpu_init.c | 1 +

  3 files changed, 7 insertions(+)
cpu-target.c +1 -0
index 5eecd7ea2d..b0f6deb13b 100644
@@ -291,6 +291,7 @@ static void cpu_list_entry(gpointer data, gpointer 
 user_data)
       const char *typename = object_class_get_name(OBJECT_CLASS(data));
       g_autofree char *model = cpu_model_from_type(typename);
+    assert(model);
       if (cc->deprecation_note) {
           qemu_printf("  %s (deprecated)\n", model);
       } else {
hw/core/machine.c +1 -0
index fc239101f9..730ec10328 100644
@@ -1422,16 +1422,21 @@ static bool is_cpu_type_supported(const 
 MachineState *machine, Error **errp)
           /* The user specified CPU type isn't valid */
           if (!mc->valid_cpu_types[i]) {
               g_autofree char *requested = cpu_model_from_type(machine->cpu_type);
+            assert(requested);
               error_setg(errp, "Invalid CPU model: %s", requested);
               if (!mc->valid_cpu_types[1]) {

                  g_autofree char *model = cpu_model_from_type(

mc->valid_cpu_types[0]);
+                assert(model);
                  error_append_hint(errp, "The only valid type is: %s\n", model); } else {
                  error_append_hint(errp, "The valid models are: "); for (i = 0; mc->valid_cpu_types[i]; i++) { g_autofree char *model = cpu_model_from_type(

Quoted 6 lines
Quoted 4 lines

mc->valid_cpu_types[i]);
+                    if (!model) {
+                        continue;
+                    }

Shall we assert(model) for this case, to be consistent with other cases? :)

No, this is the "(null)" cases displayed in the example.

IOW, mc->valid_cpu_types[] contains a CPU type which isn't registered, so we just skip it.

Quoted 7 lines
Quoted 4 lines

                      error_append_hint(errp, "%s%s",
                                        model, mc->valid_cpu_types[i + 1] ? ", " : "");

Otherwise, the separator here need to be adjusted because it's uncertain that mc->valid_cpu_types[i+1] ... mc->valid_cpu_types[END] are valid.

Here we know mc->valid_cpu_types[i] is *not* NULL, but mc->valid_cpu_types[i + 1] might be (signaling the end of the array).

This seems correct to me, but I might be missing something.

Quoted 17 lines

target/ppc/cpu_init.c static void ppc_cpu_list_entry(gpointer data, · line 7043

Quoted 13 lines
target/ppc/cpu_init.c +1 -0
index 344196a8ce..58f0c1e30e 100644
@@ -7037,6 +7037,7 @@ static void ppc_cpu_list_entry(gpointer data, 
 gpointer user_data)
       }
       name = cpu_model_from_type(typename);
+    assert(name);
       qemu_printf("PowerPC %-16s PVR %08x\n", name, pcc->pvr);
       for (i = 0; ppc_cpu_aliases[i].alias != NULL; i++) {
           PowerPCCPUAlias *alias = &ppc_cpu_aliases[i];

Thanks,
Gavin

#4 lore raw

Hi Phil,

Quoted 61 lines · Philippe Mathieu-Daudé
Quoted 51 lines · Gavin Shan
Quoted 50 lines · Philippe Mathieu-Daudé

Per cpu_model_from_type() docstring (added in commit 445946f4dd):

   * Returns: CPU model name or NULL if the CPU class doesn't exist

We must check the return value in order to avoid surprises, i.e.:

  $ qemu-system-arm -machine virt -cpu cortex-a9
   qemu-system-arm: Invalid CPU model: cortex-a9
   The valid models are: cortex-a7, cortex-a15, (null), (null), (null), (null), (null), (null), (null), (null), (null), (null), (null), max

Add assertions when the call can not fail (because the CPU type must be registered).

Fixes 5422d2a8fa ("machine: Print CPU model name instead of CPU type")

---
  cpu-target.c          | 1 +
  hw/core/machine.c     | 5 +++++
  target/ppc/cpu_init.c | 1 +

  3 files changed, 7 insertions(+)
cpu-target.c +1 -0
index 5eecd7ea2d..b0f6deb13b 100644
@@ -291,6 +291,7 @@ static void cpu_list_entry(gpointer data, gpointer user_data)
       const char *typename = object_class_get_name(OBJECT_CLASS(data));
       g_autofree char *model = cpu_model_from_type(typename);
+    assert(model);
       if (cc->deprecation_note) {
           qemu_printf("  %s (deprecated)\n", model);
       } else {
hw/core/machine.c +1 -0
index fc239101f9..730ec10328 100644
@@ -1422,16 +1422,21 @@ static bool is_cpu_type_supported(const MachineState *machine, Error **errp)
           /* The user specified CPU type isn't valid */
           if (!mc->valid_cpu_types[i]) {
               g_autofree char *requested = cpu_model_from_type(machine->cpu_type);
+            assert(requested);
               error_setg(errp, "Invalid CPU model: %s", requested);
               if (!mc->valid_cpu_types[1]) {

                  g_autofree char *model = cpu_model_from_type( mc->valid_cpu_types[0]);
+                assert(model);
                  error_append_hint(errp, "The only valid type is: %s\n", model); } else {
                  error_append_hint(errp, "The valid models are: ");

Quoted 8 lines
Quoted 6 lines

                  for (i = 0; mc->valid_cpu_types[i]; i++) { g_autofree char *model = cpu_model_from_type( mc->valid_cpu_types[i]);
+                    if (!model) {
+                        continue;
+                    }

Shall we assert(model) for this case, to be consistent with other cases? :)

Quoted 5 lines

No, this is the "(null)" cases displayed in the example.

IOW, mc->valid_cpu_types[] contains a CPU type which isn't registered, so we just skip it.

I thought this should be fixed by correcting mc->valid_cpu_types[] in hw/arm/virt.c. It means the consistent mc->valid_cpu_types[] needs to be provided by the specific board. Otherwise, the logic is incorrect from the code level at least. For example, "cortex-a9" isn't available to qemu-system-arm but it has been wrongly declared as supported in hw/arm/virt.c

I've posted one patch against it:

https://lists.nongnu.org/archive/html/qemu-arm/2024-01/msg00531.html

Quoted 5 lines
Quoted 5 lines
Quoted 4 lines

                      error_append_hint(errp, "%s%s",
                                        model, mc->valid_cpu_types[i + 1] ? ", " : "");

Quoted 9 lines
Quoted 2 lines

Otherwise, the separator here need to be adjusted because it's uncertain that mc->valid_cpu_types[i+1] ... mc->valid_cpu_types[END] are valid.

Here we know mc->valid_cpu_types[i] is *not* NULL, but mc->valid_cpu_types[i + 1] might be (signaling the end of the array).

This seems correct to me, but I might be missing something.

When the class for mc->valid_cpu_types[i + 1] isn't registered, we will skip the entry. it's possible that the class of mc->valid_cpu_types[i + 2] isn't registered either. mc->valid_cpu_types[i + 3] to mc->valid_cpu_types[END - 1] have the similar situations.

In order to correct the separator, we need to invalidate the return value
from cpu_model_from_type(mc->valid_cpu_types[i + 1]) ... cpu_model_from_type(mc->valid_cpu_types[END - 1]). Too much complex for that and it's another reason why I suggested assert(model) as above

Quoted 13 lines
Quoted 13 lines

target/ppc/cpu_init.c static void ppc_cpu_list_entry(gpointer data, gpointer user_data) · line 7042

Quoted 12 lines
target/ppc/cpu_init.c +1 -0
index 344196a8ce..58f0c1e30e 100644
@@ -7037,6 +7037,7 @@ static void ppc_cpu_list_entry(gpointer data, gpointer user_data)
       }
       name = cpu_model_from_type(typename);
+    assert(name);
       qemu_printf("PowerPC %-16s PVR %08x\n", name, pcc->pvr);
       for (i = 0; ppc_cpu_aliases[i].alias != NULL; i++) {
           PowerPCCPUAlias *alias = &ppc_cpu_aliases[i];

Thanks,
Gavin