Skip to contentSkip to conversation

qga: clean up command source locations and conditionals

PATCH Superseded Daniel P. Berrangé 6 participants 70 messages Jun 13, 2024 – Jul 12, 2024 Arm/Amlogic Meson SoC support

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

v2 cover qga: clean up command source locations and conditionals

This series is a side effect of other work I started, to attempt to make the QGA safe to use in confidential VMs by automatically restricting the permitted commands. Since this cleanup stands on its own, I'm sending it now.

The QGA codebase has a very complicated maze of #ifdefs to create stubs for the various commands that cannot be implemented on certain platforms. It then has further logic to dynamically disable the stub commands at runtime, except this is not consistently applied, so some commands remain enabled despite being merely stubs.

The resulting code is hard to follow, when trying to understand exactly what commands are available under what circumstances, and when changing impls it is easy to get the #ifdefs wrong, resulting in stubs getting missed on platforms without a real impl. In some cases, we have multiple stubs for the same command, due to the maze of #ifdefs.

The QAPI schema language has support for many years for expressing conditions against commands when declaring them. This results in the QAPI code generator omitting their implementation entirely at build time. This has mutliple benefits

* The unsupported commands are guaranteed to not exist at runtime
* No stubs need ever be defined in the code
* The generated QAPI reference manual documents the build conditions

This series is broadly split into three parts

* Moving tonnes of Linux only commands out of commands-posix.c into commands-linux.c to remove many #ifdefs.
* Adding 'if' conditions in the QAPI schema to reflect the build conditions, removing many more #ifdefs
* Sanitizing the logic for disabling/enabling commands at runtime to guarantee consistency

Changed in v2:

- Make FSFreeze error reporting distinguish inability to enable VSS from user config choice

- Fully remove ga_command_init_blockedrpcs() methods. No more special case disabling of commands. Either they're disabled at build time, or disabled by user config, or by well defined rule ie not permitted during FS freeze.

- Apply rules later in startup to avoid crash from NULL config pointer

- Document changed error messages in commit messages

- Add -c / --config command line parameter

- Fix mistaken enabling of fsfreeze hooks on win32

- Remove pointless 'blockrpcs_key' variable

- Allow concurrent setting of allow and block lists for RPC commands

Daniel P. Berrangé (22):
qga: drop blocking of guest-get-memory-block-size command
qga: move linux vcpu command impls to commands-linux.c
qga: move linux suspend command impls to commands-linux.c
qga: move linux fs/disk command impls to commands-linux.c
qga: move linux disk/cpu stats command impls to commands-linux.c
qga: move linux memory block command impls to commands-linux.c
qga: move CONFIG_FSFREEZE/TRIM to be meson defined options
qga: conditionalize schema for commands unsupported on Windows
qga: conditionalize schema for commands unsupported on non-Linux POSIX
qga: conditionalize schema for commands requiring getifaddrs
qga: conditionalize schema for commands requiring linux/win32
qga: conditionalize schema for commands only supported on Windows
qga: conditionalize schema for commands requiring fsfreeze
qga: conditionalize schema for commands requiring fstrim
qga: conditionalize schema for commands requiring libudev
qga: conditionalize schema for commands requiring utmpx
qga: conditionalize schema for commands not supported on other UNIX
qga: don't disable fsfreeze commands if vss_init fails
qga: move declare of QGAConfig struct to top of file
qga: remove pointless 'blockrpcs_key' variable
qga: allow configuration file path via the cli
qga: centralize logic for disabling/enabling commands

 docs/interop/qemu-ga.rst |   19 +
 meson.build              |   16 +
 qga/commands-bsd.c       |   24 -
 qga/commands-common.h    |    9 -
 qga/commands-linux.c     | 1805 +++++++++++++++++++++++++++++
 qga/commands-posix.c     | 2373 +++-----------------------------------
 qga/commands-win32.c     |   78 +-
 qga/main.c               |  216 ++--
 qga/qapi-schema.json     |  153 ++-
 9 files changed, 2234 insertions(+), 2459 deletions(-)
Signature · 1 line
2.45.1
#1 lore raw

PATCH v2 1/22 qga: drop blocking of guest-get-memory-block-size command

This command has never existed in tree, since it was renamed to guest-get-memory-block-info before being merged.

---
 qga/commands-posix.c | 2 +-
 qga/commands-win32.c | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)
qga/commands-posix.c +1 -1
index 7f05996495..76af98ba32 100644
@@ -3099,7 +3099,7 @@ GList *ga_command_init_blockedrpcs(GList *blockedrpcs)
             "guest-suspend-disk", "guest-suspend-ram",
             "guest-suspend-hybrid", "guest-get-vcpus", "guest-set-vcpus",
             "guest-get-memory-blocks", "guest-set-memory-blocks",
-            "guest-get-memory-block-size", "guest-get-memory-block-info",
+            "guest-get-memory-block-info",
             NULL};
         char **p = (char **)list;
 
qga/commands-win32.c +1 -1
index 0d1b836e87..9fe670d5b4 100644
@@ -1995,7 +1995,7 @@ GList *ga_command_init_blockedrpcs(GList *blockedrpcs)
         "guest-suspend-hybrid",
         "guest-set-vcpus",
         "guest-get-memory-blocks", "guest-set-memory-blocks",
-        "guest-get-memory-block-size", "guest-get-memory-block-info",
+        "guest-get-memory-block-info",
         NULL};
     char **p = (char **)list_unsupported;
 
Signature · 1 line
2.45.1
Reviewed-by on v2 1/22
#2 lore raw

Re: [PATCH v2 01/22] qga: drop blocking of guest-get-memory-block-size command

Quoted 42 lines · Daniel P. Berrangé

This command has never existed in tree, since it was renamed to guest-get-memory-block-info before being merged.

---
 qga/commands-posix.c | 2 +-
 qga/commands-win32.c | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)
qga/commands-posix.c +1 -1
index 7f05996495..76af98ba32 100644
@@ -3099,7 +3099,7 @@ GList *ga_command_init_blockedrpcs(GList
 *blockedrpcs)
             "guest-suspend-disk", "guest-suspend-ram",
             "guest-suspend-hybrid", "guest-get-vcpus", "guest-set-vcpus",
             "guest-get-memory-blocks", "guest-set-memory-blocks",
-            "guest-get-memory-block-size", "guest-get-memory-block-info",
+            "guest-get-memory-block-info",
             NULL};
         char **p = (char **)list;
qga/commands-win32.c +1 -2
index 0d1b836e87..9fe670d5b4 100644
@@ -1995,7 +1995,7 @@ GList *ga_command_init_blockedrpcs(GList
 *blockedrpcs)
         "guest-suspend-hybrid",
         "guest-set-vcpus",
         "guest-get-memory-blocks", "guest-set-memory-blocks",
-        "guest-get-memory-block-size", "guest-get-memory-block-info",
+        "guest-get-memory-block-info",
         NULL};
     char **p = (char **)list_unsupported;
 
--

2.45.1

#3 lore raw

PATCH v2 2/22 qga: move linux vcpu command impls to commands-linux.c

In reply to Daniel P. Berrangé

The qmp_guest_set_vcpus and qmp_guest_get_vcpus command impls in commands-posix.c are surrounded by '#ifdef __linux__' so should instead live in commands-linux.c

---
 qga/commands-linux.c | 141 +++++++++++++++++++++++++++++++++++++++++++
 qga/commands-posix.c | 139 ------------------------------------------
 2 files changed, 141 insertions(+), 139 deletions(-)
qga/commands-linux.c +141 -0
index 214e408fcd..78580ac39d 100644
@@ -13,6 +13,7 @@
 
 #include "qemu/osdep.h"
 #include "qapi/error.h"
+#include "qga-qapi-commands.h"
 #include "commands-common.h"
 #include "cutils.h"
 #include <mntent.h>
@@ -284,3 +285,143 @@ int qmp_guest_fsfreeze_do_thaw(Error **errp)
     return i;
 }
 #endif /* CONFIG_FSFREEZE */
+
+/* Transfer online/offline status between @vcpu and the guest system.
+ *
+ * On input either @errp or *@errp must be NULL.
+ *
+ * In system-to-@vcpu direction, the following @vcpu fields are accessed:
+ * - R: vcpu->logical_id
+ * - W: vcpu->online
+ * - W: vcpu->can_offline
+ *
+ * In @vcpu-to-system direction, the following @vcpu fields are accessed:
+ * - R: vcpu->logical_id
+ * - R: vcpu->online
+ *
+ * Written members remain unmodified on error.
+ */
+static void transfer_vcpu(GuestLogicalProcessor *vcpu, bool sys2vcpu,
+                          char *dirpath, Error **errp)
+{
+    int fd;
+    int res;
+    int dirfd;
+    static const char fn[] = "online";
+
+    dirfd = open(dirpath, O_RDONLY | O_DIRECTORY);
+    if (dirfd == -1) {
+        error_setg_errno(errp, errno, "open(\"%s\")", dirpath);
+        return;
+    }
+
+    fd = openat(dirfd, fn, sys2vcpu ? O_RDONLY : O_RDWR);
+    if (fd == -1) {
+        if (errno != ENOENT) {
+            error_setg_errno(errp, errno, "open(\"%s/%s\")", dirpath, fn);
+        } else if (sys2vcpu) {
+            vcpu->online = true;
+            vcpu->can_offline = false;
+        } else if (!vcpu->online) {
+            error_setg(errp, "logical processor #%" PRId64 " can't be "
+                       "offlined", vcpu->logical_id);
+        } /* otherwise pretend successful re-onlining */
+    } else {
+        unsigned char status;
+
+        res = pread(fd, &status, 1, 0);
+        if (res == -1) {
+            error_setg_errno(errp, errno, "pread(\"%s/%s\")", dirpath, fn);
+        } else if (res == 0) {
+            error_setg(errp, "pread(\"%s/%s\"): unexpected EOF", dirpath,
+                       fn);
+        } else if (sys2vcpu) {
+            vcpu->online = (status != '0');
+            vcpu->can_offline = true;
+        } else if (vcpu->online != (status != '0')) {
+            status = '0' + vcpu->online;
+            if (pwrite(fd, &status, 1, 0) == -1) {
+                error_setg_errno(errp, errno, "pwrite(\"%s/%s\")", dirpath,
+                                 fn);
+            }
+        } /* otherwise pretend successful re-(on|off)-lining */
+
+        res = close(fd);
+        g_assert(res == 0);
+    }
+
+    res = close(dirfd);
+    g_assert(res == 0);
+}
+
+GuestLogicalProcessorList *qmp_guest_get_vcpus(Error **errp)
+{
+    GuestLogicalProcessorList *head, **tail;
+    const char *cpu_dir = "/sys/devices/system/cpu";
+    const gchar *line;
+    g_autoptr(GDir) cpu_gdir = NULL;
+    Error *local_err = NULL;
+
+    head = NULL;
+    tail = &head;
+    cpu_gdir = g_dir_open(cpu_dir, 0, NULL);
+
+    if (cpu_gdir == NULL) {
+        error_setg_errno(errp, errno, "failed to list entries: %s", cpu_dir);
+        return NULL;
+    }
+
+    while (local_err == NULL && (line = g_dir_read_name(cpu_gdir)) != NULL) {
+        GuestLogicalProcessor *vcpu;
+        int64_t id;
+        if (sscanf(line, "cpu%" PRId64, &id)) {
+            g_autofree char *path = g_strdup_printf("/sys/devices/system/cpu/"
+                                                    "cpu%" PRId64 "/", id);
+            vcpu = g_malloc0(sizeof *vcpu);
+            vcpu->logical_id = id;
+            vcpu->has_can_offline = true; /* lolspeak ftw */
+            transfer_vcpu(vcpu, true, path, &local_err);
+            QAPI_LIST_APPEND(tail, vcpu);
+        }
+    }
+
+    if (local_err == NULL) {
+        /* there's no guest with zero VCPUs */
+        g_assert(head != NULL);
+        return head;
+    }
+
+    qapi_free_GuestLogicalProcessorList(head);
+    error_propagate(errp, local_err);
+    return NULL;
+}
+
+int64_t qmp_guest_set_vcpus(GuestLogicalProcessorList *vcpus, Error **errp)
+{
+    int64_t processed;
+    Error *local_err = NULL;
+
+    processed = 0;
+    while (vcpus != NULL) {
+        char *path = g_strdup_printf("/sys/devices/system/cpu/cpu%" PRId64 "/",
+                                     vcpus->value->logical_id);
+
+        transfer_vcpu(vcpus->value, false, path, &local_err);
+        g_free(path);
+        if (local_err != NULL) {
+            break;
+        }
+        ++processed;
+        vcpus = vcpus->next;
+    }
+
+    if (local_err != NULL) {
+        if (processed == 0) {
+            error_propagate(errp, local_err);
+        } else {
+            error_free(local_err);
+        }
+    }
+
+    return processed;
+}
qga/commands-posix.c +0 -139

145 lines not on this page · raw message

Signature · 1 line
2.45.1
Reviewed-by on v2 2/22
#4 lore raw

Re: [PATCH v2 02/22] qga: move linux vcpu command impls to commands-linux.c

Quoted 10 lines · Daniel P. Berrangé

The qmp_guest_set_vcpus and qmp_guest_get_vcpus command impls in commands-posix.c are surrounded by '#ifdef __linux__' so should instead live in commands-linux.c

---
  qga/commands-linux.c | 141 +++++++++++++++++++++++++++++++++++++++++++
  qga/commands-posix.c | 139 ------------------------------------------
  2 files changed, 141 insertions(+), 139 deletions(-)
Reviewed-by on v2 2/22
#5 lore raw

Re: [PATCH v2 02/22] qga: move linux vcpu command impls to commands-linux.c

In reply to Daniel P. Berrangé
Quoted 337 lines · Daniel P. Berrangé

The qmp_guest_set_vcpus and qmp_guest_get_vcpus command impls in commands-posix.c are surrounded by '#ifdef __linux__' so should instead live in commands-linux.c

---
 qga/commands-linux.c | 141 +++++++++++++++++++++++++++++++++++++++++++
 qga/commands-posix.c | 139 ------------------------------------------
 2 files changed, 141 insertions(+), 139 deletions(-)
qga/commands-linux.c +48 -0
index 214e408fcd..78580ac39d 100644
@@ -13,6 +13,7 @@
 
 #include "qemu/osdep.h"
 #include "qapi/error.h"
+#include "qga-qapi-commands.h"
 #include "commands-common.h"
 #include "cutils.h"
 #include <mntent.h>
@@ -284,3 +285,143 @@ int qmp_guest_fsfreeze_do_thaw(Error **errp)
     return i;
 }
 #endif /* CONFIG_FSFREEZE */
+
+/* Transfer online/offline status between @vcpu and the guest system.
+ *
+ * On input either @errp or *@errp must be NULL.
+ *
+ * In system-to-@vcpu direction, the following @vcpu fields are accessed:
+ * - R: vcpu->logical_id
+ * - W: vcpu->online
+ * - W: vcpu->can_offline
+ *
+ * In @vcpu-to-system direction, the following @vcpu fields are accessed:
+ * - R: vcpu->logical_id
+ * - R: vcpu->online
+ *
+ * Written members remain unmodified on error.
+ */
+static void transfer_vcpu(GuestLogicalProcessor *vcpu, bool sys2vcpu,
+                          char *dirpath, Error **errp)
+{
+    int fd;
+    int res;
+    int dirfd;
+    static const char fn[] = "online";
+
+    dirfd = open(dirpath, O_RDONLY | O_DIRECTORY);
+    if (dirfd == -1) {
+        error_setg_errno(errp, errno, "open(\"%s\")", dirpath);
+        return;
+    }
+
+    fd = openat(dirfd, fn, sys2vcpu ? O_RDONLY : O_RDWR);
+    if (fd == -1) {
+        if (errno != ENOENT) {
+            error_setg_errno(errp, errno, "open(\"%s/%s\")", dirpath, fn);
+        } else if (sys2vcpu) {
+            vcpu->online = true;
+            vcpu->can_offline = false;
+        } else if (!vcpu->online) {
+            error_setg(errp, "logical processor #%" PRId64 " can't be "
+                       "offlined", vcpu->logical_id);
+        } /* otherwise pretend successful re-onlining */
+    } else {
+        unsigned char status;
+
+        res = pread(fd, &status, 1, 0);
+        if (res == -1) {
+            error_setg_errno(errp, errno, "pread(\"%s/%s\")", dirpath,

fn);

qga/commands-linux.c +10 -0
+        } else if (res == 0) {
+            error_setg(errp, "pread(\"%s/%s\"): unexpected EOF", dirpath,
+                       fn);
+        } else if (sys2vcpu) {
+            vcpu->online = (status != '0');
+            vcpu->can_offline = true;
+        } else if (vcpu->online != (status != '0')) {
+            status = '0' + vcpu->online;
+            if (pwrite(fd, &status, 1, 0) == -1) {
+                error_setg_errno(errp, errno, "pwrite(\"%s/%s\")",

dirpath,

qga/commands-linux.c +26 -0
+                                 fn);
+            }
+        } /* otherwise pretend successful re-(on|off)-lining */
+
+        res = close(fd);
+        g_assert(res == 0);
+    }
+
+    res = close(dirfd);
+    g_assert(res == 0);
+}
+
+GuestLogicalProcessorList *qmp_guest_get_vcpus(Error **errp)
+{
+    GuestLogicalProcessorList *head, **tail;
+    const char *cpu_dir = "/sys/devices/system/cpu";
+    const gchar *line;
+    g_autoptr(GDir) cpu_gdir = NULL;
+    Error *local_err = NULL;
+
+    head = NULL;
+    tail = &head;
+    cpu_gdir = g_dir_open(cpu_dir, 0, NULL);
+
+    if (cpu_gdir == NULL) {
+        error_setg_errno(errp, errno, "failed to list entries: %s",

cpu_dir);

qga/commands-linux.c +4 -0
+        return NULL;
+    }
+
+    while (local_err == NULL && (line = g_dir_read_name(cpu_gdir)) !=

NULL) {

qga/commands-linux.c +4 -0
+        GuestLogicalProcessor *vcpu;
+        int64_t id;
+        if (sscanf(line, "cpu%" PRId64, &id)) {
+            g_autofree char *path =

g_strdup_printf("/sys/devices/system/cpu/"
+ "cpu%" PRId64 "/",
id);

qga/commands-linux.c +20 -0
+            vcpu = g_malloc0(sizeof *vcpu);
+            vcpu->logical_id = id;
+            vcpu->has_can_offline = true; /* lolspeak ftw */
+            transfer_vcpu(vcpu, true, path, &local_err);
+            QAPI_LIST_APPEND(tail, vcpu);
+        }
+    }
+
+    if (local_err == NULL) {
+        /* there's no guest with zero VCPUs */
+        g_assert(head != NULL);
+        return head;
+    }
+
+    qapi_free_GuestLogicalProcessorList(head);
+    error_propagate(errp, local_err);
+    return NULL;
+}
+
+int64_t qmp_guest_set_vcpus(GuestLogicalProcessorList *vcpus, Error

**errp)

qga/commands-linux.c +7 -0
+{
+    int64_t processed;
+    Error *local_err = NULL;
+
+    processed = 0;
+    while (vcpus != NULL) {
+        char *path = g_strdup_printf("/sys/devices/system/cpu/cpu%"

PRId64 "/",

qga/commands-linux.c +21 -0
+                                     vcpus->value->logical_id);
+
+        transfer_vcpu(vcpus->value, false, path, &local_err);
+        g_free(path);
+        if (local_err != NULL) {
+            break;
+        }
+        ++processed;
+        vcpus = vcpus->next;
+    }
+
+    if (local_err != NULL) {
+        if (processed == 0) {
+            error_propagate(errp, local_err);
+        } else {
+            error_free(local_err);
+        }
+    }
+
+    return processed;
+}
qga/commands-posix.c +0 -86
index 76af98ba32..a8ef41f175 100644
@@ -2003,145 +2003,6 @@ void qmp_guest_suspend_hybrid(Error **errp)
     guest_suspend(SUSPEND_MODE_HYBRID, errp);
 }
 
-/* Transfer online/offline status between @vcpu and the guest system.
- *
- * On input either @errp or *@errp must be NULL.
- *
- * In system-to-@vcpu direction, the following @vcpu fields are accessed:
- * - R: vcpu->logical_id
- * - W: vcpu->online
- * - W: vcpu->can_offline
- *
- * In @vcpu-to-system direction, the following @vcpu fields are accessed:
- * - R: vcpu->logical_id
- * - R: vcpu->online
- *
- * Written members remain unmodified on error.
- */
-static void transfer_vcpu(GuestLogicalProcessor *vcpu, bool sys2vcpu,
-                          char *dirpath, Error **errp)
-{
-    int fd;
-    int res;
-    int dirfd;
-    static const char fn[] = "online";
-
-    dirfd = open(dirpath, O_RDONLY | O_DIRECTORY);
-    if (dirfd == -1) {
-        error_setg_errno(errp, errno, "open(\"%s\")", dirpath);
-        return;
-    }
-
-    fd = openat(dirfd, fn, sys2vcpu ? O_RDONLY : O_RDWR);
-    if (fd == -1) {
-        if (errno != ENOENT) {
-            error_setg_errno(errp, errno, "open(\"%s/%s\")", dirpath, fn);
-        } else if (sys2vcpu) {
-            vcpu->online = true;
-            vcpu->can_offline = false;
-        } else if (!vcpu->online) {
-            error_setg(errp, "logical processor #%" PRId64 " can't be "
-                       "offlined", vcpu->logical_id);
-        } /* otherwise pretend successful re-onlining */
-    } else {
-        unsigned char status;
-
-        res = pread(fd, &status, 1, 0);
-        if (res == -1) {
-            error_setg_errno(errp, errno, "pread(\"%s/%s\")", dirpath,
 fn);
-        } else if (res == 0) {
-            error_setg(errp, "pread(\"%s/%s\"): unexpected EOF", dirpath,
-                       fn);
-        } else if (sys2vcpu) {
-            vcpu->online = (status != '0');
-            vcpu->can_offline = true;
-        } else if (vcpu->online != (status != '0')) {
-            status = '0' + vcpu->online;
-            if (pwrite(fd, &status, 1, 0) == -1) {
-                error_setg_errno(errp, errno, "pwrite(\"%s/%s\")",
 dirpath,
-                                 fn);
-            }
-        } /* otherwise pretend successful re-(on|off)-lining */
-
-        res = close(fd);
-        g_assert(res == 0);
-    }
-
-    res = close(dirfd);
-    g_assert(res == 0);
-}
-
-GuestLogicalProcessorList *qmp_guest_get_vcpus(Error **errp)
-{
-    GuestLogicalProcessorList *head, **tail;
-    const char *cpu_dir = "/sys/devices/system/cpu";
-    const gchar *line;
-    g_autoptr(GDir) cpu_gdir = NULL;
-    Error *local_err = NULL;
-
-    head = NULL;
-    tail = &head;
-    cpu_gdir = g_dir_open(cpu_dir, 0, NULL);
-
-    if (cpu_gdir == NULL) {
-        error_setg_errno(errp, errno, "failed to list entries: %s",
 cpu_dir);
-        return NULL;
-    }
-
-    while (local_err == NULL && (line = g_dir_read_name(cpu_gdir)) !=

NULL) {

qga/commands-posix.c +0 -4
-        GuestLogicalProcessor *vcpu;
-        int64_t id;
-        if (sscanf(line, "cpu%" PRId64, &id)) {
-            g_autofree char *path =

g_strdup_printf("/sys/devices/system/cpu/"
- "cpu%" PRId64 "/",
id);

qga/commands-posix.c +0 -20
-            vcpu = g_malloc0(sizeof *vcpu);
-            vcpu->logical_id = id;
-            vcpu->has_can_offline = true; /* lolspeak ftw */
-            transfer_vcpu(vcpu, true, path, &local_err);
-            QAPI_LIST_APPEND(tail, vcpu);
-        }
-    }
-
-    if (local_err == NULL) {
-        /* there's no guest with zero VCPUs */
-        g_assert(head != NULL);
-        return head;
-    }
-
-    qapi_free_GuestLogicalProcessorList(head);
-    error_propagate(errp, local_err);
-    return NULL;
-}
-
-int64_t qmp_guest_set_vcpus(GuestLogicalProcessorList *vcpus, Error

**errp)

qga/commands-posix.c +0 -7
-{
-    int64_t processed;
-    Error *local_err = NULL;
-
-    processed = 0;
-    while (vcpus != NULL) {
-        char *path = g_strdup_printf("/sys/devices/system/cpu/cpu%"

PRId64 "/",

qga/commands-posix.c +0 -22
-                                     vcpus->value->logical_id);
-
-        transfer_vcpu(vcpus->value, false, path, &local_err);
-        g_free(path);
-        if (local_err != NULL) {
-            break;
-        }
-        ++processed;
-        vcpus = vcpus->next;
-    }
-
-    if (local_err != NULL) {
-        if (processed == 0) {
-            error_propagate(errp, local_err);
-        } else {
-            error_free(local_err);
-        }
-    }
-
-    return processed;
-}
 #endif /* __linux__ */
 
 #if defined(__linux__) || defined(__FreeBSD__)
--

2.45.1

#6 lore raw

PATCH v2 3/22 qga: move linux suspend command impls to commands-linux.c

In reply to Daniel P. Berrangé

The qmp_guest_suspend_{disk,ram,hybrid} command impls in commands-posix.c are surrounded by '#ifdef __linux__' so should instead live in commands-linux.c

---
 qga/commands-linux.c | 265 +++++++++++++++++++++++++++++++++++++++++++
 qga/commands-posix.c | 265 -------------------------------------------
 2 files changed, 265 insertions(+), 265 deletions(-)
qga/commands-linux.c +265 -0
index 78580ac39d..3fabf54882 100644
@@ -286,6 +286,271 @@ int qmp_guest_fsfreeze_do_thaw(Error **errp)
 }
 #endif /* CONFIG_FSFREEZE */
 
+
+#define LINUX_SYS_STATE_FILE "/sys/power/state"
+#define SUSPEND_SUPPORTED 0
+#define SUSPEND_NOT_SUPPORTED 1
+
+typedef enum {
+    SUSPEND_MODE_DISK = 0,
+    SUSPEND_MODE_RAM = 1,
+    SUSPEND_MODE_HYBRID = 2,
+} SuspendMode;
+
+/*
+ * Executes a command in a child process using g_spawn_sync,
+ * returning an int >= 0 representing the exit status of the
+ * process.
+ *
+ * If the program wasn't found in path, returns -1.
+ *
+ * If a problem happened when creating the child process,
+ * returns -1 and errp is set.
+ */
+static int run_process_child(const char *command[], Error **errp)
+{
+    int exit_status, spawn_flag;
+    GError *g_err = NULL;
+    bool success;
+
+    spawn_flag = G_SPAWN_SEARCH_PATH | G_SPAWN_STDOUT_TO_DEV_NULL |
+                 G_SPAWN_STDERR_TO_DEV_NULL;
+
+    success =  g_spawn_sync(NULL, (char **)command, NULL, spawn_flag,
+                            NULL, NULL, NULL, NULL,
+                            &exit_status, &g_err);
+
+    if (success) {
+        return WEXITSTATUS(exit_status);
+    }
+
+    if (g_err && (g_err->code != G_SPAWN_ERROR_NOENT)) {
+        error_setg(errp, "failed to create child process, error '%s'",
+                   g_err->message);
+    }
+
+    g_error_free(g_err);
+    return -1;
+}
+
+static bool systemd_supports_mode(SuspendMode mode, Error **errp)
+{
+    const char *systemctl_args[3] = {"systemd-hibernate", "systemd-suspend",
+                                     "systemd-hybrid-sleep"};
+    const char *cmd[4] = {"systemctl", "status", systemctl_args[mode], NULL};
+    int status;
+
+    status = run_process_child(cmd, errp);
+
+    /*
+     * systemctl status uses LSB return codes so we can expect
+     * status > 0 and be ok. To assert if the guest has support
+     * for the selected suspend mode, status should be < 4. 4 is
+     * the code for unknown service status, the return value when
+     * the service does not exist. A common value is status = 3
+     * (program is not running).
+     */
+    if (status > 0 && status < 4) {
+        return true;
+    }
+
+    return false;
+}
+
+static void systemd_suspend(SuspendMode mode, Error **errp)
+{
+    Error *local_err = NULL;
+    const char *systemctl_args[3] = {"hibernate", "suspend", "hybrid-sleep"};
+    const char *cmd[3] = {"systemctl", systemctl_args[mode], NULL};
+    int status;
+
+    status = run_process_child(cmd, &local_err);
+
+    if (status == 0) {
+        return;
+    }
+
+    if ((status == -1) && !local_err) {
+        error_setg(errp, "the helper program 'systemctl %s' was not found",
+                   systemctl_args[mode]);
+        return;
+    }
+
+    if (local_err) {
+        error_propagate(errp, local_err);
+    } else {
+        error_setg(errp, "the helper program 'systemctl %s' returned an "
+                   "unexpected exit status code (%d)",
+                   systemctl_args[mode], status);
+    }
+}
+
+static bool pmutils_supports_mode(SuspendMode mode, Error **errp)
+{
+    Error *local_err = NULL;
+    const char *pmutils_args[3] = {"--hibernate", "--suspend",
+                                   "--suspend-hybrid"};
+    const char *cmd[3] = {"pm-is-supported", pmutils_args[mode], NULL};
+    int status;
+
+    status = run_process_child(cmd, &local_err);
+
+    if (status == SUSPEND_SUPPORTED) {
+        return true;
+    }
+
+    if ((status == -1) && !local_err) {
+        return false;
+    }
+
+    if (local_err) {
+        error_propagate(errp, local_err);
+    } else {
+        error_setg(errp,
+                   "the helper program '%s' returned an unexpected exit"
+                   " status code (%d)", "pm-is-supported", status);
+    }
+
+    return false;
+}
+
+static void pmutils_suspend(SuspendMode mode, Error **errp)
+{
+    Error *local_err = NULL;
+    const char *pmutils_binaries[3] = {"pm-hibernate", "pm-suspend",
+                                       "pm-suspend-hybrid"};
+    const char *cmd[2] = {pmutils_binaries[mode], NULL};
+    int status;
+
+    status = run_process_child(cmd, &local_err);
+
+    if (status == 0) {
+        return;
+    }
+
+    if ((status == -1) && !local_err) {
+        error_setg(errp, "the helper program '%s' was not found",
+                   pmutils_binaries[mode]);
+        return;
+    }
+
+    if (local_err) {
+        error_propagate(errp, local_err);
+    } else {
+        error_setg(errp,
+                   "the helper program '%s' returned an unexpected exit"
+                   " status code (%d)", pmutils_binaries[mode], status);
+    }
+}
+
+static bool linux_sys_state_supports_mode(SuspendMode mode, Error **errp)
+{
+    const char *sysfile_strs[3] = {"disk", "mem", NULL};
+    const char *sysfile_str = sysfile_strs[mode];
+    char buf[32]; /* hopefully big enough */
+    int fd;
+    ssize_t ret;
+
+    if (!sysfile_str) {
+        error_setg(errp, "unknown guest suspend mode");
+        return false;
+    }
+
+    fd = open(LINUX_SYS_STATE_FILE, O_RDONLY);
+    if (fd < 0) {
+        return false;
+    }
+
+    ret = read(fd, buf, sizeof(buf) - 1);
+    close(fd);
+    if (ret <= 0) {
+        return false;
+    }
+    buf[ret] = '\0';
+
+    if (strstr(buf, sysfile_str)) {
+        return true;
+    }
+    return false;
+}
+
+static void linux_sys_state_suspend(SuspendMode mode, Error **errp)
+{
+    g_autoptr(GError) local_gerr = NULL;
+    const char *sysfile_strs[3] = {"disk", "mem", NULL};
+    const char *sysfile_str = sysfile_strs[mode];
+
+    if (!sysfile_str) {
+        error_setg(errp, "unknown guest suspend mode");
+        return;
+    }
+
+    if (!g_file_set_contents(LINUX_SYS_STATE_FILE, sysfile_str,
+                             -1, &local_gerr)) {
+        error_setg(errp, "suspend: cannot write to '%s': %s",
+                   LINUX_SYS_STATE_FILE, local_gerr->message);
+        return;
+    }
+}
+
+static void guest_suspend(SuspendMode mode, Error **errp)
+{
+    Error *local_err = NULL;
+    bool mode_supported = false;
+
+    if (systemd_supports_mode(mode, &local_err)) {
+        mode_supported = true;
+        systemd_suspend(mode, &local_err);
+
+        if (!local_err) {
+            return;
+        }
+    }
+
+    error_free(local_err);
+    local_err = NULL;
+
+    if (pmutils_supports_mode(mode, &local_err)) {
+        mode_supported = true;
+        pmutils_suspend(mode, &local_err);
+
+        if (!local_err) {
+            return;
+        }
+    }
+
+    error_free(local_err);
+    local_err = NULL;
+
+    if (linux_sys_state_supports_mode(mode, &local_err)) {
+        mode_supported = true;
+        linux_sys_state_suspend(mode, &local_err);
+    }
+
+    if (!mode_supported) {
+        error_free(local_err);
+        error_setg(errp,
+                   "the requested suspend mode is not supported by the guest");
+    } else {
+        error_propagate(errp, local_err);
+    }
+}
+
+void qmp_guest_suspend_disk(Error **errp)
+{
+    guest_suspend(SUSPEND_MODE_DISK, errp);
+}
+
+void qmp_guest_suspend_ram(Error **errp)
+{
+    guest_suspend(SUSPEND_MODE_RAM, errp);
+}
+
+void qmp_guest_suspend_hybrid(Error **errp)
+{
+    guest_suspend(SUSPEND_MODE_HYBRID, errp);
+}
+
 /* Transfer online/offline status between @vcpu and the guest system.
  *
  * On input either @errp or *@errp must be NULL.
qga/commands-posix.c +0 -265

271 lines not on this page · raw message

Signature · 1 line
2.45.1
Reviewed-by on v2 3/22
#7 lore raw

Re: [PATCH v2 03/22] qga: move linux suspend command impls to commands-linux.c

Quoted 10 lines · Daniel P. Berrangé

The qmp_guest_suspend_{disk,ram,hybrid} command impls in commands-posix.c are surrounded by '#ifdef __linux__' so should instead live in commands-linux.c

---
  qga/commands-linux.c | 265 +++++++++++++++++++++++++++++++++++++++++++
  qga/commands-posix.c | 265 -------------------------------------------
  2 files changed, 265 insertions(+), 265 deletions(-)
Reviewed-by on v2 3/22
#8 lore raw

Re: [PATCH v2 03/22] qga: move linux suspend command impls to commands-linux.c

In reply to Daniel P. Berrangé
Quoted 578 lines · Daniel P. Berrangé

The qmp_guest_suspend_{disk,ram,hybrid} command impls in commands-posix.c are surrounded by '#ifdef __linux__' so should instead live in commands-linux.c

---
 qga/commands-linux.c | 265 +++++++++++++++++++++++++++++++++++++++++++
 qga/commands-posix.c | 265 -------------------------------------------
 2 files changed, 265 insertions(+), 265 deletions(-)
qga/commands-linux.c +86 -0
index 78580ac39d..3fabf54882 100644
@@ -286,6 +286,271 @@ int qmp_guest_fsfreeze_do_thaw(Error **errp)
 }
 #endif /* CONFIG_FSFREEZE */
 
+
+#define LINUX_SYS_STATE_FILE "/sys/power/state"
+#define SUSPEND_SUPPORTED 0
+#define SUSPEND_NOT_SUPPORTED 1
+
+typedef enum {
+    SUSPEND_MODE_DISK = 0,
+    SUSPEND_MODE_RAM = 1,
+    SUSPEND_MODE_HYBRID = 2,
+} SuspendMode;
+
+/*
+ * Executes a command in a child process using g_spawn_sync,
+ * returning an int >= 0 representing the exit status of the
+ * process.
+ *
+ * If the program wasn't found in path, returns -1.
+ *
+ * If a problem happened when creating the child process,
+ * returns -1 and errp is set.
+ */
+static int run_process_child(const char *command[], Error **errp)
+{
+    int exit_status, spawn_flag;
+    GError *g_err = NULL;
+    bool success;
+
+    spawn_flag = G_SPAWN_SEARCH_PATH | G_SPAWN_STDOUT_TO_DEV_NULL |
+                 G_SPAWN_STDERR_TO_DEV_NULL;
+
+    success =  g_spawn_sync(NULL, (char **)command, NULL, spawn_flag,
+                            NULL, NULL, NULL, NULL,
+                            &exit_status, &g_err);
+
+    if (success) {
+        return WEXITSTATUS(exit_status);
+    }
+
+    if (g_err && (g_err->code != G_SPAWN_ERROR_NOENT)) {
+        error_setg(errp, "failed to create child process, error '%s'",
+                   g_err->message);
+    }
+
+    g_error_free(g_err);
+    return -1;
+}
+
+static bool systemd_supports_mode(SuspendMode mode, Error **errp)
+{
+    const char *systemctl_args[3] = {"systemd-hibernate",
 "systemd-suspend",
+                                     "systemd-hybrid-sleep"};
+    const char *cmd[4] = {"systemctl", "status", systemctl_args[mode],
 NULL};
+    int status;
+
+    status = run_process_child(cmd, errp);
+
+    /*
+     * systemctl status uses LSB return codes so we can expect
+     * status > 0 and be ok. To assert if the guest has support
+     * for the selected suspend mode, status should be < 4. 4 is
+     * the code for unknown service status, the return value when
+     * the service does not exist. A common value is status = 3
+     * (program is not running).
+     */
+    if (status > 0 && status < 4) {
+        return true;
+    }
+
+    return false;
+}
+
+static void systemd_suspend(SuspendMode mode, Error **errp)
+{
+    Error *local_err = NULL;
+    const char *systemctl_args[3] = {"hibernate", "suspend",
 "hybrid-sleep"};
+    const char *cmd[3] = {"systemctl", systemctl_args[mode], NULL};
+    int status;
+
+    status = run_process_child(cmd, &local_err);
+
+    if (status == 0) {
+        return;
+    }
+
+    if ((status == -1) && !local_err) {
+        error_setg(errp, "the helper program 'systemctl %s' was not

found",

qga/commands-linux.c +159 -0
+                   systemctl_args[mode]);
+        return;
+    }
+
+    if (local_err) {
+        error_propagate(errp, local_err);
+    } else {
+        error_setg(errp, "the helper program 'systemctl %s' returned an "
+                   "unexpected exit status code (%d)",
+                   systemctl_args[mode], status);
+    }
+}
+
+static bool pmutils_supports_mode(SuspendMode mode, Error **errp)
+{
+    Error *local_err = NULL;
+    const char *pmutils_args[3] = {"--hibernate", "--suspend",
+                                   "--suspend-hybrid"};
+    const char *cmd[3] = {"pm-is-supported", pmutils_args[mode], NULL};
+    int status;
+
+    status = run_process_child(cmd, &local_err);
+
+    if (status == SUSPEND_SUPPORTED) {
+        return true;
+    }
+
+    if ((status == -1) && !local_err) {
+        return false;
+    }
+
+    if (local_err) {
+        error_propagate(errp, local_err);
+    } else {
+        error_setg(errp,
+                   "the helper program '%s' returned an unexpected exit"
+                   " status code (%d)", "pm-is-supported", status);
+    }
+
+    return false;
+}
+
+static void pmutils_suspend(SuspendMode mode, Error **errp)
+{
+    Error *local_err = NULL;
+    const char *pmutils_binaries[3] = {"pm-hibernate", "pm-suspend",
+                                       "pm-suspend-hybrid"};
+    const char *cmd[2] = {pmutils_binaries[mode], NULL};
+    int status;
+
+    status = run_process_child(cmd, &local_err);
+
+    if (status == 0) {
+        return;
+    }
+
+    if ((status == -1) && !local_err) {
+        error_setg(errp, "the helper program '%s' was not found",
+                   pmutils_binaries[mode]);
+        return;
+    }
+
+    if (local_err) {
+        error_propagate(errp, local_err);
+    } else {
+        error_setg(errp,
+                   "the helper program '%s' returned an unexpected exit"
+                   " status code (%d)", pmutils_binaries[mode], status);
+    }
+}
+
+static bool linux_sys_state_supports_mode(SuspendMode mode, Error **errp)
+{
+    const char *sysfile_strs[3] = {"disk", "mem", NULL};
+    const char *sysfile_str = sysfile_strs[mode];
+    char buf[32]; /* hopefully big enough */
+    int fd;
+    ssize_t ret;
+
+    if (!sysfile_str) {
+        error_setg(errp, "unknown guest suspend mode");
+        return false;
+    }
+
+    fd = open(LINUX_SYS_STATE_FILE, O_RDONLY);
+    if (fd < 0) {
+        return false;
+    }
+
+    ret = read(fd, buf, sizeof(buf) - 1);
+    close(fd);
+    if (ret <= 0) {
+        return false;
+    }
+    buf[ret] = '\0';
+
+    if (strstr(buf, sysfile_str)) {
+        return true;
+    }
+    return false;
+}
+
+static void linux_sys_state_suspend(SuspendMode mode, Error **errp)
+{
+    g_autoptr(GError) local_gerr = NULL;
+    const char *sysfile_strs[3] = {"disk", "mem", NULL};
+    const char *sysfile_str = sysfile_strs[mode];
+
+    if (!sysfile_str) {
+        error_setg(errp, "unknown guest suspend mode");
+        return;
+    }
+
+    if (!g_file_set_contents(LINUX_SYS_STATE_FILE, sysfile_str,
+                             -1, &local_gerr)) {
+        error_setg(errp, "suspend: cannot write to '%s': %s",
+                   LINUX_SYS_STATE_FILE, local_gerr->message);
+        return;
+    }
+}
+
+static void guest_suspend(SuspendMode mode, Error **errp)
+{
+    Error *local_err = NULL;
+    bool mode_supported = false;
+
+    if (systemd_supports_mode(mode, &local_err)) {
+        mode_supported = true;
+        systemd_suspend(mode, &local_err);
+
+        if (!local_err) {
+            return;
+        }
+    }
+
+    error_free(local_err);
+    local_err = NULL;
+
+    if (pmutils_supports_mode(mode, &local_err)) {
+        mode_supported = true;
+        pmutils_suspend(mode, &local_err);
+
+        if (!local_err) {
+            return;
+        }
+    }
+
+    error_free(local_err);
+    local_err = NULL;
+
+    if (linux_sys_state_supports_mode(mode, &local_err)) {
+        mode_supported = true;
+        linux_sys_state_suspend(mode, &local_err);
+    }
+
+    if (!mode_supported) {
+        error_free(local_err);
+        error_setg(errp,
+                   "the requested suspend mode is not supported by the

guest");

qga/commands-linux.c +20 -0
+    } else {
+        error_propagate(errp, local_err);
+    }
+}
+
+void qmp_guest_suspend_disk(Error **errp)
+{
+    guest_suspend(SUSPEND_MODE_DISK, errp);
+}
+
+void qmp_guest_suspend_ram(Error **errp)
+{
+    guest_suspend(SUSPEND_MODE_RAM, errp);
+}
+
+void qmp_guest_suspend_hybrid(Error **errp)
+{
+    guest_suspend(SUSPEND_MODE_HYBRID, errp);
+}
+
 /* Transfer online/offline status between @vcpu and the guest system.
  *
  * On input either @errp or *@errp must be NULL.
qga/commands-posix.c +0 -75
index a8ef41f175..ef21da63be 100644
@@ -1738,271 +1738,6 @@ qmp_guest_fstrim(bool has_minimum, int64_t
 minimum, Error **errp)
 }
 #endif /* CONFIG_FSTRIM */
 
-
-#define LINUX_SYS_STATE_FILE "/sys/power/state"
-#define SUSPEND_SUPPORTED 0
-#define SUSPEND_NOT_SUPPORTED 1
-
-typedef enum {
-    SUSPEND_MODE_DISK = 0,
-    SUSPEND_MODE_RAM = 1,
-    SUSPEND_MODE_HYBRID = 2,
-} SuspendMode;
-
-/*
- * Executes a command in a child process using g_spawn_sync,
- * returning an int >= 0 representing the exit status of the
- * process.
- *
- * If the program wasn't found in path, returns -1.
- *
- * If a problem happened when creating the child process,
- * returns -1 and errp is set.
- */
-static int run_process_child(const char *command[], Error **errp)
-{
-    int exit_status, spawn_flag;
-    GError *g_err = NULL;
-    bool success;
-
-    spawn_flag = G_SPAWN_SEARCH_PATH | G_SPAWN_STDOUT_TO_DEV_NULL |
-                 G_SPAWN_STDERR_TO_DEV_NULL;
-
-    success =  g_spawn_sync(NULL, (char **)command, NULL, spawn_flag,
-                            NULL, NULL, NULL, NULL,
-                            &exit_status, &g_err);
-
-    if (success) {
-        return WEXITSTATUS(exit_status);
-    }
-
-    if (g_err && (g_err->code != G_SPAWN_ERROR_NOENT)) {
-        error_setg(errp, "failed to create child process, error '%s'",
-                   g_err->message);
-    }
-
-    g_error_free(g_err);
-    return -1;
-}
-
-static bool systemd_supports_mode(SuspendMode mode, Error **errp)
-{
-    const char *systemctl_args[3] = {"systemd-hibernate",
 "systemd-suspend",
-                                     "systemd-hybrid-sleep"};
-    const char *cmd[4] = {"systemctl", "status", systemctl_args[mode],
 NULL};
-    int status;
-
-    status = run_process_child(cmd, errp);
-
-    /*
-     * systemctl status uses LSB return codes so we can expect
-     * status > 0 and be ok. To assert if the guest has support
-     * for the selected suspend mode, status should be < 4. 4 is
-     * the code for unknown service status, the return value when
-     * the service does not exist. A common value is status = 3
-     * (program is not running).
-     */
-    if (status > 0 && status < 4) {
-        return true;
-    }
-
-    return false;
-}
-
-static void systemd_suspend(SuspendMode mode, Error **errp)
-{
-    Error *local_err = NULL;
-    const char *systemctl_args[3] = {"hibernate", "suspend",

"hybrid-sleep"};

qga/commands-posix.c +0 -11
-    const char *cmd[3] = {"systemctl", systemctl_args[mode], NULL};
-    int status;
-
-    status = run_process_child(cmd, &local_err);
-
-    if (status == 0) {
-        return;
-    }
-
-    if ((status == -1) && !local_err) {
-        error_setg(errp, "the helper program 'systemctl %s' was not

found",

qga/commands-posix.c +0 -159
-                   systemctl_args[mode]);
-        return;
-    }
-
-    if (local_err) {
-        error_propagate(errp, local_err);
-    } else {
-        error_setg(errp, "the helper program 'systemctl %s' returned an "
-                   "unexpected exit status code (%d)",
-                   systemctl_args[mode], status);
-    }
-}
-
-static bool pmutils_supports_mode(SuspendMode mode, Error **errp)
-{
-    Error *local_err = NULL;
-    const char *pmutils_args[3] = {"--hibernate", "--suspend",
-                                   "--suspend-hybrid"};
-    const char *cmd[3] = {"pm-is-supported", pmutils_args[mode], NULL};
-    int status;
-
-    status = run_process_child(cmd, &local_err);
-
-    if (status == SUSPEND_SUPPORTED) {
-        return true;
-    }
-
-    if ((status == -1) && !local_err) {
-        return false;
-    }
-
-    if (local_err) {
-        error_propagate(errp, local_err);
-    } else {
-        error_setg(errp,
-                   "the helper program '%s' returned an unexpected exit"
-                   " status code (%d)", "pm-is-supported", status);
-    }
-
-    return false;
-}
-
-static void pmutils_suspend(SuspendMode mode, Error **errp)
-{
-    Error *local_err = NULL;
-    const char *pmutils_binaries[3] = {"pm-hibernate", "pm-suspend",
-                                       "pm-suspend-hybrid"};
-    const char *cmd[2] = {pmutils_binaries[mode], NULL};
-    int status;
-
-    status = run_process_child(cmd, &local_err);
-
-    if (status == 0) {
-        return;
-    }
-
-    if ((status == -1) && !local_err) {
-        error_setg(errp, "the helper program '%s' was not found",
-                   pmutils_binaries[mode]);
-        return;
-    }
-
-    if (local_err) {
-        error_propagate(errp, local_err);
-    } else {
-        error_setg(errp,
-                   "the helper program '%s' returned an unexpected exit"
-                   " status code (%d)", pmutils_binaries[mode], status);
-    }
-}
-
-static bool linux_sys_state_supports_mode(SuspendMode mode, Error **errp)
-{
-    const char *sysfile_strs[3] = {"disk", "mem", NULL};
-    const char *sysfile_str = sysfile_strs[mode];
-    char buf[32]; /* hopefully big enough */
-    int fd;
-    ssize_t ret;
-
-    if (!sysfile_str) {
-        error_setg(errp, "unknown guest suspend mode");
-        return false;
-    }
-
-    fd = open(LINUX_SYS_STATE_FILE, O_RDONLY);
-    if (fd < 0) {
-        return false;
-    }
-
-    ret = read(fd, buf, sizeof(buf) - 1);
-    close(fd);
-    if (ret <= 0) {
-        return false;
-    }
-    buf[ret] = '\0';
-
-    if (strstr(buf, sysfile_str)) {
-        return true;
-    }
-    return false;
-}
-
-static void linux_sys_state_suspend(SuspendMode mode, Error **errp)
-{
-    g_autoptr(GError) local_gerr = NULL;
-    const char *sysfile_strs[3] = {"disk", "mem", NULL};
-    const char *sysfile_str = sysfile_strs[mode];
-
-    if (!sysfile_str) {
-        error_setg(errp, "unknown guest suspend mode");
-        return;
-    }
-
-    if (!g_file_set_contents(LINUX_SYS_STATE_FILE, sysfile_str,
-                             -1, &local_gerr)) {
-        error_setg(errp, "suspend: cannot write to '%s': %s",
-                   LINUX_SYS_STATE_FILE, local_gerr->message);
-        return;
-    }
-}
-
-static void guest_suspend(SuspendMode mode, Error **errp)
-{
-    Error *local_err = NULL;
-    bool mode_supported = false;
-
-    if (systemd_supports_mode(mode, &local_err)) {
-        mode_supported = true;
-        systemd_suspend(mode, &local_err);
-
-        if (!local_err) {
-            return;
-        }
-    }
-
-    error_free(local_err);
-    local_err = NULL;
-
-    if (pmutils_supports_mode(mode, &local_err)) {
-        mode_supported = true;
-        pmutils_suspend(mode, &local_err);
-
-        if (!local_err) {
-            return;
-        }
-    }
-
-    error_free(local_err);
-    local_err = NULL;
-
-    if (linux_sys_state_supports_mode(mode, &local_err)) {
-        mode_supported = true;
-        linux_sys_state_suspend(mode, &local_err);
-    }
-
-    if (!mode_supported) {
-        error_free(local_err);
-        error_setg(errp,
-                   "the requested suspend mode is not supported by the

guest");

qga/commands-posix.c +0 -21
-    } else {
-        error_propagate(errp, local_err);
-    }
-}
-
-void qmp_guest_suspend_disk(Error **errp)
-{
-    guest_suspend(SUSPEND_MODE_DISK, errp);
-}
-
-void qmp_guest_suspend_ram(Error **errp)
-{
-    guest_suspend(SUSPEND_MODE_RAM, errp);
-}
-
-void qmp_guest_suspend_hybrid(Error **errp)
-{
-    guest_suspend(SUSPEND_MODE_HYBRID, errp);
-}
-
 #endif /* __linux__ */
 
 #if defined(__linux__) || defined(__FreeBSD__)
--

2.45.1

#9 lore raw

PATCH v2 4/22 qga: move linux fs/disk command impls to commands-linux.c

In reply to Daniel P. Berrangé

The qmp_guest_{fstrim, get_fsinfo, get_disks} command impls in commands-posix.c are surrounded by '#ifdef __linux__' so should instead live in commands-linux.c

---
 qga/commands-linux.c | 904 ++++++++++++++++++++++++++++++++++++++++++
 qga/commands-posix.c | 909 -------------------------------------------
 2 files changed, 904 insertions(+), 909 deletions(-)
qga/commands-linux.c +904 -0
index 3fabf54882..084e6c9e85 100644
@@ -14,10 +14,21 @@
 #include "qemu/osdep.h"
 #include "qapi/error.h"
 #include "qga-qapi-commands.h"
+#include "qapi/error.h"
+#include "qapi/qmp/qerror.h"
 #include "commands-common.h"
 #include "cutils.h"
 #include <mntent.h>
 #include <sys/ioctl.h>
+#include <mntent.h>
+#include <linux/nvme_ioctl.h>
+#include "block/nvme.h"
+
+#ifdef CONFIG_LIBUDEV
+#include <libudev.h>
+#endif
+
+#include <sys/statvfs.h>
 
 #if defined(CONFIG_FSFREEZE) || defined(CONFIG_FSTRIM)
 static int dev_major_minor(const char *devpath,
@@ -286,6 +297,899 @@ int qmp_guest_fsfreeze_do_thaw(Error **errp)
 }
 #endif /* CONFIG_FSFREEZE */
 
+#if defined(CONFIG_FSFREEZE)
+
+static char *get_pci_driver(char const *syspath, int pathlen, Error **errp)
+{
+    char *path;
+    char *dpath;
+    char *driver = NULL;
+    char buf[PATH_MAX];
+    ssize_t len;
+
+    path = g_strndup(syspath, pathlen);
+    dpath = g_strdup_printf("%s/driver", path);
+    len = readlink(dpath, buf, sizeof(buf) - 1);
+    if (len != -1) {
+        buf[len] = 0;
+        driver = g_path_get_basename(buf);
+    }
+    g_free(dpath);
+    g_free(path);
+    return driver;
+}
+
+static int compare_uint(const void *_a, const void *_b)
+{
+    unsigned int a = *(unsigned int *)_a;
+    unsigned int b = *(unsigned int *)_b;
+
+    return a < b ? -1 : a > b ? 1 : 0;
+}
+
+/* Walk the specified sysfs and build a sorted list of host or ata numbers */
+static int build_hosts(char const *syspath, char const *host, bool ata,
+                       unsigned int *hosts, int hosts_max, Error **errp)
+{
+    char *path;
+    DIR *dir;
+    struct dirent *entry;
+    int i = 0;
+
+    path = g_strndup(syspath, host - syspath);
+    dir = opendir(path);
+    if (!dir) {
+        error_setg_errno(errp, errno, "opendir(\"%s\")", path);
+        g_free(path);
+        return -1;
+    }
+
+    while (i < hosts_max) {
+        entry = readdir(dir);
+        if (!entry) {
+            break;
+        }
+        if (ata && sscanf(entry->d_name, "ata%d", hosts + i) == 1) {
+            ++i;
+        } else if (!ata && sscanf(entry->d_name, "host%d", hosts + i) == 1) {
+            ++i;
+        }
+    }
+
+    qsort(hosts, i, sizeof(hosts[0]), compare_uint);
+
+    g_free(path);
+    closedir(dir);
+    return i;
+}
+
+/*
+ * Store disk device info for devices on the PCI bus.
+ * Returns true if information has been stored, or false for failure.
+ */
+static bool build_guest_fsinfo_for_pci_dev(char const *syspath,
+                                           GuestDiskAddress *disk,
+                                           Error **errp)
+{
+    unsigned int pci[4], host, hosts[8], tgt[3];
+    int i, nhosts = 0, pcilen;
+    GuestPCIAddress *pciaddr = disk->pci_controller;
+    bool has_ata = false, has_host = false, has_tgt = false;
+    char *p, *q, *driver = NULL;
+    bool ret = false;
+
+    p = strstr(syspath, "/devices/pci");
+    if (!p || sscanf(p + 12, "%*x:%*x/%x:%x:%x.%x%n",
+                     pci, pci + 1, pci + 2, pci + 3, &pcilen) < 4) {
+        g_debug("only pci device is supported: sysfs path '%s'", syspath);
+        return false;
+    }
+
+    p += 12 + pcilen;
+    while (true) {
+        driver = get_pci_driver(syspath, p - syspath, errp);
+        if (driver && (g_str_equal(driver, "ata_piix") ||
+                       g_str_equal(driver, "sym53c8xx") ||
+                       g_str_equal(driver, "virtio-pci") ||
+                       g_str_equal(driver, "ahci") ||
+                       g_str_equal(driver, "nvme") ||
+                       g_str_equal(driver, "xhci_hcd") ||
+                       g_str_equal(driver, "ehci-pci"))) {
+            break;
+        }
+
+        g_free(driver);
+        if (sscanf(p, "/%x:%x:%x.%x%n",
+                          pci, pci + 1, pci + 2, pci + 3, &pcilen) == 4) {
+            p += pcilen;
+            continue;
+        }
+
+        g_debug("unsupported driver or sysfs path '%s'", syspath);
+        return false;
+    }
+
+    p = strstr(syspath, "/target");
+    if (p && sscanf(p + 7, "%*u:%*u:%*u/%*u:%u:%u:%u",
+                    tgt, tgt + 1, tgt + 2) == 3) {
+        has_tgt = true;
+    }
+
+    p = strstr(syspath, "/ata");
+    if (p) {
+        q = p + 4;
+        has_ata = true;
+    } else {
+        p = strstr(syspath, "/host");
+        q = p + 5;
+    }
+    if (p && sscanf(q, "%u", &host) == 1) {
+        has_host = true;
+        nhosts = build_hosts(syspath, p, has_ata, hosts,
+                             ARRAY_SIZE(hosts), errp);
+        if (nhosts < 0) {
+            goto cleanup;
+        }
+    }
+
+    pciaddr->domain = pci[0];
+    pciaddr->bus = pci[1];
+    pciaddr->slot = pci[2];
+    pciaddr->function = pci[3];
+
+    if (strcmp(driver, "ata_piix") == 0) {
+        /* a host per ide bus, target*:0:<unit>:0 */
+        if (!has_host || !has_tgt) {
+            g_debug("invalid sysfs path '%s' (driver '%s')", syspath, driver);
+            goto cleanup;
+        }
+        for (i = 0; i < nhosts; i++) {
+            if (host == hosts[i]) {
+                disk->bus_type = GUEST_DISK_BUS_TYPE_IDE;
+                disk->bus = i;
+                disk->unit = tgt[1];
+                break;
+            }
+        }
+        if (i >= nhosts) {
+            g_debug("no host for '%s' (driver '%s')", syspath, driver);
+            goto cleanup;
+        }
+    } else if (strcmp(driver, "sym53c8xx") == 0) {
+        /* scsi(LSI Logic): target*:0:<unit>:0 */
+        if (!has_tgt) {
+            g_debug("invalid sysfs path '%s' (driver '%s')", syspath, driver);
+            goto cleanup;
+        }
+        disk->bus_type = GUEST_DISK_BUS_TYPE_SCSI;
+        disk->unit = tgt[1];
+    } else if (strcmp(driver, "virtio-pci") == 0) {
+        if (has_tgt) {
+            /* virtio-scsi: target*:0:0:<unit> */
+            disk->bus_type = GUEST_DISK_BUS_TYPE_SCSI;
+            disk->unit = tgt[2];
+        } else {
+            /* virtio-blk: 1 disk per 1 device */
+            disk->bus_type = GUEST_DISK_BUS_TYPE_VIRTIO;
+        }
+    } else if (strcmp(driver, "ahci") == 0) {
+        /* ahci: 1 host per 1 unit */
+        if (!has_host || !has_tgt) {
+            g_debug("invalid sysfs path '%s' (driver '%s')", syspath, driver);
+            goto cleanup;
+        }
+        for (i = 0; i < nhosts; i++) {
+            if (host == hosts[i]) {
+                disk->unit = i;
+                disk->bus_type = GUEST_DISK_BUS_TYPE_SATA;
+                break;
+            }
+        }
+        if (i >= nhosts) {
+            g_debug("no host for '%s' (driver '%s')", syspath, driver);
+            goto cleanup;
+        }
+    } else if (strcmp(driver, "nvme") == 0) {
+        disk->bus_type = GUEST_DISK_BUS_TYPE_NVME;
+    } else if (strcmp(driver, "ehci-pci") == 0 || strcmp(driver, "xhci_hcd") == 0) {
+        disk->bus_type = GUEST_DISK_BUS_TYPE_USB;
+    } else {
+        g_debug("unknown driver '%s' (sysfs path '%s')", driver, syspath);
+        goto cleanup;
+    }
+
+    ret = true;
+
+cleanup:
+    g_free(driver);
+    return ret;
+}
+
+/*
+ * Store disk device info for non-PCI virtio devices (for example s390x
+ * channel I/O devices). Returns true if information has been stored, or
+ * false for failure.
+ */
+static bool build_guest_fsinfo_for_nonpci_virtio(char const *syspath,
+                                                 GuestDiskAddress *disk,
+                                                 Error **errp)
+{
+    unsigned int tgt[3];
+    char *p;
+
+    if (!strstr(syspath, "/virtio") || !strstr(syspath, "/block")) {
+        g_debug("Unsupported virtio device '%s'", syspath);
+        return false;
+    }
+
+    p = strstr(syspath, "/target");
+    if (p && sscanf(p + 7, "%*u:%*u:%*u/%*u:%u:%u:%u",
+                    &tgt[0], &tgt[1], &tgt[2]) == 3) {
+        /* virtio-scsi: target*:0:<target>:<unit> */
+        disk->bus_type = GUEST_DISK_BUS_TYPE_SCSI;
+        disk->bus = tgt[0];
+        disk->target = tgt[1];
+        disk->unit = tgt[2];
+    } else {
+        /* virtio-blk: 1 disk per 1 device */
+        disk->bus_type = GUEST_DISK_BUS_TYPE_VIRTIO;
+    }
+
+    return true;
+}
+
+/*
+ * Store disk device info for CCW devices (s390x channel I/O devices).
+ * Returns true if information has been stored, or false for failure.
+ */
+static bool build_guest_fsinfo_for_ccw_dev(char const *syspath,
+                                           GuestDiskAddress *disk,
+                                           Error **errp)
+{
+    unsigned int cssid, ssid, subchno, devno;
+    char *p;
+
+    p = strstr(syspath, "/devices/css");
+    if (!p || sscanf(p + 12, "%*x/%x.%x.%x/%*x.%*x.%x/",
+                     &cssid, &ssid, &subchno, &devno) < 4) {
+        g_debug("could not parse ccw device sysfs path: %s", syspath);
+        return false;
+    }
+
+    disk->ccw_address = g_new0(GuestCCWAddress, 1);
+    disk->ccw_address->cssid = cssid;
+    disk->ccw_address->ssid = ssid;
+    disk->ccw_address->subchno = subchno;
+    disk->ccw_address->devno = devno;
+
+    if (strstr(p, "/virtio")) {
+        build_guest_fsinfo_for_nonpci_virtio(syspath, disk, errp);
+    }
+
+    return true;
+}
+
+/* Store disk device info specified by @sysfs into @fs */
+static void build_guest_fsinfo_for_real_device(char const *syspath,
+                                               GuestFilesystemInfo *fs,
+                                               Error **errp)
+{
+    GuestDiskAddress *disk;
+    GuestPCIAddress *pciaddr;
+    bool has_hwinf;
+#ifdef CONFIG_LIBUDEV
+    struct udev *udev = NULL;
+    struct udev_device *udevice = NULL;
+#endif
+
+    pciaddr = g_new0(GuestPCIAddress, 1);
+    pciaddr->domain = -1;                       /* -1 means field is invalid */
+    pciaddr->bus = -1;
+    pciaddr->slot = -1;
+    pciaddr->function = -1;
+
+    disk = g_new0(GuestDiskAddress, 1);
+    disk->pci_controller = pciaddr;
+    disk->bus_type = GUEST_DISK_BUS_TYPE_UNKNOWN;
+
+#ifdef CONFIG_LIBUDEV
+    udev = udev_new();
+    udevice = udev_device_new_from_syspath(udev, syspath);
+    if (udev == NULL || udevice == NULL) {
+        g_debug("failed to query udev");
+    } else {
+        const char *devnode, *serial;
+        devnode = udev_device_get_devnode(udevice);
+        if (devnode != NULL) {
+            disk->dev = g_strdup(devnode);
+        }
+        serial = udev_device_get_property_value(udevice, "ID_SERIAL");
+        if (serial != NULL && *serial != 0) {
+            disk->serial = g_strdup(serial);
+        }
+    }
+
+    udev_unref(udev);
+    udev_device_unref(udevice);
+#endif
+
+    if (strstr(syspath, "/devices/pci")) {
+        has_hwinf = build_guest_fsinfo_for_pci_dev(syspath, disk, errp);
+    } else if (strstr(syspath, "/devices/css")) {
+        has_hwinf = build_guest_fsinfo_for_ccw_dev(syspath, disk, errp);
+    } else if (strstr(syspath, "/virtio")) {
+        has_hwinf = build_guest_fsinfo_for_nonpci_virtio(syspath, disk, errp);
+    } else {
+        g_debug("Unsupported device type for '%s'", syspath);
+        has_hwinf = false;
+    }
+
+    if (has_hwinf || disk->dev || disk->serial) {
+        QAPI_LIST_PREPEND(fs->disk, disk);
+    } else {
+        qapi_free_GuestDiskAddress(disk);
+    }
+}
+
+static void build_guest_fsinfo_for_device(char const *devpath,
+                                          GuestFilesystemInfo *fs,
+                                          Error **errp);
+
+/* Store a list of slave devices of virtual volume specified by @syspath into
+ * @fs */
+static void build_guest_fsinfo_for_virtual_device(char const *syspath,
+                                                  GuestFilesystemInfo *fs,
+                                                  Error **errp)
+{
+    Error *err = NULL;
+    DIR *dir;
+    char *dirpath;
+    struct dirent *entry;
+
+    dirpath = g_strdup_printf("%s/slaves", syspath);
+    dir = opendir(dirpath);
+    if (!dir) {
+        if (errno != ENOENT) {
+            error_setg_errno(errp, errno, "opendir(\"%s\")", dirpath);
+        }
+        g_free(dirpath);
+        return;
+    }
+
+    for (;;) {
+        errno = 0;
+        entry = readdir(dir);
+        if (entry == NULL) {
+            if (errno) {
+                error_setg_errno(errp, errno, "readdir(\"%s\")", dirpath);
+            }
+            break;
+        }
+
+        if (entry->d_type == DT_LNK) {
+            char *path;
+
+            g_debug(" slave device '%s'", entry->d_name);
+            path = g_strdup_printf("%s/slaves/%s", syspath, entry->d_name);
+            build_guest_fsinfo_for_device(path, fs, &err);
+            g_free(path);
+
+            if (err) {
+                error_propagate(errp, err);
+                break;
+            }
+        }
+    }
+
+    g_free(dirpath);
+    closedir(dir);
+}
+
+static bool is_disk_virtual(const char *devpath, Error **errp)
+{
+    g_autofree char *syspath = realpath(devpath, NULL);
+
+    if (!syspath) {
+        error_setg_errno(errp, errno, "realpath(\"%s\")", devpath);
+        return false;
+    }
+    return strstr(syspath, "/devices/virtual/block/") != NULL;
+}
+
+/* Dispatch to functions for virtual/real device */
+static void build_guest_fsinfo_for_device(char const *devpath,
+                                          GuestFilesystemInfo *fs,
+                                          Error **errp)
+{
+    ERRP_GUARD();
+    g_autofree char *syspath = NULL;
+    bool is_virtual = false;
+
+    syspath = realpath(devpath, NULL);
+    if (!syspath) {
+        if (errno != ENOENT) {
+            error_setg_errno(errp, errno, "realpath(\"%s\")", devpath);
+            return;
+        }
+
+        /* ENOENT: This devpath may not exist because of container config */
+        if (!fs->name) {
+            fs->name = g_path_get_basename(devpath);
+        }
+        return;
+    }
+
+    if (!fs->name) {
+        fs->name = g_path_get_basename(syspath);
+    }
+
+    g_debug("  parse sysfs path '%s'", syspath);
+    is_virtual = is_disk_virtual(syspath, errp);
+    if (*errp != NULL) {
+        return;
+    }
+    if (is_virtual) {
+        build_guest_fsinfo_for_virtual_device(syspath, fs, errp);
+    } else {
+        build_guest_fsinfo_for_real_device(syspath, fs, errp);
+    }
+}
+
+#ifdef CONFIG_LIBUDEV
+
+/*
+ * Wrapper around build_guest_fsinfo_for_device() for getting just
+ * the disk address.
+ */
+static GuestDiskAddress *get_disk_address(const char *syspath, Error **errp)
+{
+    g_autoptr(GuestFilesystemInfo) fs = NULL;
+
+    fs = g_new0(GuestFilesystemInfo, 1);
+    build_guest_fsinfo_for_device(syspath, fs, errp);
+    if (fs->disk != NULL) {
+        return g_steal_pointer(&fs->disk->value);
+    }
+    return NULL;
+}
+
+static char *get_alias_for_syspath(const char *syspath)
+{
+    struct udev *udev = NULL;
+    struct udev_device *udevice = NULL;
+    char *ret = NULL;
+
+    udev = udev_new();
+    if (udev == NULL) {
+        g_debug("failed to query udev");
+        goto out;
+    }
+    udevice = udev_device_new_from_syspath(udev, syspath);
+    if (udevice == NULL) {
+        g_debug("failed to query udev for path: %s", syspath);
+        goto out;
+    } else {
+        const char *alias = udev_device_get_property_value(
+            udevice, "DM_NAME");
+        /*
+         * NULL means there was an error and empty string means there is no
+         * alias. In case of no alias we return NULL instead of empty string.
+         */
+        if (alias == NULL) {
+            g_debug("failed to query udev for device alias for: %s",
+                syspath);
+        } else if (*alias != 0) {
+            ret = g_strdup(alias);
+        }
+    }
+
+out:
+    udev_unref(udev);
+    udev_device_unref(udevice);
+    return ret;
+}
+
+static char *get_device_for_syspath(const char *syspath)
+{
+    struct udev *udev = NULL;
+    struct udev_device *udevice = NULL;
+    char *ret = NULL;
+
+    udev = udev_new();
+    if (udev == NULL) {
+        g_debug("failed to query udev");
+        goto out;
+    }
+    udevice = udev_device_new_from_syspath(udev, syspath);
+    if (udevice == NULL) {
+        g_debug("failed to query udev for path: %s", syspath);
+        goto out;
+    } else {
+        ret = g_strdup(udev_device_get_devnode(udevice));
+    }
+
+out:
+    udev_unref(udev);
+    udev_device_unref(udevice);
+    return ret;
+}
+
+static void get_disk_deps(const char *disk_dir, GuestDiskInfo *disk)
+{
+    g_autofree char *deps_dir = NULL;
+    const gchar *dep;
+    GDir *dp_deps = NULL;
+
+    /* List dependent disks */
+    deps_dir = g_strdup_printf("%s/slaves", disk_dir);
+    g_debug("  listing entries in: %s", deps_dir);
+    dp_deps = g_dir_open(deps_dir, 0, NULL);
+    if (dp_deps == NULL) {
+        g_debug("failed to list entries in %s", deps_dir);
+        return;
+    }
+    disk->has_dependencies = true;
+    while ((dep = g_dir_read_name(dp_deps)) != NULL) {
+        g_autofree char *dep_dir = NULL;
+        char *dev_name;
+
+        /* Add dependent disks */
+        dep_dir = g_strdup_printf("%s/%s", deps_dir, dep);
+        dev_name = get_device_for_syspath(dep_dir);
+        if (dev_name != NULL) {
+            g_debug("  adding dependent device: %s", dev_name);
+            QAPI_LIST_PREPEND(disk->dependencies, dev_name);
+        }
+    }
+    g_dir_close(dp_deps);
+}
+
+/*
+ * Detect partitions subdirectory, name is "<disk_name><number>" or
+ * "<disk_name>p<number>"
+ *
+ * @disk_name -- last component of /sys path (e.g. sda)
+ * @disk_dir -- sys path of the disk (e.g. /sys/block/sda)
+ * @disk_dev -- device node of the disk (e.g. /dev/sda)
+ */
+static GuestDiskInfoList *get_disk_partitions(
+    GuestDiskInfoList *list,
+    const char *disk_name, const char *disk_dir,
+    const char *disk_dev)
+{
+    GuestDiskInfoList *ret = list;
+    struct dirent *de_disk;
+    DIR *dp_disk = NULL;
+    size_t len = strlen(disk_name);
+
+    dp_disk = opendir(disk_dir);
+    while ((de_disk = readdir(dp_disk)) != NULL) {
+        g_autofree char *partition_dir = NULL;
+        char *dev_name;
+        GuestDiskInfo *partition;
+
+        if (!(de_disk->d_type & DT_DIR)) {
+            continue;
+        }
+
+        if (!(strncmp(disk_name, de_disk->d_name, len) == 0 &&
+            ((*(de_disk->d_name + len) == 'p' &&
+            isdigit(*(de_disk->d_name + len + 1))) ||
+                isdigit(*(de_disk->d_name + len))))) {
+            continue;
+        }
+
+        partition_dir = g_strdup_printf("%s/%s",
+            disk_dir, de_disk->d_name);
+        dev_name = get_device_for_syspath(partition_dir);
+        if (dev_name == NULL) {
+            g_debug("Failed to get device name for syspath: %s",
+                disk_dir);
+            continue;
+        }
+        partition = g_new0(GuestDiskInfo, 1);
+        partition->name = dev_name;
+        partition->partition = true;
+        partition->has_dependencies = true;
+        /* Add parent disk as dependent for easier tracking of hierarchy */
+        QAPI_LIST_PREPEND(partition->dependencies, g_strdup(disk_dev));
+
+        QAPI_LIST_PREPEND(ret, partition);
+    }
+    closedir(dp_disk);
+
+    return ret;
+}
+
+static void get_nvme_smart(GuestDiskInfo *disk)
+{
+    int fd;
+    GuestNVMeSmart *smart;
+    NvmeSmartLog log = {0};
+    struct nvme_admin_cmd cmd = {
+        .opcode = NVME_ADM_CMD_GET_LOG_PAGE,
+        .nsid = NVME_NSID_BROADCAST,
+        .addr = (uintptr_t)&log,
+        .data_len = sizeof(log),
+        .cdw10 = NVME_LOG_SMART_INFO | (1 << 15) /* RAE bit */
+                 | (((sizeof(log) >> 2) - 1) << 16)
+    };
+
+    fd = qga_open_cloexec(disk->name, O_RDONLY, 0);
+    if (fd == -1) {
+        g_debug("Failed to open device: %s: %s", disk->name, g_strerror(errno));
+        return;
+    }
+
+    if (ioctl(fd, NVME_IOCTL_ADMIN_CMD, &cmd)) {
+        g_debug("Failed to get smart: %s: %s", disk->name, g_strerror(errno));
+        close(fd);
+        return;
+    }
+
+    disk->smart = g_new0(GuestDiskSmart, 1);
+    disk->smart->type = GUEST_DISK_BUS_TYPE_NVME;
+
+    smart = &disk->smart->u.nvme;
+    smart->critical_warning = log.critical_warning;
+    smart->temperature = lduw_le_p(&log.temperature); /* unaligned field */
+    smart->available_spare = log.available_spare;
+    smart->available_spare_threshold = log.available_spare_threshold;
+    smart->percentage_used = log.percentage_used;
+    smart->data_units_read_lo = le64_to_cpu(log.data_units_read[0]);
+    smart->data_units_read_hi = le64_to_cpu(log.data_units_read[1]);
+    smart->data_units_written_lo = le64_to_cpu(log.data_units_written[0]);
+    smart->data_units_written_hi = le64_to_cpu(log.data_units_written[1]);
+    smart->host_read_commands_lo = le64_to_cpu(log.host_read_commands[0]);
+    smart->host_read_commands_hi = le64_to_cpu(log.host_read_commands[1]);
+    smart->host_write_commands_lo = le64_to_cpu(log.host_write_commands[0]);
+    smart->host_write_commands_hi = le64_to_cpu(log.host_write_commands[1]);
+    smart->controller_busy_time_lo = le64_to_cpu(log.controller_busy_time[0]);
+    smart->controller_busy_time_hi = le64_to_cpu(log.controller_busy_time[1]);
+    smart->power_cycles_lo = le64_to_cpu(log.power_cycles[0]);
+    smart->power_cycles_hi = le64_to_cpu(log.power_cycles[1]);
+    smart->power_on_hours_lo = le64_to_cpu(log.power_on_hours[0]);
+    smart->power_on_hours_hi = le64_to_cpu(log.power_on_hours[1]);
+    smart->unsafe_shutdowns_lo = le64_to_cpu(log.unsafe_shutdowns[0]);
+    smart->unsafe_shutdowns_hi = le64_to_cpu(log.unsafe_shutdowns[1]);
+    smart->media_errors_lo = le64_to_cpu(log.media_errors[0]);
+    smart->media_errors_hi = le64_to_cpu(log.media_errors[1]);
+    smart->number_of_error_log_entries_lo =
+        le64_to_cpu(log.number_of_error_log_entries[0]);
+    smart->number_of_error_log_entries_hi =
+        le64_to_cpu(log.number_of_error_log_entries[1]);
+
+    close(fd);
+}
+
+static void get_disk_smart(GuestDiskInfo *disk)
+{
+    if (disk->address
+        && (disk->address->bus_type == GUEST_DISK_BUS_TYPE_NVME)) {
+        get_nvme_smart(disk);
+    }
+}
+
+GuestDiskInfoList *qmp_guest_get_disks(Error **errp)
+{
+    GuestDiskInfoList *ret = NULL;
+    GuestDiskInfo *disk;
+    DIR *dp = NULL;
+    struct dirent *de = NULL;
+
+    g_debug("listing /sys/block directory");
+    dp = opendir("/sys/block");
+    if (dp == NULL) {
+        error_setg_errno(errp, errno, "Can't open directory \"/sys/block\"");
+        return NULL;
+    }
+    while ((de = readdir(dp)) != NULL) {
+        g_autofree char *disk_dir = NULL, *line = NULL,
+            *size_path = NULL;
+        char *dev_name;
+        Error *local_err = NULL;
+        if (de->d_type != DT_LNK) {
+            g_debug("  skipping entry: %s", de->d_name);
+            continue;
+        }
+
+        /* Check size and skip zero-sized disks */
+        g_debug("  checking disk size");
+        size_path = g_strdup_printf("/sys/block/%s/size", de->d_name);
+        if (!g_file_get_contents(size_path, &line, NULL, NULL)) {
+            g_debug("  failed to read disk size");
+            continue;
+        }
+        if (g_strcmp0(line, "0\n") == 0) {
+            g_debug("  skipping zero-sized disk");
+            continue;
+        }
+
+        g_debug("  adding %s", de->d_name);
+        disk_dir = g_strdup_printf("/sys/block/%s", de->d_name);
+        dev_name = get_device_for_syspath(disk_dir);
+        if (dev_name == NULL) {
+            g_debug("Failed to get device name for syspath: %s",
+                disk_dir);
+            continue;
+        }
+        disk = g_new0(GuestDiskInfo, 1);
+        disk->name = dev_name;
+        disk->partition = false;
+        disk->alias = get_alias_for_syspath(disk_dir);
+        QAPI_LIST_PREPEND(ret, disk);
+
+        /* Get address for non-virtual devices */
+        bool is_virtual = is_disk_virtual(disk_dir, &local_err);
+        if (local_err != NULL) {
+            g_debug("  failed to check disk path, ignoring error: %s",
+                error_get_pretty(local_err));
+            error_free(local_err);
+            local_err = NULL;
+            /* Don't try to get the address */
+            is_virtual = true;
+        }
+        if (!is_virtual) {
+            disk->address = get_disk_address(disk_dir, &local_err);
+            if (local_err != NULL) {
+                g_debug("  failed to get device info, ignoring error: %s",
+                    error_get_pretty(local_err));
+                error_free(local_err);
+                local_err = NULL;
+            }
+        }
+
+        get_disk_deps(disk_dir, disk);
+        get_disk_smart(disk);
+        ret = get_disk_partitions(ret, de->d_name, disk_dir, dev_name);
+    }
+
+    closedir(dp);
+
+    return ret;
+}
+
+#else
+
+GuestDiskInfoList *qmp_guest_get_disks(Error **errp)
+{
+    error_setg(errp, QERR_UNSUPPORTED);
+    return NULL;
+}
+
+#endif
+
+/* Return a list of the disk device(s)' info which @mount lies on */
+static GuestFilesystemInfo *build_guest_fsinfo(struct FsMount *mount,
+                                               Error **errp)
+{
+    GuestFilesystemInfo *fs = g_malloc0(sizeof(*fs));
+    struct statvfs buf;
+    unsigned long used, nonroot_total, fr_size;
+    char *devpath = g_strdup_printf("/sys/dev/block/%u:%u",
+                                    mount->devmajor, mount->devminor);
+
+    fs->mountpoint = g_strdup(mount->dirname);
+    fs->type = g_strdup(mount->devtype);
+    build_guest_fsinfo_for_device(devpath, fs, errp);
+
+    if (statvfs(fs->mountpoint, &buf) == 0) {
+        fr_size = buf.f_frsize;
+        used = buf.f_blocks - buf.f_bfree;
+        nonroot_total = used + buf.f_bavail;
+        fs->used_bytes = used * fr_size;
+        fs->total_bytes = nonroot_total * fr_size;
+        fs->total_bytes_privileged = buf.f_blocks * fr_size;
+
+        fs->has_total_bytes = true;
+        fs->has_total_bytes_privileged = true;
+        fs->has_used_bytes = true;
+    }
+
+    g_free(devpath);
+
+    return fs;
+}
+
+GuestFilesystemInfoList *qmp_guest_get_fsinfo(Error **errp)
+{
+    FsMountList mounts;
+    struct FsMount *mount;
+    GuestFilesystemInfoList *ret = NULL;
+    Error *local_err = NULL;
+
+    QTAILQ_INIT(&mounts);
+    if (!build_fs_mount_list(&mounts, &local_err)) {
+        error_propagate(errp, local_err);
+        return NULL;
+    }
+
+    QTAILQ_FOREACH(mount, &mounts, next) {
+        g_debug("Building guest fsinfo for '%s'", mount->dirname);
+
+        QAPI_LIST_PREPEND(ret, build_guest_fsinfo(mount, &local_err));
+        if (local_err) {
+            error_propagate(errp, local_err);
+            qapi_free_GuestFilesystemInfoList(ret);
+            ret = NULL;
+            break;
+        }
+    }
+
+    free_fs_mount_list(&mounts);
+    return ret;
+}
+#endif /* CONFIG_FSFREEZE */
+
+#if defined(CONFIG_FSTRIM)
+/*
+ * Walk list of mounted file systems in the guest, and trim them.
+ */
+GuestFilesystemTrimResponse *
+qmp_guest_fstrim(bool has_minimum, int64_t minimum, Error **errp)
+{
+    GuestFilesystemTrimResponse *response;
+    GuestFilesystemTrimResult *result;
+    int ret = 0;
+    FsMountList mounts;
+    struct FsMount *mount;
+    int fd;
+    struct fstrim_range r;
+
+    slog("guest-fstrim called");
+
+    QTAILQ_INIT(&mounts);
+    if (!build_fs_mount_list(&mounts, errp)) {
+        return NULL;
+    }
+
+    response = g_malloc0(sizeof(*response));
+
+    QTAILQ_FOREACH(mount, &mounts, next) {
+        result = g_malloc0(sizeof(*result));
+        result->path = g_strdup(mount->dirname);
+
+        QAPI_LIST_PREPEND(response->paths, result);
+
+        fd = qga_open_cloexec(mount->dirname, O_RDONLY, 0);
+        if (fd == -1) {
+            result->error = g_strdup_printf("failed to open: %s",
+                                            strerror(errno));
+            continue;
+        }
+
+        /* We try to cull filesystems we know won't work in advance, but other
+         * filesystems may not implement fstrim for less obvious reasons.
+         * These will report EOPNOTSUPP; while in some other cases ENOTTY
+         * will be reported (e.g. CD-ROMs).
+         * Any other error means an unexpected error.
+         */
+        r.start = 0;
+        r.len = -1;
+        r.minlen = has_minimum ? minimum : 0;
+        ret = ioctl(fd, FITRIM, &r);
+        if (ret == -1) {
+            if (errno == ENOTTY || errno == EOPNOTSUPP) {
+                result->error = g_strdup("trim not supported");
+            } else {
+                result->error = g_strdup_printf("failed to trim: %s",
+                                                strerror(errno));
+            }
+            close(fd);
+            continue;
+        }
+
+        result->has_minimum = true;
+        result->minimum = r.minlen;
+        result->has_trimmed = true;
+        result->trimmed = r.len;
+        close(fd);
+    }
+
+    free_fs_mount_list(&mounts);
+    return response;
+}
+#endif /* CONFIG_FSTRIM */
 
 #define LINUX_SYS_STATE_FILE "/sys/power/state"
 #define SUSPEND_SUPPORTED 0
qga/commands-posix.c +0 -909

927 lines not on this page · raw message

Signature · 1 line
2.45.1
Reviewed-by on v2 4/22
#10 lore raw

Re: [PATCH v2 04/22] qga: move linux fs/disk command impls to commands-linux.c

Quoted 10 lines · Daniel P. Berrangé

The qmp_guest_{fstrim, get_fsinfo, get_disks} command impls in commands-posix.c are surrounded by '#ifdef __linux__' so should instead live in commands-linux.c

---
  qga/commands-linux.c | 904 ++++++++++++++++++++++++++++++++++++++++++
  qga/commands-posix.c | 909 -------------------------------------------
  2 files changed, 904 insertions(+), 909 deletions(-)
Reviewed-by on v2 4/22
#11 lore raw

Re: [PATCH v2 04/22] qga: move linux fs/disk command impls to commands-linux.c

In reply to Daniel P. Berrangé
Quoted 1,924 lines · Daniel P. Berrangé

The qmp_guest_{fstrim, get_fsinfo, get_disks} command impls in commands-posix.c are surrounded by '#ifdef __linux__' so should instead live in commands-linux.c

---
 qga/commands-linux.c | 904 ++++++++++++++++++++++++++++++++++++++++++
 qga/commands-posix.c | 909 -------------------------------------------
 2 files changed, 904 insertions(+), 909 deletions(-)
qga/commands-linux.c +42 -0
index 3fabf54882..084e6c9e85 100644
@@ -14,10 +14,21 @@
 #include "qemu/osdep.h"
 #include "qapi/error.h"
 #include "qga-qapi-commands.h"
+#include "qapi/error.h"
+#include "qapi/qmp/qerror.h"
 #include "commands-common.h"
 #include "cutils.h"
 #include <mntent.h>
 #include <sys/ioctl.h>
+#include <mntent.h>
+#include <linux/nvme_ioctl.h>
+#include "block/nvme.h"
+
+#ifdef CONFIG_LIBUDEV
+#include <libudev.h>
+#endif
+
+#include <sys/statvfs.h>
 
 #if defined(CONFIG_FSFREEZE) || defined(CONFIG_FSTRIM)
 static int dev_major_minor(const char *devpath,
@@ -286,6 +297,899 @@ int qmp_guest_fsfreeze_do_thaw(Error **errp)
 }
 #endif /* CONFIG_FSFREEZE */
 
+#if defined(CONFIG_FSFREEZE)
+
+static char *get_pci_driver(char const *syspath, int pathlen, Error
 **errp)
+{
+    char *path;
+    char *dpath;
+    char *driver = NULL;
+    char buf[PATH_MAX];
+    ssize_t len;
+
+    path = g_strndup(syspath, pathlen);
+    dpath = g_strdup_printf("%s/driver", path);
+    len = readlink(dpath, buf, sizeof(buf) - 1);
+    if (len != -1) {
+        buf[len] = 0;
+        driver = g_path_get_basename(buf);
+    }
+    g_free(dpath);
+    g_free(path);
+    return driver;
+}
+
+static int compare_uint(const void *_a, const void *_b)
+{
+    unsigned int a = *(unsigned int *)_a;
+    unsigned int b = *(unsigned int *)_b;
+
+    return a < b ? -1 : a > b ? 1 : 0;
+}
+
+/* Walk the specified sysfs and build a sorted list of host or ata

numbers */

qga/commands-linux.c +24 -0
+static int build_hosts(char const *syspath, char const *host, bool ata,
+                       unsigned int *hosts, int hosts_max, Error **errp)
+{
+    char *path;
+    DIR *dir;
+    struct dirent *entry;
+    int i = 0;
+
+    path = g_strndup(syspath, host - syspath);
+    dir = opendir(path);
+    if (!dir) {
+        error_setg_errno(errp, errno, "opendir(\"%s\")", path);
+        g_free(path);
+        return -1;
+    }
+
+    while (i < hosts_max) {
+        entry = readdir(dir);
+        if (!entry) {
+            break;
+        }
+        if (ata && sscanf(entry->d_name, "ata%d", hosts + i) == 1) {
+            ++i;
+        } else if (!ata && sscanf(entry->d_name, "host%d", hosts + i) ==

1) {

qga/commands-linux.c +89 -0
+            ++i;
+        }
+    }
+
+    qsort(hosts, i, sizeof(hosts[0]), compare_uint);
+
+    g_free(path);
+    closedir(dir);
+    return i;
+}
+
+/*
+ * Store disk device info for devices on the PCI bus.
+ * Returns true if information has been stored, or false for failure.
+ */
+static bool build_guest_fsinfo_for_pci_dev(char const *syspath,
+                                           GuestDiskAddress *disk,
+                                           Error **errp)
+{
+    unsigned int pci[4], host, hosts[8], tgt[3];
+    int i, nhosts = 0, pcilen;
+    GuestPCIAddress *pciaddr = disk->pci_controller;
+    bool has_ata = false, has_host = false, has_tgt = false;
+    char *p, *q, *driver = NULL;
+    bool ret = false;
+
+    p = strstr(syspath, "/devices/pci");
+    if (!p || sscanf(p + 12, "%*x:%*x/%x:%x:%x.%x%n",
+                     pci, pci + 1, pci + 2, pci + 3, &pcilen) < 4) {
+        g_debug("only pci device is supported: sysfs path '%s'", syspath);
+        return false;
+    }
+
+    p += 12 + pcilen;
+    while (true) {
+        driver = get_pci_driver(syspath, p - syspath, errp);
+        if (driver && (g_str_equal(driver, "ata_piix") ||
+                       g_str_equal(driver, "sym53c8xx") ||
+                       g_str_equal(driver, "virtio-pci") ||
+                       g_str_equal(driver, "ahci") ||
+                       g_str_equal(driver, "nvme") ||
+                       g_str_equal(driver, "xhci_hcd") ||
+                       g_str_equal(driver, "ehci-pci"))) {
+            break;
+        }
+
+        g_free(driver);
+        if (sscanf(p, "/%x:%x:%x.%x%n",
+                          pci, pci + 1, pci + 2, pci + 3, &pcilen) == 4) {
+            p += pcilen;
+            continue;
+        }
+
+        g_debug("unsupported driver or sysfs path '%s'", syspath);
+        return false;
+    }
+
+    p = strstr(syspath, "/target");
+    if (p && sscanf(p + 7, "%*u:%*u:%*u/%*u:%u:%u:%u",
+                    tgt, tgt + 1, tgt + 2) == 3) {
+        has_tgt = true;
+    }
+
+    p = strstr(syspath, "/ata");
+    if (p) {
+        q = p + 4;
+        has_ata = true;
+    } else {
+        p = strstr(syspath, "/host");
+        q = p + 5;
+    }
+    if (p && sscanf(q, "%u", &host) == 1) {
+        has_host = true;
+        nhosts = build_hosts(syspath, p, has_ata, hosts,
+                             ARRAY_SIZE(hosts), errp);
+        if (nhosts < 0) {
+            goto cleanup;
+        }
+    }
+
+    pciaddr->domain = pci[0];
+    pciaddr->bus = pci[1];
+    pciaddr->slot = pci[2];
+    pciaddr->function = pci[3];
+
+    if (strcmp(driver, "ata_piix") == 0) {
+        /* a host per ide bus, target*:0:<unit>:0 */
+        if (!has_host || !has_tgt) {
+            g_debug("invalid sysfs path '%s' (driver '%s')", syspath,

driver);

qga/commands-linux.c +18 -0
+            goto cleanup;
+        }
+        for (i = 0; i < nhosts; i++) {
+            if (host == hosts[i]) {
+                disk->bus_type = GUEST_DISK_BUS_TYPE_IDE;
+                disk->bus = i;
+                disk->unit = tgt[1];
+                break;
+            }
+        }
+        if (i >= nhosts) {
+            g_debug("no host for '%s' (driver '%s')", syspath, driver);
+            goto cleanup;
+        }
+    } else if (strcmp(driver, "sym53c8xx") == 0) {
+        /* scsi(LSI Logic): target*:0:<unit>:0 */
+        if (!has_tgt) {
+            g_debug("invalid sysfs path '%s' (driver '%s')", syspath,

driver);

qga/commands-linux.c +17 -0
+            goto cleanup;
+        }
+        disk->bus_type = GUEST_DISK_BUS_TYPE_SCSI;
+        disk->unit = tgt[1];
+    } else if (strcmp(driver, "virtio-pci") == 0) {
+        if (has_tgt) {
+            /* virtio-scsi: target*:0:0:<unit> */
+            disk->bus_type = GUEST_DISK_BUS_TYPE_SCSI;
+            disk->unit = tgt[2];
+        } else {
+            /* virtio-blk: 1 disk per 1 device */
+            disk->bus_type = GUEST_DISK_BUS_TYPE_VIRTIO;
+        }
+    } else if (strcmp(driver, "ahci") == 0) {
+        /* ahci: 1 host per 1 unit */
+        if (!has_host || !has_tgt) {
+            g_debug("invalid sysfs path '%s' (driver '%s')", syspath,

driver);

qga/commands-linux.c +16 -0
+            goto cleanup;
+        }
+        for (i = 0; i < nhosts; i++) {
+            if (host == hosts[i]) {
+                disk->unit = i;
+                disk->bus_type = GUEST_DISK_BUS_TYPE_SATA;
+                break;
+            }
+        }
+        if (i >= nhosts) {
+            g_debug("no host for '%s' (driver '%s')", syspath, driver);
+            goto cleanup;
+        }
+    } else if (strcmp(driver, "nvme") == 0) {
+        disk->bus_type = GUEST_DISK_BUS_TYPE_NVME;
+    } else if (strcmp(driver, "ehci-pci") == 0 || strcmp(driver,

"xhci_hcd") == 0) {

qga/commands-linux.c +92 -0
+        disk->bus_type = GUEST_DISK_BUS_TYPE_USB;
+    } else {
+        g_debug("unknown driver '%s' (sysfs path '%s')", driver, syspath);
+        goto cleanup;
+    }
+
+    ret = true;
+
+cleanup:
+    g_free(driver);
+    return ret;
+}
+
+/*
+ * Store disk device info for non-PCI virtio devices (for example s390x
+ * channel I/O devices). Returns true if information has been stored, or
+ * false for failure.
+ */
+static bool build_guest_fsinfo_for_nonpci_virtio(char const *syspath,
+                                                 GuestDiskAddress *disk,
+                                                 Error **errp)
+{
+    unsigned int tgt[3];
+    char *p;
+
+    if (!strstr(syspath, "/virtio") || !strstr(syspath, "/block")) {
+        g_debug("Unsupported virtio device '%s'", syspath);
+        return false;
+    }
+
+    p = strstr(syspath, "/target");
+    if (p && sscanf(p + 7, "%*u:%*u:%*u/%*u:%u:%u:%u",
+                    &tgt[0], &tgt[1], &tgt[2]) == 3) {
+        /* virtio-scsi: target*:0:<target>:<unit> */
+        disk->bus_type = GUEST_DISK_BUS_TYPE_SCSI;
+        disk->bus = tgt[0];
+        disk->target = tgt[1];
+        disk->unit = tgt[2];
+    } else {
+        /* virtio-blk: 1 disk per 1 device */
+        disk->bus_type = GUEST_DISK_BUS_TYPE_VIRTIO;
+    }
+
+    return true;
+}
+
+/*
+ * Store disk device info for CCW devices (s390x channel I/O devices).
+ * Returns true if information has been stored, or false for failure.
+ */
+static bool build_guest_fsinfo_for_ccw_dev(char const *syspath,
+                                           GuestDiskAddress *disk,
+                                           Error **errp)
+{
+    unsigned int cssid, ssid, subchno, devno;
+    char *p;
+
+    p = strstr(syspath, "/devices/css");
+    if (!p || sscanf(p + 12, "%*x/%x.%x.%x/%*x.%*x.%x/",
+                     &cssid, &ssid, &subchno, &devno) < 4) {
+        g_debug("could not parse ccw device sysfs path: %s", syspath);
+        return false;
+    }
+
+    disk->ccw_address = g_new0(GuestCCWAddress, 1);
+    disk->ccw_address->cssid = cssid;
+    disk->ccw_address->ssid = ssid;
+    disk->ccw_address->subchno = subchno;
+    disk->ccw_address->devno = devno;
+
+    if (strstr(p, "/virtio")) {
+        build_guest_fsinfo_for_nonpci_virtio(syspath, disk, errp);
+    }
+
+    return true;
+}
+
+/* Store disk device info specified by @sysfs into @fs */
+static void build_guest_fsinfo_for_real_device(char const *syspath,
+                                               GuestFilesystemInfo *fs,
+                                               Error **errp)
+{
+    GuestDiskAddress *disk;
+    GuestPCIAddress *pciaddr;
+    bool has_hwinf;
+#ifdef CONFIG_LIBUDEV
+    struct udev *udev = NULL;
+    struct udev_device *udevice = NULL;
+#endif
+
+    pciaddr = g_new0(GuestPCIAddress, 1);
+    pciaddr->domain = -1;                       /* -1 means field is

invalid */

qga/commands-linux.c +35 -0
+    pciaddr->bus = -1;
+    pciaddr->slot = -1;
+    pciaddr->function = -1;
+
+    disk = g_new0(GuestDiskAddress, 1);
+    disk->pci_controller = pciaddr;
+    disk->bus_type = GUEST_DISK_BUS_TYPE_UNKNOWN;
+
+#ifdef CONFIG_LIBUDEV
+    udev = udev_new();
+    udevice = udev_device_new_from_syspath(udev, syspath);
+    if (udev == NULL || udevice == NULL) {
+        g_debug("failed to query udev");
+    } else {
+        const char *devnode, *serial;
+        devnode = udev_device_get_devnode(udevice);
+        if (devnode != NULL) {
+            disk->dev = g_strdup(devnode);
+        }
+        serial = udev_device_get_property_value(udevice, "ID_SERIAL");
+        if (serial != NULL && *serial != 0) {
+            disk->serial = g_strdup(serial);
+        }
+    }
+
+    udev_unref(udev);
+    udev_device_unref(udevice);
+#endif
+
+    if (strstr(syspath, "/devices/pci")) {
+        has_hwinf = build_guest_fsinfo_for_pci_dev(syspath, disk, errp);
+    } else if (strstr(syspath, "/devices/css")) {
+        has_hwinf = build_guest_fsinfo_for_ccw_dev(syspath, disk, errp);
+    } else if (strstr(syspath, "/virtio")) {
+        has_hwinf = build_guest_fsinfo_for_nonpci_virtio(syspath, disk,

errp);

qga/commands-linux.c +17 -0
+    } else {
+        g_debug("Unsupported device type for '%s'", syspath);
+        has_hwinf = false;
+    }
+
+    if (has_hwinf || disk->dev || disk->serial) {
+        QAPI_LIST_PREPEND(fs->disk, disk);
+    } else {
+        qapi_free_GuestDiskAddress(disk);
+    }
+}
+
+static void build_guest_fsinfo_for_device(char const *devpath,
+                                          GuestFilesystemInfo *fs,
+                                          Error **errp);
+
+/* Store a list of slave devices of virtual volume specified by @syspath

into
+ * @fs */

qga/commands-linux.c +34 -0
+static void build_guest_fsinfo_for_virtual_device(char const *syspath,
+                                                  GuestFilesystemInfo *fs,
+                                                  Error **errp)
+{
+    Error *err = NULL;
+    DIR *dir;
+    char *dirpath;
+    struct dirent *entry;
+
+    dirpath = g_strdup_printf("%s/slaves", syspath);
+    dir = opendir(dirpath);
+    if (!dir) {
+        if (errno != ENOENT) {
+            error_setg_errno(errp, errno, "opendir(\"%s\")", dirpath);
+        }
+        g_free(dirpath);
+        return;
+    }
+
+    for (;;) {
+        errno = 0;
+        entry = readdir(dir);
+        if (entry == NULL) {
+            if (errno) {
+                error_setg_errno(errp, errno, "readdir(\"%s\")", dirpath);
+            }
+            break;
+        }
+
+        if (entry->d_type == DT_LNK) {
+            char *path;
+
+            g_debug(" slave device '%s'", entry->d_name);
+            path = g_strdup_printf("%s/slaves/%s", syspath,

entry->d_name);

qga/commands-linux.c +42 -0
+            build_guest_fsinfo_for_device(path, fs, &err);
+            g_free(path);
+
+            if (err) {
+                error_propagate(errp, err);
+                break;
+            }
+        }
+    }
+
+    g_free(dirpath);
+    closedir(dir);
+}
+
+static bool is_disk_virtual(const char *devpath, Error **errp)
+{
+    g_autofree char *syspath = realpath(devpath, NULL);
+
+    if (!syspath) {
+        error_setg_errno(errp, errno, "realpath(\"%s\")", devpath);
+        return false;
+    }
+    return strstr(syspath, "/devices/virtual/block/") != NULL;
+}
+
+/* Dispatch to functions for virtual/real device */
+static void build_guest_fsinfo_for_device(char const *devpath,
+                                          GuestFilesystemInfo *fs,
+                                          Error **errp)
+{
+    ERRP_GUARD();
+    g_autofree char *syspath = NULL;
+    bool is_virtual = false;
+
+    syspath = realpath(devpath, NULL);
+    if (!syspath) {
+        if (errno != ENOENT) {
+            error_setg_errno(errp, errno, "realpath(\"%s\")", devpath);
+            return;
+        }
+
+        /* ENOENT: This devpath may not exist because of container config

*/

qga/commands-linux.c +29 -0
+        if (!fs->name) {
+            fs->name = g_path_get_basename(devpath);
+        }
+        return;
+    }
+
+    if (!fs->name) {
+        fs->name = g_path_get_basename(syspath);
+    }
+
+    g_debug("  parse sysfs path '%s'", syspath);
+    is_virtual = is_disk_virtual(syspath, errp);
+    if (*errp != NULL) {
+        return;
+    }
+    if (is_virtual) {
+        build_guest_fsinfo_for_virtual_device(syspath, fs, errp);
+    } else {
+        build_guest_fsinfo_for_real_device(syspath, fs, errp);
+    }
+}
+
+#ifdef CONFIG_LIBUDEV
+
+/*
+ * Wrapper around build_guest_fsinfo_for_device() for getting just
+ * the disk address.
+ */
+static GuestDiskAddress *get_disk_address(const char *syspath, Error

**errp)

qga/commands-linux.c +31 -0
+{
+    g_autoptr(GuestFilesystemInfo) fs = NULL;
+
+    fs = g_new0(GuestFilesystemInfo, 1);
+    build_guest_fsinfo_for_device(syspath, fs, errp);
+    if (fs->disk != NULL) {
+        return g_steal_pointer(&fs->disk->value);
+    }
+    return NULL;
+}
+
+static char *get_alias_for_syspath(const char *syspath)
+{
+    struct udev *udev = NULL;
+    struct udev_device *udevice = NULL;
+    char *ret = NULL;
+
+    udev = udev_new();
+    if (udev == NULL) {
+        g_debug("failed to query udev");
+        goto out;
+    }
+    udevice = udev_device_new_from_syspath(udev, syspath);
+    if (udevice == NULL) {
+        g_debug("failed to query udev for path: %s", syspath);
+        goto out;
+    } else {
+        const char *alias = udev_device_get_property_value(
+            udevice, "DM_NAME");
+        /*
+         * NULL means there was an error and empty string means there is

no
+ * alias. In case of no alias we return NULL instead of empty
string.

qga/commands-linux.c +118 -0
+         */
+        if (alias == NULL) {
+            g_debug("failed to query udev for device alias for: %s",
+                syspath);
+        } else if (*alias != 0) {
+            ret = g_strdup(alias);
+        }
+    }
+
+out:
+    udev_unref(udev);
+    udev_device_unref(udevice);
+    return ret;
+}
+
+static char *get_device_for_syspath(const char *syspath)
+{
+    struct udev *udev = NULL;
+    struct udev_device *udevice = NULL;
+    char *ret = NULL;
+
+    udev = udev_new();
+    if (udev == NULL) {
+        g_debug("failed to query udev");
+        goto out;
+    }
+    udevice = udev_device_new_from_syspath(udev, syspath);
+    if (udevice == NULL) {
+        g_debug("failed to query udev for path: %s", syspath);
+        goto out;
+    } else {
+        ret = g_strdup(udev_device_get_devnode(udevice));
+    }
+
+out:
+    udev_unref(udev);
+    udev_device_unref(udevice);
+    return ret;
+}
+
+static void get_disk_deps(const char *disk_dir, GuestDiskInfo *disk)
+{
+    g_autofree char *deps_dir = NULL;
+    const gchar *dep;
+    GDir *dp_deps = NULL;
+
+    /* List dependent disks */
+    deps_dir = g_strdup_printf("%s/slaves", disk_dir);
+    g_debug("  listing entries in: %s", deps_dir);
+    dp_deps = g_dir_open(deps_dir, 0, NULL);
+    if (dp_deps == NULL) {
+        g_debug("failed to list entries in %s", deps_dir);
+        return;
+    }
+    disk->has_dependencies = true;
+    while ((dep = g_dir_read_name(dp_deps)) != NULL) {
+        g_autofree char *dep_dir = NULL;
+        char *dev_name;
+
+        /* Add dependent disks */
+        dep_dir = g_strdup_printf("%s/%s", deps_dir, dep);
+        dev_name = get_device_for_syspath(dep_dir);
+        if (dev_name != NULL) {
+            g_debug("  adding dependent device: %s", dev_name);
+            QAPI_LIST_PREPEND(disk->dependencies, dev_name);
+        }
+    }
+    g_dir_close(dp_deps);
+}
+
+/*
+ * Detect partitions subdirectory, name is "<disk_name><number>" or
+ * "<disk_name>p<number>"
+ *
+ * @disk_name -- last component of /sys path (e.g. sda)
+ * @disk_dir -- sys path of the disk (e.g. /sys/block/sda)
+ * @disk_dev -- device node of the disk (e.g. /dev/sda)
+ */
+static GuestDiskInfoList *get_disk_partitions(
+    GuestDiskInfoList *list,
+    const char *disk_name, const char *disk_dir,
+    const char *disk_dev)
+{
+    GuestDiskInfoList *ret = list;
+    struct dirent *de_disk;
+    DIR *dp_disk = NULL;
+    size_t len = strlen(disk_name);
+
+    dp_disk = opendir(disk_dir);
+    while ((de_disk = readdir(dp_disk)) != NULL) {
+        g_autofree char *partition_dir = NULL;
+        char *dev_name;
+        GuestDiskInfo *partition;
+
+        if (!(de_disk->d_type & DT_DIR)) {
+            continue;
+        }
+
+        if (!(strncmp(disk_name, de_disk->d_name, len) == 0 &&
+            ((*(de_disk->d_name + len) == 'p' &&
+            isdigit(*(de_disk->d_name + len + 1))) ||
+                isdigit(*(de_disk->d_name + len))))) {
+            continue;
+        }
+
+        partition_dir = g_strdup_printf("%s/%s",
+            disk_dir, de_disk->d_name);
+        dev_name = get_device_for_syspath(partition_dir);
+        if (dev_name == NULL) {
+            g_debug("Failed to get device name for syspath: %s",
+                disk_dir);
+            continue;
+        }
+        partition = g_new0(GuestDiskInfo, 1);
+        partition->name = dev_name;
+        partition->partition = true;
+        partition->has_dependencies = true;
+        /* Add parent disk as dependent for easier tracking of hierarchy

*/

qga/commands-linux.c +26 -0
+        QAPI_LIST_PREPEND(partition->dependencies, g_strdup(disk_dev));
+
+        QAPI_LIST_PREPEND(ret, partition);
+    }
+    closedir(dp_disk);
+
+    return ret;
+}
+
+static void get_nvme_smart(GuestDiskInfo *disk)
+{
+    int fd;
+    GuestNVMeSmart *smart;
+    NvmeSmartLog log = {0};
+    struct nvme_admin_cmd cmd = {
+        .opcode = NVME_ADM_CMD_GET_LOG_PAGE,
+        .nsid = NVME_NSID_BROADCAST,
+        .addr = (uintptr_t)&log,
+        .data_len = sizeof(log),
+        .cdw10 = NVME_LOG_SMART_INFO | (1 << 15) /* RAE bit */
+                 | (((sizeof(log) >> 2) - 1) << 16)
+    };
+
+    fd = qga_open_cloexec(disk->name, O_RDONLY, 0);
+    if (fd == -1) {
+        g_debug("Failed to open device: %s: %s", disk->name,

g_strerror(errno));

qga/commands-linux.c +5 -0
+        return;
+    }
+
+    if (ioctl(fd, NVME_IOCTL_ADMIN_CMD, &cmd)) {
+        g_debug("Failed to get smart: %s: %s", disk->name,

g_strerror(errno));

qga/commands-linux.c +10 -0
+        close(fd);
+        return;
+    }
+
+    disk->smart = g_new0(GuestDiskSmart, 1);
+    disk->smart->type = GUEST_DISK_BUS_TYPE_NVME;
+
+    smart = &disk->smart->u.nvme;
+    smart->critical_warning = log.critical_warning;
+    smart->temperature = lduw_le_p(&log.temperature); /* unaligned field

*/

qga/commands-linux.c +10 -0
+    smart->available_spare = log.available_spare;
+    smart->available_spare_threshold = log.available_spare_threshold;
+    smart->percentage_used = log.percentage_used;
+    smart->data_units_read_lo = le64_to_cpu(log.data_units_read[0]);
+    smart->data_units_read_hi = le64_to_cpu(log.data_units_read[1]);
+    smart->data_units_written_lo = le64_to_cpu(log.data_units_written[0]);
+    smart->data_units_written_hi = le64_to_cpu(log.data_units_written[1]);
+    smart->host_read_commands_lo = le64_to_cpu(log.host_read_commands[0]);
+    smart->host_read_commands_hi = le64_to_cpu(log.host_read_commands[1]);
+    smart->host_write_commands_lo =

le64_to_cpu(log.host_write_commands[0]);
+ smart->host_write_commands_hi =
le64_to_cpu(log.host_write_commands[1]);
+ smart->controller_busy_time_lo =
le64_to_cpu(log.controller_busy_time[0]);
+ smart->controller_busy_time_hi =
le64_to_cpu(log.controller_busy_time[1]);

qga/commands-linux.c +35 -0
+    smart->power_cycles_lo = le64_to_cpu(log.power_cycles[0]);
+    smart->power_cycles_hi = le64_to_cpu(log.power_cycles[1]);
+    smart->power_on_hours_lo = le64_to_cpu(log.power_on_hours[0]);
+    smart->power_on_hours_hi = le64_to_cpu(log.power_on_hours[1]);
+    smart->unsafe_shutdowns_lo = le64_to_cpu(log.unsafe_shutdowns[0]);
+    smart->unsafe_shutdowns_hi = le64_to_cpu(log.unsafe_shutdowns[1]);
+    smart->media_errors_lo = le64_to_cpu(log.media_errors[0]);
+    smart->media_errors_hi = le64_to_cpu(log.media_errors[1]);
+    smart->number_of_error_log_entries_lo =
+        le64_to_cpu(log.number_of_error_log_entries[0]);
+    smart->number_of_error_log_entries_hi =
+        le64_to_cpu(log.number_of_error_log_entries[1]);
+
+    close(fd);
+}
+
+static void get_disk_smart(GuestDiskInfo *disk)
+{
+    if (disk->address
+        && (disk->address->bus_type == GUEST_DISK_BUS_TYPE_NVME)) {
+        get_nvme_smart(disk);
+    }
+}
+
+GuestDiskInfoList *qmp_guest_get_disks(Error **errp)
+{
+    GuestDiskInfoList *ret = NULL;
+    GuestDiskInfo *disk;
+    DIR *dp = NULL;
+    struct dirent *de = NULL;
+
+    g_debug("listing /sys/block directory");
+    dp = opendir("/sys/block");
+    if (dp == NULL) {
+        error_setg_errno(errp, errno, "Can't open directory

\"/sys/block\"");

qga/commands-linux.c +178 -0
+        return NULL;
+    }
+    while ((de = readdir(dp)) != NULL) {
+        g_autofree char *disk_dir = NULL, *line = NULL,
+            *size_path = NULL;
+        char *dev_name;
+        Error *local_err = NULL;
+        if (de->d_type != DT_LNK) {
+            g_debug("  skipping entry: %s", de->d_name);
+            continue;
+        }
+
+        /* Check size and skip zero-sized disks */
+        g_debug("  checking disk size");
+        size_path = g_strdup_printf("/sys/block/%s/size", de->d_name);
+        if (!g_file_get_contents(size_path, &line, NULL, NULL)) {
+            g_debug("  failed to read disk size");
+            continue;
+        }
+        if (g_strcmp0(line, "0\n") == 0) {
+            g_debug("  skipping zero-sized disk");
+            continue;
+        }
+
+        g_debug("  adding %s", de->d_name);
+        disk_dir = g_strdup_printf("/sys/block/%s", de->d_name);
+        dev_name = get_device_for_syspath(disk_dir);
+        if (dev_name == NULL) {
+            g_debug("Failed to get device name for syspath: %s",
+                disk_dir);
+            continue;
+        }
+        disk = g_new0(GuestDiskInfo, 1);
+        disk->name = dev_name;
+        disk->partition = false;
+        disk->alias = get_alias_for_syspath(disk_dir);
+        QAPI_LIST_PREPEND(ret, disk);
+
+        /* Get address for non-virtual devices */
+        bool is_virtual = is_disk_virtual(disk_dir, &local_err);
+        if (local_err != NULL) {
+            g_debug("  failed to check disk path, ignoring error: %s",
+                error_get_pretty(local_err));
+            error_free(local_err);
+            local_err = NULL;
+            /* Don't try to get the address */
+            is_virtual = true;
+        }
+        if (!is_virtual) {
+            disk->address = get_disk_address(disk_dir, &local_err);
+            if (local_err != NULL) {
+                g_debug("  failed to get device info, ignoring error: %s",
+                    error_get_pretty(local_err));
+                error_free(local_err);
+                local_err = NULL;
+            }
+        }
+
+        get_disk_deps(disk_dir, disk);
+        get_disk_smart(disk);
+        ret = get_disk_partitions(ret, de->d_name, disk_dir, dev_name);
+    }
+
+    closedir(dp);
+
+    return ret;
+}
+
+#else
+
+GuestDiskInfoList *qmp_guest_get_disks(Error **errp)
+{
+    error_setg(errp, QERR_UNSUPPORTED);
+    return NULL;
+}
+
+#endif
+
+/* Return a list of the disk device(s)' info which @mount lies on */
+static GuestFilesystemInfo *build_guest_fsinfo(struct FsMount *mount,
+                                               Error **errp)
+{
+    GuestFilesystemInfo *fs = g_malloc0(sizeof(*fs));
+    struct statvfs buf;
+    unsigned long used, nonroot_total, fr_size;
+    char *devpath = g_strdup_printf("/sys/dev/block/%u:%u",
+                                    mount->devmajor, mount->devminor);
+
+    fs->mountpoint = g_strdup(mount->dirname);
+    fs->type = g_strdup(mount->devtype);
+    build_guest_fsinfo_for_device(devpath, fs, errp);
+
+    if (statvfs(fs->mountpoint, &buf) == 0) {
+        fr_size = buf.f_frsize;
+        used = buf.f_blocks - buf.f_bfree;
+        nonroot_total = used + buf.f_bavail;
+        fs->used_bytes = used * fr_size;
+        fs->total_bytes = nonroot_total * fr_size;
+        fs->total_bytes_privileged = buf.f_blocks * fr_size;
+
+        fs->has_total_bytes = true;
+        fs->has_total_bytes_privileged = true;
+        fs->has_used_bytes = true;
+    }
+
+    g_free(devpath);
+
+    return fs;
+}
+
+GuestFilesystemInfoList *qmp_guest_get_fsinfo(Error **errp)
+{
+    FsMountList mounts;
+    struct FsMount *mount;
+    GuestFilesystemInfoList *ret = NULL;
+    Error *local_err = NULL;
+
+    QTAILQ_INIT(&mounts);
+    if (!build_fs_mount_list(&mounts, &local_err)) {
+        error_propagate(errp, local_err);
+        return NULL;
+    }
+
+    QTAILQ_FOREACH(mount, &mounts, next) {
+        g_debug("Building guest fsinfo for '%s'", mount->dirname);
+
+        QAPI_LIST_PREPEND(ret, build_guest_fsinfo(mount, &local_err));
+        if (local_err) {
+            error_propagate(errp, local_err);
+            qapi_free_GuestFilesystemInfoList(ret);
+            ret = NULL;
+            break;
+        }
+    }
+
+    free_fs_mount_list(&mounts);
+    return ret;
+}
+#endif /* CONFIG_FSFREEZE */
+
+#if defined(CONFIG_FSTRIM)
+/*
+ * Walk list of mounted file systems in the guest, and trim them.
+ */
+GuestFilesystemTrimResponse *
+qmp_guest_fstrim(bool has_minimum, int64_t minimum, Error **errp)
+{
+    GuestFilesystemTrimResponse *response;
+    GuestFilesystemTrimResult *result;
+    int ret = 0;
+    FsMountList mounts;
+    struct FsMount *mount;
+    int fd;
+    struct fstrim_range r;
+
+    slog("guest-fstrim called");
+
+    QTAILQ_INIT(&mounts);
+    if (!build_fs_mount_list(&mounts, errp)) {
+        return NULL;
+    }
+
+    response = g_malloc0(sizeof(*response));
+
+    QTAILQ_FOREACH(mount, &mounts, next) {
+        result = g_malloc0(sizeof(*result));
+        result->path = g_strdup(mount->dirname);
+
+        QAPI_LIST_PREPEND(response->paths, result);
+
+        fd = qga_open_cloexec(mount->dirname, O_RDONLY, 0);
+        if (fd == -1) {
+            result->error = g_strdup_printf("failed to open: %s",
+                                            strerror(errno));
+            continue;
+        }
+
+        /* We try to cull filesystems we know won't work in advance, but

other

qga/commands-linux.c +31 -0
+         * filesystems may not implement fstrim for less obvious reasons.
+         * These will report EOPNOTSUPP; while in some other cases ENOTTY
+         * will be reported (e.g. CD-ROMs).
+         * Any other error means an unexpected error.
+         */
+        r.start = 0;
+        r.len = -1;
+        r.minlen = has_minimum ? minimum : 0;
+        ret = ioctl(fd, FITRIM, &r);
+        if (ret == -1) {
+            if (errno == ENOTTY || errno == EOPNOTSUPP) {
+                result->error = g_strdup("trim not supported");
+            } else {
+                result->error = g_strdup_printf("failed to trim: %s",
+                                                strerror(errno));
+            }
+            close(fd);
+            continue;
+        }
+
+        result->has_minimum = true;
+        result->minimum = r.minlen;
+        result->has_trimmed = true;
+        result->trimmed = r.len;
+        close(fd);
+    }
+
+    free_fs_mount_list(&mounts);
+    return response;
+}
+#endif /* CONFIG_FSTRIM */
 
 #define LINUX_SYS_STATE_FILE "/sys/power/state"
 #define SUSPEND_SUPPORTED 0
qga/commands-posix.c +0 -44
index ef21da63be..98aafc45f3 100644
@@ -24,23 +24,12 @@
 #include "qemu/base64.h"
 #include "qemu/cutils.h"
 #include "commands-common.h"
-#include "block/nvme.h"
 #include "cutils.h"
 
 #ifdef HAVE_UTMPX
 #include <utmpx.h>
 #endif
 
-#if defined(__linux__)
-#include <mntent.h>
-#include <sys/statvfs.h>
-#include <linux/nvme_ioctl.h>
-
-#ifdef CONFIG_LIBUDEV
-#include <libudev.h>
-#endif
-#endif
-
 #ifdef HAVE_GETIFADDRS
 #include <arpa/inet.h>
 #include <sys/socket.h>
@@ -842,904 +831,6 @@ static void guest_fsfreeze_cleanup(void)
 }
 #endif
 
-/* linux-specific implementations. avoid this if at all possible. */
-#if defined(__linux__)
-#if defined(CONFIG_FSFREEZE)
-
-static char *get_pci_driver(char const *syspath, int pathlen, Error
 **errp)
-{
-    char *path;
-    char *dpath;
-    char *driver = NULL;
-    char buf[PATH_MAX];
-    ssize_t len;
-
-    path = g_strndup(syspath, pathlen);
-    dpath = g_strdup_printf("%s/driver", path);
-    len = readlink(dpath, buf, sizeof(buf) - 1);
-    if (len != -1) {
-        buf[len] = 0;
-        driver = g_path_get_basename(buf);
-    }
-    g_free(dpath);
-    g_free(path);
-    return driver;
-}
-
-static int compare_uint(const void *_a, const void *_b)
-{
-    unsigned int a = *(unsigned int *)_a;
-    unsigned int b = *(unsigned int *)_b;
-
-    return a < b ? -1 : a > b ? 1 : 0;
-}
-
-/* Walk the specified sysfs and build a sorted list of host or ata

numbers */

qga/commands-posix.c +0 -24
-static int build_hosts(char const *syspath, char const *host, bool ata,
-                       unsigned int *hosts, int hosts_max, Error **errp)
-{
-    char *path;
-    DIR *dir;
-    struct dirent *entry;
-    int i = 0;
-
-    path = g_strndup(syspath, host - syspath);
-    dir = opendir(path);
-    if (!dir) {
-        error_setg_errno(errp, errno, "opendir(\"%s\")", path);
-        g_free(path);
-        return -1;
-    }
-
-    while (i < hosts_max) {
-        entry = readdir(dir);
-        if (!entry) {
-            break;
-        }
-        if (ata && sscanf(entry->d_name, "ata%d", hosts + i) == 1) {
-            ++i;
-        } else if (!ata && sscanf(entry->d_name, "host%d", hosts + i) ==

1) {

qga/commands-posix.c +0 -89
-            ++i;
-        }
-    }
-
-    qsort(hosts, i, sizeof(hosts[0]), compare_uint);
-
-    g_free(path);
-    closedir(dir);
-    return i;
-}
-
-/*
- * Store disk device info for devices on the PCI bus.
- * Returns true if information has been stored, or false for failure.
- */
-static bool build_guest_fsinfo_for_pci_dev(char const *syspath,
-                                           GuestDiskAddress *disk,
-                                           Error **errp)
-{
-    unsigned int pci[4], host, hosts[8], tgt[3];
-    int i, nhosts = 0, pcilen;
-    GuestPCIAddress *pciaddr = disk->pci_controller;
-    bool has_ata = false, has_host = false, has_tgt = false;
-    char *p, *q, *driver = NULL;
-    bool ret = false;
-
-    p = strstr(syspath, "/devices/pci");
-    if (!p || sscanf(p + 12, "%*x:%*x/%x:%x:%x.%x%n",
-                     pci, pci + 1, pci + 2, pci + 3, &pcilen) < 4) {
-        g_debug("only pci device is supported: sysfs path '%s'", syspath);
-        return false;
-    }
-
-    p += 12 + pcilen;
-    while (true) {
-        driver = get_pci_driver(syspath, p - syspath, errp);
-        if (driver && (g_str_equal(driver, "ata_piix") ||
-                       g_str_equal(driver, "sym53c8xx") ||
-                       g_str_equal(driver, "virtio-pci") ||
-                       g_str_equal(driver, "ahci") ||
-                       g_str_equal(driver, "nvme") ||
-                       g_str_equal(driver, "xhci_hcd") ||
-                       g_str_equal(driver, "ehci-pci"))) {
-            break;
-        }
-
-        g_free(driver);
-        if (sscanf(p, "/%x:%x:%x.%x%n",
-                          pci, pci + 1, pci + 2, pci + 3, &pcilen) == 4) {
-            p += pcilen;
-            continue;
-        }
-
-        g_debug("unsupported driver or sysfs path '%s'", syspath);
-        return false;
-    }
-
-    p = strstr(syspath, "/target");
-    if (p && sscanf(p + 7, "%*u:%*u:%*u/%*u:%u:%u:%u",
-                    tgt, tgt + 1, tgt + 2) == 3) {
-        has_tgt = true;
-    }
-
-    p = strstr(syspath, "/ata");
-    if (p) {
-        q = p + 4;
-        has_ata = true;
-    } else {
-        p = strstr(syspath, "/host");
-        q = p + 5;
-    }
-    if (p && sscanf(q, "%u", &host) == 1) {
-        has_host = true;
-        nhosts = build_hosts(syspath, p, has_ata, hosts,
-                             ARRAY_SIZE(hosts), errp);
-        if (nhosts < 0) {
-            goto cleanup;
-        }
-    }
-
-    pciaddr->domain = pci[0];
-    pciaddr->bus = pci[1];
-    pciaddr->slot = pci[2];
-    pciaddr->function = pci[3];
-
-    if (strcmp(driver, "ata_piix") == 0) {
-        /* a host per ide bus, target*:0:<unit>:0 */
-        if (!has_host || !has_tgt) {
-            g_debug("invalid sysfs path '%s' (driver '%s')", syspath,

driver);

qga/commands-posix.c +0 -18
-            goto cleanup;
-        }
-        for (i = 0; i < nhosts; i++) {
-            if (host == hosts[i]) {
-                disk->bus_type = GUEST_DISK_BUS_TYPE_IDE;
-                disk->bus = i;
-                disk->unit = tgt[1];
-                break;
-            }
-        }
-        if (i >= nhosts) {
-            g_debug("no host for '%s' (driver '%s')", syspath, driver);
-            goto cleanup;
-        }
-    } else if (strcmp(driver, "sym53c8xx") == 0) {
-        /* scsi(LSI Logic): target*:0:<unit>:0 */
-        if (!has_tgt) {
-            g_debug("invalid sysfs path '%s' (driver '%s')", syspath,

driver);

qga/commands-posix.c +0 -17
-            goto cleanup;
-        }
-        disk->bus_type = GUEST_DISK_BUS_TYPE_SCSI;
-        disk->unit = tgt[1];
-    } else if (strcmp(driver, "virtio-pci") == 0) {
-        if (has_tgt) {
-            /* virtio-scsi: target*:0:0:<unit> */
-            disk->bus_type = GUEST_DISK_BUS_TYPE_SCSI;
-            disk->unit = tgt[2];
-        } else {
-            /* virtio-blk: 1 disk per 1 device */
-            disk->bus_type = GUEST_DISK_BUS_TYPE_VIRTIO;
-        }
-    } else if (strcmp(driver, "ahci") == 0) {
-        /* ahci: 1 host per 1 unit */
-        if (!has_host || !has_tgt) {
-            g_debug("invalid sysfs path '%s' (driver '%s')", syspath,

driver);

qga/commands-posix.c +0 -16
-            goto cleanup;
-        }
-        for (i = 0; i < nhosts; i++) {
-            if (host == hosts[i]) {
-                disk->unit = i;
-                disk->bus_type = GUEST_DISK_BUS_TYPE_SATA;
-                break;
-            }
-        }
-        if (i >= nhosts) {
-            g_debug("no host for '%s' (driver '%s')", syspath, driver);
-            goto cleanup;
-        }
-    } else if (strcmp(driver, "nvme") == 0) {
-        disk->bus_type = GUEST_DISK_BUS_TYPE_NVME;
-    } else if (strcmp(driver, "ehci-pci") == 0 || strcmp(driver,

"xhci_hcd") == 0) {

qga/commands-posix.c +0 -92
-        disk->bus_type = GUEST_DISK_BUS_TYPE_USB;
-    } else {
-        g_debug("unknown driver '%s' (sysfs path '%s')", driver, syspath);
-        goto cleanup;
-    }
-
-    ret = true;
-
-cleanup:
-    g_free(driver);
-    return ret;
-}
-
-/*
- * Store disk device info for non-PCI virtio devices (for example s390x
- * channel I/O devices). Returns true if information has been stored, or
- * false for failure.
- */
-static bool build_guest_fsinfo_for_nonpci_virtio(char const *syspath,
-                                                 GuestDiskAddress *disk,
-                                                 Error **errp)
-{
-    unsigned int tgt[3];
-    char *p;
-
-    if (!strstr(syspath, "/virtio") || !strstr(syspath, "/block")) {
-        g_debug("Unsupported virtio device '%s'", syspath);
-        return false;
-    }
-
-    p = strstr(syspath, "/target");
-    if (p && sscanf(p + 7, "%*u:%*u:%*u/%*u:%u:%u:%u",
-                    &tgt[0], &tgt[1], &tgt[2]) == 3) {
-        /* virtio-scsi: target*:0:<target>:<unit> */
-        disk->bus_type = GUEST_DISK_BUS_TYPE_SCSI;
-        disk->bus = tgt[0];
-        disk->target = tgt[1];
-        disk->unit = tgt[2];
-    } else {
-        /* virtio-blk: 1 disk per 1 device */
-        disk->bus_type = GUEST_DISK_BUS_TYPE_VIRTIO;
-    }
-
-    return true;
-}
-
-/*
- * Store disk device info for CCW devices (s390x channel I/O devices).
- * Returns true if information has been stored, or false for failure.
- */
-static bool build_guest_fsinfo_for_ccw_dev(char const *syspath,
-                                           GuestDiskAddress *disk,
-                                           Error **errp)
-{
-    unsigned int cssid, ssid, subchno, devno;
-    char *p;
-
-    p = strstr(syspath, "/devices/css");
-    if (!p || sscanf(p + 12, "%*x/%x.%x.%x/%*x.%*x.%x/",
-                     &cssid, &ssid, &subchno, &devno) < 4) {
-        g_debug("could not parse ccw device sysfs path: %s", syspath);
-        return false;
-    }
-
-    disk->ccw_address = g_new0(GuestCCWAddress, 1);
-    disk->ccw_address->cssid = cssid;
-    disk->ccw_address->ssid = ssid;
-    disk->ccw_address->subchno = subchno;
-    disk->ccw_address->devno = devno;
-
-    if (strstr(p, "/virtio")) {
-        build_guest_fsinfo_for_nonpci_virtio(syspath, disk, errp);
-    }
-
-    return true;
-}
-
-/* Store disk device info specified by @sysfs into @fs */
-static void build_guest_fsinfo_for_real_device(char const *syspath,
-                                               GuestFilesystemInfo *fs,
-                                               Error **errp)
-{
-    GuestDiskAddress *disk;
-    GuestPCIAddress *pciaddr;
-    bool has_hwinf;
-#ifdef CONFIG_LIBUDEV
-    struct udev *udev = NULL;
-    struct udev_device *udevice = NULL;
-#endif
-
-    pciaddr = g_new0(GuestPCIAddress, 1);
-    pciaddr->domain = -1;                       /* -1 means field is

invalid */

qga/commands-posix.c +0 -35
-    pciaddr->bus = -1;
-    pciaddr->slot = -1;
-    pciaddr->function = -1;
-
-    disk = g_new0(GuestDiskAddress, 1);
-    disk->pci_controller = pciaddr;
-    disk->bus_type = GUEST_DISK_BUS_TYPE_UNKNOWN;
-
-#ifdef CONFIG_LIBUDEV
-    udev = udev_new();
-    udevice = udev_device_new_from_syspath(udev, syspath);
-    if (udev == NULL || udevice == NULL) {
-        g_debug("failed to query udev");
-    } else {
-        const char *devnode, *serial;
-        devnode = udev_device_get_devnode(udevice);
-        if (devnode != NULL) {
-            disk->dev = g_strdup(devnode);
-        }
-        serial = udev_device_get_property_value(udevice, "ID_SERIAL");
-        if (serial != NULL && *serial != 0) {
-            disk->serial = g_strdup(serial);
-        }
-    }
-
-    udev_unref(udev);
-    udev_device_unref(udevice);
-#endif
-
-    if (strstr(syspath, "/devices/pci")) {
-        has_hwinf = build_guest_fsinfo_for_pci_dev(syspath, disk, errp);
-    } else if (strstr(syspath, "/devices/css")) {
-        has_hwinf = build_guest_fsinfo_for_ccw_dev(syspath, disk, errp);
-    } else if (strstr(syspath, "/virtio")) {
-        has_hwinf = build_guest_fsinfo_for_nonpci_virtio(syspath, disk,

errp);

qga/commands-posix.c +0 -17
-    } else {
-        g_debug("Unsupported device type for '%s'", syspath);
-        has_hwinf = false;
-    }
-
-    if (has_hwinf || disk->dev || disk->serial) {
-        QAPI_LIST_PREPEND(fs->disk, disk);
-    } else {
-        qapi_free_GuestDiskAddress(disk);
-    }
-}
-
-static void build_guest_fsinfo_for_device(char const *devpath,
-                                          GuestFilesystemInfo *fs,
-                                          Error **errp);
-
-/* Store a list of slave devices of virtual volume specified by @syspath

into
- * @fs */

qga/commands-posix.c +0 -34
-static void build_guest_fsinfo_for_virtual_device(char const *syspath,
-                                                  GuestFilesystemInfo *fs,
-                                                  Error **errp)
-{
-    Error *err = NULL;
-    DIR *dir;
-    char *dirpath;
-    struct dirent *entry;
-
-    dirpath = g_strdup_printf("%s/slaves", syspath);
-    dir = opendir(dirpath);
-    if (!dir) {
-        if (errno != ENOENT) {
-            error_setg_errno(errp, errno, "opendir(\"%s\")", dirpath);
-        }
-        g_free(dirpath);
-        return;
-    }
-
-    for (;;) {
-        errno = 0;
-        entry = readdir(dir);
-        if (entry == NULL) {
-            if (errno) {
-                error_setg_errno(errp, errno, "readdir(\"%s\")", dirpath);
-            }
-            break;
-        }
-
-        if (entry->d_type == DT_LNK) {
-            char *path;
-
-            g_debug(" slave device '%s'", entry->d_name);
-            path = g_strdup_printf("%s/slaves/%s", syspath,

entry->d_name);

qga/commands-posix.c +0 -42
-            build_guest_fsinfo_for_device(path, fs, &err);
-            g_free(path);
-
-            if (err) {
-                error_propagate(errp, err);
-                break;
-            }
-        }
-    }
-
-    g_free(dirpath);
-    closedir(dir);
-}
-
-static bool is_disk_virtual(const char *devpath, Error **errp)
-{
-    g_autofree char *syspath = realpath(devpath, NULL);
-
-    if (!syspath) {
-        error_setg_errno(errp, errno, "realpath(\"%s\")", devpath);
-        return false;
-    }
-    return strstr(syspath, "/devices/virtual/block/") != NULL;
-}
-
-/* Dispatch to functions for virtual/real device */
-static void build_guest_fsinfo_for_device(char const *devpath,
-                                          GuestFilesystemInfo *fs,
-                                          Error **errp)
-{
-    ERRP_GUARD();
-    g_autofree char *syspath = NULL;
-    bool is_virtual = false;
-
-    syspath = realpath(devpath, NULL);
-    if (!syspath) {
-        if (errno != ENOENT) {
-            error_setg_errno(errp, errno, "realpath(\"%s\")", devpath);
-            return;
-        }
-
-        /* ENOENT: This devpath may not exist because of container config

*/

qga/commands-posix.c +0 -29
-        if (!fs->name) {
-            fs->name = g_path_get_basename(devpath);
-        }
-        return;
-    }
-
-    if (!fs->name) {
-        fs->name = g_path_get_basename(syspath);
-    }
-
-    g_debug("  parse sysfs path '%s'", syspath);
-    is_virtual = is_disk_virtual(syspath, errp);
-    if (*errp != NULL) {
-        return;
-    }
-    if (is_virtual) {
-        build_guest_fsinfo_for_virtual_device(syspath, fs, errp);
-    } else {
-        build_guest_fsinfo_for_real_device(syspath, fs, errp);
-    }
-}
-
-#ifdef CONFIG_LIBUDEV
-
-/*
- * Wrapper around build_guest_fsinfo_for_device() for getting just
- * the disk address.
- */
-static GuestDiskAddress *get_disk_address(const char *syspath, Error

**errp)

qga/commands-posix.c +0 -31
-{
-    g_autoptr(GuestFilesystemInfo) fs = NULL;
-
-    fs = g_new0(GuestFilesystemInfo, 1);
-    build_guest_fsinfo_for_device(syspath, fs, errp);
-    if (fs->disk != NULL) {
-        return g_steal_pointer(&fs->disk->value);
-    }
-    return NULL;
-}
-
-static char *get_alias_for_syspath(const char *syspath)
-{
-    struct udev *udev = NULL;
-    struct udev_device *udevice = NULL;
-    char *ret = NULL;
-
-    udev = udev_new();
-    if (udev == NULL) {
-        g_debug("failed to query udev");
-        goto out;
-    }
-    udevice = udev_device_new_from_syspath(udev, syspath);
-    if (udevice == NULL) {
-        g_debug("failed to query udev for path: %s", syspath);
-        goto out;
-    } else {
-        const char *alias = udev_device_get_property_value(
-            udevice, "DM_NAME");
-        /*
-         * NULL means there was an error and empty string means there is

no
- * alias. In case of no alias we return NULL instead of empty
string.

qga/commands-posix.c +0 -118
-         */
-        if (alias == NULL) {
-            g_debug("failed to query udev for device alias for: %s",
-                syspath);
-        } else if (*alias != 0) {
-            ret = g_strdup(alias);
-        }
-    }
-
-out:
-    udev_unref(udev);
-    udev_device_unref(udevice);
-    return ret;
-}
-
-static char *get_device_for_syspath(const char *syspath)
-{
-    struct udev *udev = NULL;
-    struct udev_device *udevice = NULL;
-    char *ret = NULL;
-
-    udev = udev_new();
-    if (udev == NULL) {
-        g_debug("failed to query udev");
-        goto out;
-    }
-    udevice = udev_device_new_from_syspath(udev, syspath);
-    if (udevice == NULL) {
-        g_debug("failed to query udev for path: %s", syspath);
-        goto out;
-    } else {
-        ret = g_strdup(udev_device_get_devnode(udevice));
-    }
-
-out:
-    udev_unref(udev);
-    udev_device_unref(udevice);
-    return ret;
-}
-
-static void get_disk_deps(const char *disk_dir, GuestDiskInfo *disk)
-{
-    g_autofree char *deps_dir = NULL;
-    const gchar *dep;
-    GDir *dp_deps = NULL;
-
-    /* List dependent disks */
-    deps_dir = g_strdup_printf("%s/slaves", disk_dir);
-    g_debug("  listing entries in: %s", deps_dir);
-    dp_deps = g_dir_open(deps_dir, 0, NULL);
-    if (dp_deps == NULL) {
-        g_debug("failed to list entries in %s", deps_dir);
-        return;
-    }
-    disk->has_dependencies = true;
-    while ((dep = g_dir_read_name(dp_deps)) != NULL) {
-        g_autofree char *dep_dir = NULL;
-        char *dev_name;
-
-        /* Add dependent disks */
-        dep_dir = g_strdup_printf("%s/%s", deps_dir, dep);
-        dev_name = get_device_for_syspath(dep_dir);
-        if (dev_name != NULL) {
-            g_debug("  adding dependent device: %s", dev_name);
-            QAPI_LIST_PREPEND(disk->dependencies, dev_name);
-        }
-    }
-    g_dir_close(dp_deps);
-}
-
-/*
- * Detect partitions subdirectory, name is "<disk_name><number>" or
- * "<disk_name>p<number>"
- *
- * @disk_name -- last component of /sys path (e.g. sda)
- * @disk_dir -- sys path of the disk (e.g. /sys/block/sda)
- * @disk_dev -- device node of the disk (e.g. /dev/sda)
- */
-static GuestDiskInfoList *get_disk_partitions(
-    GuestDiskInfoList *list,
-    const char *disk_name, const char *disk_dir,
-    const char *disk_dev)
-{
-    GuestDiskInfoList *ret = list;
-    struct dirent *de_disk;
-    DIR *dp_disk = NULL;
-    size_t len = strlen(disk_name);
-
-    dp_disk = opendir(disk_dir);
-    while ((de_disk = readdir(dp_disk)) != NULL) {
-        g_autofree char *partition_dir = NULL;
-        char *dev_name;
-        GuestDiskInfo *partition;
-
-        if (!(de_disk->d_type & DT_DIR)) {
-            continue;
-        }
-
-        if (!(strncmp(disk_name, de_disk->d_name, len) == 0 &&
-            ((*(de_disk->d_name + len) == 'p' &&
-            isdigit(*(de_disk->d_name + len + 1))) ||
-                isdigit(*(de_disk->d_name + len))))) {
-            continue;
-        }
-
-        partition_dir = g_strdup_printf("%s/%s",
-            disk_dir, de_disk->d_name);
-        dev_name = get_device_for_syspath(partition_dir);
-        if (dev_name == NULL) {
-            g_debug("Failed to get device name for syspath: %s",
-                disk_dir);
-            continue;
-        }
-        partition = g_new0(GuestDiskInfo, 1);
-        partition->name = dev_name;
-        partition->partition = true;
-        partition->has_dependencies = true;
-        /* Add parent disk as dependent for easier tracking of hierarchy

*/

qga/commands-posix.c +0 -26
-        QAPI_LIST_PREPEND(partition->dependencies, g_strdup(disk_dev));
-
-        QAPI_LIST_PREPEND(ret, partition);
-    }
-    closedir(dp_disk);
-
-    return ret;
-}
-
-static void get_nvme_smart(GuestDiskInfo *disk)
-{
-    int fd;
-    GuestNVMeSmart *smart;
-    NvmeSmartLog log = {0};
-    struct nvme_admin_cmd cmd = {
-        .opcode = NVME_ADM_CMD_GET_LOG_PAGE,
-        .nsid = NVME_NSID_BROADCAST,
-        .addr = (uintptr_t)&log,
-        .data_len = sizeof(log),
-        .cdw10 = NVME_LOG_SMART_INFO | (1 << 15) /* RAE bit */
-                 | (((sizeof(log) >> 2) - 1) << 16)
-    };
-
-    fd = qga_open_cloexec(disk->name, O_RDONLY, 0);
-    if (fd == -1) {
-        g_debug("Failed to open device: %s: %s", disk->name,

g_strerror(errno));

qga/commands-posix.c +0 -5
-        return;
-    }
-
-    if (ioctl(fd, NVME_IOCTL_ADMIN_CMD, &cmd)) {
-        g_debug("Failed to get smart: %s: %s", disk->name,

g_strerror(errno));

qga/commands-posix.c +0 -10
-        close(fd);
-        return;
-    }
-
-    disk->smart = g_new0(GuestDiskSmart, 1);
-    disk->smart->type = GUEST_DISK_BUS_TYPE_NVME;
-
-    smart = &disk->smart->u.nvme;
-    smart->critical_warning = log.critical_warning;
-    smart->temperature = lduw_le_p(&log.temperature); /* unaligned field

*/

qga/commands-posix.c +0 -10
-    smart->available_spare = log.available_spare;
-    smart->available_spare_threshold = log.available_spare_threshold;
-    smart->percentage_used = log.percentage_used;
-    smart->data_units_read_lo = le64_to_cpu(log.data_units_read[0]);
-    smart->data_units_read_hi = le64_to_cpu(log.data_units_read[1]);
-    smart->data_units_written_lo = le64_to_cpu(log.data_units_written[0]);
-    smart->data_units_written_hi = le64_to_cpu(log.data_units_written[1]);
-    smart->host_read_commands_lo = le64_to_cpu(log.host_read_commands[0]);
-    smart->host_read_commands_hi = le64_to_cpu(log.host_read_commands[1]);
-    smart->host_write_commands_lo =

le64_to_cpu(log.host_write_commands[0]);
- smart->host_write_commands_hi =
le64_to_cpu(log.host_write_commands[1]);
- smart->controller_busy_time_lo =
le64_to_cpu(log.controller_busy_time[0]);
- smart->controller_busy_time_hi =
le64_to_cpu(log.controller_busy_time[1]);

qga/commands-posix.c +0 -35
-    smart->power_cycles_lo = le64_to_cpu(log.power_cycles[0]);
-    smart->power_cycles_hi = le64_to_cpu(log.power_cycles[1]);
-    smart->power_on_hours_lo = le64_to_cpu(log.power_on_hours[0]);
-    smart->power_on_hours_hi = le64_to_cpu(log.power_on_hours[1]);
-    smart->unsafe_shutdowns_lo = le64_to_cpu(log.unsafe_shutdowns[0]);
-    smart->unsafe_shutdowns_hi = le64_to_cpu(log.unsafe_shutdowns[1]);
-    smart->media_errors_lo = le64_to_cpu(log.media_errors[0]);
-    smart->media_errors_hi = le64_to_cpu(log.media_errors[1]);
-    smart->number_of_error_log_entries_lo =
-        le64_to_cpu(log.number_of_error_log_entries[0]);
-    smart->number_of_error_log_entries_hi =
-        le64_to_cpu(log.number_of_error_log_entries[1]);
-
-    close(fd);
-}
-
-static void get_disk_smart(GuestDiskInfo *disk)
-{
-    if (disk->address
-        && (disk->address->bus_type == GUEST_DISK_BUS_TYPE_NVME)) {
-        get_nvme_smart(disk);
-    }
-}
-
-GuestDiskInfoList *qmp_guest_get_disks(Error **errp)
-{
-    GuestDiskInfoList *ret = NULL;
-    GuestDiskInfo *disk;
-    DIR *dp = NULL;
-    struct dirent *de = NULL;
-
-    g_debug("listing /sys/block directory");
-    dp = opendir("/sys/block");
-    if (dp == NULL) {
-        error_setg_errno(errp, errno, "Can't open directory

\"/sys/block\"");

qga/commands-posix.c +0 -178
-        return NULL;
-    }
-    while ((de = readdir(dp)) != NULL) {
-        g_autofree char *disk_dir = NULL, *line = NULL,
-            *size_path = NULL;
-        char *dev_name;
-        Error *local_err = NULL;
-        if (de->d_type != DT_LNK) {
-            g_debug("  skipping entry: %s", de->d_name);
-            continue;
-        }
-
-        /* Check size and skip zero-sized disks */
-        g_debug("  checking disk size");
-        size_path = g_strdup_printf("/sys/block/%s/size", de->d_name);
-        if (!g_file_get_contents(size_path, &line, NULL, NULL)) {
-            g_debug("  failed to read disk size");
-            continue;
-        }
-        if (g_strcmp0(line, "0\n") == 0) {
-            g_debug("  skipping zero-sized disk");
-            continue;
-        }
-
-        g_debug("  adding %s", de->d_name);
-        disk_dir = g_strdup_printf("/sys/block/%s", de->d_name);
-        dev_name = get_device_for_syspath(disk_dir);
-        if (dev_name == NULL) {
-            g_debug("Failed to get device name for syspath: %s",
-                disk_dir);
-            continue;
-        }
-        disk = g_new0(GuestDiskInfo, 1);
-        disk->name = dev_name;
-        disk->partition = false;
-        disk->alias = get_alias_for_syspath(disk_dir);
-        QAPI_LIST_PREPEND(ret, disk);
-
-        /* Get address for non-virtual devices */
-        bool is_virtual = is_disk_virtual(disk_dir, &local_err);
-        if (local_err != NULL) {
-            g_debug("  failed to check disk path, ignoring error: %s",
-                error_get_pretty(local_err));
-            error_free(local_err);
-            local_err = NULL;
-            /* Don't try to get the address */
-            is_virtual = true;
-        }
-        if (!is_virtual) {
-            disk->address = get_disk_address(disk_dir, &local_err);
-            if (local_err != NULL) {
-                g_debug("  failed to get device info, ignoring error: %s",
-                    error_get_pretty(local_err));
-                error_free(local_err);
-                local_err = NULL;
-            }
-        }
-
-        get_disk_deps(disk_dir, disk);
-        get_disk_smart(disk);
-        ret = get_disk_partitions(ret, de->d_name, disk_dir, dev_name);
-    }
-
-    closedir(dp);
-
-    return ret;
-}
-
-#else
-
-GuestDiskInfoList *qmp_guest_get_disks(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
-
-#endif
-
-/* Return a list of the disk device(s)' info which @mount lies on */
-static GuestFilesystemInfo *build_guest_fsinfo(struct FsMount *mount,
-                                               Error **errp)
-{
-    GuestFilesystemInfo *fs = g_malloc0(sizeof(*fs));
-    struct statvfs buf;
-    unsigned long used, nonroot_total, fr_size;
-    char *devpath = g_strdup_printf("/sys/dev/block/%u:%u",
-                                    mount->devmajor, mount->devminor);
-
-    fs->mountpoint = g_strdup(mount->dirname);
-    fs->type = g_strdup(mount->devtype);
-    build_guest_fsinfo_for_device(devpath, fs, errp);
-
-    if (statvfs(fs->mountpoint, &buf) == 0) {
-        fr_size = buf.f_frsize;
-        used = buf.f_blocks - buf.f_bfree;
-        nonroot_total = used + buf.f_bavail;
-        fs->used_bytes = used * fr_size;
-        fs->total_bytes = nonroot_total * fr_size;
-        fs->total_bytes_privileged = buf.f_blocks * fr_size;
-
-        fs->has_total_bytes = true;
-        fs->has_total_bytes_privileged = true;
-        fs->has_used_bytes = true;
-    }
-
-    g_free(devpath);
-
-    return fs;
-}
-
-GuestFilesystemInfoList *qmp_guest_get_fsinfo(Error **errp)
-{
-    FsMountList mounts;
-    struct FsMount *mount;
-    GuestFilesystemInfoList *ret = NULL;
-    Error *local_err = NULL;
-
-    QTAILQ_INIT(&mounts);
-    if (!build_fs_mount_list(&mounts, &local_err)) {
-        error_propagate(errp, local_err);
-        return NULL;
-    }
-
-    QTAILQ_FOREACH(mount, &mounts, next) {
-        g_debug("Building guest fsinfo for '%s'", mount->dirname);
-
-        QAPI_LIST_PREPEND(ret, build_guest_fsinfo(mount, &local_err));
-        if (local_err) {
-            error_propagate(errp, local_err);
-            qapi_free_GuestFilesystemInfoList(ret);
-            ret = NULL;
-            break;
-        }
-    }
-
-    free_fs_mount_list(&mounts);
-    return ret;
-}
-#endif /* CONFIG_FSFREEZE */
-
-#if defined(CONFIG_FSTRIM)
-/*
- * Walk list of mounted file systems in the guest, and trim them.
- */
-GuestFilesystemTrimResponse *
-qmp_guest_fstrim(bool has_minimum, int64_t minimum, Error **errp)
-{
-    GuestFilesystemTrimResponse *response;
-    GuestFilesystemTrimResult *result;
-    int ret = 0;
-    FsMountList mounts;
-    struct FsMount *mount;
-    int fd;
-    struct fstrim_range r;
-
-    slog("guest-fstrim called");
-
-    QTAILQ_INIT(&mounts);
-    if (!build_fs_mount_list(&mounts, errp)) {
-        return NULL;
-    }
-
-    response = g_malloc0(sizeof(*response));
-
-    QTAILQ_FOREACH(mount, &mounts, next) {
-        result = g_malloc0(sizeof(*result));
-        result->path = g_strdup(mount->dirname);
-
-        QAPI_LIST_PREPEND(response->paths, result);
-
-        fd = qga_open_cloexec(mount->dirname, O_RDONLY, 0);
-        if (fd == -1) {
-            result->error = g_strdup_printf("failed to open: %s",
-                                            strerror(errno));
-            continue;
-        }
-
-        /* We try to cull filesystems we know won't work in advance, but

other

qga/commands-posix.c +0 -35
-         * filesystems may not implement fstrim for less obvious reasons.
-         * These will report EOPNOTSUPP; while in some other cases ENOTTY
-         * will be reported (e.g. CD-ROMs).
-         * Any other error means an unexpected error.
-         */
-        r.start = 0;
-        r.len = -1;
-        r.minlen = has_minimum ? minimum : 0;
-        ret = ioctl(fd, FITRIM, &r);
-        if (ret == -1) {
-            if (errno == ENOTTY || errno == EOPNOTSUPP) {
-                result->error = g_strdup("trim not supported");
-            } else {
-                result->error = g_strdup_printf("failed to trim: %s",
-                                                strerror(errno));
-            }
-            close(fd);
-            continue;
-        }
-
-        result->has_minimum = true;
-        result->minimum = r.minlen;
-        result->has_trimmed = true;
-        result->trimmed = r.len;
-        close(fd);
-    }
-
-    free_fs_mount_list(&mounts);
-    return response;
-}
-#endif /* CONFIG_FSTRIM */
-
-#endif /* __linux__ */
-
 #if defined(__linux__) || defined(__FreeBSD__)
 void qmp_guest_set_user_password(const char *username,
                                  const char *password,
--

2.45.1

#12 lore raw

PATCH v2 5/22 qga: move linux disk/cpu stats command impls to commands-linux.c

In reply to Daniel P. Berrangé

The qmp_guest_{diskstats,cpustats} command impls in commands-posix.c are surrounded by '#ifdef __linux__' so should instead live in commands-linux.c

This also removes a "#ifdef CONFIG_LINUX" that was nested inside a "#ifdef __linux__".

---
 qga/commands-linux.c | 195 ++++++++++++++++++++++++++++++++++++++++++
 qga/commands-posix.c | 199 -------------------------------------------
 2 files changed, 195 insertions(+), 199 deletions(-)
qga/commands-linux.c +195 -0
index 084e6c9e85..c0e8bd4062 100644
@@ -1594,3 +1594,198 @@ int64_t qmp_guest_set_vcpus(GuestLogicalProcessorList *vcpus, Error **errp)
 
     return processed;
 }
+
+#define MAX_NAME_LEN 128
+static GuestDiskStatsInfoList *guest_get_diskstats(Error **errp)
+{
+    GuestDiskStatsInfoList *head = NULL, **tail = &head;
+    const char *diskstats = "/proc/diskstats";
+    FILE *fp;
+    size_t n;
+    char *line = NULL;
+
+    fp = fopen(diskstats, "r");
+    if (fp  == NULL) {
+        error_setg_errno(errp, errno, "open(\"%s\")", diskstats);
+        return NULL;
+    }
+
+    while (getline(&line, &n, fp) != -1) {
+        g_autofree GuestDiskStatsInfo *diskstatinfo = NULL;
+        g_autofree GuestDiskStats *diskstat = NULL;
+        char dev_name[MAX_NAME_LEN];
+        unsigned int ios_pgr, tot_ticks, rq_ticks, wr_ticks, dc_ticks, fl_ticks;
+        unsigned long rd_ios, rd_merges_or_rd_sec, rd_ticks_or_wr_sec, wr_ios;
+        unsigned long wr_merges, rd_sec_or_wr_ios, wr_sec;
+        unsigned long dc_ios, dc_merges, dc_sec, fl_ios;
+        unsigned int major, minor;
+        int i;
+
+        i = sscanf(line, "%u %u %s %lu %lu %lu"
+                   "%lu %lu %lu %lu %u %u %u %u"
+                   "%lu %lu %lu %u %lu %u",
+                   &major, &minor, dev_name,
+                   &rd_ios, &rd_merges_or_rd_sec, &rd_sec_or_wr_ios,
+                   &rd_ticks_or_wr_sec, &wr_ios, &wr_merges, &wr_sec,
+                   &wr_ticks, &ios_pgr, &tot_ticks, &rq_ticks,
+                   &dc_ios, &dc_merges, &dc_sec, &dc_ticks,
+                   &fl_ios, &fl_ticks);
+
+        if (i < 7) {
+            continue;
+        }
+
+        diskstatinfo = g_new0(GuestDiskStatsInfo, 1);
+        diskstatinfo->name = g_strdup(dev_name);
+        diskstatinfo->major = major;
+        diskstatinfo->minor = minor;
+
+        diskstat = g_new0(GuestDiskStats, 1);
+        if (i == 7) {
+            diskstat->has_read_ios = true;
+            diskstat->read_ios = rd_ios;
+            diskstat->has_read_sectors = true;
+            diskstat->read_sectors = rd_merges_or_rd_sec;
+            diskstat->has_write_ios = true;
+            diskstat->write_ios = rd_sec_or_wr_ios;
+            diskstat->has_write_sectors = true;
+            diskstat->write_sectors = rd_ticks_or_wr_sec;
+        }
+        if (i >= 14) {
+            diskstat->has_read_ios = true;
+            diskstat->read_ios = rd_ios;
+            diskstat->has_read_sectors = true;
+            diskstat->read_sectors = rd_sec_or_wr_ios;
+            diskstat->has_read_merges = true;
+            diskstat->read_merges = rd_merges_or_rd_sec;
+            diskstat->has_read_ticks = true;
+            diskstat->read_ticks = rd_ticks_or_wr_sec;
+            diskstat->has_write_ios = true;
+            diskstat->write_ios = wr_ios;
+            diskstat->has_write_sectors = true;
+            diskstat->write_sectors = wr_sec;
+            diskstat->has_write_merges = true;
+            diskstat->write_merges = wr_merges;
+            diskstat->has_write_ticks = true;
+            diskstat->write_ticks = wr_ticks;
+            diskstat->has_ios_pgr = true;
+            diskstat->ios_pgr = ios_pgr;
+            diskstat->has_total_ticks = true;
+            diskstat->total_ticks = tot_ticks;
+            diskstat->has_weight_ticks = true;
+            diskstat->weight_ticks = rq_ticks;
+        }
+        if (i >= 18) {
+            diskstat->has_discard_ios = true;
+            diskstat->discard_ios = dc_ios;
+            diskstat->has_discard_merges = true;
+            diskstat->discard_merges = dc_merges;
+            diskstat->has_discard_sectors = true;
+            diskstat->discard_sectors = dc_sec;
+            diskstat->has_discard_ticks = true;
+            diskstat->discard_ticks = dc_ticks;
+        }
+        if (i >= 20) {
+            diskstat->has_flush_ios = true;
+            diskstat->flush_ios = fl_ios;
+            diskstat->has_flush_ticks = true;
+            diskstat->flush_ticks = fl_ticks;
+        }
+
+        diskstatinfo->stats = g_steal_pointer(&diskstat);
+        QAPI_LIST_APPEND(tail, diskstatinfo);
+        diskstatinfo = NULL;
+    }
+    free(line);
+    fclose(fp);
+    return head;
+}
+
+GuestDiskStatsInfoList *qmp_guest_get_diskstats(Error **errp)
+{
+    return guest_get_diskstats(errp);
+}
+
+GuestCpuStatsList *qmp_guest_get_cpustats(Error **errp)
+{
+    GuestCpuStatsList *head = NULL, **tail = &head;
+    const char *cpustats = "/proc/stat";
+    int clk_tck = sysconf(_SC_CLK_TCK);
+    FILE *fp;
+    size_t n;
+    char *line = NULL;
+
+    fp = fopen(cpustats, "r");
+    if (fp  == NULL) {
+        error_setg_errno(errp, errno, "open(\"%s\")", cpustats);
+        return NULL;
+    }
+
+    while (getline(&line, &n, fp) != -1) {
+        GuestCpuStats *cpustat = NULL;
+        GuestLinuxCpuStats *linuxcpustat;
+        int i;
+        unsigned long user, system, idle, iowait, irq, softirq, steal, guest;
+        unsigned long nice, guest_nice;
+        char name[64];
+
+        i = sscanf(line, "%s %lu %lu %lu %lu %lu %lu %lu %lu %lu %lu",
+                   name, &user, &nice, &system, &idle, &iowait, &irq, &softirq,
+                   &steal, &guest, &guest_nice);
+
+        /* drop "cpu 1 2 3 ...", get "cpuX 1 2 3 ..." only */
+        if ((i == EOF) || strncmp(name, "cpu", 3) || (name[3] == '\0')) {
+            continue;
+        }
+
+        if (i < 5) {
+            slog("Parsing cpu stat from %s failed, see \"man proc\"", cpustats);
+            break;
+        }
+
+        cpustat = g_new0(GuestCpuStats, 1);
+        cpustat->type = GUEST_CPU_STATS_TYPE_LINUX;
+
+        linuxcpustat = &cpustat->u.q_linux;
+        linuxcpustat->cpu = atoi(&name[3]);
+        linuxcpustat->user = user * 1000 / clk_tck;
+        linuxcpustat->nice = nice * 1000 / clk_tck;
+        linuxcpustat->system = system * 1000 / clk_tck;
+        linuxcpustat->idle = idle * 1000 / clk_tck;
+
+        if (i > 5) {
+            linuxcpustat->has_iowait = true;
+            linuxcpustat->iowait = iowait * 1000 / clk_tck;
+        }
+
+        if (i > 6) {
+            linuxcpustat->has_irq = true;
+            linuxcpustat->irq = irq * 1000 / clk_tck;
+            linuxcpustat->has_softirq = true;
+            linuxcpustat->softirq = softirq * 1000 / clk_tck;
+        }
+
+        if (i > 8) {
+            linuxcpustat->has_steal = true;
+            linuxcpustat->steal = steal * 1000 / clk_tck;
+        }
+
+        if (i > 9) {
+            linuxcpustat->has_guest = true;
+            linuxcpustat->guest = guest * 1000 / clk_tck;
+        }
+
+        if (i > 10) {
+            linuxcpustat->has_guest = true;
+            linuxcpustat->guest = guest * 1000 / clk_tck;
+            linuxcpustat->has_guestnice = true;
+            linuxcpustat->guestnice = guest_nice * 1000 / clk_tck;
+        }
+
+        QAPI_LIST_APPEND(tail, cpustat);
+    }
+
+    free(line);
+    fclose(fp);
+    return head;
+}
qga/commands-posix.c +0 -199

205 lines not on this page · raw message

Signature · 1 line
2.45.1
Reviewed-by on v2 5/22
#13 lore raw

Re: [PATCH v2 05/22] qga: move linux disk/cpu stats command impls to commands-linux.c

Quoted 4 lines · Daniel P. Berrangé

The qmp_guest_{diskstats,cpustats} command impls in commands-posix.c are surrounded by '#ifdef __linux__' so should instead live in commands-linux.c

Quoted 8 lines

This also removes a "#ifdef CONFIG_LINUX" that was nested inside a "#ifdef __linux__".

---
  qga/commands-linux.c | 195 ++++++++++++++++++++++++++++++++++++++++++
  qga/commands-posix.c | 199 -------------------------------------------
  2 files changed, 195 insertions(+), 199 deletions(-)

Easy to review using 'git-diff --color-moved=dimmed-zebra'.

Reviewed-by on v2 5/22
#14 lore raw

Re: [PATCH v2 05/22] qga: move linux disk/cpu stats command impls to commands-linux.c

In reply to Daniel P. Berrangé
Quoted 442 lines · Daniel P. Berrangé

The qmp_guest_{diskstats,cpustats} command impls in commands-posix.c are surrounded by '#ifdef __linux__' so should instead live in commands-linux.c

This also removes a "#ifdef CONFIG_LINUX" that was nested inside a "#ifdef __linux__".

---
 qga/commands-linux.c | 195 ++++++++++++++++++++++++++++++++++++++++++
 qga/commands-posix.c | 199 -------------------------------------------
 2 files changed, 195 insertions(+), 199 deletions(-)
qga/commands-linux.c +21 -0
index 084e6c9e85..c0e8bd4062 100644
@@ -1594,3 +1594,198 @@ int64_t
 qmp_guest_set_vcpus(GuestLogicalProcessorList *vcpus, Error **errp)
 
     return processed;
 }
+
+#define MAX_NAME_LEN 128
+static GuestDiskStatsInfoList *guest_get_diskstats(Error **errp)
+{
+    GuestDiskStatsInfoList *head = NULL, **tail = &head;
+    const char *diskstats = "/proc/diskstats";
+    FILE *fp;
+    size_t n;
+    char *line = NULL;
+
+    fp = fopen(diskstats, "r");
+    if (fp  == NULL) {
+        error_setg_errno(errp, errno, "open(\"%s\")", diskstats);
+        return NULL;
+    }
+
+    while (getline(&line, &n, fp) != -1) {
+        g_autofree GuestDiskStatsInfo *diskstatinfo = NULL;
+        g_autofree GuestDiskStats *diskstat = NULL;
+        char dev_name[MAX_NAME_LEN];
+        unsigned int ios_pgr, tot_ticks, rq_ticks, wr_ticks, dc_ticks,

fl_ticks;
+ unsigned long rd_ios, rd_merges_or_rd_sec, rd_ticks_or_wr_sec,
wr_ios;

qga/commands-linux.c +110 -0
+        unsigned long wr_merges, rd_sec_or_wr_ios, wr_sec;
+        unsigned long dc_ios, dc_merges, dc_sec, fl_ios;
+        unsigned int major, minor;
+        int i;
+
+        i = sscanf(line, "%u %u %s %lu %lu %lu"
+                   "%lu %lu %lu %lu %u %u %u %u"
+                   "%lu %lu %lu %u %lu %u",
+                   &major, &minor, dev_name,
+                   &rd_ios, &rd_merges_or_rd_sec, &rd_sec_or_wr_ios,
+                   &rd_ticks_or_wr_sec, &wr_ios, &wr_merges, &wr_sec,
+                   &wr_ticks, &ios_pgr, &tot_ticks, &rq_ticks,
+                   &dc_ios, &dc_merges, &dc_sec, &dc_ticks,
+                   &fl_ios, &fl_ticks);
+
+        if (i < 7) {
+            continue;
+        }
+
+        diskstatinfo = g_new0(GuestDiskStatsInfo, 1);
+        diskstatinfo->name = g_strdup(dev_name);
+        diskstatinfo->major = major;
+        diskstatinfo->minor = minor;
+
+        diskstat = g_new0(GuestDiskStats, 1);
+        if (i == 7) {
+            diskstat->has_read_ios = true;
+            diskstat->read_ios = rd_ios;
+            diskstat->has_read_sectors = true;
+            diskstat->read_sectors = rd_merges_or_rd_sec;
+            diskstat->has_write_ios = true;
+            diskstat->write_ios = rd_sec_or_wr_ios;
+            diskstat->has_write_sectors = true;
+            diskstat->write_sectors = rd_ticks_or_wr_sec;
+        }
+        if (i >= 14) {
+            diskstat->has_read_ios = true;
+            diskstat->read_ios = rd_ios;
+            diskstat->has_read_sectors = true;
+            diskstat->read_sectors = rd_sec_or_wr_ios;
+            diskstat->has_read_merges = true;
+            diskstat->read_merges = rd_merges_or_rd_sec;
+            diskstat->has_read_ticks = true;
+            diskstat->read_ticks = rd_ticks_or_wr_sec;
+            diskstat->has_write_ios = true;
+            diskstat->write_ios = wr_ios;
+            diskstat->has_write_sectors = true;
+            diskstat->write_sectors = wr_sec;
+            diskstat->has_write_merges = true;
+            diskstat->write_merges = wr_merges;
+            diskstat->has_write_ticks = true;
+            diskstat->write_ticks = wr_ticks;
+            diskstat->has_ios_pgr = true;
+            diskstat->ios_pgr = ios_pgr;
+            diskstat->has_total_ticks = true;
+            diskstat->total_ticks = tot_ticks;
+            diskstat->has_weight_ticks = true;
+            diskstat->weight_ticks = rq_ticks;
+        }
+        if (i >= 18) {
+            diskstat->has_discard_ios = true;
+            diskstat->discard_ios = dc_ios;
+            diskstat->has_discard_merges = true;
+            diskstat->discard_merges = dc_merges;
+            diskstat->has_discard_sectors = true;
+            diskstat->discard_sectors = dc_sec;
+            diskstat->has_discard_ticks = true;
+            diskstat->discard_ticks = dc_ticks;
+        }
+        if (i >= 20) {
+            diskstat->has_flush_ios = true;
+            diskstat->flush_ios = fl_ios;
+            diskstat->has_flush_ticks = true;
+            diskstat->flush_ticks = fl_ticks;
+        }
+
+        diskstatinfo->stats = g_steal_pointer(&diskstat);
+        QAPI_LIST_APPEND(tail, diskstatinfo);
+        diskstatinfo = NULL;
+    }
+    free(line);
+    fclose(fp);
+    return head;
+}
+
+GuestDiskStatsInfoList *qmp_guest_get_diskstats(Error **errp)
+{
+    return guest_get_diskstats(errp);
+}
+
+GuestCpuStatsList *qmp_guest_get_cpustats(Error **errp)
+{
+    GuestCpuStatsList *head = NULL, **tail = &head;
+    const char *cpustats = "/proc/stat";
+    int clk_tck = sysconf(_SC_CLK_TCK);
+    FILE *fp;
+    size_t n;
+    char *line = NULL;
+
+    fp = fopen(cpustats, "r");
+    if (fp  == NULL) {
+        error_setg_errno(errp, errno, "open(\"%s\")", cpustats);
+        return NULL;
+    }
+
+    while (getline(&line, &n, fp) != -1) {
+        GuestCpuStats *cpustat = NULL;
+        GuestLinuxCpuStats *linuxcpustat;
+        int i;
+        unsigned long user, system, idle, iowait, irq, softirq, steal,

guest;

qga/commands-linux.c +5 -0
+        unsigned long nice, guest_nice;
+        char name[64];
+
+        i = sscanf(line, "%s %lu %lu %lu %lu %lu %lu %lu %lu %lu %lu",
+                   name, &user, &nice, &system, &idle, &iowait, &irq,

&softirq,

qga/commands-linux.c +9 -0
+                   &steal, &guest, &guest_nice);
+
+        /* drop "cpu 1 2 3 ...", get "cpuX 1 2 3 ..." only */
+        if ((i == EOF) || strncmp(name, "cpu", 3) || (name[3] == '\0')) {
+            continue;
+        }
+
+        if (i < 5) {
+            slog("Parsing cpu stat from %s failed, see \"man proc\"",

cpustats);

qga/commands-linux.c +49 -0
+            break;
+        }
+
+        cpustat = g_new0(GuestCpuStats, 1);
+        cpustat->type = GUEST_CPU_STATS_TYPE_LINUX;
+
+        linuxcpustat = &cpustat->u.q_linux;
+        linuxcpustat->cpu = atoi(&name[3]);
+        linuxcpustat->user = user * 1000 / clk_tck;
+        linuxcpustat->nice = nice * 1000 / clk_tck;
+        linuxcpustat->system = system * 1000 / clk_tck;
+        linuxcpustat->idle = idle * 1000 / clk_tck;
+
+        if (i > 5) {
+            linuxcpustat->has_iowait = true;
+            linuxcpustat->iowait = iowait * 1000 / clk_tck;
+        }
+
+        if (i > 6) {
+            linuxcpustat->has_irq = true;
+            linuxcpustat->irq = irq * 1000 / clk_tck;
+            linuxcpustat->has_softirq = true;
+            linuxcpustat->softirq = softirq * 1000 / clk_tck;
+        }
+
+        if (i > 8) {
+            linuxcpustat->has_steal = true;
+            linuxcpustat->steal = steal * 1000 / clk_tck;
+        }
+
+        if (i > 9) {
+            linuxcpustat->has_guest = true;
+            linuxcpustat->guest = guest * 1000 / clk_tck;
+        }
+
+        if (i > 10) {
+            linuxcpustat->has_guest = true;
+            linuxcpustat->guest = guest * 1000 / clk_tck;
+            linuxcpustat->has_guestnice = true;
+            linuxcpustat->guestnice = guest_nice * 1000 / clk_tck;
+        }
+
+        QAPI_LIST_APPEND(tail, cpustat);
+    }
+
+    free(line);
+    fclose(fp);
+    return head;
+}
qga/commands-posix.c +0 -136
index 98aafc45f3..5da60e65ab 100644
@@ -1195,205 +1195,6 @@ GuestMemoryBlockInfo
 *qmp_guest_get_memory_block_info(Error **errp)
     return info;
 }
 
-#define MAX_NAME_LEN 128
-static GuestDiskStatsInfoList *guest_get_diskstats(Error **errp)
-{
-#ifdef CONFIG_LINUX
-    GuestDiskStatsInfoList *head = NULL, **tail = &head;
-    const char *diskstats = "/proc/diskstats";
-    FILE *fp;
-    size_t n;
-    char *line = NULL;
-
-    fp = fopen(diskstats, "r");
-    if (fp  == NULL) {
-        error_setg_errno(errp, errno, "open(\"%s\")", diskstats);
-        return NULL;
-    }
-
-    while (getline(&line, &n, fp) != -1) {
-        g_autofree GuestDiskStatsInfo *diskstatinfo = NULL;
-        g_autofree GuestDiskStats *diskstat = NULL;
-        char dev_name[MAX_NAME_LEN];
-        unsigned int ios_pgr, tot_ticks, rq_ticks, wr_ticks, dc_ticks,
 fl_ticks;
-        unsigned long rd_ios, rd_merges_or_rd_sec, rd_ticks_or_wr_sec,
 wr_ios;
-        unsigned long wr_merges, rd_sec_or_wr_ios, wr_sec;
-        unsigned long dc_ios, dc_merges, dc_sec, fl_ios;
-        unsigned int major, minor;
-        int i;
-
-        i = sscanf(line, "%u %u %s %lu %lu %lu"
-                   "%lu %lu %lu %lu %u %u %u %u"
-                   "%lu %lu %lu %u %lu %u",
-                   &major, &minor, dev_name,
-                   &rd_ios, &rd_merges_or_rd_sec, &rd_sec_or_wr_ios,
-                   &rd_ticks_or_wr_sec, &wr_ios, &wr_merges, &wr_sec,
-                   &wr_ticks, &ios_pgr, &tot_ticks, &rq_ticks,
-                   &dc_ios, &dc_merges, &dc_sec, &dc_ticks,
-                   &fl_ios, &fl_ticks);
-
-        if (i < 7) {
-            continue;
-        }
-
-        diskstatinfo = g_new0(GuestDiskStatsInfo, 1);
-        diskstatinfo->name = g_strdup(dev_name);
-        diskstatinfo->major = major;
-        diskstatinfo->minor = minor;
-
-        diskstat = g_new0(GuestDiskStats, 1);
-        if (i == 7) {
-            diskstat->has_read_ios = true;
-            diskstat->read_ios = rd_ios;
-            diskstat->has_read_sectors = true;
-            diskstat->read_sectors = rd_merges_or_rd_sec;
-            diskstat->has_write_ios = true;
-            diskstat->write_ios = rd_sec_or_wr_ios;
-            diskstat->has_write_sectors = true;
-            diskstat->write_sectors = rd_ticks_or_wr_sec;
-        }
-        if (i >= 14) {
-            diskstat->has_read_ios = true;
-            diskstat->read_ios = rd_ios;
-            diskstat->has_read_sectors = true;
-            diskstat->read_sectors = rd_sec_or_wr_ios;
-            diskstat->has_read_merges = true;
-            diskstat->read_merges = rd_merges_or_rd_sec;
-            diskstat->has_read_ticks = true;
-            diskstat->read_ticks = rd_ticks_or_wr_sec;
-            diskstat->has_write_ios = true;
-            diskstat->write_ios = wr_ios;
-            diskstat->has_write_sectors = true;
-            diskstat->write_sectors = wr_sec;
-            diskstat->has_write_merges = true;
-            diskstat->write_merges = wr_merges;
-            diskstat->has_write_ticks = true;
-            diskstat->write_ticks = wr_ticks;
-            diskstat->has_ios_pgr = true;
-            diskstat->ios_pgr = ios_pgr;
-            diskstat->has_total_ticks = true;
-            diskstat->total_ticks = tot_ticks;
-            diskstat->has_weight_ticks = true;
-            diskstat->weight_ticks = rq_ticks;
-        }
-        if (i >= 18) {
-            diskstat->has_discard_ios = true;
-            diskstat->discard_ios = dc_ios;
-            diskstat->has_discard_merges = true;
-            diskstat->discard_merges = dc_merges;
-            diskstat->has_discard_sectors = true;
-            diskstat->discard_sectors = dc_sec;
-            diskstat->has_discard_ticks = true;
-            diskstat->discard_ticks = dc_ticks;
-        }
-        if (i >= 20) {
-            diskstat->has_flush_ios = true;
-            diskstat->flush_ios = fl_ios;
-            diskstat->has_flush_ticks = true;
-            diskstat->flush_ticks = fl_ticks;
-        }
-
-        diskstatinfo->stats = g_steal_pointer(&diskstat);
-        QAPI_LIST_APPEND(tail, diskstatinfo);
-        diskstatinfo = NULL;
-    }
-    free(line);
-    fclose(fp);
-    return head;
-#else
-    g_debug("disk stats reporting available only for Linux");
-    return NULL;
-#endif
-}
-
-GuestDiskStatsInfoList *qmp_guest_get_diskstats(Error **errp)
-{
-    return guest_get_diskstats(errp);
-}
-
-GuestCpuStatsList *qmp_guest_get_cpustats(Error **errp)
-{
-    GuestCpuStatsList *head = NULL, **tail = &head;
-    const char *cpustats = "/proc/stat";
-    int clk_tck = sysconf(_SC_CLK_TCK);
-    FILE *fp;
-    size_t n;
-    char *line = NULL;
-
-    fp = fopen(cpustats, "r");
-    if (fp  == NULL) {
-        error_setg_errno(errp, errno, "open(\"%s\")", cpustats);
-        return NULL;
-    }
-
-    while (getline(&line, &n, fp) != -1) {
-        GuestCpuStats *cpustat = NULL;
-        GuestLinuxCpuStats *linuxcpustat;
-        int i;
-        unsigned long user, system, idle, iowait, irq, softirq, steal,

guest;

qga/commands-posix.c +0 -5
-        unsigned long nice, guest_nice;
-        char name[64];
-
-        i = sscanf(line, "%s %lu %lu %lu %lu %lu %lu %lu %lu %lu %lu",
-                   name, &user, &nice, &system, &idle, &iowait, &irq,

&softirq,

qga/commands-posix.c +0 -9
-                   &steal, &guest, &guest_nice);
-
-        /* drop "cpu 1 2 3 ...", get "cpuX 1 2 3 ..." only */
-        if ((i == EOF) || strncmp(name, "cpu", 3) || (name[3] == '\0')) {
-            continue;
-        }
-
-        if (i < 5) {
-            slog("Parsing cpu stat from %s failed, see \"man proc\"",

cpustats);

qga/commands-posix.c +0 -50
-            break;
-        }
-
-        cpustat = g_new0(GuestCpuStats, 1);
-        cpustat->type = GUEST_CPU_STATS_TYPE_LINUX;
-
-        linuxcpustat = &cpustat->u.q_linux;
-        linuxcpustat->cpu = atoi(&name[3]);
-        linuxcpustat->user = user * 1000 / clk_tck;
-        linuxcpustat->nice = nice * 1000 / clk_tck;
-        linuxcpustat->system = system * 1000 / clk_tck;
-        linuxcpustat->idle = idle * 1000 / clk_tck;
-
-        if (i > 5) {
-            linuxcpustat->has_iowait = true;
-            linuxcpustat->iowait = iowait * 1000 / clk_tck;
-        }
-
-        if (i > 6) {
-            linuxcpustat->has_irq = true;
-            linuxcpustat->irq = irq * 1000 / clk_tck;
-            linuxcpustat->has_softirq = true;
-            linuxcpustat->softirq = softirq * 1000 / clk_tck;
-        }
-
-        if (i > 8) {
-            linuxcpustat->has_steal = true;
-            linuxcpustat->steal = steal * 1000 / clk_tck;
-        }
-
-        if (i > 9) {
-            linuxcpustat->has_guest = true;
-            linuxcpustat->guest = guest * 1000 / clk_tck;
-        }
-
-        if (i > 10) {
-            linuxcpustat->has_guest = true;
-            linuxcpustat->guest = guest * 1000 / clk_tck;
-            linuxcpustat->has_guestnice = true;
-            linuxcpustat->guestnice = guest_nice * 1000 / clk_tck;
-        }
-
-        QAPI_LIST_APPEND(tail, cpustat);
-    }
-
-    free(line);
-    fclose(fp);
-    return head;
-}
 
 #else /* defined(__linux__) */
 
--

2.45.1

#15 lore raw

PATCH v2 6/22 qga: move linux memory block command impls to commands-linux.c

In reply to Daniel P. Berrangé

The qmp_guest_{set,get}_{memory_blocks,block_info} command impls in commands-posix.c are surrounded by '#ifdef __linux__' so should instead live in commands-linux.c

This also removes a "#ifdef CONFIG_LINUX" that was nested inside a "#ifdef __linux__".

---
 qga/commands-linux.c | 308 ++++++++++++++++++++++++++++++++++++++++++
 qga/commands-posix.c | 311 +------------------------------------------
 2 files changed, 309 insertions(+), 310 deletions(-)
qga/commands-linux.c +308 -0
index c0e8bd4062..73b13fbaf6 100644
@@ -1595,6 +1595,314 @@ int64_t qmp_guest_set_vcpus(GuestLogicalProcessorList *vcpus, Error **errp)
     return processed;
 }
 
+
+static void ga_read_sysfs_file(int dirfd, const char *pathname, char *buf,
+                               int size, Error **errp)
+{
+    int fd;
+    int res;
+
+    errno = 0;
+    fd = openat(dirfd, pathname, O_RDONLY);
+    if (fd == -1) {
+        error_setg_errno(errp, errno, "open sysfs file \"%s\"", pathname);
+        return;
+    }
+
+    res = pread(fd, buf, size, 0);
+    if (res == -1) {
+        error_setg_errno(errp, errno, "pread sysfs file \"%s\"", pathname);
+    } else if (res == 0) {
+        error_setg(errp, "pread sysfs file \"%s\": unexpected EOF", pathname);
+    }
+    close(fd);
+}
+
+static void ga_write_sysfs_file(int dirfd, const char *pathname,
+                                const char *buf, int size, Error **errp)
+{
+    int fd;
+
+    errno = 0;
+    fd = openat(dirfd, pathname, O_WRONLY);
+    if (fd == -1) {
+        error_setg_errno(errp, errno, "open sysfs file \"%s\"", pathname);
+        return;
+    }
+
+    if (pwrite(fd, buf, size, 0) == -1) {
+        error_setg_errno(errp, errno, "pwrite sysfs file \"%s\"", pathname);
+    }
+
+    close(fd);
+}
+
+/* Transfer online/offline status between @mem_blk and the guest system.
+ *
+ * On input either @errp or *@errp must be NULL.
+ *
+ * In system-to-@mem_blk direction, the following @mem_blk fields are accessed:
+ * - R: mem_blk->phys_index
+ * - W: mem_blk->online
+ * - W: mem_blk->can_offline
+ *
+ * In @mem_blk-to-system direction, the following @mem_blk fields are accessed:
+ * - R: mem_blk->phys_index
+ * - R: mem_blk->online
+ *-  R: mem_blk->can_offline
+ * Written members remain unmodified on error.
+ */
+static void transfer_memory_block(GuestMemoryBlock *mem_blk, bool sys2memblk,
+                                  GuestMemoryBlockResponse *result,
+                                  Error **errp)
+{
+    char *dirpath;
+    int dirfd;
+    char *status;
+    Error *local_err = NULL;
+
+    if (!sys2memblk) {
+        DIR *dp;
+
+        if (!result) {
+            error_setg(errp, "Internal error, 'result' should not be NULL");
+            return;
+        }
+        errno = 0;
+        dp = opendir("/sys/devices/system/memory/");
+         /* if there is no 'memory' directory in sysfs,
+         * we think this VM does not support online/offline memory block,
+         * any other solution?
+         */
+        if (!dp) {
+            if (errno == ENOENT) {
+                result->response =
+                    GUEST_MEMORY_BLOCK_RESPONSE_TYPE_OPERATION_NOT_SUPPORTED;
+            }
+            goto out1;
+        }
+        closedir(dp);
+    }
+
+    dirpath = g_strdup_printf("/sys/devices/system/memory/memory%" PRId64 "/",
+                              mem_blk->phys_index);
+    dirfd = open(dirpath, O_RDONLY | O_DIRECTORY);
+    if (dirfd == -1) {
+        if (sys2memblk) {
+            error_setg_errno(errp, errno, "open(\"%s\")", dirpath);
+        } else {
+            if (errno == ENOENT) {
+                result->response = GUEST_MEMORY_BLOCK_RESPONSE_TYPE_NOT_FOUND;
+            } else {
+                result->response =
+                    GUEST_MEMORY_BLOCK_RESPONSE_TYPE_OPERATION_FAILED;
+            }
+        }
+        g_free(dirpath);
+        goto out1;
+    }
+    g_free(dirpath);
+
+    status = g_malloc0(10);
+    ga_read_sysfs_file(dirfd, "state", status, 10, &local_err);
+    if (local_err) {
+        /* treat with sysfs file that not exist in old kernel */
+        if (errno == ENOENT) {
+            error_free(local_err);
+            if (sys2memblk) {
+                mem_blk->online = true;
+                mem_blk->can_offline = false;
+            } else if (!mem_blk->online) {
+                result->response =
+                    GUEST_MEMORY_BLOCK_RESPONSE_TYPE_OPERATION_NOT_SUPPORTED;
+            }
+        } else {
+            if (sys2memblk) {
+                error_propagate(errp, local_err);
+            } else {
+                error_free(local_err);
+                result->response =
+                    GUEST_MEMORY_BLOCK_RESPONSE_TYPE_OPERATION_FAILED;
+            }
+        }
+        goto out2;
+    }
+
+    if (sys2memblk) {
+        char removable = '0';
+
+        mem_blk->online = (strncmp(status, "online", 6) == 0);
+
+        ga_read_sysfs_file(dirfd, "removable", &removable, 1, &local_err);
+        if (local_err) {
+            /* if no 'removable' file, it doesn't support offline mem blk */
+            if (errno == ENOENT) {
+                error_free(local_err);
+                mem_blk->can_offline = false;
+            } else {
+                error_propagate(errp, local_err);
+            }
+        } else {
+            mem_blk->can_offline = (removable != '0');
+        }
+    } else {
+        if (mem_blk->online != (strncmp(status, "online", 6) == 0)) {
+            const char *new_state = mem_blk->online ? "online" : "offline";
+
+            ga_write_sysfs_file(dirfd, "state", new_state, strlen(new_state),
+                                &local_err);
+            if (local_err) {
+                error_free(local_err);
+                result->response =
+                    GUEST_MEMORY_BLOCK_RESPONSE_TYPE_OPERATION_FAILED;
+                goto out2;
+            }
+
+            result->response = GUEST_MEMORY_BLOCK_RESPONSE_TYPE_SUCCESS;
+            result->has_error_code = false;
+        } /* otherwise pretend successful re-(on|off)-lining */
+    }
+    g_free(status);
+    close(dirfd);
+    return;
+
+out2:
+    g_free(status);
+    close(dirfd);
+out1:
+    if (!sys2memblk) {
+        result->has_error_code = true;
+        result->error_code = errno;
+    }
+}
+
+GuestMemoryBlockList *qmp_guest_get_memory_blocks(Error **errp)
+{
+    GuestMemoryBlockList *head, **tail;
+    Error *local_err = NULL;
+    struct dirent *de;
+    DIR *dp;
+
+    head = NULL;
+    tail = &head;
+
+    dp = opendir("/sys/devices/system/memory/");
+    if (!dp) {
+        /* it's ok if this happens to be a system that doesn't expose
+         * memory blocks via sysfs, but otherwise we should report
+         * an error
+         */
+        if (errno != ENOENT) {
+            error_setg_errno(errp, errno, "Can't open directory"
+                             "\"/sys/devices/system/memory/\"");
+        }
+        return NULL;
+    }
+
+    /* Note: the phys_index of memory block may be discontinuous,
+     * this is because a memblk is the unit of the Sparse Memory design, which
+     * allows discontinuous memory ranges (ex. NUMA), so here we should
+     * traverse the memory block directory.
+     */
+    while ((de = readdir(dp)) != NULL) {
+        GuestMemoryBlock *mem_blk;
+
+        if ((strncmp(de->d_name, "memory", 6) != 0) ||
+            !(de->d_type & DT_DIR)) {
+            continue;
+        }
+
+        mem_blk = g_malloc0(sizeof *mem_blk);
+        /* The d_name is "memoryXXX",  phys_index is block id, same as XXX */
+        mem_blk->phys_index = strtoul(&de->d_name[6], NULL, 10);
+        mem_blk->has_can_offline = true; /* lolspeak ftw */
+        transfer_memory_block(mem_blk, true, NULL, &local_err);
+        if (local_err) {
+            break;
+        }
+
+        QAPI_LIST_APPEND(tail, mem_blk);
+    }
+
+    closedir(dp);
+    if (local_err == NULL) {
+        /* there's no guest with zero memory blocks */
+        if (head == NULL) {
+            error_setg(errp, "guest reported zero memory blocks!");
+        }
+        return head;
+    }
+
+    qapi_free_GuestMemoryBlockList(head);
+    error_propagate(errp, local_err);
+    return NULL;
+}
+
+GuestMemoryBlockResponseList *
+qmp_guest_set_memory_blocks(GuestMemoryBlockList *mem_blks, Error **errp)
+{
+    GuestMemoryBlockResponseList *head, **tail;
+    Error *local_err = NULL;
+
+    head = NULL;
+    tail = &head;
+
+    while (mem_blks != NULL) {
+        GuestMemoryBlockResponse *result;
+        GuestMemoryBlock *current_mem_blk = mem_blks->value;
+
+        result = g_malloc0(sizeof(*result));
+        result->phys_index = current_mem_blk->phys_index;
+        transfer_memory_block(current_mem_blk, false, result, &local_err);
+        if (local_err) { /* should never happen */
+            goto err;
+        }
+
+        QAPI_LIST_APPEND(tail, result);
+        mem_blks = mem_blks->next;
+    }
+
+    return head;
+err:
+    qapi_free_GuestMemoryBlockResponseList(head);
+    error_propagate(errp, local_err);
+    return NULL;
+}
+
+GuestMemoryBlockInfo *qmp_guest_get_memory_block_info(Error **errp)
+{
+    Error *local_err = NULL;
+    char *dirpath;
+    int dirfd;
+    char *buf;
+    GuestMemoryBlockInfo *info;
+
+    dirpath = g_strdup_printf("/sys/devices/system/memory/");
+    dirfd = open(dirpath, O_RDONLY | O_DIRECTORY);
+    if (dirfd == -1) {
+        error_setg_errno(errp, errno, "open(\"%s\")", dirpath);
+        g_free(dirpath);
+        return NULL;
+    }
+    g_free(dirpath);
+
+    buf = g_malloc0(20);
+    ga_read_sysfs_file(dirfd, "block_size_bytes", buf, 20, &local_err);
+    close(dirfd);
+    if (local_err) {
+        g_free(buf);
+        error_propagate(errp, local_err);
+        return NULL;
+    }
+
+    info = g_new0(GuestMemoryBlockInfo, 1);
+    info->size = strtol(buf, NULL, 16); /* the unit is bytes */
+
+    g_free(buf);
+
+    return info;
+}
+
 #define MAX_NAME_LEN 128
 static GuestDiskStatsInfoList *guest_get_diskstats(Error **errp)
 {
qga/commands-posix.c +1 -310

317 lines not on this page · raw message

Signature · 1 line
2.45.1
#16 lore raw

PATCH v2 7/22 qga: move CONFIG_FSFREEZE/TRIM to be meson defined options

Defining these at the meson level allows them to be used a conditional tests in the QAPI schemas.

---
 meson.build           | 15 +++++++++++++++
 qga/commands-common.h |  9 ---------
 2 files changed, 15 insertions(+), 9 deletions(-)
meson.build +15 -0
index 91278667ea..d9f3349b0a 100644
@@ -2145,6 +2145,19 @@ have_virtfs_proxy_helper = get_option('virtfs_proxy_helper') \
     .require(libcap_ng.found(), error_message: 'the virtfs proxy helper requires libcap-ng') \
     .allowed()
 
+qga_fsfreeze = false
+qga_fstrim = false
+if host_os == 'linux'
+    if cc.has_header_symbol('linux/fs.h', 'FIFREEZE')
+        qga_fsfreeze = true
+    endif
+    if cc.has_header_symbol('linux/fs.h', 'FITRIM')
+        qga_fstrim = true
+    endif
+elif host_os == 'freebsd' and cc.has_header_symbol('ufs/ffs/fs.h', 'UFSSUSPEND')
+    qga_fsfreeze = true
+endif
+
 if get_option('block_drv_ro_whitelist') == ''
   config_host_data.set('CONFIG_BDRV_RO_WHITELIST', '')
 else
@@ -2379,6 +2392,8 @@ config_host_data.set('CONFIG_DEBUG_TCG', get_option('debug_tcg'))
 config_host_data.set('CONFIG_DEBUG_REMAP', get_option('debug_remap'))
 config_host_data.set('CONFIG_QOM_CAST_DEBUG', get_option('qom_cast_debug'))
 config_host_data.set('CONFIG_REPLICATION', get_option('replication').allowed())
+config_host_data.set('CONFIG_FSFREEZE', qga_fsfreeze)
+config_host_data.set('CONFIG_FSTRIM', qga_fstrim)
 
 # has_header
 config_host_data.set('CONFIG_EPOLL', cc.has_header('sys/epoll.h'))
qga/commands-common.h +0 -9
index 8c1c56aac9..263e7c0525 100644
@@ -15,19 +15,10 @@
 
 #if defined(__linux__)
 #include <linux/fs.h>
-#ifdef FIFREEZE
-#define CONFIG_FSFREEZE
-#endif
-#ifdef FITRIM
-#define CONFIG_FSTRIM
-#endif
 #endif /* __linux__ */
 
 #ifdef __FreeBSD__
 #include <ufs/ffs/fs.h>
-#ifdef UFSSUSPEND
-#define CONFIG_FSFREEZE
-#endif
 #endif /* __FreeBSD__ */
 
 #if defined(CONFIG_FSFREEZE) || defined(CONFIG_FSTRIM)
Signature · 1 line
2.45.1
#17 lore raw

PATCH v2 8/22 qga: conditionalize schema for commands unsupported on Windows

In reply to Daniel P. Berrangé

Rather than creating stubs for every command that just return QERR_UNSUPPORTED, use 'if' conditions in the QAPI schema to fully exclude generation of the commands on Windows.

The command will be rejected at QMP dispatch time instead, avoiding reimplementing rejection by blocking the stub commands. This changes the error message for affected commands from

    {"class": "CommandNotFound", "desc": "Command FOO has been disabled"}

to

    {"class": "CommandNotFound", "desc": "The command FOO has not been found"}

This also fixes an accidental inconsistency where some commands (guest-get-diskstats & guest-get-cpustats) are implemented as stubs, yet not added to the blockedrpc list. Those change their error message from

    {"class": "GenericError, "desc": "this feature or command is not currently supported"}

to

    {"class": "CommandNotFound", "desc": "The command FOO has not been found"}

The final additional benefit is that the QGA protocol reference now documents what conditions enable use of the command.

---
 qga/commands-posix.c |  2 +-
 qga/commands-win32.c | 56 +-------------------------------------------
 qga/qapi-schema.json | 45 +++++++++++++++++++++++------------
 3 files changed, 32 insertions(+), 71 deletions(-)
qga/commands-posix.c +1 -1
index 2a3bef7445..0dd8555867 100644
@@ -1280,7 +1280,7 @@ GList *ga_command_init_blockedrpcs(GList *blockedrpcs)
             "guest-get-memory-blocks", "guest-set-memory-blocks",
             "guest-get-memory-block-info",
             NULL};
-        char **p = (char **)list;
+        const char **p = list;
 
         while (*p) {
             blockedrpcs = g_list_append(blockedrpcs, g_strdup(*p++));
qga/commands-win32.c +1 -55

85 lines not on this page · raw message

qga/qapi-schema.json +30 -15

133 lines not on this page · raw message

Signature · 1 line
2.45.1
Reviewed-by on v2 8/22
#18 lore raw

Re: [PATCH v2 08/22] qga: conditionalize schema for commands unsupported on Windows

Quoted 28 lines · Daniel P. Berrangé

Rather than creating stubs for every command that just return QERR_UNSUPPORTED, use 'if' conditions in the QAPI schema to fully exclude generation of the commands on Windows.

The command will be rejected at QMP dispatch time instead, avoiding reimplementing rejection by blocking the stub commands. This changes the error message for affected commands from

     {"class": "CommandNotFound", "desc": "Command FOO has been disabled"}

to

     {"class": "CommandNotFound", "desc": "The command FOO has not been found"}

This also fixes an accidental inconsistency where some commands (guest-get-diskstats & guest-get-cpustats) are implemented as stubs, yet not added to the blockedrpc list. Those change their error message from

     {"class": "GenericError, "desc": "this feature or command is not currently supported"}

to

     {"class": "CommandNotFound", "desc": "The command FOO has not been found"}

The final additional benefit is that the QGA protocol reference now documents what conditions enable use of the command.

Quoted 6 lines
---
  qga/commands-posix.c |  2 +-
  qga/commands-win32.c | 56 +-------------------------------------------
  qga/qapi-schema.json | 45 +++++++++++++++++++++++------------
  3 files changed, 32 insertions(+), 71 deletions(-)
Reviewed-by on v2 8/22
#19 lore raw

Re: [PATCH v2 08/22] qga: conditionalize schema for commands unsupported on Windows

In reply to Daniel P. Berrangé
Quoted 303 lines · Daniel P. Berrangé

Rather than creating stubs for every command that just return QERR_UNSUPPORTED, use 'if' conditions in the QAPI schema to fully exclude generation of the commands on Windows.

The command will be rejected at QMP dispatch time instead, avoiding reimplementing rejection by blocking the stub commands. This changes the error message for affected commands from

    {"class": "CommandNotFound", "desc": "Command FOO has been disabled"}

to

{"class": "CommandNotFound", "desc": "The command FOO has not been
found"}

This also fixes an accidental inconsistency where some commands (guest-get-diskstats & guest-get-cpustats) are implemented as stubs, yet not added to the blockedrpc list. Those change their error message from

{"class": "GenericError, "desc": "this feature or command is not
currently supported"}

to

{"class": "CommandNotFound", "desc": "The command FOO has not been
found"}

The final additional benefit is that the QGA protocol reference now documents what conditions enable use of the command.

---
 qga/commands-posix.c |  2 +-
 qga/commands-win32.c | 56 +-------------------------------------------
 qga/qapi-schema.json | 45 +++++++++++++++++++++++------------
 3 files changed, 32 insertions(+), 71 deletions(-)
qga/commands-posix.c +1 -1
index 2a3bef7445..0dd8555867 100644
@@ -1280,7 +1280,7 @@ GList *ga_command_init_blockedrpcs(GList
 *blockedrpcs)
             "guest-get-memory-blocks", "guest-set-memory-blocks",
             "guest-get-memory-block-info",
             NULL};
-        char **p = (char **)list;
+        const char **p = list;
 
         while (*p) {
             blockedrpcs = g_list_append(blockedrpcs, g_strdup(*p++));
qga/commands-win32.c +0 -11
index 9fe670d5b4..2533e4c748 100644
@@ -1494,11 +1494,6 @@ out:
     }
 }
 
-void qmp_guest_suspend_hybrid(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-}
-
 static IP_ADAPTER_ADDRESSES *guest_get_adapters_addresses(Error **errp)
 {
     IP_ADAPTER_ADDRESSES *adptr_addrs = NULL;
@@ -1862,12 +1857,6 @@ GuestLogicalProcessorList
 *qmp_guest_get_vcpus(Error **errp)
     return NULL;
 }
 
-int64_t qmp_guest_set_vcpus(GuestLogicalProcessorList *vcpus, Error
 **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return -1;
-}
-
 static gchar *

get_net_error_message(gint error)
{

qga/commands-win32.c +1 -44
@@ -1969,46 +1958,15 @@ done:
     g_free(rawpasswddata);
 }
 
-GuestMemoryBlockList *qmp_guest_get_memory_blocks(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
-
-GuestMemoryBlockResponseList *
-qmp_guest_set_memory_blocks(GuestMemoryBlockList *mem_blks, Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
-
-GuestMemoryBlockInfo *qmp_guest_get_memory_block_info(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
-
 /* add unsupported commands to the list of blocked RPCs */
 GList *ga_command_init_blockedrpcs(GList *blockedrpcs)
 {
-    const char *list_unsupported[] = {
-        "guest-suspend-hybrid",
-        "guest-set-vcpus",
-        "guest-get-memory-blocks", "guest-set-memory-blocks",
-        "guest-get-memory-block-info",
-        NULL};
-    char **p = (char **)list_unsupported;
-
-    while (*p) {
-        blockedrpcs = g_list_append(blockedrpcs, g_strdup(*p++));
-    }
-
     if (!vss_init(true)) {
         g_debug("vss_init failed, vss commands are going to be disabled");
         const char *list[] = {
             "guest-get-fsinfo", "guest-fsfreeze-status",
             "guest-fsfreeze-freeze", "guest-fsfreeze-thaw", NULL};
-        p = (char **)list;
+        char **p = (char **)list;
 
         while (*p) {
             blockedrpcs = g_list_append(blockedrpcs, g_strdup(*p++));
@@ -2505,15 +2463,3 @@ char *qga_get_host_name(Error **errp)
 
     return g_utf16_to_utf8(tmp, size, NULL, NULL, NULL);
 }
-
-GuestDiskStatsInfoList *qmp_guest_get_diskstats(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
-
-GuestCpuStatsList *qmp_guest_get_cpustats(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
qga/qapi-schema.json +30 -16
index b3de1fb6b3..b91456e9ad 100644
@@ -636,7 +636,8 @@
 #
 # Since: 1.1
 ##
-{ 'command': 'guest-suspend-hybrid', 'success-response': false }
+{ 'command': 'guest-suspend-hybrid', 'success-response': false,
+  'if': 'CONFIG_POSIX' }
 
 ##
 # @GuestIpAddressType:
@@ -806,7 +807,8 @@
 ##
 { 'command': 'guest-set-vcpus',
   'data':    {'vcpus': ['GuestLogicalProcessor'] },
-  'returns': 'int' }
+  'returns': 'int',
+  'if': 'CONFIG_POSIX' }
 
 ##
 # @GuestDiskBusType:
@@ -1099,7 +1101,8 @@
 { 'struct': 'GuestMemoryBlock',
   'data': {'phys-index': 'uint64',
            'online': 'bool',
-           '*can-offline': 'bool'} }
+           '*can-offline': 'bool'},
+  'if': 'CONFIG_POSIX' }
 
 ##
 # @guest-get-memory-blocks:
@@ -1115,7 +1118,8 @@
 # Since: 2.3
 ##
 { 'command': 'guest-get-memory-blocks',
-  'returns': ['GuestMemoryBlock'] }
+  'returns': ['GuestMemoryBlock'],
+  'if': 'CONFIG_POSIX' }
 
 ##
 # @GuestMemoryBlockResponseType:
@@ -1138,7 +1142,8 @@
 ##
 { 'enum': 'GuestMemoryBlockResponseType',
   'data': ['success', 'not-found', 'operation-not-supported',
-           'operation-failed'] }
+           'operation-failed'],
+  'if': 'CONFIG_POSIX' }
 
 ##
 # @GuestMemoryBlockResponse:
@@ -1156,7 +1161,8 @@
 { 'struct': 'GuestMemoryBlockResponse',
   'data': { 'phys-index': 'uint64',
             'response': 'GuestMemoryBlockResponseType',
-            '*error-code': 'int' }}
+            '*error-code': 'int' },
+  'if': 'CONFIG_POSIX'}
 
 ##
 # @guest-set-memory-blocks:
@@ -1187,7 +1193,8 @@
 ##
 { 'command': 'guest-set-memory-blocks',
   'data':    {'mem-blks': ['GuestMemoryBlock'] },
-  'returns': ['GuestMemoryBlockResponse'] }
+  'returns': ['GuestMemoryBlockResponse'],
+  'if': 'CONFIG_POSIX' }
 
 ##
 # @GuestMemoryBlockInfo:
@@ -1199,7 +1206,8 @@
 # Since: 2.3
 ##
 { 'struct': 'GuestMemoryBlockInfo',
-  'data': {'size': 'uint64'} }
+  'data': {'size': 'uint64'},
+  'if': 'CONFIG_POSIX' }
 
 ##
 # @guest-get-memory-block-info:
@@ -1211,7 +1219,8 @@
 # Since: 2.3
 ##
 { 'command': 'guest-get-memory-block-info',
-  'returns': 'GuestMemoryBlockInfo' }
+  'returns': 'GuestMemoryBlockInfo',
+  'if': 'CONFIG_POSIX' }
 
 ##
 # @GuestExecStatus:
@@ -1702,7 +1711,8 @@
   'data': {'name': 'str',
            'major': 'uint64',
            'minor': 'uint64',
-           'stats': 'GuestDiskStats' } }
+           'stats': 'GuestDiskStats' },
+  'if': 'CONFIG_POSIX' }
 
 ##
 # @guest-get-diskstats:
@@ -1714,7 +1724,8 @@
 # Since: 7.1
 ##
 { 'command': 'guest-get-diskstats',
-  'returns': ['GuestDiskStatsInfo']
+  'returns': ['GuestDiskStatsInfo'],
+  'if': 'CONFIG_POSIX'
 }
 
 ##
@@ -1727,7 +1738,8 @@
 # Since: 7.1
 ##
 { 'enum': 'GuestCpuStatsType',
-  'data': [ 'linux' ] }
+  'data': [ 'linux' ],
+  'if': 'CONFIG_POSIX' }
 
 
 ##
@@ -1772,7 +1784,8 @@
            '*steal': 'uint64',
            '*guest': 'uint64',
            '*guestnice': 'uint64'
-           } }
+           },
+  'if': 'CONFIG_POSIX' }
 
 ##
 # @GuestCpuStats:
@@ -1786,7 +1799,8 @@
 { 'union': 'GuestCpuStats',
   'base': { 'type': 'GuestCpuStatsType' },
   'discriminator': 'type',
-  'data': { 'linux': 'GuestLinuxCpuStats' } }
+  'data': { 'linux': 'GuestLinuxCpuStats' },
+  'if': 'CONFIG_POSIX' }
 
 ##
 # @guest-get-cpustats:
@@ -1798,5 +1812,6 @@
 # Since: 7.1
 ##
 { 'command': 'guest-get-cpustats',
-  'returns': ['GuestCpuStats']
+  'returns': ['GuestCpuStats'],
+  'if': 'CONFIG_POSIX'
 }
--

2.45.1

#20 lore raw

PATCH v2 9/22 qga: conditionalize schema for commands unsupported on non-Linux POSIX

In reply to Daniel P. Berrangé

Rather than creating stubs for every command that just return QERR_UNSUPPORTED, use 'if' conditions in the QAPI schema to fully exclude generation of the commands on non-Linux POSIX platforms

The command will be rejected at QMP dispatch time instead, avoiding reimplementing rejection by blocking the stub commands. This changes the error message for affected commands from

    {"class": "CommandNotFound", "desc": "Command FOO has been disabled"}

to

    {"class": "CommandNotFound", "desc": "The command FOO has not been found"}

This has the additional benefit that the QGA protocol reference now documents what conditions enable use of the command.

---
 qga/commands-posix.c | 66 --------------------------------------------
 qga/qapi-schema.json | 30 +++++++++++---------
 2 files changed, 17 insertions(+), 79 deletions(-)
qga/commands-posix.c +0 -66
index 0dd8555867..559d71ffae 100644
@@ -887,56 +887,6 @@ void qmp_guest_set_user_password(const char *username,
 }
 #endif /* __linux__ || __FreeBSD__ */
 
-#ifndef __linux__
-
-void qmp_guest_suspend_disk(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-}
-
-void qmp_guest_suspend_ram(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-}
-
-void qmp_guest_suspend_hybrid(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-}
-
-GuestLogicalProcessorList *qmp_guest_get_vcpus(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
-
-int64_t qmp_guest_set_vcpus(GuestLogicalProcessorList *vcpus, Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return -1;
-}
-
-GuestMemoryBlockList *qmp_guest_get_memory_blocks(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
-
-GuestMemoryBlockResponseList *
-qmp_guest_set_memory_blocks(GuestMemoryBlockList *mem_blks, Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
-
-GuestMemoryBlockInfo *qmp_guest_get_memory_block_info(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
-
-#endif
-
 #ifdef HAVE_GETIFADDRS
 static GuestNetworkInterface *
 guest_find_interface(GuestNetworkInterfaceList *head,
@@ -1272,22 +1222,6 @@ qmp_guest_fstrim(bool has_minimum, int64_t minimum, Error **errp)
 /* add unsupported commands to the list of blocked RPCs */
 GList *ga_command_init_blockedrpcs(GList *blockedrpcs)
 {
-#if !defined(__linux__)
-    {
-        const char *list[] = {
-            "guest-suspend-disk", "guest-suspend-ram",
-            "guest-suspend-hybrid", "guest-get-vcpus", "guest-set-vcpus",
-            "guest-get-memory-blocks", "guest-set-memory-blocks",
-            "guest-get-memory-block-info",
-            NULL};
-        const char **p = list;
-
-        while (*p) {
-            blockedrpcs = g_list_append(blockedrpcs, g_strdup(*p++));
-        }
-    }
-#endif
-
 #if !defined(HAVE_GETIFADDRS)
     blockedrpcs = g_list_append(blockedrpcs,
                               g_strdup("guest-network-get-interfaces"));
qga/qapi-schema.json +17 -13
index b91456e9ad..d164c30ec3 100644
@@ -565,7 +565,8 @@
 #
 # Since: 1.1
 ##
-{ 'command': 'guest-suspend-disk', 'success-response': false }
+{ 'command': 'guest-suspend-disk', 'success-response': false,
+  'if': { 'any': ['CONFIG_LINUX', 'CONFIG_WIN32'] } }
 
 ##
 # @guest-suspend-ram:
@@ -601,7 +602,8 @@
 #
 # Since: 1.1
 ##
-{ 'command': 'guest-suspend-ram', 'success-response': false }
+{ 'command': 'guest-suspend-ram', 'success-response': false,
+  'if': { 'any': ['CONFIG_LINUX', 'CONFIG_WIN32'] } }
 
 ##
 # @guest-suspend-hybrid:
@@ -637,7 +639,7 @@
 # Since: 1.1
 ##
 { 'command': 'guest-suspend-hybrid', 'success-response': false,
-  'if': 'CONFIG_POSIX' }
+  'if': 'CONFIG_LINUX' }
 
 ##
 # @GuestIpAddressType:
@@ -750,7 +752,8 @@
 { 'struct': 'GuestLogicalProcessor',
   'data': {'logical-id': 'int',
            'online': 'bool',
-           '*can-offline': 'bool'} }
+           '*can-offline': 'bool'},
+  'if': { 'any': ['CONFIG_LINUX', 'CONFIG_WIN32'] } }
 
 ##
 # @guest-get-vcpus:
@@ -765,7 +768,8 @@
 # Since: 1.5
 ##
 { 'command': 'guest-get-vcpus',
-  'returns': ['GuestLogicalProcessor'] }
+  'returns': ['GuestLogicalProcessor'],
+  'if': { 'any': ['CONFIG_LINUX', 'CONFIG_WIN32'] } }
 
 ##
 # @guest-set-vcpus:
@@ -808,7 +812,7 @@
 { 'command': 'guest-set-vcpus',
   'data':    {'vcpus': ['GuestLogicalProcessor'] },
   'returns': 'int',
-  'if': 'CONFIG_POSIX' }
+  'if': 'CONFIG_LINUX' }
 
 ##
 # @GuestDiskBusType:
@@ -1102,7 +1106,7 @@
   'data': {'phys-index': 'uint64',
            'online': 'bool',
            '*can-offline': 'bool'},
-  'if': 'CONFIG_POSIX' }
+  'if': 'CONFIG_LINUX' }
 
 ##
 # @guest-get-memory-blocks:
@@ -1119,7 +1123,7 @@
 ##
 { 'command': 'guest-get-memory-blocks',
   'returns': ['GuestMemoryBlock'],
-  'if': 'CONFIG_POSIX' }
+  'if': 'CONFIG_LINUX' }
 
 ##
 # @GuestMemoryBlockResponseType:
@@ -1143,7 +1147,7 @@
 { 'enum': 'GuestMemoryBlockResponseType',
   'data': ['success', 'not-found', 'operation-not-supported',
            'operation-failed'],
-  'if': 'CONFIG_POSIX' }
+  'if': 'CONFIG_LINUX' }
 
 ##
 # @GuestMemoryBlockResponse:
@@ -1162,7 +1166,7 @@
   'data': { 'phys-index': 'uint64',
             'response': 'GuestMemoryBlockResponseType',
             '*error-code': 'int' },
-  'if': 'CONFIG_POSIX'}
+  'if': 'CONFIG_LINUX'}
 
 ##
 # @guest-set-memory-blocks:
@@ -1194,7 +1198,7 @@
 { 'command': 'guest-set-memory-blocks',
   'data':    {'mem-blks': ['GuestMemoryBlock'] },
   'returns': ['GuestMemoryBlockResponse'],
-  'if': 'CONFIG_POSIX' }
+  'if': 'CONFIG_LINUX' }
 
 ##
 # @GuestMemoryBlockInfo:
@@ -1207,7 +1211,7 @@
 ##
 { 'struct': 'GuestMemoryBlockInfo',
   'data': {'size': 'uint64'},
-  'if': 'CONFIG_POSIX' }
+  'if': 'CONFIG_LINUX' }
 
 ##
 # @guest-get-memory-block-info:
@@ -1220,7 +1224,7 @@
 ##
 { 'command': 'guest-get-memory-block-info',
   'returns': 'GuestMemoryBlockInfo',
-  'if': 'CONFIG_POSIX' }
+  'if': 'CONFIG_LINUX' }
 
 ##
 # @GuestExecStatus:
Signature · 1 line
2.45.1
Reviewed-by on v2 9/22
#21 lore raw

Re: [PATCH v2 09/22] qga: conditionalize schema for commands unsupported on non-Linux POSIX

Quoted 15 lines · Daniel P. Berrangé

Rather than creating stubs for every command that just return QERR_UNSUPPORTED, use 'if' conditions in the QAPI schema to fully exclude generation of the commands on non-Linux POSIX platforms

The command will be rejected at QMP dispatch time instead, avoiding reimplementing rejection by blocking the stub commands. This changes the error message for affected commands from

     {"class": "CommandNotFound", "desc": "Command FOO has been disabled"}

to

     {"class": "CommandNotFound", "desc": "The command FOO has not been found"}
Quoted 8 lines

This has the additional benefit that the QGA protocol reference now documents what conditions enable use of the command.

---
  qga/commands-posix.c | 66 --------------------------------------------
  qga/qapi-schema.json | 30 +++++++++++---------
  2 files changed, 17 insertions(+), 79 deletions(-)
Reviewed-by on v2 9/22
#22 lore raw

Re: [PATCH v2 09/22] qga: conditionalize schema for commands unsupported on non-Linux POSIX

In reply to Daniel P. Berrangé
Quoted 240 lines · Daniel P. Berrangé

Rather than creating stubs for every command that just return QERR_UNSUPPORTED, use 'if' conditions in the QAPI schema to fully exclude generation of the commands on non-Linux POSIX platforms

The command will be rejected at QMP dispatch time instead, avoiding reimplementing rejection by blocking the stub commands. This changes the error message for affected commands from

    {"class": "CommandNotFound", "desc": "Command FOO has been disabled"}

to

{"class": "CommandNotFound", "desc": "The command FOO has not been
found"}

This has the additional benefit that the QGA protocol reference now documents what conditions enable use of the command.

---
 qga/commands-posix.c | 66 --------------------------------------------
 qga/qapi-schema.json | 30 +++++++++++---------
 2 files changed, 17 insertions(+), 79 deletions(-)
qga/commands-posix.c +0 -50
index 0dd8555867..559d71ffae 100644
@@ -887,56 +887,6 @@ void qmp_guest_set_user_password(const char *username,
 }
 #endif /* __linux__ || __FreeBSD__ */
 
-#ifndef __linux__
-
-void qmp_guest_suspend_disk(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-}
-
-void qmp_guest_suspend_ram(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-}
-
-void qmp_guest_suspend_hybrid(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-}
-
-GuestLogicalProcessorList *qmp_guest_get_vcpus(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
-
-int64_t qmp_guest_set_vcpus(GuestLogicalProcessorList *vcpus, Error
 **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return -1;
-}
-
-GuestMemoryBlockList *qmp_guest_get_memory_blocks(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
-
-GuestMemoryBlockResponseList *
-qmp_guest_set_memory_blocks(GuestMemoryBlockList *mem_blks, Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
-
-GuestMemoryBlockInfo *qmp_guest_get_memory_block_info(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
-
-#endif
-
 #ifdef HAVE_GETIFADDRS
 static GuestNetworkInterface *

guest_find_interface(GuestNetworkInterfaceList *head,

qga/commands-posix.c +0 -16
@@ -1272,22 +1222,6 @@ qmp_guest_fstrim(bool has_minimum, int64_t minimum,
 Error **errp)
 /* add unsupported commands to the list of blocked RPCs */
 GList *ga_command_init_blockedrpcs(GList *blockedrpcs)
 {
-#if !defined(__linux__)
-    {
-        const char *list[] = {
-            "guest-suspend-disk", "guest-suspend-ram",
-            "guest-suspend-hybrid", "guest-get-vcpus", "guest-set-vcpus",
-            "guest-get-memory-blocks", "guest-set-memory-blocks",
-            "guest-get-memory-block-info",
-            NULL};
-        const char **p = list;
-
-        while (*p) {
-            blockedrpcs = g_list_append(blockedrpcs, g_strdup(*p++));
-        }
-    }
-#endif
-
 #if !defined(HAVE_GETIFADDRS)
     blockedrpcs = g_list_append(blockedrpcs,
                               g_strdup("guest-network-get-interfaces"));
qga/qapi-schema.json +17 -14
index b91456e9ad..d164c30ec3 100644
@@ -565,7 +565,8 @@
 #
 # Since: 1.1
 ##
-{ 'command': 'guest-suspend-disk', 'success-response': false }
+{ 'command': 'guest-suspend-disk', 'success-response': false,
+  'if': { 'any': ['CONFIG_LINUX', 'CONFIG_WIN32'] } }
 
 ##
 # @guest-suspend-ram:
@@ -601,7 +602,8 @@
 #
 # Since: 1.1
 ##
-{ 'command': 'guest-suspend-ram', 'success-response': false }
+{ 'command': 'guest-suspend-ram', 'success-response': false,
+  'if': { 'any': ['CONFIG_LINUX', 'CONFIG_WIN32'] } }
 
 ##
 # @guest-suspend-hybrid:
@@ -637,7 +639,7 @@
 # Since: 1.1
 ##
 { 'command': 'guest-suspend-hybrid', 'success-response': false,
-  'if': 'CONFIG_POSIX' }
+  'if': 'CONFIG_LINUX' }
 
 ##
 # @GuestIpAddressType:
@@ -750,7 +752,8 @@
 { 'struct': 'GuestLogicalProcessor',
   'data': {'logical-id': 'int',
            'online': 'bool',
-           '*can-offline': 'bool'} }
+           '*can-offline': 'bool'},
+  'if': { 'any': ['CONFIG_LINUX', 'CONFIG_WIN32'] } }
 
 ##
 # @guest-get-vcpus:
@@ -765,7 +768,8 @@
 # Since: 1.5
 ##
 { 'command': 'guest-get-vcpus',
-  'returns': ['GuestLogicalProcessor'] }
+  'returns': ['GuestLogicalProcessor'],
+  'if': { 'any': ['CONFIG_LINUX', 'CONFIG_WIN32'] } }
 
 ##
 # @guest-set-vcpus:
@@ -808,7 +812,7 @@
 { 'command': 'guest-set-vcpus',
   'data':    {'vcpus': ['GuestLogicalProcessor'] },
   'returns': 'int',
-  'if': 'CONFIG_POSIX' }
+  'if': 'CONFIG_LINUX' }
 
 ##
 # @GuestDiskBusType:
@@ -1102,7 +1106,7 @@
   'data': {'phys-index': 'uint64',
            'online': 'bool',
            '*can-offline': 'bool'},
-  'if': 'CONFIG_POSIX' }
+  'if': 'CONFIG_LINUX' }
 
 ##
 # @guest-get-memory-blocks:
@@ -1119,7 +1123,7 @@
 ##
 { 'command': 'guest-get-memory-blocks',
   'returns': ['GuestMemoryBlock'],
-  'if': 'CONFIG_POSIX' }
+  'if': 'CONFIG_LINUX' }
 
 ##
 # @GuestMemoryBlockResponseType:
@@ -1143,7 +1147,7 @@
 { 'enum': 'GuestMemoryBlockResponseType',
   'data': ['success', 'not-found', 'operation-not-supported',
            'operation-failed'],
-  'if': 'CONFIG_POSIX' }
+  'if': 'CONFIG_LINUX' }
 
 ##
 # @GuestMemoryBlockResponse:
@@ -1162,7 +1166,7 @@
   'data': { 'phys-index': 'uint64',
             'response': 'GuestMemoryBlockResponseType',
             '*error-code': 'int' },
-  'if': 'CONFIG_POSIX'}
+  'if': 'CONFIG_LINUX'}
 
 ##
 # @guest-set-memory-blocks:
@@ -1194,7 +1198,7 @@
 { 'command': 'guest-set-memory-blocks',
   'data':    {'mem-blks': ['GuestMemoryBlock'] },
   'returns': ['GuestMemoryBlockResponse'],
-  'if': 'CONFIG_POSIX' }
+  'if': 'CONFIG_LINUX' }
 
 ##
 # @GuestMemoryBlockInfo:
@@ -1207,7 +1211,7 @@
 ##
 { 'struct': 'GuestMemoryBlockInfo',
   'data': {'size': 'uint64'},
-  'if': 'CONFIG_POSIX' }
+  'if': 'CONFIG_LINUX' }
 
 ##
 # @guest-get-memory-block-info:
@@ -1220,7 +1224,7 @@
 ##
 { 'command': 'guest-get-memory-block-info',
   'returns': 'GuestMemoryBlockInfo',
-  'if': 'CONFIG_POSIX' }
+  'if': 'CONFIG_LINUX' }
 
 ##
 # @GuestExecStatus:
--

2.45.1

#23 lore raw

PATCH v2 10/22 qga: conditionalize schema for commands requiring getifaddrs

In reply to Daniel P. Berrangé

Rather than creating stubs for every comamnd that just return QERR_UNSUPPORTED, use 'if' conditions in the QAPI schema to fully exclude generation of the network interface command on POSIX platforms lacking getifaddrs().

The command will be rejected at QMP dispatch time instead, avoiding reimplementing rejection by blocking the stub commands. This changes the error message for affected commands from

{"class": "CommandNotFound", "desc": "Command FOO has been disabled"}

to

    {"class": "CommandNotFound", "desc": "The command FOO has not been found"}

This has the additional benefit that the QGA protocol reference now documents what conditions enable use of the command.

---
 qga/commands-posix.c | 13 -------------
 qga/qapi-schema.json | 15 ++++++++++-----
 2 files changed, 10 insertions(+), 18 deletions(-)
qga/commands-posix.c +0 -13
index 559d71ffae..09d08ee2ca 100644
@@ -1142,14 +1142,6 @@ error:
     return NULL;
 }
 
-#else
-
-GuestNetworkInterfaceList *qmp_guest_network_get_interfaces(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
-
 #endif /* HAVE_GETIFADDRS */
 
 #if !defined(CONFIG_FSFREEZE)
@@ -1222,11 +1214,6 @@ qmp_guest_fstrim(bool has_minimum, int64_t minimum, Error **errp)
 /* add unsupported commands to the list of blocked RPCs */
 GList *ga_command_init_blockedrpcs(GList *blockedrpcs)
 {
-#if !defined(HAVE_GETIFADDRS)
-    blockedrpcs = g_list_append(blockedrpcs,
-                              g_strdup("guest-network-get-interfaces"));
-#endif
-
 #if !defined(CONFIG_FSFREEZE)
     {
         const char *list[] = {
qga/qapi-schema.json +10 -5
index d164c30ec3..c37c904aae 100644
@@ -653,7 +653,8 @@
 # Since: 1.1
 ##
 { 'enum': 'GuestIpAddressType',
-  'data': [ 'ipv4', 'ipv6' ] }
+  'data': [ 'ipv4', 'ipv6' ],
+  'if': { 'any': ['CONFIG_WIN32', 'HAVE_GETIFADDRS'] } }
 
 ##
 # @GuestIpAddress:
@@ -669,7 +670,8 @@
 { 'struct': 'GuestIpAddress',
   'data': {'ip-address': 'str',
            'ip-address-type': 'GuestIpAddressType',
-           'prefix': 'int'} }
+           'prefix': 'int'},
+  'if': { 'any': ['CONFIG_WIN32', 'HAVE_GETIFADDRS'] } }
 
 ##
 # @GuestNetworkInterfaceStat:
@@ -701,7 +703,8 @@
             'tx-packets': 'uint64',
             'tx-errs': 'uint64',
             'tx-dropped': 'uint64'
-           } }
+           },
+  'if': { 'any': ['CONFIG_WIN32', 'HAVE_GETIFADDRS'] } }
 
 ##
 # @GuestNetworkInterface:
@@ -721,7 +724,8 @@
   'data': {'name': 'str',
            '*hardware-address': 'str',
            '*ip-addresses': ['GuestIpAddress'],
-           '*statistics': 'GuestNetworkInterfaceStat' } }
+           '*statistics': 'GuestNetworkInterfaceStat' },
+  'if': { 'any': ['CONFIG_WIN32', 'HAVE_GETIFADDRS'] } }
 
 ##
 # @guest-network-get-interfaces:
@@ -733,7 +737,8 @@
 # Since: 1.1
 ##
 { 'command': 'guest-network-get-interfaces',
-  'returns': ['GuestNetworkInterface'] }
+  'returns': ['GuestNetworkInterface'],
+  'if': { 'any': ['CONFIG_WIN32', 'HAVE_GETIFADDRS'] } }
 
 ##
 # @GuestLogicalProcessor:
Signature · 1 line
2.45.1
Reviewed-by on v2 10/22
#24 lore raw

Re: [PATCH v2 10/22] qga: conditionalize schema for commands requiring getifaddrs

Quoted 15 lines · Daniel P. Berrangé

Rather than creating stubs for every comamnd that just return QERR_UNSUPPORTED, use 'if' conditions in the QAPI schema to fully exclude generation of the network interface command on POSIX platforms lacking getifaddrs().

The command will be rejected at QMP dispatch time instead, avoiding reimplementing rejection by blocking the stub commands. This changes the error message for affected commands from

    {"class": "CommandNotFound", "desc": "Command FOO has been disabled"}

to

     {"class": "CommandNotFound", "desc": "The command FOO has not been found"}
Quoted 8 lines

This has the additional benefit that the QGA protocol reference now documents what conditions enable use of the command.

---
  qga/commands-posix.c | 13 -------------
  qga/qapi-schema.json | 15 ++++++++++-----
  2 files changed, 10 insertions(+), 18 deletions(-)
Reviewed-by on v2 10/22
#25 lore raw

Re: [PATCH v2 10/22] qga: conditionalize schema for commands requiring getifaddrs

In reply to Daniel P. Berrangé
Quoted 115 lines · Daniel P. Berrangé

Rather than creating stubs for every comamnd that just return QERR_UNSUPPORTED, use 'if' conditions in the QAPI schema to fully exclude generation of the network interface command on POSIX platforms lacking getifaddrs().

The command will be rejected at QMP dispatch time instead, avoiding reimplementing rejection by blocking the stub commands. This changes the error message for affected commands from

   {"class": "CommandNotFound", "desc": "Command FOO has been disabled"}

to

{"class": "CommandNotFound", "desc": "The command FOO has not been
found"}

This has the additional benefit that the QGA protocol reference now documents what conditions enable use of the command.

---
 qga/commands-posix.c | 13 -------------
 qga/qapi-schema.json | 15 ++++++++++-----
 2 files changed, 10 insertions(+), 18 deletions(-)
qga/commands-posix.c +0 -13
index 559d71ffae..09d08ee2ca 100644
@@ -1142,14 +1142,6 @@ error:
     return NULL;
 }
 
-#else
-
-GuestNetworkInterfaceList *qmp_guest_network_get_interfaces(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
-
 #endif /* HAVE_GETIFADDRS */
 
 #if !defined(CONFIG_FSFREEZE)
@@ -1222,11 +1214,6 @@ qmp_guest_fstrim(bool has_minimum, int64_t minimum,
 Error **errp)
 /* add unsupported commands to the list of blocked RPCs */
 GList *ga_command_init_blockedrpcs(GList *blockedrpcs)
 {
-#if !defined(HAVE_GETIFADDRS)
-    blockedrpcs = g_list_append(blockedrpcs,
-                              g_strdup("guest-network-get-interfaces"));
-#endif
-
 #if !defined(CONFIG_FSFREEZE)
     {
         const char *list[] = {
qga/qapi-schema.json +10 -6
index d164c30ec3..c37c904aae 100644
@@ -653,7 +653,8 @@
 # Since: 1.1
 ##
 { 'enum': 'GuestIpAddressType',
-  'data': [ 'ipv4', 'ipv6' ] }
+  'data': [ 'ipv4', 'ipv6' ],
+  'if': { 'any': ['CONFIG_WIN32', 'HAVE_GETIFADDRS'] } }
 
 ##
 # @GuestIpAddress:
@@ -669,7 +670,8 @@
 { 'struct': 'GuestIpAddress',
   'data': {'ip-address': 'str',
            'ip-address-type': 'GuestIpAddressType',
-           'prefix': 'int'} }
+           'prefix': 'int'},
+  'if': { 'any': ['CONFIG_WIN32', 'HAVE_GETIFADDRS'] } }
 
 ##
 # @GuestNetworkInterfaceStat:
@@ -701,7 +703,8 @@
             'tx-packets': 'uint64',
             'tx-errs': 'uint64',
             'tx-dropped': 'uint64'
-           } }
+           },
+  'if': { 'any': ['CONFIG_WIN32', 'HAVE_GETIFADDRS'] } }
 
 ##
 # @GuestNetworkInterface:
@@ -721,7 +724,8 @@
   'data': {'name': 'str',
            '*hardware-address': 'str',
            '*ip-addresses': ['GuestIpAddress'],
-           '*statistics': 'GuestNetworkInterfaceStat' } }
+           '*statistics': 'GuestNetworkInterfaceStat' },
+  'if': { 'any': ['CONFIG_WIN32', 'HAVE_GETIFADDRS'] } }
 
 ##
 # @guest-network-get-interfaces:
@@ -733,7 +737,8 @@
 # Since: 1.1
 ##
 { 'command': 'guest-network-get-interfaces',
-  'returns': ['GuestNetworkInterface'] }
+  'returns': ['GuestNetworkInterface'],
+  'if': { 'any': ['CONFIG_WIN32', 'HAVE_GETIFADDRS'] } }
 
 ##
 # @GuestLogicalProcessor:
--

2.45.1

#26 lore raw

PATCH v2 11/22 qga: conditionalize schema for commands requiring linux/win32

In reply to Daniel P. Berrangé

Some commands were blocked based on CONFIG_FSFREEZE, but their impl had nothing todo with CONFIG_FSFREEZE, and were instead either Linux-only, or Win+Linux-only.

Rather than creating stubs for every command that just return QERR_UNSUPPORTED, use 'if' conditions in the QAPI schema to fully exclude generation of the stats and fsinfo commands on platforms that can't support them.

The command will be rejected at QMP dispatch time instead, avoiding reimplementing rejection by blocking the stub commands. This changes the error message for affected commands from

    {"class": "CommandNotFound", "desc": "Command FOO has been disabled"}

to

    {"class": "CommandNotFound", "desc": "The command FOO has not been found"}

This has the additional benefit that the QGA protocol reference now documents what conditions enable use of the command.

---
 qga/commands-bsd.c   | 24 -----------------------
 qga/commands-posix.c | 30 ++---------------------------
 qga/qapi-schema.json | 45 +++++++++++++++++++++++++++-----------------
 3 files changed, 30 insertions(+), 69 deletions(-)
qga/commands-bsd.c +0 -24
index 17bddda1cf..9ce48af311 100644
@@ -149,30 +149,6 @@ int qmp_guest_fsfreeze_do_thaw(Error **errp)
     }
     return ret;
 }
-
-GuestFilesystemInfoList *qmp_guest_get_fsinfo(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
-
-GuestDiskInfoList *qmp_guest_get_disks(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
-
-GuestDiskStatsInfoList *qmp_guest_get_diskstats(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
-
-GuestCpuStatsList *qmp_guest_get_cpustats(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
 #endif /* CONFIG_FSFREEZE */
 
 #ifdef HAVE_GETIFADDRS
qga/commands-posix.c +2 -28
index 09d08ee2ca..838dc3cf98 100644
@@ -1146,12 +1146,6 @@ error:
 
 #if !defined(CONFIG_FSFREEZE)
 
-GuestFilesystemInfoList *qmp_guest_get_fsinfo(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
-
 GuestFsfreezeStatus qmp_guest_fsfreeze_status(Error **errp)
 {
     error_setg(errp, QERR_UNSUPPORTED);
@@ -1181,25 +1175,6 @@ int64_t qmp_guest_fsfreeze_thaw(Error **errp)
 
     return 0;
 }
-
-GuestDiskInfoList *qmp_guest_get_disks(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
-
-GuestDiskStatsInfoList *qmp_guest_get_diskstats(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
-
-GuestCpuStatsList *qmp_guest_get_cpustats(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
-
 #endif /* CONFIG_FSFREEZE */
 
 #if !defined(CONFIG_FSTRIM)
@@ -1217,10 +1192,9 @@ GList *ga_command_init_blockedrpcs(GList *blockedrpcs)
 #if !defined(CONFIG_FSFREEZE)
     {
         const char *list[] = {
-            "guest-get-fsinfo", "guest-fsfreeze-status",
+            "guest-fsfreeze-status",
             "guest-fsfreeze-freeze", "guest-fsfreeze-freeze-list",
-            "guest-fsfreeze-thaw", "guest-get-fsinfo",
-            "guest-get-disks", NULL};
+            "guest-fsfreeze-thaw", NULL};
         char **p = (char **)list;
 
         while (*p) {
qga/qapi-schema.json +28 -17

145 lines not on this page · raw message

Signature · 1 line
2.45.1
#27 lore raw

PATCH v2 12/22 qga: conditionalize schema for commands only supported on Windows

In reply to Daniel P. Berrangé

Rather than creating stubs for every command that just return QERR_UNSUPPORTED, use 'if' conditions in the QAPI schema to fully exclude generation of the commands on non-Windows.

The command will be rejected at QMP dispatch time instead, avoiding reimplementing rejection by blocking the stub commands. This changes the error message for affected commands from

    {"class": "CommandNotFound", "desc": "Command FOO has been disabled"}

to

    {"class": "CommandNotFound", "desc": "The command FOO has not been found"}

This has the additional benefit that the QGA protocol reference now documents what conditions enable use of the command.

---
 qga/commands-posix.c |  9 ---------
 qga/qapi-schema.json | 15 ++++++++++-----
 2 files changed, 10 insertions(+), 14 deletions(-)
qga/commands-posix.c +0 -9
index 838dc3cf98..b7f96aa005 100644
@@ -1207,8 +1207,6 @@ GList *ga_command_init_blockedrpcs(GList *blockedrpcs)
     blockedrpcs = g_list_append(blockedrpcs, g_strdup("guest-fstrim"));
 #endif
 
-    blockedrpcs = g_list_append(blockedrpcs, g_strdup("guest-get-devices"));
-
     return blockedrpcs;
 }
 
@@ -1419,13 +1417,6 @@ GuestOSInfo *qmp_guest_get_osinfo(Error **errp)
     return info;
 }
 
-GuestDeviceInfoList *qmp_guest_get_devices(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-
-    return NULL;
-}
-
 #ifndef HOST_NAME_MAX
 # ifdef _POSIX_HOST_NAME_MAX
 #  define HOST_NAME_MAX _POSIX_HOST_NAME_MAX
qga/qapi-schema.json +10 -5
index 700c5baa5a..2704b814ab 100644
@@ -1527,7 +1527,8 @@
 # @pci: PCI device
 ##
 { 'enum': 'GuestDeviceType',
-  'data': [ 'pci' ] }
+  'data': [ 'pci' ],
+  'if': 'CONFIG_WIN32' }
 
 ##
 # @GuestDeviceIdPCI:
@@ -1539,7 +1540,8 @@
 # Since: 5.2
 ##
 { 'struct': 'GuestDeviceIdPCI',
-  'data': { 'vendor-id': 'uint16', 'device-id': 'uint16' } }
+  'data': { 'vendor-id': 'uint16', 'device-id': 'uint16' },
+  'if': 'CONFIG_WIN32' }
 
 ##
 # @GuestDeviceId:
@@ -1553,7 +1555,8 @@
 { 'union': 'GuestDeviceId',
   'base': { 'type': 'GuestDeviceType' },
   'discriminator': 'type',
-  'data': { 'pci': 'GuestDeviceIdPCI' } }
+  'data': { 'pci': 'GuestDeviceIdPCI' },
+  'if': 'CONFIG_WIN32' }
 
 ##
 # @GuestDeviceInfo:
@@ -1574,7 +1577,8 @@
       '*driver-date': 'int',
       '*driver-version': 'str',
       '*id': 'GuestDeviceId'
-  } }
+  },
+  'if': 'CONFIG_WIN32' }
 
 ##
 # @guest-get-devices:
@@ -1586,7 +1590,8 @@
 # Since: 5.2
 ##
 { 'command': 'guest-get-devices',
-  'returns': ['GuestDeviceInfo'] }
+  'returns': ['GuestDeviceInfo'],
+  'if': 'CONFIG_WIN32' }
 
 ##
 # @GuestAuthorizedKeys:
Signature · 1 line
2.45.1
Reviewed-by on v2 12/22
#28 lore raw

Re: [PATCH v2 12/22] qga: conditionalize schema for commands only supported on Windows

Quoted 14 lines · Daniel P. Berrangé

Rather than creating stubs for every command that just return QERR_UNSUPPORTED, use 'if' conditions in the QAPI schema to fully exclude generation of the commands on non-Windows.

The command will be rejected at QMP dispatch time instead, avoiding reimplementing rejection by blocking the stub commands. This changes the error message for affected commands from

     {"class": "CommandNotFound", "desc": "Command FOO has been disabled"}

to

     {"class": "CommandNotFound", "desc": "The command FOO has not been found"}
Quoted 8 lines

This has the additional benefit that the QGA protocol reference now documents what conditions enable use of the command.

---
  qga/commands-posix.c |  9 ---------
  qga/qapi-schema.json | 15 ++++++++++-----
  2 files changed, 10 insertions(+), 14 deletions(-)
Reviewed-by on v2 12/22
#29 lore raw

Re: [PATCH v2 12/22] qga: conditionalize schema for commands only supported on Windows

In reply to Daniel P. Berrangé
Quoted 111 lines · Daniel P. Berrangé

Rather than creating stubs for every command that just return QERR_UNSUPPORTED, use 'if' conditions in the QAPI schema to fully exclude generation of the commands on non-Windows.

The command will be rejected at QMP dispatch time instead, avoiding reimplementing rejection by blocking the stub commands. This changes the error message for affected commands from

    {"class": "CommandNotFound", "desc": "Command FOO has been disabled"}

to

{"class": "CommandNotFound", "desc": "The command FOO has not been
found"}

This has the additional benefit that the QGA protocol reference now documents what conditions enable use of the command.

---
 qga/commands-posix.c |  9 ---------
 qga/qapi-schema.json | 15 ++++++++++-----
 2 files changed, 10 insertions(+), 14 deletions(-)
qga/commands-posix.c +0 -2
index 838dc3cf98..b7f96aa005 100644
@@ -1207,8 +1207,6 @@ GList *ga_command_init_blockedrpcs(GList
 *blockedrpcs)
     blockedrpcs = g_list_append(blockedrpcs, g_strdup("guest-fstrim"));
 #endif
 
-    blockedrpcs = g_list_append(blockedrpcs,
 g_strdup("guest-get-devices"));
-
     return blockedrpcs;

}

qga/commands-posix.c +0 -7
@@ -1419,13 +1417,6 @@ GuestOSInfo *qmp_guest_get_osinfo(Error **errp)
     return info;
 }
 
-GuestDeviceInfoList *qmp_guest_get_devices(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-
-    return NULL;
-}
-
 #ifndef HOST_NAME_MAX
 # ifdef _POSIX_HOST_NAME_MAX
 #  define HOST_NAME_MAX _POSIX_HOST_NAME_MAX
qga/qapi-schema.json +10 -6
index 700c5baa5a..2704b814ab 100644
@@ -1527,7 +1527,8 @@
 # @pci: PCI device
 ##
 { 'enum': 'GuestDeviceType',
-  'data': [ 'pci' ] }
+  'data': [ 'pci' ],
+  'if': 'CONFIG_WIN32' }
 
 ##
 # @GuestDeviceIdPCI:
@@ -1539,7 +1540,8 @@
 # Since: 5.2
 ##
 { 'struct': 'GuestDeviceIdPCI',
-  'data': { 'vendor-id': 'uint16', 'device-id': 'uint16' } }
+  'data': { 'vendor-id': 'uint16', 'device-id': 'uint16' },
+  'if': 'CONFIG_WIN32' }
 
 ##
 # @GuestDeviceId:
@@ -1553,7 +1555,8 @@
 { 'union': 'GuestDeviceId',
   'base': { 'type': 'GuestDeviceType' },
   'discriminator': 'type',
-  'data': { 'pci': 'GuestDeviceIdPCI' } }
+  'data': { 'pci': 'GuestDeviceIdPCI' },
+  'if': 'CONFIG_WIN32' }
 
 ##
 # @GuestDeviceInfo:
@@ -1574,7 +1577,8 @@
       '*driver-date': 'int',
       '*driver-version': 'str',
       '*id': 'GuestDeviceId'
-  } }
+  },
+  'if': 'CONFIG_WIN32' }
 
 ##
 # @guest-get-devices:
@@ -1586,7 +1590,8 @@
 # Since: 5.2
 ##
 { 'command': 'guest-get-devices',
-  'returns': ['GuestDeviceInfo'] }
+  'returns': ['GuestDeviceInfo'],
+  'if': 'CONFIG_WIN32' }
 
 ##
 # @GuestAuthorizedKeys:
--

2.45.1

#30 lore raw

PATCH v2 13/22 qga: conditionalize schema for commands requiring fsfreeze

In reply to Daniel P. Berrangé

Rather than creating stubs for every command that just return QERR_UNSUPPORTED, use 'if' conditions in the schema to fully exclude generation of the filesystem freezing commands on POSIX platforms lacking the required APIs.

The command will be rejected at QMP dispatch time instead, avoiding reimplementing rejection by blocking the stub commands. This changes the error message for affected commands from

    {"class": "CommandNotFound", "desc": "Command FOO has been disabled"}

to

    {"class": "CommandNotFound", "desc": "The command FOO has not been found"}

This has the additional benefit that the QGA protocol reference now documents what conditions enable use of the command.

---
 qga/commands-posix.c | 47 --------------------------------------------
 qga/qapi-schema.json | 15 +++++++++-----
 2 files changed, 10 insertions(+), 52 deletions(-)
qga/commands-posix.c +0 -47
index b7f96aa005..9207cb7a8f 100644
@@ -1144,39 +1144,6 @@ error:
 
 #endif /* HAVE_GETIFADDRS */
 
-#if !defined(CONFIG_FSFREEZE)
-
-GuestFsfreezeStatus qmp_guest_fsfreeze_status(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-
-    return 0;
-}
-
-int64_t qmp_guest_fsfreeze_freeze(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-
-    return 0;
-}
-
-int64_t qmp_guest_fsfreeze_freeze_list(bool has_mountpoints,
-                                       strList *mountpoints,
-                                       Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-
-    return 0;
-}
-
-int64_t qmp_guest_fsfreeze_thaw(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-
-    return 0;
-}
-#endif /* CONFIG_FSFREEZE */
-
 #if !defined(CONFIG_FSTRIM)
 GuestFilesystemTrimResponse *
 qmp_guest_fstrim(bool has_minimum, int64_t minimum, Error **errp)
@@ -1189,20 +1156,6 @@ qmp_guest_fstrim(bool has_minimum, int64_t minimum, Error **errp)
 /* add unsupported commands to the list of blocked RPCs */
 GList *ga_command_init_blockedrpcs(GList *blockedrpcs)
 {
-#if !defined(CONFIG_FSFREEZE)
-    {
-        const char *list[] = {
-            "guest-fsfreeze-status",
-            "guest-fsfreeze-freeze", "guest-fsfreeze-freeze-list",
-            "guest-fsfreeze-thaw", NULL};
-        char **p = (char **)list;
-
-        while (*p) {
-            blockedrpcs = g_list_append(blockedrpcs, g_strdup(*p++));
-        }
-    }
-#endif
-
 #if !defined(CONFIG_FSTRIM)
     blockedrpcs = g_list_append(blockedrpcs, g_strdup("guest-fstrim"));
 #endif
qga/qapi-schema.json +10 -5
index 2704b814ab..098fa7a08b 100644
@@ -412,7 +412,8 @@
 # Since: 0.15.0
 ##
 { 'enum': 'GuestFsfreezeStatus',
-  'data': [ 'thawed', 'frozen' ] }
+  'data': [ 'thawed', 'frozen' ],
+  'if': { 'any': ['CONFIG_WIN32', 'CONFIG_FSFREEZE'] } }
 
 ##
 # @guest-fsfreeze-status:
@@ -428,7 +429,8 @@
 # Since: 0.15.0
 ##
 { 'command': 'guest-fsfreeze-status',
-  'returns': 'GuestFsfreezeStatus' }
+  'returns': 'GuestFsfreezeStatus',
+  'if': { 'any': ['CONFIG_WIN32', 'CONFIG_FSFREEZE'] } }
 
 ##
 # @guest-fsfreeze-freeze:
@@ -450,7 +452,8 @@
 # Since: 0.15.0
 ##
 { 'command': 'guest-fsfreeze-freeze',
-  'returns': 'int' }
+  'returns': 'int',
+  'if': { 'any': ['CONFIG_WIN32', 'CONFIG_FSFREEZE'] } }
 
 ##
 # @guest-fsfreeze-freeze-list:
@@ -470,7 +473,8 @@
 ##
 { 'command': 'guest-fsfreeze-freeze-list',
   'data':    { '*mountpoints': ['str'] },
-  'returns': 'int' }
+  'returns': 'int',
+  'if': { 'any': ['CONFIG_WIN32', 'CONFIG_FSFREEZE'] } }
 
 ##
 # @guest-fsfreeze-thaw:
@@ -487,7 +491,8 @@
 # Since: 0.15.0
 ##
 { 'command': 'guest-fsfreeze-thaw',
-  'returns': 'int' }
+  'returns': 'int',
+  'if': { 'any': ['CONFIG_WIN32', 'CONFIG_FSFREEZE'] } }
 
 ##
 # @GuestFilesystemTrimResult:
Signature · 1 line
2.45.1
Reviewed-by on v2 13/22
#31 lore raw

Re: [PATCH v2 13/22] qga: conditionalize schema for commands requiring fsfreeze

Quoted 15 lines · Daniel P. Berrangé

Rather than creating stubs for every command that just return QERR_UNSUPPORTED, use 'if' conditions in the schema to fully exclude generation of the filesystem freezing commands on POSIX platforms lacking the required APIs.

The command will be rejected at QMP dispatch time instead, avoiding reimplementing rejection by blocking the stub commands. This changes the error message for affected commands from

     {"class": "CommandNotFound", "desc": "Command FOO has been disabled"}

to

     {"class": "CommandNotFound", "desc": "The command FOO has not been found"}
Quoted 8 lines

This has the additional benefit that the QGA protocol reference now documents what conditions enable use of the command.

---
  qga/commands-posix.c | 47 --------------------------------------------
  qga/qapi-schema.json | 15 +++++++++-----
  2 files changed, 10 insertions(+), 52 deletions(-)
Reviewed-by on v2 13/22
#32 lore raw

Re: [PATCH v2 13/22] qga: conditionalize schema for commands requiring fsfreeze

In reply to Daniel P. Berrangé
Quoted 149 lines · Daniel P. Berrangé

Rather than creating stubs for every command that just return QERR_UNSUPPORTED, use 'if' conditions in the schema to fully exclude generation of the filesystem freezing commands on POSIX platforms lacking the required APIs.

The command will be rejected at QMP dispatch time instead, avoiding reimplementing rejection by blocking the stub commands. This changes the error message for affected commands from

    {"class": "CommandNotFound", "desc": "Command FOO has been disabled"}

to

{"class": "CommandNotFound", "desc": "The command FOO has not been
found"}

This has the additional benefit that the QGA protocol reference now documents what conditions enable use of the command.

---
 qga/commands-posix.c | 47 --------------------------------------------
 qga/qapi-schema.json | 15 +++++++++-----
 2 files changed, 10 insertions(+), 52 deletions(-)
qga/commands-posix.c +0 -47
index b7f96aa005..9207cb7a8f 100644
@@ -1144,39 +1144,6 @@ error:
 
 #endif /* HAVE_GETIFADDRS */
 
-#if !defined(CONFIG_FSFREEZE)
-
-GuestFsfreezeStatus qmp_guest_fsfreeze_status(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-
-    return 0;
-}
-
-int64_t qmp_guest_fsfreeze_freeze(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-
-    return 0;
-}
-
-int64_t qmp_guest_fsfreeze_freeze_list(bool has_mountpoints,
-                                       strList *mountpoints,
-                                       Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-
-    return 0;
-}
-
-int64_t qmp_guest_fsfreeze_thaw(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-
-    return 0;
-}
-#endif /* CONFIG_FSFREEZE */
-
 #if !defined(CONFIG_FSTRIM)
 GuestFilesystemTrimResponse *
 qmp_guest_fstrim(bool has_minimum, int64_t minimum, Error **errp)
@@ -1189,20 +1156,6 @@ qmp_guest_fstrim(bool has_minimum, int64_t minimum,
 Error **errp)
 /* add unsupported commands to the list of blocked RPCs */
 GList *ga_command_init_blockedrpcs(GList *blockedrpcs)
 {
-#if !defined(CONFIG_FSFREEZE)
-    {
-        const char *list[] = {
-            "guest-fsfreeze-status",
-            "guest-fsfreeze-freeze", "guest-fsfreeze-freeze-list",
-            "guest-fsfreeze-thaw", NULL};
-        char **p = (char **)list;
-
-        while (*p) {
-            blockedrpcs = g_list_append(blockedrpcs, g_strdup(*p++));
-        }
-    }
-#endif
-
 #if !defined(CONFIG_FSTRIM)
     blockedrpcs = g_list_append(blockedrpcs, g_strdup("guest-fstrim"));

#endif

qga/qapi-schema.json +10 -6
index 2704b814ab..098fa7a08b 100644
@@ -412,7 +412,8 @@
 # Since: 0.15.0
 ##
 { 'enum': 'GuestFsfreezeStatus',
-  'data': [ 'thawed', 'frozen' ] }
+  'data': [ 'thawed', 'frozen' ],
+  'if': { 'any': ['CONFIG_WIN32', 'CONFIG_FSFREEZE'] } }
 
 ##
 # @guest-fsfreeze-status:
@@ -428,7 +429,8 @@
 # Since: 0.15.0
 ##
 { 'command': 'guest-fsfreeze-status',
-  'returns': 'GuestFsfreezeStatus' }
+  'returns': 'GuestFsfreezeStatus',
+  'if': { 'any': ['CONFIG_WIN32', 'CONFIG_FSFREEZE'] } }
 
 ##
 # @guest-fsfreeze-freeze:
@@ -450,7 +452,8 @@
 # Since: 0.15.0
 ##
 { 'command': 'guest-fsfreeze-freeze',
-  'returns': 'int' }
+  'returns': 'int',
+  'if': { 'any': ['CONFIG_WIN32', 'CONFIG_FSFREEZE'] } }
 
 ##
 # @guest-fsfreeze-freeze-list:
@@ -470,7 +473,8 @@
 ##
 { 'command': 'guest-fsfreeze-freeze-list',
   'data':    { '*mountpoints': ['str'] },
-  'returns': 'int' }
+  'returns': 'int',
+  'if': { 'any': ['CONFIG_WIN32', 'CONFIG_FSFREEZE'] } }
 
 ##
 # @guest-fsfreeze-thaw:
@@ -487,7 +491,8 @@
 # Since: 0.15.0
 ##
 { 'command': 'guest-fsfreeze-thaw',
-  'returns': 'int' }
+  'returns': 'int',
+  'if': { 'any': ['CONFIG_WIN32', 'CONFIG_FSFREEZE'] } }
 
 ##
 # @GuestFilesystemTrimResult:
--

2.45.1

#33 lore raw

PATCH v2 14/22 qga: conditionalize schema for commands requiring fstrim

In reply to Daniel P. Berrangé

Rather than creating stubs for every command that just return QERR_UNSUPPORTED, use 'if' conditions in the QAPI schema to fully exclude generation of the filesystem trimming commands on POSIX platforms lacking required APIs.

The command will be rejected at QMP dispatch time instead, avoiding reimplementing rejection by blocking the stub commands. This changes the error message for affected commands from

    {"class": "CommandNotFound", "desc": "Command FOO has been disabled"}

to

    {"class": "CommandNotFound", "desc": "The command FOO has not been found"}

This has the additional benefit that the QGA protocol reference now documents what conditions enable use of the command.

---
 qga/commands-posix.c | 13 -------------
 qga/qapi-schema.json |  9 ++++++---
 2 files changed, 6 insertions(+), 16 deletions(-)
qga/commands-posix.c +0 -13
index 9207cb7a8f..d92fa0ec87 100644
@@ -1144,22 +1144,9 @@ error:
 
 #endif /* HAVE_GETIFADDRS */
 
-#if !defined(CONFIG_FSTRIM)
-GuestFilesystemTrimResponse *
-qmp_guest_fstrim(bool has_minimum, int64_t minimum, Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
-#endif
-
 /* add unsupported commands to the list of blocked RPCs */
 GList *ga_command_init_blockedrpcs(GList *blockedrpcs)
 {
-#if !defined(CONFIG_FSTRIM)
-    blockedrpcs = g_list_append(blockedrpcs, g_strdup("guest-fstrim"));
-#endif
-
     return blockedrpcs;
 }
 
qga/qapi-schema.json +6 -3
index 098fa7a08b..0f27375ea0 100644
@@ -509,7 +509,8 @@
 ##
 { 'struct': 'GuestFilesystemTrimResult',
   'data': {'path': 'str',
-           '*trimmed': 'int', '*minimum': 'int', '*error': 'str'} }
+           '*trimmed': 'int', '*minimum': 'int', '*error': 'str'},
+  'if': { 'any': ['CONFIG_WIN32', 'CONFIG_FSTRIM'] } }
 
 ##
 # @GuestFilesystemTrimResponse:
@@ -519,7 +520,8 @@
 # Since: 2.4
 ##
 { 'struct': 'GuestFilesystemTrimResponse',
-  'data': {'paths': ['GuestFilesystemTrimResult']} }
+  'data': {'paths': ['GuestFilesystemTrimResult']},
+  'if': { 'any': ['CONFIG_WIN32', 'CONFIG_FSTRIM'] } }
 
 ##
 # @guest-fstrim:
@@ -541,7 +543,8 @@
 ##
 { 'command': 'guest-fstrim',
   'data': { '*minimum': 'int' },
-  'returns': 'GuestFilesystemTrimResponse' }
+  'returns': 'GuestFilesystemTrimResponse',
+  'if': { 'any': ['CONFIG_WIN32', 'CONFIG_FSTRIM'] } }
 
 ##
 # @guest-suspend-disk:
Signature · 1 line
2.45.1
Reviewed-by on v2 14/22
#34 lore raw

Re: [PATCH v2 14/22] qga: conditionalize schema for commands requiring fstrim

Quoted 15 lines · Daniel P. Berrangé

Rather than creating stubs for every command that just return QERR_UNSUPPORTED, use 'if' conditions in the QAPI schema to fully exclude generation of the filesystem trimming commands on POSIX platforms lacking required APIs.

The command will be rejected at QMP dispatch time instead, avoiding reimplementing rejection by blocking the stub commands. This changes the error message for affected commands from

     {"class": "CommandNotFound", "desc": "Command FOO has been disabled"}

to

     {"class": "CommandNotFound", "desc": "The command FOO has not been found"}
Quoted 8 lines

This has the additional benefit that the QGA protocol reference now documents what conditions enable use of the command.

---
  qga/commands-posix.c | 13 -------------
  qga/qapi-schema.json |  9 ++++++---
  2 files changed, 6 insertions(+), 16 deletions(-)
Reviewed-by on v2 14/22
#35 lore raw

Re: [PATCH v2 14/22] qga: conditionalize schema for commands requiring fstrim

In reply to Daniel P. Berrangé
Quoted 90 lines · Daniel P. Berrangé

Rather than creating stubs for every command that just return QERR_UNSUPPORTED, use 'if' conditions in the QAPI schema to fully exclude generation of the filesystem trimming commands on POSIX platforms lacking required APIs.

The command will be rejected at QMP dispatch time instead, avoiding reimplementing rejection by blocking the stub commands. This changes the error message for affected commands from

    {"class": "CommandNotFound", "desc": "Command FOO has been disabled"}

to

{"class": "CommandNotFound", "desc": "The command FOO has not been
found"}

This has the additional benefit that the QGA protocol reference now documents what conditions enable use of the command.

---
 qga/commands-posix.c | 13 -------------
 qga/qapi-schema.json |  9 ++++++---
 2 files changed, 6 insertions(+), 16 deletions(-)
qga/commands-posix.c +0 -13
index 9207cb7a8f..d92fa0ec87 100644
@@ -1144,22 +1144,9 @@ error:
 
 #endif /* HAVE_GETIFADDRS */
 
-#if !defined(CONFIG_FSTRIM)
-GuestFilesystemTrimResponse *
-qmp_guest_fstrim(bool has_minimum, int64_t minimum, Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
-#endif
-
 /* add unsupported commands to the list of blocked RPCs */
 GList *ga_command_init_blockedrpcs(GList *blockedrpcs)
 {
-#if !defined(CONFIG_FSTRIM)
-    blockedrpcs = g_list_append(blockedrpcs, g_strdup("guest-fstrim"));
-#endif
-
     return blockedrpcs;
 }
 
qga/qapi-schema.json +6 -4
index 098fa7a08b..0f27375ea0 100644
@@ -509,7 +509,8 @@
 ##
 { 'struct': 'GuestFilesystemTrimResult',
   'data': {'path': 'str',
-           '*trimmed': 'int', '*minimum': 'int', '*error': 'str'} }
+           '*trimmed': 'int', '*minimum': 'int', '*error': 'str'},
+  'if': { 'any': ['CONFIG_WIN32', 'CONFIG_FSTRIM'] } }
 
 ##
 # @GuestFilesystemTrimResponse:
@@ -519,7 +520,8 @@
 # Since: 2.4
 ##
 { 'struct': 'GuestFilesystemTrimResponse',
-  'data': {'paths': ['GuestFilesystemTrimResult']} }
+  'data': {'paths': ['GuestFilesystemTrimResult']},
+  'if': { 'any': ['CONFIG_WIN32', 'CONFIG_FSTRIM'] } }
 
 ##
 # @guest-fstrim:
@@ -541,7 +543,8 @@
 ##
 { 'command': 'guest-fstrim',
   'data': { '*minimum': 'int' },
-  'returns': 'GuestFilesystemTrimResponse' }
+  'returns': 'GuestFilesystemTrimResponse',
+  'if': { 'any': ['CONFIG_WIN32', 'CONFIG_FSTRIM'] } }
 
 ##
 # @guest-suspend-disk:
--

2.45.1

#36 lore raw

PATCH v2 15/22 qga: conditionalize schema for commands requiring libudev

In reply to Daniel P. Berrangé

Rather than creating stubs for every command that just return QERR_UNSUPPORTED, use 'if' conditions in the schema to fully exclude generation of the filesystem trimming commands on POSIX platforms lacking required APIs.

The command will be rejected at QMP dispatch time instead, avoiding reimplementing rejection by blocking the stub commands. This changes the error message for affected commands from

    {"class": "CommandNotFound", "desc": "Command FOO has been disabled"}

to

    {"class": "CommandNotFound", "desc": "The command FOO has not been found"}

This has the additional benefit that the QGA protocol reference now documents what conditions enable use of the command.

---
 qga/commands-linux.c | 8 --------
 qga/qapi-schema.json | 8 ++++----
 2 files changed, 4 insertions(+), 12 deletions(-)
qga/commands-linux.c +0 -8
index 73b13fbaf6..89bdcded01 100644
@@ -1049,14 +1049,6 @@ GuestDiskInfoList *qmp_guest_get_disks(Error **errp)
     return ret;
 }
 
-#else
-
-GuestDiskInfoList *qmp_guest_get_disks(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
-
 #endif
 
 /* Return a list of the disk device(s)' info which @mount lies on */
qga/qapi-schema.json +4 -4
index 0f27375ea0..0b7f911ca5 100644
@@ -985,7 +985,7 @@
            'media-errors-hi': 'uint64',
            'number-of-error-log-entries-lo': 'uint64',
            'number-of-error-log-entries-hi': 'uint64' },
-  'if': { 'any': [ 'CONFIG_WIN32', 'CONFIG_LINUX' ] } }
+  'if': { 'any': [ 'CONFIG_WIN32', 'CONFIG_LIBUDEV' ] } }
 
 ##
 # @GuestDiskSmart:
@@ -1000,7 +1000,7 @@
   'base': { 'type': 'GuestDiskBusType' },
   'discriminator': 'type',
   'data': { 'nvme': 'GuestNVMeSmart' },
-  'if': { 'any': [ 'CONFIG_WIN32', 'CONFIG_LINUX' ] } }
+  'if': { 'any': [ 'CONFIG_WIN32', 'CONFIG_LIBUDEV' ] } }
 
 ##
 # @GuestDiskInfo:
@@ -1026,7 +1026,7 @@
   'data': {'name': 'str', 'partition': 'bool', '*dependencies': ['str'],
            '*address': 'GuestDiskAddress', '*alias': 'str',
            '*smart': 'GuestDiskSmart'},
-  'if': { 'any': [ 'CONFIG_WIN32', 'CONFIG_LINUX' ] } }
+  'if': { 'any': [ 'CONFIG_WIN32', 'CONFIG_LIBUDEV' ] } }
 
 ##
 # @guest-get-disks:
@@ -1040,7 +1040,7 @@
 ##
 { 'command': 'guest-get-disks',
   'returns': ['GuestDiskInfo'],
-  'if': { 'any': [ 'CONFIG_WIN32', 'CONFIG_LINUX' ] } }
+  'if': { 'any': [ 'CONFIG_WIN32', 'CONFIG_LIBUDEV' ] } }
 
 ##
 # @GuestFilesystemInfo:
Signature · 1 line
2.45.1
Reviewed-by on v2 15/22
#37 lore raw

Re: [PATCH v2 15/22] qga: conditionalize schema for commands requiring libudev

Quoted 15 lines · Daniel P. Berrangé

Rather than creating stubs for every command that just return QERR_UNSUPPORTED, use 'if' conditions in the schema to fully exclude generation of the filesystem trimming commands on POSIX platforms lacking required APIs.

The command will be rejected at QMP dispatch time instead, avoiding reimplementing rejection by blocking the stub commands. This changes the error message for affected commands from

     {"class": "CommandNotFound", "desc": "Command FOO has been disabled"}

to

     {"class": "CommandNotFound", "desc": "The command FOO has not been found"}
Quoted 8 lines

This has the additional benefit that the QGA protocol reference now documents what conditions enable use of the command.

---
  qga/commands-linux.c | 8 --------
  qga/qapi-schema.json | 8 ++++----
  2 files changed, 4 insertions(+), 12 deletions(-)
Reviewed-by on v2 15/22
#38 lore raw

Re: [PATCH v2 15/22] qga: conditionalize schema for commands requiring libudev

In reply to Daniel P. Berrangé
Quoted 88 lines · Daniel P. Berrangé

Rather than creating stubs for every command that just return QERR_UNSUPPORTED, use 'if' conditions in the schema to fully exclude generation of the filesystem trimming commands on POSIX platforms lacking required APIs.

The command will be rejected at QMP dispatch time instead, avoiding reimplementing rejection by blocking the stub commands. This changes the error message for affected commands from

    {"class": "CommandNotFound", "desc": "Command FOO has been disabled"}

to

{"class": "CommandNotFound", "desc": "The command FOO has not been
found"}

This has the additional benefit that the QGA protocol reference now documents what conditions enable use of the command.

---
 qga/commands-linux.c | 8 --------
 qga/qapi-schema.json | 8 ++++----
 2 files changed, 4 insertions(+), 12 deletions(-)
qga/commands-linux.c +0 -8
index 73b13fbaf6..89bdcded01 100644
@@ -1049,14 +1049,6 @@ GuestDiskInfoList *qmp_guest_get_disks(Error **errp)
     return ret;
 }
 
-#else
-
-GuestDiskInfoList *qmp_guest_get_disks(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
-
 #endif
 
 /* Return a list of the disk device(s)' info which @mount lies on */
qga/qapi-schema.json +4 -5
index 0f27375ea0..0b7f911ca5 100644
@@ -985,7 +985,7 @@
            'media-errors-hi': 'uint64',
            'number-of-error-log-entries-lo': 'uint64',
            'number-of-error-log-entries-hi': 'uint64' },
-  'if': { 'any': [ 'CONFIG_WIN32', 'CONFIG_LINUX' ] } }
+  'if': { 'any': [ 'CONFIG_WIN32', 'CONFIG_LIBUDEV' ] } }
 
 ##
 # @GuestDiskSmart:
@@ -1000,7 +1000,7 @@
   'base': { 'type': 'GuestDiskBusType' },
   'discriminator': 'type',
   'data': { 'nvme': 'GuestNVMeSmart' },
-  'if': { 'any': [ 'CONFIG_WIN32', 'CONFIG_LINUX' ] } }
+  'if': { 'any': [ 'CONFIG_WIN32', 'CONFIG_LIBUDEV' ] } }
 
 ##
 # @GuestDiskInfo:
@@ -1026,7 +1026,7 @@
   'data': {'name': 'str', 'partition': 'bool', '*dependencies': ['str'],
            '*address': 'GuestDiskAddress', '*alias': 'str',
            '*smart': 'GuestDiskSmart'},
-  'if': { 'any': [ 'CONFIG_WIN32', 'CONFIG_LINUX' ] } }
+  'if': { 'any': [ 'CONFIG_WIN32', 'CONFIG_LIBUDEV' ] } }
 
 ##
 # @guest-get-disks:
@@ -1040,7 +1040,7 @@
 ##
 { 'command': 'guest-get-disks',
   'returns': ['GuestDiskInfo'],
-  'if': { 'any': [ 'CONFIG_WIN32', 'CONFIG_LINUX' ] } }
+  'if': { 'any': [ 'CONFIG_WIN32', 'CONFIG_LIBUDEV' ] } }
 
 ##
 # @GuestFilesystemInfo:
--

2.45.1

#39 lore raw

PATCH v2 16/22 qga: conditionalize schema for commands requiring utmpx

In reply to Daniel P. Berrangé

Rather than creating stubs for every command that just return QERR_UNSUPPORTED, use 'if' conditions in the QAPI schema to fully exclude generation of the get-users command on POSIX platforms lacking required APIs.

The command will be rejected at QMP dispatch time instead, avoiding reimplementing rejection by blocking the stub commands. This changes the error message for affected commands from

    {"class": "CommandNotFound", "desc": "Command FOO has been disabled"}

to

    {"class": "CommandNotFound", "desc": "The command FOO has not been found"}

This has the additional benefit that the QGA protocol reference now documents what conditions enable use of the command.

---
 qga/commands-posix.c | 10 +---------
 qga/qapi-schema.json |  6 ++++--
 2 files changed, 5 insertions(+), 11 deletions(-)
qga/commands-posix.c +1 -9
index d92fa0ec87..a353f64ae6 100644
@@ -1212,15 +1212,7 @@ GuestUserList *qmp_guest_get_users(Error **errp)
     return head;
 }
 
-#else
-
-GuestUserList *qmp_guest_get_users(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
-
-#endif
+#endif /* HAVE_UTMPX */
 
 /* Replace escaped special characters with their real values. The replacement
  * is done in place -- returned value is in the original string.
qga/qapi-schema.json +4 -2
index 0b7f911ca5..70d4f173ad 100644
@@ -1413,7 +1413,8 @@
 # Since: 2.10
 ##
 { 'struct': 'GuestUser',
-  'data': { 'user': 'str', 'login-time': 'number', '*domain': 'str' } }
+  'data': { 'user': 'str', 'login-time': 'number', '*domain': 'str' },
+  'if': { 'any': ['CONFIG_WIN32', 'HAVE_UTMPX' ] } }
 
 ##
 # @guest-get-users:
@@ -1425,7 +1426,8 @@
 # Since: 2.10
 ##
 { 'command': 'guest-get-users',
-  'returns': ['GuestUser'] }
+  'returns': ['GuestUser'],
+  'if': { 'any': ['CONFIG_WIN32', 'HAVE_UTMPX' ] } }
 
 ##
 # @GuestTimezone:
Signature · 1 line
2.45.1
Reviewed-by on v2 16/22
#40 lore raw

Re: [PATCH v2 16/22] qga: conditionalize schema for commands requiring utmpx

Quoted 15 lines · Daniel P. Berrangé

Rather than creating stubs for every command that just return QERR_UNSUPPORTED, use 'if' conditions in the QAPI schema to fully exclude generation of the get-users command on POSIX platforms lacking required APIs.

The command will be rejected at QMP dispatch time instead, avoiding reimplementing rejection by blocking the stub commands. This changes the error message for affected commands from

     {"class": "CommandNotFound", "desc": "Command FOO has been disabled"}

to

     {"class": "CommandNotFound", "desc": "The command FOO has not been found"}
Quoted 8 lines

This has the additional benefit that the QGA protocol reference now documents what conditions enable use of the command.

---
  qga/commands-posix.c | 10 +---------
  qga/qapi-schema.json |  6 ++++--
  2 files changed, 5 insertions(+), 11 deletions(-)
Reviewed-by on v2 16/22
#41 lore raw

Re: [PATCH v2 16/22] qga: conditionalize schema for commands requiring utmpx

In reply to Daniel P. Berrangé
Quoted 75 lines · Daniel P. Berrangé

Rather than creating stubs for every command that just return QERR_UNSUPPORTED, use 'if' conditions in the QAPI schema to fully exclude generation of the get-users command on POSIX platforms lacking required APIs.

The command will be rejected at QMP dispatch time instead, avoiding reimplementing rejection by blocking the stub commands. This changes the error message for affected commands from

    {"class": "CommandNotFound", "desc": "Command FOO has been disabled"}

to

{"class": "CommandNotFound", "desc": "The command FOO has not been
found"}

This has the additional benefit that the QGA protocol reference now documents what conditions enable use of the command.

---
 qga/commands-posix.c | 10 +---------
 qga/qapi-schema.json |  6 ++++--
 2 files changed, 5 insertions(+), 11 deletions(-)
qga/commands-posix.c +1 -9
index d92fa0ec87..a353f64ae6 100644
@@ -1212,15 +1212,7 @@ GuestUserList *qmp_guest_get_users(Error **errp)
     return head;
 }
 
-#else
-
-GuestUserList *qmp_guest_get_users(Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-    return NULL;
-}
-
-#endif
+#endif /* HAVE_UTMPX */
 
 /* Replace escaped special characters with their real values. The

replacement
* is done in place -- returned value is in the original string.

qga/qapi-schema.json +4 -3
index 0b7f911ca5..70d4f173ad 100644
@@ -1413,7 +1413,8 @@
 # Since: 2.10
 ##
 { 'struct': 'GuestUser',
-  'data': { 'user': 'str', 'login-time': 'number', '*domain': 'str' } }
+  'data': { 'user': 'str', 'login-time': 'number', '*domain': 'str' },
+  'if': { 'any': ['CONFIG_WIN32', 'HAVE_UTMPX' ] } }
 
 ##
 # @guest-get-users:
@@ -1425,7 +1426,8 @@
 # Since: 2.10
 ##
 { 'command': 'guest-get-users',
-  'returns': ['GuestUser'] }
+  'returns': ['GuestUser'],
+  'if': { 'any': ['CONFIG_WIN32', 'HAVE_UTMPX' ] } }
 
 ##
 # @GuestTimezone:
--

2.45.1

#42 lore raw

PATCH v2 17/22 qga: conditionalize schema for commands not supported on other UNIX

In reply to Daniel P. Berrangé

Rather than creating stubs for every command that just return QERR_UNSUPPORTED, use 'if' conditions in the QAPI schema to fully exclude generation of the commands on other UNIX.

The command will be rejected at QMP dispatch time instead, avoiding reimplementing rejection by blocking the stub commands. This changes the error message for affected commands from

    {"class": "CommandNotFound", "desc": "Command FOO has been disabled"}

to

    {"class": "CommandNotFound", "desc": "The command FOO has not been found"}

This has the additional benefit that the QGA protocol reference now documents what conditions enable use of the command.

---
 meson.build          | 1 +
 qga/commands-posix.c | 8 --------
 qga/qapi-schema.json | 3 ++-
 3 files changed, 3 insertions(+), 9 deletions(-)
meson.build +1 -0
index d9f3349b0a..92b8c02582 100644
@@ -2234,6 +2234,7 @@ config_host_data.set('CONFIG_ATTR', libattr.found())
 config_host_data.set('CONFIG_BDRV_WHITELIST_TOOLS', get_option('block_drv_whitelist_in_tools'))
 config_host_data.set('CONFIG_BRLAPI', brlapi.found())
 config_host_data.set('CONFIG_BSD', host_os in bsd_oses)
+config_host_data.set('CONFIG_FREEBSD', host_os == 'freebsd')
 config_host_data.set('CONFIG_CAPSTONE', capstone.found())
 config_host_data.set('CONFIG_COCOA', cocoa.found())
 config_host_data.set('CONFIG_DARWIN', host_os == 'darwin')
qga/commands-posix.c +0 -8
index a353f64ae6..f4104f2760 100644
@@ -877,14 +877,6 @@ void qmp_guest_set_user_password(const char *username,
         return;
     }
 }
-#else /* __linux__ || __FreeBSD__ */
-void qmp_guest_set_user_password(const char *username,
-                                 const char *password,
-                                 bool crypted,
-                                 Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-}
 #endif /* __linux__ || __FreeBSD__ */
 
 #ifdef HAVE_GETIFADDRS
qga/qapi-schema.json +2 -1
index 70d4f173ad..571be3a914 100644
@@ -1108,7 +1108,8 @@
 # Since: 2.3
 ##
 { 'command': 'guest-set-user-password',
-  'data': { 'username': 'str', 'password': 'str', 'crypted': 'bool' } }
+  'data': { 'username': 'str', 'password': 'str', 'crypted': 'bool' },
+  'if': { 'any': [ 'CONFIG_WIN32', 'CONFIG_LINUX', 'CONFIG_FREEBSD'] } }
 
 ##
 # @GuestMemoryBlock:
Signature · 1 line
2.45.1
Reviewed-by on v2 17/22
#43 lore raw

Re: [PATCH v2 17/22] qga: conditionalize schema for commands not supported on other UNIX

Quoted 17 lines · Daniel P. Berrangé

Rather than creating stubs for every command that just return QERR_UNSUPPORTED, use 'if' conditions in the QAPI schema to fully exclude generation of the commands on other UNIX.

The command will be rejected at QMP dispatch time instead, avoiding reimplementing rejection by blocking the stub commands. This changes the error message for affected commands from

     {"class": "CommandNotFound", "desc": "Command FOO has been disabled"}

to

     {"class": "CommandNotFound", "desc": "The command FOO has not been found"}

This has the additional benefit that the QGA protocol reference now documents what conditions enable use of the command.

Quoted 6 lines
---
  meson.build          | 1 +
  qga/commands-posix.c | 8 --------
  qga/qapi-schema.json | 3 ++-
  3 files changed, 3 insertions(+), 9 deletions(-)
Reviewed-by on v2 17/22
#44 lore raw

Re: [PATCH v2 17/22] qga: conditionalize schema for commands not supported on other UNIX

In reply to Daniel P. Berrangé
Quoted 75 lines · Daniel P. Berrangé

Rather than creating stubs for every command that just return QERR_UNSUPPORTED, use 'if' conditions in the QAPI schema to fully exclude generation of the commands on other UNIX.

The command will be rejected at QMP dispatch time instead, avoiding reimplementing rejection by blocking the stub commands. This changes the error message for affected commands from

    {"class": "CommandNotFound", "desc": "Command FOO has been disabled"}

to

{"class": "CommandNotFound", "desc": "The command FOO has not been
found"}

This has the additional benefit that the QGA protocol reference now documents what conditions enable use of the command.

---
 meson.build          | 1 +
 qga/commands-posix.c | 8 --------
 qga/qapi-schema.json | 3 ++-
 3 files changed, 3 insertions(+), 9 deletions(-)
meson.build +1 -0
index d9f3349b0a..92b8c02582 100644
@@ -2234,6 +2234,7 @@ config_host_data.set('CONFIG_ATTR', libattr.found())
 config_host_data.set('CONFIG_BDRV_WHITELIST_TOOLS',
 get_option('block_drv_whitelist_in_tools'))
 config_host_data.set('CONFIG_BRLAPI', brlapi.found())
 config_host_data.set('CONFIG_BSD', host_os in bsd_oses)
+config_host_data.set('CONFIG_FREEBSD', host_os == 'freebsd')
 config_host_data.set('CONFIG_CAPSTONE', capstone.found())
 config_host_data.set('CONFIG_COCOA', cocoa.found())

config_host_data.set('CONFIG_DARWIN', host_os == 'darwin')

qga/commands-posix.c +0 -8
index a353f64ae6..f4104f2760 100644
@@ -877,14 +877,6 @@ void qmp_guest_set_user_password(const char *username,
         return;
     }
 }
-#else /* __linux__ || __FreeBSD__ */
-void qmp_guest_set_user_password(const char *username,
-                                 const char *password,
-                                 bool crypted,
-                                 Error **errp)
-{
-    error_setg(errp, QERR_UNSUPPORTED);
-}
 #endif /* __linux__ || __FreeBSD__ */
 
 #ifdef HAVE_GETIFADDRS
qga/qapi-schema.json +2 -2
index 70d4f173ad..571be3a914 100644
@@ -1108,7 +1108,8 @@
 # Since: 2.3
 ##
 { 'command': 'guest-set-user-password',
-  'data': { 'username': 'str', 'password': 'str', 'crypted': 'bool' } }
+  'data': { 'username': 'str', 'password': 'str', 'crypted': 'bool' },
+  'if': { 'any': [ 'CONFIG_WIN32', 'CONFIG_LINUX', 'CONFIG_FREEBSD'] } }
 
 ##
 # @GuestMemoryBlock:
--

2.45.1

#45 lore raw

PATCH v2 18/22 qga: don't disable fsfreeze commands if vss_init fails

In reply to Daniel P. Berrangé

The fsfreeze commands are already written to report an error if vss_init() fails. Reporting a more specific error message is more helpful than a generic "command is disabled" message, which cannot beteween an admin config decision and lack of platform support.

---
 qga/commands-win32.c | 18 +++---------------
 qga/main.c           |  4 ++++
 2 files changed, 7 insertions(+), 15 deletions(-)
qga/commands-win32.c +3 -15
index 2533e4c748..5866cc2e3c 100644
@@ -1203,7 +1203,7 @@ GuestFilesystemInfoList *qmp_guest_get_fsinfo(Error **errp)
 GuestFsfreezeStatus qmp_guest_fsfreeze_status(Error **errp)
 {
     if (!vss_initialized()) {
-        error_setg(errp, QERR_UNSUPPORTED);
+        error_setg(errp, "fsfreeze not possible as VSS failed to initialize");
         return 0;
     }
 
@@ -1231,7 +1231,7 @@ int64_t qmp_guest_fsfreeze_freeze_list(bool has_mountpoints,
     Error *local_err = NULL;
 
     if (!vss_initialized()) {
-        error_setg(errp, QERR_UNSUPPORTED);
+        error_setg(errp, "fsfreeze not possible as VSS failed to initialize");
         return 0;
     }
 
@@ -1266,7 +1266,7 @@ int64_t qmp_guest_fsfreeze_thaw(Error **errp)
     int i;
 
     if (!vss_initialized()) {
-        error_setg(errp, QERR_UNSUPPORTED);
+        error_setg(errp, "fsfreeze not possible as VSS failed to initialize");
         return 0;
     }
 
@@ -1961,18 +1961,6 @@ done:
 /* add unsupported commands to the list of blocked RPCs */
 GList *ga_command_init_blockedrpcs(GList *blockedrpcs)
 {
-    if (!vss_init(true)) {
-        g_debug("vss_init failed, vss commands are going to be disabled");
-        const char *list[] = {
-            "guest-get-fsinfo", "guest-fsfreeze-status",
-            "guest-fsfreeze-freeze", "guest-fsfreeze-thaw", NULL};
-        char **p = (char **)list;
-
-        while (*p) {
-            blockedrpcs = g_list_append(blockedrpcs, g_strdup(*p++));
-        }
-    }
-
     return blockedrpcs;
 }
 
qga/main.c +4 -0
index f4d5f15bb3..17b6ce18ac 100644
@@ -1395,6 +1395,10 @@ static GAState *initialize_agent(GAConfig *config, int socket_activation)
                    " '%s': %s", config->state_dir, strerror(errno));
         return NULL;
     }
+
+    if (!vss_init(true)) {
+        g_debug("vss_init failed, vss commands will not function");
+    }
 #endif
 
     if (ga_is_frozen(s)) {
Signature · 1 line
2.45.1
Reviewed-by on v2 18/22
#46 lore raw

Re: [PATCH v2 18/22] qga: don't disable fsfreeze commands if vss_init fails

Quoted 4 lines · Daniel P. Berrangé

The fsfreeze commands are already written to report an error if vss_init() fails. Reporting a more specific error message is more helpful than a generic "command is disabled" message, which cannot beteween an admin config decision and lack of platform support.

s/beteween/between

Quoted 12 lines

---
qga/commands-win32.c | 18 +++---------------
qga/main.c | 4 ++++

2 files changed, 7 insertions(+), 15 deletions(-)
qga/commands-win32.c +0 -0
index 2533e4c748..5866cc2e3c 100644

qga/commands-win32.c GuestFilesystemInfoList *qmp_guest_get_fsinfo(Error **errp) · line 1207

Quoted 6 lines
qga/commands-win32.c +1 -1
@@ -1203,7 +1203,7 @@ GuestFilesystemInfoList *qmp_guest_get_fsinfo(Error **errp)
 GuestFsfreezeStatus qmp_guest_fsfreeze_status(Error **errp)
 {
    if (!vss_initialized()) {
-        error_setg(errp, QERR_UNSUPPORTED);
+        error_setg(errp, "fsfreeze not possible as VSS failed to initialize");
        return 0;

Should this be return -1 by the way?

Quoted 58 lines
qga/commands-win32.c +2 -14
    }
 
@@ -1231,7 +1231,7 @@ int64_t qmp_guest_fsfreeze_freeze_list(bool has_mountpoints,
    Error *local_err = NULL;
 
    if (!vss_initialized()) {
-        error_setg(errp, QERR_UNSUPPORTED);
+        error_setg(errp, "fsfreeze not possible as VSS failed to initialize");
        return 0;
    }
 
@@ -1266,7 +1266,7 @@ int64_t qmp_guest_fsfreeze_thaw(Error **errp)
    int i;
 
    if (!vss_initialized()) {
-        error_setg(errp, QERR_UNSUPPORTED);
+        error_setg(errp, "fsfreeze not possible as VSS failed to initialize");
        return 0;
    }
 
@@ -1961,18 +1961,6 @@ done:
 /* add unsupported commands to the list of blocked RPCs */
 GList *ga_command_init_blockedrpcs(GList *blockedrpcs)
 {
-    if (!vss_init(true)) {
-        g_debug("vss_init failed, vss commands are going to be disabled");
-        const char *list[] = {
-            "guest-get-fsinfo", "guest-fsfreeze-status",
-            "guest-fsfreeze-freeze", "guest-fsfreeze-thaw", NULL};
-        char **p = (char **)list;
-
-        while (*p) {
-            blockedrpcs = g_list_append(blockedrpcs, g_strdup(*p++));
-        }
-    }
-
    return blockedrpcs;
 }
 
qga/main.c +4 -0
index f4d5f15bb3..17b6ce18ac 100644
@@ -1395,6 +1395,10 @@ static GAState *initialize_agent(GAConfig *config, int socket_activation)
                   " '%s': %s", config->state_dir, strerror(errno));
        return NULL;
    }
+
+    if (!vss_init(true)) {
+        g_debug("vss_init failed, vss commands will not function");
+    }
 #endif
 
    if (ga_is_frozen(s)) {
Signature · 1 line
2.45.1

Otherwise LGTM:

#47 lore raw

Re: [PATCH v2 18/22] qga: don't disable fsfreeze commands if vss_init fails

Quoted 15 lines · Manos Pitsidianakis
Quoted 4 lines · Daniel P. Berrangé

The fsfreeze commands are already written to report an error if vss_init() fails. Reporting a more specific error message is more helpful than a generic "command is disabled" message, which cannot beteween an admin config decision and lack of platform support.

s/beteween/between

Quoted 7 lines

---
qga/commands-win32.c | 18 +++---------------
qga/main.c | 4 ++++

2 files changed, 7 insertions(+), 15 deletions(-)
Quoted 13 lines

qga/commands-win32.c GuestFilesystemInfoList *qmp_guest_get_fsinfo(Error **errp) · line 1207

Quoted 11 lines
qga/commands-win32.c +1 -1
index 2533e4c748..5866cc2e3c 100644
@@ -1203,7 +1203,7 @@ GuestFilesystemInfoList *qmp_guest_get_fsinfo(Error **errp)
 GuestFsfreezeStatus qmp_guest_fsfreeze_status(Error **errp)
 {
    if (!vss_initialized()) {
-        error_setg(errp, QERR_UNSUPPORTED);
+        error_setg(errp, "fsfreeze not possible as VSS failed to initialize");
        return 0;

Should this be return -1 by the way?

This method has to return GuestFsfreezeStatus and
-1 isn't valid. Not a problem though, as the
QAPI code will check for *errp != NULL and not
consider the return value for error detection.

With regards,
Daniel

Signature · 3 lines
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|
#48 lore raw

PATCH v2 19/22 qga: move declare of QGAConfig struct to top of file

In reply to Daniel P. Berrangé

It is referenced by QGAState already, and it is clearer to declare all data types at the top of the file, rather than have them mixed with code later.

---
 qga/main.c | 44 ++++++++++++++++++++++----------------------
 1 file changed, 22 insertions(+), 22 deletions(-)
qga/main.c +22 -22
index 17b6ce18ac..647d27037c 100644
@@ -70,6 +70,28 @@ typedef struct GAPersistentState {
 
 typedef struct GAConfig GAConfig;
 
+struct GAConfig {
+    char *channel_path;
+    char *method;
+    char *log_filepath;
+    char *pid_filepath;
+#ifdef CONFIG_FSFREEZE
+    char *fsfreeze_hook;
+#endif
+    char *state_dir;
+#ifdef _WIN32
+    const char *service;
+#endif
+    gchar *bliststr; /* blockedrpcs may point to this string */
+    gchar *aliststr; /* allowedrpcs may point to this string */
+    GList *blockedrpcs;
+    GList *allowedrpcs;
+    int daemonize;
+    GLogLevelFlags log_level;
+    int dumpconf;
+    bool retry_path;
+};
+
 struct GAState {
     JSONMessageParser parser;
     GMainLoop *main_loop;
@@ -996,28 +1018,6 @@ static GList *split_list(const gchar *str, const gchar *delim)
     return list;
 }
 
-struct GAConfig {
-    char *channel_path;
-    char *method;
-    char *log_filepath;
-    char *pid_filepath;
-#ifdef CONFIG_FSFREEZE
-    char *fsfreeze_hook;
-#endif
-    char *state_dir;
-#ifdef _WIN32
-    const char *service;
-#endif
-    gchar *bliststr; /* blockedrpcs may point to this string */
-    gchar *aliststr; /* allowedrpcs may point to this string */
-    GList *blockedrpcs;
-    GList *allowedrpcs;
-    int daemonize;
-    GLogLevelFlags log_level;
-    int dumpconf;
-    bool retry_path;
-};
-
 static void config_load(GAConfig *config)
 {
     GError *gerr = NULL;
Signature · 1 line
2.45.1
Reviewed-by on v2 19/22
#49 lore raw

Re: [PATCH v2 19/22] qga: move declare of QGAConfig struct to top of file

Quoted 14 lines · Daniel P. Berrangé

It is referenced by QGAState already, and it is clearer to declare all data types at the top of the file, rather than have them mixed with code later.

---
  qga/main.c | 44 ++++++++++++++++++++++----------------------
  1 file changed, 22 insertions(+), 22 deletions(-)
qga/main.c +0 -0
index 17b6ce18ac..647d27037c 100644

qga/main.c typedef struct GAPersistentState { · line 71

Quoted 2 lines
qga/main.c +0 -0
@@ -70,6 +70,28 @@ typedef struct GAPersistentState {
 
  typedef struct GAConfig GAConfig;

Matter of style, personally I'd squash within the
typedef.

Quoted 11 lines
qga/main.c +11 -0
+struct GAConfig {
+    char *channel_path;
+    char *method;
+    char *log_filepath;
+    char *pid_filepath;
+#ifdef CONFIG_FSFREEZE
+    char *fsfreeze_hook;
+#endif
+    char *state_dir;
+#ifdef _WIN32
+    const char *service;

qga/main.c typedef struct GAPersistentState { · line 93

Quoted 10 lines
qga/main.c +10 -0
+#endif
+    gchar *bliststr; /* blockedrpcs may point to this string */
+    gchar *aliststr; /* allowedrpcs may point to this string */
+    GList *blockedrpcs;
+    GList *allowedrpcs;
+    int daemonize;
+    GLogLevelFlags log_level;
+    int dumpconf;
+    bool retry_path;
+};
Reviewed-by on v2 19/22
#50 lore raw

Re: [PATCH v2 19/22] qga: move declare of QGAConfig struct to top of file

In reply to Daniel P. Berrangé
Quoted 76 lines · Daniel P. Berrangé

It is referenced by QGAState already, and it is clearer to declare all data types at the top of the file, rather than have them mixed with code later.

---
 qga/main.c | 44 ++++++++++++++++++++++----------------------
 1 file changed, 22 insertions(+), 22 deletions(-)
qga/main.c +22 -22
index 17b6ce18ac..647d27037c 100644
@@ -70,6 +70,28 @@ typedef struct GAPersistentState {
 
 typedef struct GAConfig GAConfig;
 
+struct GAConfig {
+    char *channel_path;
+    char *method;
+    char *log_filepath;
+    char *pid_filepath;
+#ifdef CONFIG_FSFREEZE
+    char *fsfreeze_hook;
+#endif
+    char *state_dir;
+#ifdef _WIN32
+    const char *service;
+#endif
+    gchar *bliststr; /* blockedrpcs may point to this string */
+    gchar *aliststr; /* allowedrpcs may point to this string */
+    GList *blockedrpcs;
+    GList *allowedrpcs;
+    int daemonize;
+    GLogLevelFlags log_level;
+    int dumpconf;
+    bool retry_path;
+};
+
 struct GAState {
     JSONMessageParser parser;
     GMainLoop *main_loop;
@@ -996,28 +1018,6 @@ static GList *split_list(const gchar *str, const
 gchar *delim)
     return list;
 }
 
-struct GAConfig {
-    char *channel_path;
-    char *method;
-    char *log_filepath;
-    char *pid_filepath;
-#ifdef CONFIG_FSFREEZE
-    char *fsfreeze_hook;
-#endif
-    char *state_dir;
-#ifdef _WIN32
-    const char *service;
-#endif
-    gchar *bliststr; /* blockedrpcs may point to this string */
-    gchar *aliststr; /* allowedrpcs may point to this string */
-    GList *blockedrpcs;
-    GList *allowedrpcs;
-    int daemonize;
-    GLogLevelFlags log_level;
-    int dumpconf;
-    bool retry_path;
-};
-
 static void config_load(GAConfig *config)
 {
     GError *gerr = NULL;

--
2.45.1

#51 lore raw

PATCH v2 20/22 qga: remove pointless 'blockrpcs_key' variable

In reply to Daniel P. Berrangé

This variable was used to support back compat for the old config file key name, and became redundant after the following change:

commit a7a2d636ae4549ef0551134d4bf8e084a14431c4
Author: Philippe Mathieu-Daudé <[email protected]>
Date: Thu May 30 08:36:43 2024 +0200

qga: Remove deprecated 'blacklist' argument / config key

---
 qga/main.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)
qga/main.c +3 -4
index 647d27037c..6ff022a85d 100644
@@ -1023,7 +1023,6 @@ static void config_load(GAConfig *config)
     GError *gerr = NULL;
     GKeyFile *keyfile;
     g_autofree char *conf = g_strdup(g_getenv("QGA_CONF")) ?: get_relocated_path(QGA_CONF_DEFAULT);
-    const gchar *blockrpcs_key = "block-rpcs";
 
     /* read system config */
     keyfile = g_key_file_new();
@@ -1071,9 +1070,9 @@ static void config_load(GAConfig *config)
             g_key_file_get_boolean(keyfile, "general", "retry-path", &gerr);
     }
 
-    if (g_key_file_has_key(keyfile, "general", blockrpcs_key, NULL)) {
+    if (g_key_file_has_key(keyfile, "general", "block-rpcs", NULL)) {
         config->bliststr =
-            g_key_file_get_string(keyfile, "general", blockrpcs_key, &gerr);
+            g_key_file_get_string(keyfile, "general", "block-rpcs", &gerr);
         config->blockedrpcs = g_list_concat(config->blockedrpcs,
                                           split_list(config->bliststr, ","));
     }
@@ -1084,7 +1083,7 @@ static void config_load(GAConfig *config)
                                           split_list(config->aliststr, ","));
     }
 
-    if (g_key_file_has_key(keyfile, "general", blockrpcs_key, NULL) &&
+    if (g_key_file_has_key(keyfile, "general", "block-rpcs", NULL) &&
         g_key_file_has_key(keyfile, "general", "allow-rpcs", NULL)) {
         g_critical("wrong config, using 'block-rpcs' and 'allow-rpcs' keys at"
                    " the same time is not allowed");
Signature · 1 line
2.45.1
Reviewed-by on v2 20/22
#52 lore raw

Re: [PATCH v2 20/22] qga: remove pointless 'blockrpcs_key' variable

Quoted 7 lines · Daniel P. Berrangé

This variable was used to support back compat for the old config file key name, and became redundant after the following change:

commit a7a2d636ae4549ef0551134d4bf8e084a14431c4
Author: Philippe Mathieu-Daudé <[email protected]>
Date: Thu May 30 08:36:43 2024 +0200

Quoted 6 lines

qga: Remove deprecated 'blacklist' argument / config key

---
  qga/main.c | 7 +++----
  1 file changed, 3 insertions(+), 4 deletions(-)
Reviewed-by on v2 20/22
#53 lore raw

Re: [PATCH v2 20/22] qga: remove pointless 'blockrpcs_key' variable

In reply to Daniel P. Berrangé
Quoted 58 lines · Daniel P. Berrangé

This variable was used to support back compat for the old config file key name, and became redundant after the following change:

commit a7a2d636ae4549ef0551134d4bf8e084a14431c4
Author: Philippe Mathieu-Daudé <[email protected]>
Date: Thu May 30 08:36:43 2024 +0200

qga: Remove deprecated 'blacklist' argument / config key

---
 qga/main.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)
qga/main.c +0 -1
index 647d27037c..6ff022a85d 100644
@@ -1023,7 +1023,6 @@ static void config_load(GAConfig *config)
     GError *gerr = NULL;
     GKeyFile *keyfile;
     g_autofree char *conf = g_strdup(g_getenv("QGA_CONF")) ?:
 get_relocated_path(QGA_CONF_DEFAULT);
-    const gchar *blockrpcs_key = "block-rpcs";
 
     /* read system config */
     keyfile = g_key_file_new();
qga/main.c +2 -2
@@ -1071,9 +1070,9 @@ static void config_load(GAConfig *config)
             g_key_file_get_boolean(keyfile, "general", "retry-path",
 &gerr);
     }
 
-    if (g_key_file_has_key(keyfile, "general", blockrpcs_key, NULL)) {
+    if (g_key_file_has_key(keyfile, "general", "block-rpcs", NULL)) {
         config->bliststr =
-            g_key_file_get_string(keyfile, "general", blockrpcs_key,
 &gerr);
+            g_key_file_get_string(keyfile, "general", "block-rpcs",
 &gerr);
         config->blockedrpcs = g_list_concat(config->blockedrpcs,
                                           split_list(config->bliststr,
","));
     }
qga/main.c +1 -1
@@ -1084,7 +1083,7 @@ static void config_load(GAConfig *config)
                                           split_list(config->aliststr,
 ","));
     }
 
-    if (g_key_file_has_key(keyfile, "general", blockrpcs_key, NULL) &&
+    if (g_key_file_has_key(keyfile, "general", "block-rpcs", NULL) &&
         g_key_file_has_key(keyfile, "general", "allow-rpcs", NULL)) {
         g_critical("wrong config, using 'block-rpcs' and 'allow-rpcs'

keys at"

                    " the same time is not allowed");

--
2.45.1

#54 lore raw

PATCH v2 21/22 qga: allow configuration file path via the cli

In reply to Daniel P. Berrangé

Allowing the user to set the QGA_CONF environment variable to change the default configuration file path is very unusual practice, made more obscure since this ability is not documented.

This introduces the more normal '-c PATH' / '--config=PATH' command line argument approach. This requires that we parse the comamnd line twice, since we want the command line arguments to take priority over the configuration file settings in general.

---
 docs/interop/qemu-ga.rst |  5 +++++
 qga/main.c               | 35 +++++++++++++++++++++++++++--------
 2 files changed, 32 insertions(+), 8 deletions(-)
docs/interop/qemu-ga.rst +5 -0
index 72fb75a6f5..e42b370319 100644
@@ -33,6 +33,11 @@ Options
 
 .. program:: qemu-ga
 
+.. option:: -c, --config=PATH
+
+  Configuration file path (the default is |CONFDIR|\ ``/qemu-ga.conf``,
+  unless overriden by the QGA_CONF environment variable)
+
 .. option:: -m, --method=METHOD
 
   Transport method: one of ``unix-listen``, ``virtio-serial``, or
qga/main.c +27 -8
index 6ff022a85d..f68a32bf7b 100644
@@ -1018,15 +1018,14 @@ static GList *split_list(const gchar *str, const gchar *delim)
     return list;
 }
 
-static void config_load(GAConfig *config)
+static void config_load(GAConfig *config, const char *confpath, bool required)
 {
     GError *gerr = NULL;
     GKeyFile *keyfile;
-    g_autofree char *conf = g_strdup(g_getenv("QGA_CONF")) ?: get_relocated_path(QGA_CONF_DEFAULT);
 
     /* read system config */
     keyfile = g_key_file_new();
-    if (!g_key_file_load_from_file(keyfile, conf, 0, &gerr)) {
+    if (!g_key_file_load_from_file(keyfile, confpath, 0, &gerr)) {
         goto end;
     }
     if (g_key_file_has_key(keyfile, "general", "daemon", NULL)) {
@@ -1092,10 +1091,10 @@ static void config_load(GAConfig *config)
 
 end:
     g_key_file_free(keyfile);
-    if (gerr &&
-        !(gerr->domain == G_FILE_ERROR && gerr->code == G_FILE_ERROR_NOENT)) {
+    if (gerr && (required ||
+                 !(gerr->domain == G_FILE_ERROR && gerr->code == G_FILE_ERROR_NOENT))) {
         g_critical("error loading configuration from path: %s, %s",
-                   conf, gerr->message);
+                   confpath, gerr->message);
         exit(EXIT_FAILURE);
     }
     g_clear_error(&gerr);
@@ -1167,12 +1166,13 @@ static void config_dump(GAConfig *config)
 
 static void config_parse(GAConfig *config, int argc, char **argv)
 {
-    const char *sopt = "hVvdm:p:l:f:F::b:a:s:t:Dr";
+    const char *sopt = "hVvdc:m:p:l:f:F::b:a:s:t:Dr";
     int opt_ind = 0, ch;
     bool block_rpcs = false, allow_rpcs = false;
     const struct option lopt[] = {
         { "help", 0, NULL, 'h' },
         { "version", 0, NULL, 'V' },
+        { "config", 1, NULL, 'c' },
         { "dump-conf", 0, NULL, 'D' },
         { "logfile", 1, NULL, 'l' },
         { "pidfile", 1, NULL, 'f' },
@@ -1192,6 +1192,26 @@ static void config_parse(GAConfig *config, int argc, char **argv)
         { "retry-path", 0, NULL, 'r' },
         { NULL, 0, NULL, 0 }
     };
+    g_autofree char *confpath = g_strdup(g_getenv("QGA_CONF")) ?:
+        get_relocated_path(QGA_CONF_DEFAULT);
+    bool confrequired = false;
+
+    while ((ch = getopt_long(argc, argv, sopt, lopt, NULL)) != -1) {
+        switch (ch) {
+        case 'c':
+            g_free(confpath);
+            confpath = g_strdup(optarg);
+            confrequired = true;
+            break;
+        default:
+            break;
+        }
+    }
+
+    config_load(config, confpath, confrequired);
+
+    /* Reset for second pass */
+    optind = 1;
 
     while ((ch = getopt_long(argc, argv, sopt, lopt, &opt_ind)) != -1) {
         switch (ch) {
@@ -1582,7 +1602,6 @@ int main(int argc, char **argv)
     qga_qmp_init_marshal(&ga_commands);
 
     init_dfl_pathnames();
-    config_load(config);
     config_parse(config, argc, argv);
 
     if (config->pid_filepath == NULL) {
Signature · 1 line
2.45.1
#55 lore raw

Re: [PATCH v2 21/22] qga: allow configuration file path via the cli

Quoted 110 lines · Daniel P. Berrangé

Allowing the user to set the QGA_CONF environment variable to change the default configuration file path is very unusual practice, made more obscure since this ability is not documented.

This introduces the more normal '-c PATH' / '--config=PATH' command line argument approach. This requires that we parse the comamnd line twice, since we want the command line arguments to take priority over the configuration file settings in general.

---
  docs/interop/qemu-ga.rst |  5 +++++
  qga/main.c               | 35 +++++++++++++++++++++++++++--------
  2 files changed, 32 insertions(+), 8 deletions(-)
docs/interop/qemu-ga.rst +5 -0
index 72fb75a6f5..e42b370319 100644
@@ -33,6 +33,11 @@ Options
 
  .. program:: qemu-ga
 
+.. option:: -c, --config=PATH
+
+  Configuration file path (the default is |CONFDIR|\ ``/qemu-ga.conf``,
+  unless overriden by the QGA_CONF environment variable)
+
  .. option:: -m, --method=METHOD
 
    Transport method: one of ``unix-listen``, ``virtio-serial``, or
qga/main.c +27 -7
index 6ff022a85d..f68a32bf7b 100644
@@ -1018,15 +1018,14 @@ static GList *split_list(const gchar *str, const gchar *delim)
      return list;
  }
 
-static void config_load(GAConfig *config)
+static void config_load(GAConfig *config, const char *confpath, bool required)
  {
      GError *gerr = NULL;
      GKeyFile *keyfile;
-    g_autofree char *conf = g_strdup(g_getenv("QGA_CONF")) ?: get_relocated_path(QGA_CONF_DEFAULT);
 
      /* read system config */
      keyfile = g_key_file_new();
-    if (!g_key_file_load_from_file(keyfile, conf, 0, &gerr)) {
+    if (!g_key_file_load_from_file(keyfile, confpath, 0, &gerr)) {
          goto end;
      }
      if (g_key_file_has_key(keyfile, "general", "daemon", NULL)) {
@@ -1092,10 +1091,10 @@ static void config_load(GAConfig *config)
 
  end:
      g_key_file_free(keyfile);
-    if (gerr &&
-        !(gerr->domain == G_FILE_ERROR && gerr->code == G_FILE_ERROR_NOENT)) {
+    if (gerr && (required ||
+                 !(gerr->domain == G_FILE_ERROR && gerr->code == G_FILE_ERROR_NOENT))) {
          g_critical("error loading configuration from path: %s, %s",
-                   conf, gerr->message);
+                   confpath, gerr->message);
          exit(EXIT_FAILURE);
      }
      g_clear_error(&gerr);
@@ -1167,12 +1166,13 @@ static void config_dump(GAConfig *config)
 
  static void config_parse(GAConfig *config, int argc, char **argv)
  {
-    const char *sopt = "hVvdm:p:l:f:F::b:a:s:t:Dr";
+    const char *sopt = "hVvdc:m:p:l:f:F::b:a:s:t:Dr";
      int opt_ind = 0, ch;
      bool block_rpcs = false, allow_rpcs = false;
      const struct option lopt[] = {
          { "help", 0, NULL, 'h' },
          { "version", 0, NULL, 'V' },
+        { "config", 1, NULL, 'c' },
          { "dump-conf", 0, NULL, 'D' },
          { "logfile", 1, NULL, 'l' },
          { "pidfile", 1, NULL, 'f' },
@@ -1192,6 +1192,26 @@ static void config_parse(GAConfig *config, int argc, char **argv)
          { "retry-path", 0, NULL, 'r' },
          { NULL, 0, NULL, 0 }
      };
+    g_autofree char *confpath = g_strdup(g_getenv("QGA_CONF")) ?:
+        get_relocated_path(QGA_CONF_DEFAULT);
+    bool confrequired = false;
+
+    while ((ch = getopt_long(argc, argv, sopt, lopt, NULL)) != -1) {
+        switch (ch) {
+        case 'c':
+            g_free(confpath);
+            confpath = g_strdup(optarg);
+            confrequired = true;
+            break;
+        default:
+            break;
+        }
+    }
+
+    config_load(config, confpath, confrequired);
+
+    /* Reset for second pass */
+    optind = 1;
 
      while ((ch = getopt_long(argc, argv, sopt, lopt, &opt_ind)) != -1) {
          switch (ch) {

qga/main.c int main(int argc, char **argv) · line 1607

Quoted 7 lines
qga/main.c +0 -1
@@ -1582,7 +1602,6 @@ int main(int argc, char **argv)
      qga_qmp_init_marshal(&ga_commands);
 
      init_dfl_pathnames();
-    config_load(config);
      config_parse(config, argc, argv);
 
      if (config->pid_filepath == NULL) {

This looks like a trivial (correct) CLI change, but I don't feel confident anymore reviewing this area, so will let others have a look.

Regards,

Phil.

#56 lore raw

Re: [PATCH v2 21/22] qga: allow configuration file path via the cli

In reply to Daniel P. Berrangé
Quoted 20 lines · Daniel P. Berrangé

Allowing the user to set the QGA_CONF environment variable to change the default configuration file path is very unusual practice, made more obscure since this ability is not documented.

This introduces the more normal '-c PATH' / '--config=PATH' command line argument approach. This requires that we parse the comamnd line twice, since we want the command line arguments to take priority over the configuration file settings in general.

---
 docs/interop/qemu-ga.rst |  5 +++++
 qga/main.c               | 35 +++++++++++++++++++++++++++--------
 2 files changed, 32 insertions(+), 8 deletions(-)
docs/interop/qemu-ga.rst +0 -0
index 72fb75a6f5..e42b370319 100644

docs/interop/qemu-ga.rst Options · line 42

Quoted 10 lines
docs/interop/qemu-ga.rst +5 -0
@@ -33,6 +33,11 @@ Options
 
 .. program:: qemu-ga
 
+.. option:: -c, --config=PATH
+
+  Configuration file path (the default is |CONFDIR|\ ``/qemu-ga.conf``,
+  unless overriden by the QGA_CONF environment variable)
+
 .. option:: -m, --method=METHOD
 

Please also update qga/main.c static void usage(const char *cmd)

Quoted 98 lines
docs/interop/qemu-ga.rst +0 -0
   Transport method: one of ``unix-listen``, ``virtio-serial``, or
qga/main.c +2 -3
index 6ff022a85d..f68a32bf7b 100644
@@ -1018,15 +1018,14 @@ static GList *split_list(const gchar *str, const
 gchar *delim)
     return list;
 }
 
-static void config_load(GAConfig *config)
+static void config_load(GAConfig *config, const char *confpath, bool
 required)
 {
     GError *gerr = NULL;
     GKeyFile *keyfile;
-    g_autofree char *conf = g_strdup(g_getenv("QGA_CONF")) ?:
 get_relocated_path(QGA_CONF_DEFAULT);
 
     /* read system config */
     keyfile = g_key_file_new();
-    if (!g_key_file_load_from_file(keyfile, conf, 0, &gerr)) {
+    if (!g_key_file_load_from_file(keyfile, confpath, 0, &gerr)) {
         goto end;
     }
     if (g_key_file_has_key(keyfile, "general", "daemon", NULL)) {
qga/main.c +3 -3
@@ -1092,10 +1091,10 @@ static void config_load(GAConfig *config)
 
 end:
     g_key_file_free(keyfile);
-    if (gerr &&
-        !(gerr->domain == G_FILE_ERROR && gerr->code ==
 G_FILE_ERROR_NOENT)) {
+    if (gerr && (required ||
+                 !(gerr->domain == G_FILE_ERROR && gerr->code ==
 G_FILE_ERROR_NOENT))) {
         g_critical("error loading configuration from path: %s, %s",
-                   conf, gerr->message);
+                   confpath, gerr->message);
         exit(EXIT_FAILURE);
     }
     g_clear_error(&gerr);
qga/main.c +22 -1
@@ -1167,12 +1166,13 @@ static void config_dump(GAConfig *config)
 
 static void config_parse(GAConfig *config, int argc, char **argv)
 {
-    const char *sopt = "hVvdm:p:l:f:F::b:a:s:t:Dr";
+    const char *sopt = "hVvdc:m:p:l:f:F::b:a:s:t:Dr";
     int opt_ind = 0, ch;
     bool block_rpcs = false, allow_rpcs = false;
     const struct option lopt[] = {
         { "help", 0, NULL, 'h' },
         { "version", 0, NULL, 'V' },
+        { "config", 1, NULL, 'c' },
         { "dump-conf", 0, NULL, 'D' },
         { "logfile", 1, NULL, 'l' },
         { "pidfile", 1, NULL, 'f' },
@@ -1192,6 +1192,26 @@ static void config_parse(GAConfig *config, int
 argc, char **argv)
         { "retry-path", 0, NULL, 'r' },
         { NULL, 0, NULL, 0 }
     };
+    g_autofree char *confpath = g_strdup(g_getenv("QGA_CONF")) ?:
+        get_relocated_path(QGA_CONF_DEFAULT);
+    bool confrequired = false;
+
+    while ((ch = getopt_long(argc, argv, sopt, lopt, NULL)) != -1) {
+        switch (ch) {
+        case 'c':
+            g_free(confpath);
+            confpath = g_strdup(optarg);
+            confrequired = true;
+            break;
+        default:
+            break;
+        }
+    }
+
+    config_load(config, confpath, confrequired);
+
+    /* Reset for second pass */
+    optind = 1;
 
     while ((ch = getopt_long(argc, argv, sopt, lopt, &opt_ind)) != -1) {
         switch (ch) {
qga/main.c +0 -2
@@ -1582,7 +1602,6 @@ int main(int argc, char **argv)
     qga_qmp_init_marshal(&ga_commands);
 
     init_dfl_pathnames();
-    config_load(config);
     config_parse(config, argc, argv);
 
     if (config->pid_filepath == NULL) {
--

2.45.1

#57 lore raw

Re: [PATCH v2 21/22] qga: allow configuration file path via the cli

Quoted 26 lines · Konstantin Kostiuk
Quoted 23 lines · Daniel P. Berrangé

Allowing the user to set the QGA_CONF environment variable to change the default configuration file path is very unusual practice, made more obscure since this ability is not documented.

This introduces the more normal '-c PATH' / '--config=PATH' command line argument approach. This requires that we parse the comamnd line twice, since we want the command line arguments to take priority over the configuration file settings in general.

---
 docs/interop/qemu-ga.rst |  5 +++++
 qga/main.c               | 35 +++++++++++++++++++++++++++--------
 2 files changed, 32 insertions(+), 8 deletions(-)
docs/interop/qemu-ga.rst +0 -0
index 72fb75a6f5..e42b370319 100644
@@ -33,6 +33,11 @@ Options
 
 .. program:: qemu-ga
 
Quoted 9 lines

docs/interop/qemu-ga.rst Options · line 42

Quoted 7 lines
docs/interop/qemu-ga.rst +5 -0
+.. option:: -c, --config=PATH
+
+  Configuration file path (the default is |CONFDIR|\ ``/qemu-ga.conf``,
+  unless overriden by the QGA_CONF environment variable)
+
 .. option:: -m, --method=METHOD
 

Please also update qga/main.c static void usage(const char *cmd)

Opps, yes, will do.

With regards,
Daniel

Signature · 3 lines
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|
#58 lore raw

PATCH v2 22/22 qga: centralize logic for disabling/enabling commands

In reply to Daniel P. Berrangé

It is confusing having many different pieces of code enabling and disabling commands, and it is not clear that they all have the same semantics, especially wrt prioritization of the block/allow lists. The code attempted to prevent the user from setting both the block and allow lists concurrently, however, the logic was flawed as it checked settings in the configuration file separately from the command line arguments. Thus it was possible to set a block list in the config file and an allow list via a command line argument. The --dump-conf option also creates a configuration file with both keys present, even if unset, which means it is creating a config that cannot actually be loaded again.

Centralizing the code in a single method "ga_apply_command_filters" will provide a strong guarantee of consistency and clarify the intended behaviour. With this there is no compelling technical reason to prevent concurrent setting of both the allow and block lists, so this flawed restriction is removed.

---
 docs/interop/qemu-ga.rst |  14 +++++
 qga/commands-posix.c     |   6 --
 qga/commands-win32.c     |   6 --
 qga/main.c               | 128 +++++++++++++++++----------------------
 4 files changed, 70 insertions(+), 84 deletions(-)
docs/interop/qemu-ga.rst +14 -0
index e42b370319..e35dcaf0e7 100644
@@ -28,6 +28,20 @@ configuration options on the command line. For the same key, the last
 option wins, but the lists accumulate (see below for configuration
 file format).
 
+If an allowed RPCs list is defined in the configuration, then all
+RPCs will be blocked by default, except for the allowed list.
+
+If a blocked RPCs list is defined in the configuration, then all
+RPCs will be allowed by default, except for the blocked list.
+
+If both allowed and blocked RPCs lists are defined in the configuration,
+then all RPCs will be blocked by default, and then allowed list will
+be applied, followed by the blocked list.
+
+While filesystems are frozen, all except for a designated safe set
+of RPCs will blocked, regardless of what the general configuration
+declares.
+
 Options
 -------
 
qga/commands-posix.c +0 -6
index f4104f2760..578d29f228 100644
@@ -1136,12 +1136,6 @@ error:
 
 #endif /* HAVE_GETIFADDRS */
 
-/* add unsupported commands to the list of blocked RPCs */
-GList *ga_command_init_blockedrpcs(GList *blockedrpcs)
-{
-    return blockedrpcs;
-}
-
 /* register init/cleanup routines for stateful command groups */
 void ga_command_state_init(GAState *s, GACommandState *cs)
 {
qga/commands-win32.c +0 -6
index 5866cc2e3c..61b36da469 100644
@@ -1958,12 +1958,6 @@ done:
     g_free(rawpasswddata);
 }
 
-/* add unsupported commands to the list of blocked RPCs */
-GList *ga_command_init_blockedrpcs(GList *blockedrpcs)
-{
-    return blockedrpcs;
-}
-
 /* register init/cleanup routines for stateful command groups */
 void ga_command_state_init(GAState *s, GACommandState *cs)
 {
qga/main.c +56 -72

227 lines not on this page · raw message

Signature · 1 line
2.45.1
#59 lore raw

Re: [PATCH v2 22/22] qga: centralize logic for disabling/enabling commands

Hello Daniel,

This cleanup seems like a good idea,

Quoted 34 lines · Daniel P. Berrangé

It is confusing having many different pieces of code enabling and disabling commands, and it is not clear that they all have the same semantics, especially wrt prioritization of the block/allow lists. The code attempted to prevent the user from setting both the block and allow lists concurrently, however, the logic was flawed as it checked settings in the configuration file separately from the command line arguments. Thus it was possible to set a block list in the config file and an allow list via a command line argument. The --dump-conf option also creates a configuration file with both keys present, even if unset, which means it is creating a config that cannot actually be loaded again.

Centralizing the code in a single method "ga_apply_command_filters" will provide a strong guarantee of consistency and clarify the intended behaviour. With this there is no compelling technical reason to prevent concurrent setting of both the allow and block lists, so this flawed restriction is removed.

---

docs/interop/qemu-ga.rst |  14 +++++
qga/commands-posix.c     |   6 --
qga/commands-win32.c     |   6 --
qga/main.c               | 128 +++++++++++++++++----------------------
4 files changed, 70 insertions(+), 84 deletions(-)
docs/interop/qemu-ga.rst +0 -0
index e42b370319..e35dcaf0e7 100644
@@ -28,6 +28,20 @@ configuration options on the command line. For the same key, the last

option wins, but the lists accumulate (see below for configuration file format).

docs/interop/qemu-ga.rst configuration options on the command line. For the same key, the last · line 39

Quoted 9 lines
docs/interop/qemu-ga.rst +9 -0
+If an allowed RPCs list is defined in the configuration, then all
+RPCs will be blocked by default, except for the allowed list.
+
+If a blocked RPCs list is defined in the configuration, then all
+RPCs will be allowed by default, except for the blocked list.
+
+If both allowed and blocked RPCs lists are defined in the configuration,
+then all RPCs will be blocked by default, and then allowed list will
+be applied, followed by the blocked list.

Nit: Missing an article here

-then all RPCs will be blocked by default, and then allowed list will +then all RPCs will be blocked by default, then the allowed list will

Quoted 115 lines
docs/interop/qemu-ga.rst +5 -0
+
+While filesystems are frozen, all except for a designated safe set
+of RPCs will blocked, regardless of what the general configuration
+declares.
+

Options
-------

qga/commands-posix.c +0 -6
index f4104f2760..578d29f228 100644
@@ -1136,12 +1136,6 @@ error:
 
 #endif /* HAVE_GETIFADDRS */
 
-/* add unsupported commands to the list of blocked RPCs */
-GList *ga_command_init_blockedrpcs(GList *blockedrpcs)
-{
-    return blockedrpcs;
-}
-
 /* register init/cleanup routines for stateful command groups */
 void ga_command_state_init(GAState *s, GACommandState *cs)
 {
qga/commands-win32.c +0 -6
index 5866cc2e3c..61b36da469 100644
@@ -1958,12 +1958,6 @@ done:
    g_free(rawpasswddata);
 }
 
-/* add unsupported commands to the list of blocked RPCs */
-GList *ga_command_init_blockedrpcs(GList *blockedrpcs)
-{
-    return blockedrpcs;
-}
-
 /* register init/cleanup routines for stateful command groups */
 void ga_command_state_init(GAState *s, GACommandState *cs)
 {
qga/main.c +26 -28
index f68a32bf7b..72c16fead8 100644
@@ -419,60 +419,79 @@ static gint ga_strcmp(gconstpointer str1, gconstpointer str2)
    return strcmp(str1, str2);
 }
 
-/* disable commands that aren't safe for fsfreeze */
-static void ga_disable_not_allowed_freeze(const QmpCommand *cmd, void *opaque)
+static bool ga_command_is_allowed(const QmpCommand *cmd, GAState *state)
 {
-    bool allowed = false;
    int i = 0;
+    GAConfig *config = state->config;
    const char *name = qmp_command_name(cmd);
+    /* Fallback policy is allow everything */
+    bool allowed = true;
 
-    while (ga_freeze_allowlist[i] != NULL) {
-        if (strcmp(name, ga_freeze_allowlist[i]) == 0) {
+    if (config->allowedrpcs) {
+        /*
+         * If an allow-list is given, this changes the fallback
+         * policy to deny everything
+         */
+        allowed = false;
+
+        if (g_list_find_custom(config->allowedrpcs, name, ga_strcmp) != NULL) {
            allowed = true;
        }
-        i++;
    }
-    if (!allowed) {
-        g_debug("disabling command: %s", name);
-        qmp_disable_command(&ga_commands, name, "the agent is in frozen state");
-    }
-}
 
-/* [re-]enable all commands, except those explicitly blocked by user */
-static void ga_enable_non_blocked(const QmpCommand *cmd, void *opaque)
-{
-    GAState *s = opaque;
-    GList *blockedrpcs = s->blockedrpcs;
-    GList *allowedrpcs = s->allowedrpcs;
-    const char *name = qmp_command_name(cmd);
-
-    if (g_list_find_custom(blockedrpcs, name, ga_strcmp) == NULL) {
-        if (qmp_command_is_enabled(cmd)) {
-            return;
+    /*
+     * If both allowedrpcs and blockedrpcs are set, the blocked
+     * list will take priority
+     */
+    if (config->blockedrpcs) {
+        if (g_list_find_custom(config->blockedrpcs, name, ga_strcmp) != NULL) {
+            allowed = false;
        }
+    }
 
-        if (allowedrpcs &&
-            g_list_find_custom(allowedrpcs, name, ga_strcmp) == NULL) {
-            return;
-        }
+    /*
+     * If frozen, this filtering must take priority over
+     * absolutely everything
+     */
+    if (state->frozen) {
+        allowed = false;
 
-        g_debug("enabling command: %s", name);
-        qmp_enable_command(&ga_commands, name);

qga/main.c static gint ga_strcmp(gconstpointer str1, gconstpointer str2) · line 468

Quoted 10 lines
qga/main.c +8 -0
+        while (ga_freeze_allowlist[i] != NULL) {
+            if (strcmp(name, ga_freeze_allowlist[i]) == 0) {
+                allowed = true;
+            }
+            i++;
+        }
    }
+
+    return allowed;
 }

IUUC, we can check by priority here: first check if (state->frozen), then blockedrpcs, then allowedrpcs and then return a default fallback value allowed = config->blockedrpcs != NULL && config->allowedrpcs != NULL

This way the function will sort of document what is written on the docs as well.

Quoted 7 lines
qga/main.c +2 -3
-/* disable commands that aren't allowed */
-static void ga_disable_not_allowed(const QmpCommand *cmd, void *opaque)
+static void ga_apply_command_filters_iter(const QmpCommand *cmd, void *opaque)
 {
-    GList *allowedrpcs = opaque;
+    GAState *state = opaque;

qga/main.c static gint ga_strcmp(gconstpointer str1, gconstpointer str2) · line 481

Quoted 10 lines
qga/main.c +7 -1
+    bool want = ga_command_is_allowed(cmd, state);
+    bool have = qmp_command_is_enabled(cmd);
    const char *name = qmp_command_name(cmd);
 
-    if (g_list_find_custom(allowedrpcs, name, ga_strcmp) == NULL) {
+    if (want == have) {
+        return;
+    }
+
+    if (qmp_command_is_enabled(cmd)) {

Nit:

if (have) {

Since it's already declared.

qga/main.c static gint ga_strcmp(gconstpointer str1, gconstpointer str2) · line 490

Quoted 9 lines
qga/main.c +4 -0
        g_debug("disabling command: %s", name);
        qmp_disable_command(&ga_commands, name, "the command is not allowed");
+    } else {
+        g_debug("enabling command: %s", name);
+        qmp_enable_command(&ga_commands, name);
    }
 }
 
+static void ga_apply_command_filters(GAState *state)

Nit: inline?

Quoted 137 lines
qga/main.c +7 -11
+{
+    qmp_for_each_command(&ga_commands, ga_apply_command_filters_iter, state);
+}
+
 static bool ga_create_file(const char *path)
 {
    int fd = open(path, O_CREAT | O_WRONLY, S_IWUSR | S_IRUSR);
@@ -505,15 +524,14 @@ void ga_set_frozen(GAState *s)
    if (ga_is_frozen(s)) {
        return;
    }
-    /* disable all forbidden (for frozen state) commands */
-    qmp_for_each_command(&ga_commands, ga_disable_not_allowed_freeze, NULL);
    g_warning("disabling logging due to filesystem freeze");
-    ga_disable_logging(s);
    s->frozen = true;
    if (!ga_create_file(s->state_filepath_isfrozen)) {
        g_warning("unable to create %s, fsfreeze may not function properly",
                  s->state_filepath_isfrozen);
    }
+    ga_apply_command_filters(s);
+    ga_disable_logging(s);
 }
 
 void ga_unset_frozen(GAState *s)
@@ -545,12 +563,12 @@ void ga_unset_frozen(GAState *s)
    }
 
    /* enable all disabled, non-blocked and allowed commands */
-    qmp_for_each_command(&ga_commands, ga_enable_non_blocked, s);
    s->frozen = false;
    if (!ga_delete_file(s->state_filepath_isfrozen)) {
        g_warning("unable to delete %s, fsfreeze may not function properly",
                  s->state_filepath_isfrozen);
    }
+    ga_apply_command_filters(s);
 }
 
 #ifdef CONFIG_FSFREEZE
@@ -1082,13 +1100,6 @@ static void config_load(GAConfig *config, const char *confpath, bool required)
                                          split_list(config->aliststr, ","));
    }
 
-    if (g_key_file_has_key(keyfile, "general", "block-rpcs", NULL) &&
-        g_key_file_has_key(keyfile, "general", "allow-rpcs", NULL)) {
-        g_critical("wrong config, using 'block-rpcs' and 'allow-rpcs' keys at"
-                   " the same time is not allowed");
-        exit(EXIT_FAILURE);
-    }
-

end:

    g_key_file_free(keyfile);
    if (gerr && (required ||
qga/main.c +2 -29
@@ -1168,7 +1179,6 @@ static void config_parse(GAConfig *config, int argc, char **argv)
 {
    const char *sopt = "hVvdc:m:p:l:f:F::b:a:s:t:Dr";
    int opt_ind = 0, ch;
-    bool block_rpcs = false, allow_rpcs = false;
    const struct option lopt[] = {
        { "help", 0, NULL, 'h' },
        { "version", 0, NULL, 'V' },
@@ -1264,7 +1274,6 @@ static void config_parse(GAConfig *config, int argc, char **argv)
            }
            config->blockedrpcs = g_list_concat(config->blockedrpcs,
                                                split_list(optarg, ","));
-            block_rpcs = true;
            break;
        }
        case 'a': {
@@ -1274,7 +1283,6 @@ static void config_parse(GAConfig *config, int argc, char **argv)
            }
            config->allowedrpcs = g_list_concat(config->allowedrpcs,
                                                split_list(optarg, ","));
-            allow_rpcs = true;
            break;
        }
 #ifdef _WIN32
@@ -1315,12 +1323,6 @@ static void config_parse(GAConfig *config, int argc, char **argv)
            exit(EXIT_FAILURE);
        }
    }
-
-    if (block_rpcs && allow_rpcs) {
-        g_critical("wrong commandline, using --block-rpcs and --allow-rpcs at the"
-                   " same time is not allowed");
-        exit(EXIT_FAILURE);
-    }
 }
 
 static void config_free(GAConfig *config)
@@ -1431,7 +1433,6 @@ static GAState *initialize_agent(GAConfig *config, int socket_activation)
            s->deferred_options.log_filepath = config->log_filepath;
        }
        ga_disable_logging(s);
-        qmp_for_each_command(&ga_commands, ga_disable_not_allowed_freeze, NULL);
    } else {
        if (config->daemonize) {
            become_daemon(config->pid_filepath);
@@ -1455,25 +1456,6 @@ static GAState *initialize_agent(GAConfig *config, int socket_activation)
        return NULL;
    }
 
-    if (config->allowedrpcs) {
-        qmp_for_each_command(&ga_commands, ga_disable_not_allowed, config->allowedrpcs);
-        s->allowedrpcs = config->allowedrpcs;
-    }
-
-    /*
-     * Some commands can be blocked due to system limitation.
-     * Initialize blockedrpcs list even if allowedrpcs specified.
-     */
-    config->blockedrpcs = ga_command_init_blockedrpcs(config->blockedrpcs);
-    if (config->blockedrpcs) {
-        GList *l = config->blockedrpcs;
-        s->blockedrpcs = config->blockedrpcs;
-        do {
-            g_debug("disabling command: %s", (char *)l->data);
-            qmp_disable_command(&ga_commands, l->data, NULL);
-            l = g_list_next(l);
-        } while (l);
-    }
    s->command_state = ga_command_state_new();
    ga_command_state_init(s, s->command_state);
    ga_command_state_init_all(s->command_state);
@@ -1499,6 +1481,8 @@ static GAState *initialize_agent(GAConfig *config, int socket_activation)
    }
 #endif
 
+    ga_apply_command_filters(s);
+
    ga_state = s;
    return s;
 }
Signature · 1 line
2.45.1
#60 lore raw

Re: [PATCH v2 22/22] qga: centralize logic for disabling/enabling commands

Quoted 32 lines · Manos Pitsidianakis

Hello Daniel,

This cleanup seems like a good idea,

Quoted 25 lines · Daniel P. Berrangé

It is confusing having many different pieces of code enabling and disabling commands, and it is not clear that they all have the same semantics, especially wrt prioritization of the block/allow lists. The code attempted to prevent the user from setting both the block and allow lists concurrently, however, the logic was flawed as it checked settings in the configuration file  separately from the command line arguments. Thus it was possible to set a block list in the config file and an allow list via a command line argument. The --dump-conf option also creates a configuration file with both keys present, even if unset, which means it is creating a config that cannot actually be loaded again.

Centralizing the code in a single method "ga_apply_command_filters" will provide a strong guarantee of consistency and clarify the intended behaviour. With this there is no compelling technical reason to prevent concurrent setting of both the allow and block lists, so this flawed restriction is removed.

---
docs/interop/qemu-ga.rst |  14 +++++
qga/commands-posix.c     |   6 --
qga/commands-win32.c     |   6 --
qga/main.c               | 128 +++++++++++++++++----------------------

4 files changed, 70 insertions(+), 84 deletions(-)
Quoted 3 lines
Quoted 1 line

+static void ga_apply_command_filters(GAState *state)

Nit: inline?

No, consensus is today's compilers are smart enough to notice inlining, developers shouldn't worry about this anymore.

Quoted 4 lines
Quoted 4 lines

+{
+    qmp_for_each_command(&ga_commands, ga_apply_command_filters_iter, state);
+}

#61 lore raw

Re: [PATCH v2 22/22] qga: centralize logic for disabling/enabling commands

In reply to Manos Pitsidianakis
Quoted 102 lines · Manos Pitsidianakis

Hello Daniel,

This cleanup seems like a good idea,

Quoted 97 lines · Daniel P. Berrangé

It is confusing having many different pieces of code enabling and disabling commands, and it is not clear that they all have the same semantics, especially wrt prioritization of the block/allow lists. The code attempted to prevent the user from setting both the block and allow lists concurrently, however, the logic was flawed as it checked settings in the configuration file separately from the command line arguments. Thus it was possible to set a block list in the config file and an allow list via a command line argument. The --dump-conf option also creates a configuration file with both keys present, even if unset, which means it is creating a config that cannot actually be loaded again.

Centralizing the code in a single method "ga_apply_command_filters" will provide a strong guarantee of consistency and clarify the intended behaviour. With this there is no compelling technical reason to prevent concurrent setting of both the allow and block lists, so this flawed restriction is removed.

---

docs/interop/qemu-ga.rst |  14 +++++
qga/commands-posix.c     |   6 --
qga/commands-win32.c     |   6 --
qga/main.c               | 128 +++++++++++++++++----------------------
4 files changed, 70 insertions(+), 84 deletions(-)
qga/main.c +26 -26
index f68a32bf7b..72c16fead8 100644
@@ -419,60 +419,79 @@ static gint ga_strcmp(gconstpointer str1, gconstpointer str2)
    return strcmp(str1, str2);
 }
 
-/* disable commands that aren't safe for fsfreeze */
-static void ga_disable_not_allowed_freeze(const QmpCommand *cmd, void *opaque)
+static bool ga_command_is_allowed(const QmpCommand *cmd, GAState *state)
 {
-    bool allowed = false;
    int i = 0;
+    GAConfig *config = state->config;
    const char *name = qmp_command_name(cmd);
+    /* Fallback policy is allow everything */
+    bool allowed = true;
 
-    while (ga_freeze_allowlist[i] != NULL) {
-        if (strcmp(name, ga_freeze_allowlist[i]) == 0) {
+    if (config->allowedrpcs) {
+        /*
+         * If an allow-list is given, this changes the fallback
+         * policy to deny everything
+         */
+        allowed = false;
+
+        if (g_list_find_custom(config->allowedrpcs, name, ga_strcmp) != NULL) {
            allowed = true;
        }
-        i++;
    }
-    if (!allowed) {
-        g_debug("disabling command: %s", name);
-        qmp_disable_command(&ga_commands, name, "the agent is in frozen state");
-    }
-}
 
-/* [re-]enable all commands, except those explicitly blocked by user */
-static void ga_enable_non_blocked(const QmpCommand *cmd, void *opaque)
-{
-    GAState *s = opaque;
-    GList *blockedrpcs = s->blockedrpcs;
-    GList *allowedrpcs = s->allowedrpcs;
-    const char *name = qmp_command_name(cmd);
-
-    if (g_list_find_custom(blockedrpcs, name, ga_strcmp) == NULL) {
-        if (qmp_command_is_enabled(cmd)) {
-            return;
+    /*
+     * If both allowedrpcs and blockedrpcs are set, the blocked
+     * list will take priority
+     */
+    if (config->blockedrpcs) {
+        if (g_list_find_custom(config->blockedrpcs, name, ga_strcmp) != NULL) {
+            allowed = false;
        }
+    }
 
-        if (allowedrpcs &&
-            g_list_find_custom(allowedrpcs, name, ga_strcmp) == NULL) {
-            return;
-        }
+    /*
+     * If frozen, this filtering must take priority over
+     * absolutely everything
+     */
+    if (state->frozen) {
+        allowed = false;
 
Quoted 16 lines

qga/main.c static gint ga_strcmp(gconstpointer str1, gconstpointer str2) · line 468

Quoted 12 lines
qga/main.c +8 -2
-        g_debug("enabling command: %s", name);
-        qmp_enable_command(&ga_commands, name);
+        while (ga_freeze_allowlist[i] != NULL) {
+            if (strcmp(name, ga_freeze_allowlist[i]) == 0) {
+                allowed = true;
+            }
+            i++;
+        }
    }
+
+    return allowed;
 }

IUUC, we can check by priority here: first check if (state->frozen), then blockedrpcs, then allowedrpcs and then return a default fallback value allowed = config->blockedrpcs != NULL && config->allowedrpcs != NULL

That would imply each check does an early return. When I add in the following series, I have further checks going in this method which rely on the fallthrough for overrides, which works better as it is written here.

With regards,
Daniel

Signature · 3 lines
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|
Reviewed-by on v2 6/22
#62 lore raw

Re: [PATCH v2 06/22] qga: move linux memory block command impls to commands-linux.c

In reply to Daniel P. Berrangé
Quoted 4 lines · Daniel P. Berrangé

The qmp_guest_{set,get}_{memory_blocks,block_info} command impls in commands-posix.c are surrounded by '#ifdef __linux__' so should instead live in commands-linux.c

Quoted 8 lines

This also removes a "#ifdef CONFIG_LINUX" that was nested inside a "#ifdef __linux__".

---
  qga/commands-linux.c | 308 ++++++++++++++++++++++++++++++++++++++++++
  qga/commands-posix.c | 311 +------------------------------------------
  2 files changed, 309 insertions(+), 310 deletions(-)

Reviewed using 'git-diff --color-moved=dimmed-zebra'.

Reviewed-by on v2 6/22
#63 lore raw

Re: [PATCH v2 06/22] qga: move linux memory block command impls to commands-linux.c

In reply to Daniel P. Berrangé
Quoted 692 lines · Daniel P. Berrangé

The qmp_guest_{set,get}_{memory_blocks,block_info} command impls in commands-posix.c are surrounded by '#ifdef __linux__' so should instead live in commands-linux.c

This also removes a "#ifdef CONFIG_LINUX" that was nested inside a "#ifdef __linux__".

---
 qga/commands-linux.c | 308 ++++++++++++++++++++++++++++++++++++++++++
 qga/commands-posix.c | 311 +------------------------------------------
 2 files changed, 309 insertions(+), 310 deletions(-)
qga/commands-linux.c +37 -0
index c0e8bd4062..73b13fbaf6 100644
@@ -1595,6 +1595,314 @@ int64_t
 qmp_guest_set_vcpus(GuestLogicalProcessorList *vcpus, Error **errp)
     return processed;
 }
 
+
+static void ga_read_sysfs_file(int dirfd, const char *pathname, char *buf,
+                               int size, Error **errp)
+{
+    int fd;
+    int res;
+
+    errno = 0;
+    fd = openat(dirfd, pathname, O_RDONLY);
+    if (fd == -1) {
+        error_setg_errno(errp, errno, "open sysfs file \"%s\"", pathname);
+        return;
+    }
+
+    res = pread(fd, buf, size, 0);
+    if (res == -1) {
+        error_setg_errno(errp, errno, "pread sysfs file \"%s\"",
 pathname);
+    } else if (res == 0) {
+        error_setg(errp, "pread sysfs file \"%s\": unexpected EOF",
 pathname);
+    }
+    close(fd);
+}
+
+static void ga_write_sysfs_file(int dirfd, const char *pathname,
+                                const char *buf, int size, Error **errp)
+{
+    int fd;
+
+    errno = 0;
+    fd = openat(dirfd, pathname, O_WRONLY);
+    if (fd == -1) {
+        error_setg_errno(errp, errno, "open sysfs file \"%s\"", pathname);
+        return;
+    }
+
+    if (pwrite(fd, buf, size, 0) == -1) {
+        error_setg_errno(errp, errno, "pwrite sysfs file \"%s\"",

pathname);

qga/commands-linux.c +10 -0
+    }
+
+    close(fd);
+}
+
+/* Transfer online/offline status between @mem_blk and the guest system.
+ *
+ * On input either @errp or *@errp must be NULL.
+ *
+ * In system-to-@mem_blk direction, the following @mem_blk fields are

accessed:
+ * - R: mem_blk->phys_index
+ * - W: mem_blk->online
+ * - W: mem_blk->can_offline
+ *
+ * In @mem_blk-to-system direction, the following @mem_blk fields are accessed:
+ * - R: mem_blk->phys_index
+ * - R: mem_blk->online
+ *- R: mem_blk->can_offline
+ * Written members remain unmodified on error.
+ */
+static void transfer_memory_block(GuestMemoryBlock *mem_blk, bool sys2memblk,

qga/commands-linux.c +13 -0
+                                  GuestMemoryBlockResponse *result,
+                                  Error **errp)
+{
+    char *dirpath;
+    int dirfd;
+    char *status;
+    Error *local_err = NULL;
+
+    if (!sys2memblk) {
+        DIR *dp;
+
+        if (!result) {
+            error_setg(errp, "Internal error, 'result' should not be

NULL");

qga/commands-linux.c +12 -0
+            return;
+        }
+        errno = 0;
+        dp = opendir("/sys/devices/system/memory/");
+         /* if there is no 'memory' directory in sysfs,
+         * we think this VM does not support online/offline memory block,
+         * any other solution?
+         */
+        if (!dp) {
+            if (errno == ENOENT) {
+                result->response =
+

GUEST_MEMORY_BLOCK_RESPONSE_TYPE_OPERATION_NOT_SUPPORTED;

qga/commands-linux.c +7 -0
+            }
+            goto out1;
+        }
+        closedir(dp);
+    }
+
+    dirpath = g_strdup_printf("/sys/devices/system/memory/memory%" PRId64

"/",

qga/commands-linux.c +8 -0
+                              mem_blk->phys_index);
+    dirfd = open(dirpath, O_RDONLY | O_DIRECTORY);
+    if (dirfd == -1) {
+        if (sys2memblk) {
+            error_setg_errno(errp, errno, "open(\"%s\")", dirpath);
+        } else {
+            if (errno == ENOENT) {
+                result->response =

GUEST_MEMORY_BLOCK_RESPONSE_TYPE_NOT_FOUND;

qga/commands-linux.c +22 -0
+            } else {
+                result->response =
+                    GUEST_MEMORY_BLOCK_RESPONSE_TYPE_OPERATION_FAILED;
+            }
+        }
+        g_free(dirpath);
+        goto out1;
+    }
+    g_free(dirpath);
+
+    status = g_malloc0(10);
+    ga_read_sysfs_file(dirfd, "state", status, 10, &local_err);
+    if (local_err) {
+        /* treat with sysfs file that not exist in old kernel */
+        if (errno == ENOENT) {
+            error_free(local_err);
+            if (sys2memblk) {
+                mem_blk->online = true;
+                mem_blk->can_offline = false;
+            } else if (!mem_blk->online) {
+                result->response =
+

GUEST_MEMORY_BLOCK_RESPONSE_TYPE_OPERATION_NOT_SUPPORTED;

qga/commands-linux.c +21 -0
+            }
+        } else {
+            if (sys2memblk) {
+                error_propagate(errp, local_err);
+            } else {
+                error_free(local_err);
+                result->response =
+                    GUEST_MEMORY_BLOCK_RESPONSE_TYPE_OPERATION_FAILED;
+            }
+        }
+        goto out2;
+    }
+
+    if (sys2memblk) {
+        char removable = '0';
+
+        mem_blk->online = (strncmp(status, "online", 6) == 0);
+
+        ga_read_sysfs_file(dirfd, "removable", &removable, 1, &local_err);
+        if (local_err) {
+            /* if no 'removable' file, it doesn't support offline mem blk

*/

qga/commands-linux.c +12 -0
+            if (errno == ENOENT) {
+                error_free(local_err);
+                mem_blk->can_offline = false;
+            } else {
+                error_propagate(errp, local_err);
+            }
+        } else {
+            mem_blk->can_offline = (removable != '0');
+        }
+    } else {
+        if (mem_blk->online != (strncmp(status, "online", 6) == 0)) {
+            const char *new_state = mem_blk->online ? "online" :

"offline";

qga/commands-linux.c +2 -0
+
+            ga_write_sysfs_file(dirfd, "state", new_state,

strlen(new_state),

qga/commands-linux.c +51 -0
+                                &local_err);
+            if (local_err) {
+                error_free(local_err);
+                result->response =
+                    GUEST_MEMORY_BLOCK_RESPONSE_TYPE_OPERATION_FAILED;
+                goto out2;
+            }
+
+            result->response = GUEST_MEMORY_BLOCK_RESPONSE_TYPE_SUCCESS;
+            result->has_error_code = false;
+        } /* otherwise pretend successful re-(on|off)-lining */
+    }
+    g_free(status);
+    close(dirfd);
+    return;
+
+out2:
+    g_free(status);
+    close(dirfd);
+out1:
+    if (!sys2memblk) {
+        result->has_error_code = true;
+        result->error_code = errno;
+    }
+}
+
+GuestMemoryBlockList *qmp_guest_get_memory_blocks(Error **errp)
+{
+    GuestMemoryBlockList *head, **tail;
+    Error *local_err = NULL;
+    struct dirent *de;
+    DIR *dp;
+
+    head = NULL;
+    tail = &head;
+
+    dp = opendir("/sys/devices/system/memory/");
+    if (!dp) {
+        /* it's ok if this happens to be a system that doesn't expose
+         * memory blocks via sysfs, but otherwise we should report
+         * an error
+         */
+        if (errno != ENOENT) {
+            error_setg_errno(errp, errno, "Can't open directory"
+                             "\"/sys/devices/system/memory/\"");
+        }
+        return NULL;
+    }
+
+    /* Note: the phys_index of memory block may be discontinuous,
+     * this is because a memblk is the unit of the Sparse Memory design,

which

qga/commands-linux.c +13 -0
+     * allows discontinuous memory ranges (ex. NUMA), so here we should
+     * traverse the memory block directory.
+     */
+    while ((de = readdir(dp)) != NULL) {
+        GuestMemoryBlock *mem_blk;
+
+        if ((strncmp(de->d_name, "memory", 6) != 0) ||
+            !(de->d_type & DT_DIR)) {
+            continue;
+        }
+
+        mem_blk = g_malloc0(sizeof *mem_blk);
+        /* The d_name is "memoryXXX",  phys_index is block id, same as

XXX */

qga/commands-linux.c +89 -0
+        mem_blk->phys_index = strtoul(&de->d_name[6], NULL, 10);
+        mem_blk->has_can_offline = true; /* lolspeak ftw */
+        transfer_memory_block(mem_blk, true, NULL, &local_err);
+        if (local_err) {
+            break;
+        }
+
+        QAPI_LIST_APPEND(tail, mem_blk);
+    }
+
+    closedir(dp);
+    if (local_err == NULL) {
+        /* there's no guest with zero memory blocks */
+        if (head == NULL) {
+            error_setg(errp, "guest reported zero memory blocks!");
+        }
+        return head;
+    }
+
+    qapi_free_GuestMemoryBlockList(head);
+    error_propagate(errp, local_err);
+    return NULL;
+}
+
+GuestMemoryBlockResponseList *
+qmp_guest_set_memory_blocks(GuestMemoryBlockList *mem_blks, Error **errp)
+{
+    GuestMemoryBlockResponseList *head, **tail;
+    Error *local_err = NULL;
+
+    head = NULL;
+    tail = &head;
+
+    while (mem_blks != NULL) {
+        GuestMemoryBlockResponse *result;
+        GuestMemoryBlock *current_mem_blk = mem_blks->value;
+
+        result = g_malloc0(sizeof(*result));
+        result->phys_index = current_mem_blk->phys_index;
+        transfer_memory_block(current_mem_blk, false, result, &local_err);
+        if (local_err) { /* should never happen */
+            goto err;
+        }
+
+        QAPI_LIST_APPEND(tail, result);
+        mem_blks = mem_blks->next;
+    }
+
+    return head;
+err:
+    qapi_free_GuestMemoryBlockResponseList(head);
+    error_propagate(errp, local_err);
+    return NULL;
+}
+
+GuestMemoryBlockInfo *qmp_guest_get_memory_block_info(Error **errp)
+{
+    Error *local_err = NULL;
+    char *dirpath;
+    int dirfd;
+    char *buf;
+    GuestMemoryBlockInfo *info;
+
+    dirpath = g_strdup_printf("/sys/devices/system/memory/");
+    dirfd = open(dirpath, O_RDONLY | O_DIRECTORY);
+    if (dirfd == -1) {
+        error_setg_errno(errp, errno, "open(\"%s\")", dirpath);
+        g_free(dirpath);
+        return NULL;
+    }
+    g_free(dirpath);
+
+    buf = g_malloc0(20);
+    ga_read_sysfs_file(dirfd, "block_size_bytes", buf, 20, &local_err);
+    close(dirfd);
+    if (local_err) {
+        g_free(buf);
+        error_propagate(errp, local_err);
+        return NULL;
+    }
+
+    info = g_new0(GuestMemoryBlockInfo, 1);
+    info->size = strtol(buf, NULL, 16); /* the unit is bytes */
+
+    g_free(buf);
+
+    return info;
+}
+
 #define MAX_NAME_LEN 128
 static GuestDiskStatsInfoList *guest_get_diskstats(Error **errp)
 {
qga/commands-posix.c +0 -47
index 5da60e65ab..2a3bef7445 100644
@@ -887,316 +887,7 @@ void qmp_guest_set_user_password(const char
 *username,
 }
 #endif /* __linux__ || __FreeBSD__ */
 
-#ifdef __linux__
-static void ga_read_sysfs_file(int dirfd, const char *pathname, char *buf,
-                               int size, Error **errp)
-{
-    int fd;
-    int res;
-
-    errno = 0;
-    fd = openat(dirfd, pathname, O_RDONLY);
-    if (fd == -1) {
-        error_setg_errno(errp, errno, "open sysfs file \"%s\"", pathname);
-        return;
-    }
-
-    res = pread(fd, buf, size, 0);
-    if (res == -1) {
-        error_setg_errno(errp, errno, "pread sysfs file \"%s\"",
 pathname);
-    } else if (res == 0) {
-        error_setg(errp, "pread sysfs file \"%s\": unexpected EOF",
 pathname);
-    }
-    close(fd);
-}
-
-static void ga_write_sysfs_file(int dirfd, const char *pathname,
-                                const char *buf, int size, Error **errp)
-{
-    int fd;
-
-    errno = 0;
-    fd = openat(dirfd, pathname, O_WRONLY);
-    if (fd == -1) {
-        error_setg_errno(errp, errno, "open sysfs file \"%s\"", pathname);
-        return;
-    }
-
-    if (pwrite(fd, buf, size, 0) == -1) {
-        error_setg_errno(errp, errno, "pwrite sysfs file \"%s\"",
 pathname);
-    }
-
-    close(fd);
-}
-
-/* Transfer online/offline status between @mem_blk and the guest system.
- *
- * On input either @errp or *@errp must be NULL.
- *
- * In system-to-@mem_blk direction, the following @mem_blk fields are

accessed:
- * - R: mem_blk->phys_index
- * - W: mem_blk->online
- * - W: mem_blk->can_offline
- *
- * In @mem_blk-to-system direction, the following @mem_blk fields are accessed:
- * - R: mem_blk->phys_index
- * - R: mem_blk->online
- *- R: mem_blk->can_offline
- * Written members remain unmodified on error.
- */
-static void transfer_memory_block(GuestMemoryBlock *mem_blk, bool sys2memblk,

qga/commands-posix.c +0 -13
-                                  GuestMemoryBlockResponse *result,
-                                  Error **errp)
-{
-    char *dirpath;
-    int dirfd;
-    char *status;
-    Error *local_err = NULL;
-
-    if (!sys2memblk) {
-        DIR *dp;
-
-        if (!result) {
-            error_setg(errp, "Internal error, 'result' should not be

NULL");

qga/commands-posix.c +0 -12
-            return;
-        }
-        errno = 0;
-        dp = opendir("/sys/devices/system/memory/");
-         /* if there is no 'memory' directory in sysfs,
-         * we think this VM does not support online/offline memory block,
-         * any other solution?
-         */
-        if (!dp) {
-            if (errno == ENOENT) {
-                result->response =
-

GUEST_MEMORY_BLOCK_RESPONSE_TYPE_OPERATION_NOT_SUPPORTED;

qga/commands-posix.c +0 -7
-            }
-            goto out1;
-        }
-        closedir(dp);
-    }
-
-    dirpath = g_strdup_printf("/sys/devices/system/memory/memory%" PRId64

"/",

qga/commands-posix.c +0 -8
-                              mem_blk->phys_index);
-    dirfd = open(dirpath, O_RDONLY | O_DIRECTORY);
-    if (dirfd == -1) {
-        if (sys2memblk) {
-            error_setg_errno(errp, errno, "open(\"%s\")", dirpath);
-        } else {
-            if (errno == ENOENT) {
-                result->response =

GUEST_MEMORY_BLOCK_RESPONSE_TYPE_NOT_FOUND;

qga/commands-posix.c +0 -22
-            } else {
-                result->response =
-                    GUEST_MEMORY_BLOCK_RESPONSE_TYPE_OPERATION_FAILED;
-            }
-        }
-        g_free(dirpath);
-        goto out1;
-    }
-    g_free(dirpath);
-
-    status = g_malloc0(10);
-    ga_read_sysfs_file(dirfd, "state", status, 10, &local_err);
-    if (local_err) {
-        /* treat with sysfs file that not exist in old kernel */
-        if (errno == ENOENT) {
-            error_free(local_err);
-            if (sys2memblk) {
-                mem_blk->online = true;
-                mem_blk->can_offline = false;
-            } else if (!mem_blk->online) {
-                result->response =
-

GUEST_MEMORY_BLOCK_RESPONSE_TYPE_OPERATION_NOT_SUPPORTED;

qga/commands-posix.c +0 -21
-            }
-        } else {
-            if (sys2memblk) {
-                error_propagate(errp, local_err);
-            } else {
-                error_free(local_err);
-                result->response =
-                    GUEST_MEMORY_BLOCK_RESPONSE_TYPE_OPERATION_FAILED;
-            }
-        }
-        goto out2;
-    }
-
-    if (sys2memblk) {
-        char removable = '0';
-
-        mem_blk->online = (strncmp(status, "online", 6) == 0);
-
-        ga_read_sysfs_file(dirfd, "removable", &removable, 1, &local_err);
-        if (local_err) {
-            /* if no 'removable' file, it doesn't support offline mem blk

*/

qga/commands-posix.c +0 -12
-            if (errno == ENOENT) {
-                error_free(local_err);
-                mem_blk->can_offline = false;
-            } else {
-                error_propagate(errp, local_err);
-            }
-        } else {
-            mem_blk->can_offline = (removable != '0');
-        }
-    } else {
-        if (mem_blk->online != (strncmp(status, "online", 6) == 0)) {
-            const char *new_state = mem_blk->online ? "online" :

"offline";

qga/commands-posix.c +0 -2
-
-            ga_write_sysfs_file(dirfd, "state", new_state,

strlen(new_state),

qga/commands-posix.c +0 -51
-                                &local_err);
-            if (local_err) {
-                error_free(local_err);
-                result->response =
-                    GUEST_MEMORY_BLOCK_RESPONSE_TYPE_OPERATION_FAILED;
-                goto out2;
-            }
-
-            result->response = GUEST_MEMORY_BLOCK_RESPONSE_TYPE_SUCCESS;
-            result->has_error_code = false;
-        } /* otherwise pretend successful re-(on|off)-lining */
-    }
-    g_free(status);
-    close(dirfd);
-    return;
-
-out2:
-    g_free(status);
-    close(dirfd);
-out1:
-    if (!sys2memblk) {
-        result->has_error_code = true;
-        result->error_code = errno;
-    }
-}
-
-GuestMemoryBlockList *qmp_guest_get_memory_blocks(Error **errp)
-{
-    GuestMemoryBlockList *head, **tail;
-    Error *local_err = NULL;
-    struct dirent *de;
-    DIR *dp;
-
-    head = NULL;
-    tail = &head;
-
-    dp = opendir("/sys/devices/system/memory/");
-    if (!dp) {
-        /* it's ok if this happens to be a system that doesn't expose
-         * memory blocks via sysfs, but otherwise we should report
-         * an error
-         */
-        if (errno != ENOENT) {
-            error_setg_errno(errp, errno, "Can't open directory"
-                             "\"/sys/devices/system/memory/\"");
-        }
-        return NULL;
-    }
-
-    /* Note: the phys_index of memory block may be discontinuous,
-     * this is because a memblk is the unit of the Sparse Memory design,

which

qga/commands-posix.c +0 -13
-     * allows discontinuous memory ranges (ex. NUMA), so here we should
-     * traverse the memory block directory.
-     */
-    while ((de = readdir(dp)) != NULL) {
-        GuestMemoryBlock *mem_blk;
-
-        if ((strncmp(de->d_name, "memory", 6) != 0) ||
-            !(de->d_type & DT_DIR)) {
-            continue;
-        }
-
-        mem_blk = g_malloc0(sizeof *mem_blk);
-        /* The d_name is "memoryXXX",  phys_index is block id, same as

XXX */

qga/commands-posix.c +1 -92
-        mem_blk->phys_index = strtoul(&de->d_name[6], NULL, 10);
-        mem_blk->has_can_offline = true; /* lolspeak ftw */
-        transfer_memory_block(mem_blk, true, NULL, &local_err);
-        if (local_err) {
-            break;
-        }
-
-        QAPI_LIST_APPEND(tail, mem_blk);
-    }
-
-    closedir(dp);
-    if (local_err == NULL) {
-        /* there's no guest with zero memory blocks */
-        if (head == NULL) {
-            error_setg(errp, "guest reported zero memory blocks!");
-        }
-        return head;
-    }
-
-    qapi_free_GuestMemoryBlockList(head);
-    error_propagate(errp, local_err);
-    return NULL;
-}
-
-GuestMemoryBlockResponseList *
-qmp_guest_set_memory_blocks(GuestMemoryBlockList *mem_blks, Error **errp)
-{
-    GuestMemoryBlockResponseList *head, **tail;
-    Error *local_err = NULL;
-
-    head = NULL;
-    tail = &head;
-
-    while (mem_blks != NULL) {
-        GuestMemoryBlockResponse *result;
-        GuestMemoryBlock *current_mem_blk = mem_blks->value;
-
-        result = g_malloc0(sizeof(*result));
-        result->phys_index = current_mem_blk->phys_index;
-        transfer_memory_block(current_mem_blk, false, result, &local_err);
-        if (local_err) { /* should never happen */
-            goto err;
-        }
-
-        QAPI_LIST_APPEND(tail, result);
-        mem_blks = mem_blks->next;
-    }
-
-    return head;
-err:
-    qapi_free_GuestMemoryBlockResponseList(head);
-    error_propagate(errp, local_err);
-    return NULL;
-}
-
-GuestMemoryBlockInfo *qmp_guest_get_memory_block_info(Error **errp)
-{
-    Error *local_err = NULL;
-    char *dirpath;
-    int dirfd;
-    char *buf;
-    GuestMemoryBlockInfo *info;
-
-    dirpath = g_strdup_printf("/sys/devices/system/memory/");
-    dirfd = open(dirpath, O_RDONLY | O_DIRECTORY);
-    if (dirfd == -1) {
-        error_setg_errno(errp, errno, "open(\"%s\")", dirpath);
-        g_free(dirpath);
-        return NULL;
-    }
-    g_free(dirpath);
-
-    buf = g_malloc0(20);
-    ga_read_sysfs_file(dirfd, "block_size_bytes", buf, 20, &local_err);
-    close(dirfd);
-    if (local_err) {
-        g_free(buf);
-        error_propagate(errp, local_err);
-        return NULL;
-    }
-
-    info = g_new0(GuestMemoryBlockInfo, 1);
-    info->size = strtol(buf, NULL, 16); /* the unit is bytes */
-
-    g_free(buf);
-
-    return info;
-}
-
-
-#else /* defined(__linux__) */
+#ifndef __linux__
 
 void qmp_guest_suspend_disk(Error **errp)
 {
--

2.45.1

#64 lore raw
In reply to Daniel P. Berrangé

Hi

Quoted 74 lines · Daniel P. Berrangé

This series is a side effect of other work I started, to attempt to make the QGA safe to use in confidential VMs by automatically restricting the permitted commands. Since this cleanup stands on its own, I'm sending it now.

The QGA codebase has a very complicated maze of #ifdefs to create stubs for the various commands that cannot be implemented on certain platforms. It then has further logic to dynamically disable the stub commands at runtime, except this is not consistently applied, so some commands remain enabled despite being merely stubs.

The resulting code is hard to follow, when trying to understand exactly what commands are available under what circumstances, and when changing impls it is easy to get the #ifdefs wrong, resulting in stubs getting missed on platforms without a real impl. In some cases, we have multiple stubs for the same command, due to the maze of #ifdefs.

The QAPI schema language has support for many years for expressing conditions against commands when declaring them. This results in the QAPI code generator omitting their implementation entirely at build time. This has mutliple benefits

* The unsupported commands are guaranteed to not exist at runtime
* No stubs need ever be defined in the code
* The generated QAPI reference manual documents the build conditions

This series is broadly split into three parts

* Moving tonnes of Linux only commands out of commands-posix.c into commands-linux.c to remove many #ifdefs.
* Adding 'if' conditions in the QAPI schema to reflect the build conditions, removing many more #ifdefs
* Sanitizing the logic for disabling/enabling commands at runtime to guarantee consistency

Changed in v2:

- Make FSFreeze error reporting distinguish inability to enable VSS from user config choice

- Fully remove ga_command_init_blockedrpcs() methods. No more special case disabling of commands. Either they're disabled at build time, or disabled by user config, or by well defined rule ie not permitted during FS freeze.

- Apply rules later in startup to avoid crash from NULL config pointer

- Document changed error messages in commit messages

- Add -c / --config command line parameter

- Fix mistaken enabling of fsfreeze hooks on win32

- Remove pointless 'blockrpcs_key' variable

- Allow concurrent setting of allow and block lists for RPC commands

Daniel P. Berrangé (22):
qga: drop blocking of guest-get-memory-block-size command
qga: move linux vcpu command impls to commands-linux.c
qga: move linux suspend command impls to commands-linux.c
qga: move linux fs/disk command impls to commands-linux.c
qga: move linux disk/cpu stats command impls to commands-linux.c
qga: move linux memory block command impls to commands-linux.c
qga: move CONFIG_FSFREEZE/TRIM to be meson defined options
qga: conditionalize schema for commands unsupported on Windows
qga: conditionalize schema for commands unsupported on non-Linux POSIX
qga: conditionalize schema for commands requiring getifaddrs
qga: conditionalize schema for commands requiring linux/win32
qga: conditionalize schema for commands only supported on Windows
qga: conditionalize schema for commands requiring fsfreeze
qga: conditionalize schema for commands requiring fstrim

Quoted 10 lines

qga: conditionalize schema for commands requiring libudev
qga: conditionalize schema for commands requiring utmpx
qga: conditionalize schema for commands not supported on other UNIX
qga: don't disable fsfreeze commands if vss_init fails
qga: move declare of QGAConfig struct to top of file
qga: remove pointless 'blockrpcs_key' variable
qga: allow configuration file path via the cli
qga: centralize logic for disabling/enabling commands

Something broke patchew handling:
https://patchew.org/QEMU/[email protected]/[email protected]/

Signature · 1 line
Marc-André Lureau
#65 lore raw
Quoted 89 lines · Marc-André Lureau

Hi

Quoted 74 lines · Daniel P. Berrangé

This series is a side effect of other work I started, to attempt to make the QGA safe to use in confidential VMs by automatically restricting the permitted commands. Since this cleanup stands on its own, I'm sending it now.

The QGA codebase has a very complicated maze of #ifdefs to create stubs for the various commands that cannot be implemented on certain platforms. It then has further logic to dynamically disable the stub commands at runtime, except this is not consistently applied, so some commands remain enabled despite being merely stubs.

The resulting code is hard to follow, when trying to understand exactly what commands are available under what circumstances, and when changing impls it is easy to get the #ifdefs wrong, resulting in stubs getting missed on platforms without a real impl. In some cases, we have multiple stubs for the same command, due to the maze of #ifdefs.

The QAPI schema language has support for many years for expressing conditions against commands when declaring them. This results in the QAPI code generator omitting their implementation entirely at build time. This has mutliple benefits

* The unsupported commands are guaranteed to not exist at runtime
* No stubs need ever be defined in the code
* The generated QAPI reference manual documents the build conditions

This series is broadly split into three parts

* Moving tonnes of Linux only commands out of commands-posix.c into commands-linux.c to remove many #ifdefs.
* Adding 'if' conditions in the QAPI schema to reflect the build conditions, removing many more #ifdefs
* Sanitizing the logic for disabling/enabling commands at runtime to guarantee consistency

Changed in v2:

- Make FSFreeze error reporting distinguish inability to enable VSS from user config choice

- Fully remove ga_command_init_blockedrpcs() methods. No more special case disabling of commands. Either they're disabled at build time, or disabled by user config, or by well defined rule ie not permitted during FS freeze.

- Apply rules later in startup to avoid crash from NULL config pointer

- Document changed error messages in commit messages

- Add -c / --config command line parameter

- Fix mistaken enabling of fsfreeze hooks on win32

- Remove pointless 'blockrpcs_key' variable

- Allow concurrent setting of allow and block lists for RPC commands

Daniel P. Berrangé (22):
qga: drop blocking of guest-get-memory-block-size command
qga: move linux vcpu command impls to commands-linux.c
qga: move linux suspend command impls to commands-linux.c
qga: move linux fs/disk command impls to commands-linux.c
qga: move linux disk/cpu stats command impls to commands-linux.c
qga: move linux memory block command impls to commands-linux.c
qga: move CONFIG_FSFREEZE/TRIM to be meson defined options
qga: conditionalize schema for commands unsupported on Windows
qga: conditionalize schema for commands unsupported on non-Linux POSIX
qga: conditionalize schema for commands requiring getifaddrs
qga: conditionalize schema for commands requiring linux/win32
qga: conditionalize schema for commands only supported on Windows
qga: conditionalize schema for commands requiring fsfreeze
qga: conditionalize schema for commands requiring fstrim

Quoted 10 lines

qga: conditionalize schema for commands requiring libudev
qga: conditionalize schema for commands requiring utmpx
qga: conditionalize schema for commands not supported on other UNIX
qga: don't disable fsfreeze commands if vss_init fails
qga: move declare of QGAConfig struct to top of file
qga: remove pointless 'blockrpcs_key' variable
qga: allow configuration file path via the cli
qga: centralize logic for disabling/enabling commands

Quoted 2 lines

Something broke patchew handling:
https://patchew.org/QEMU/[email protected]/[email protected]/

gmail refused further mail delivery from me part way through sending the series, so I had to send the 2nd half of it separately.

With regards,
Daniel

Signature · 3 lines
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|
#66 lore raw
In reply to Daniel P. Berrangé

Ping: for any review comments from QGA maintainers ?

Quoted 83 lines · Daniel P. Berrangé

This series is a side effect of other work I started, to attempt to make the QGA safe to use in confidential VMs by automatically restricting the permitted commands. Since this cleanup stands on its own, I'm sending it now.

The QGA codebase has a very complicated maze of #ifdefs to create stubs for the various commands that cannot be implemented on certain platforms. It then has further logic to dynamically disable the stub commands at runtime, except this is not consistently applied, so some commands remain enabled despite being merely stubs.

The resulting code is hard to follow, when trying to understand exactly what commands are available under what circumstances, and when changing impls it is easy to get the #ifdefs wrong, resulting in stubs getting missed on platforms without a real impl. In some cases, we have multiple stubs for the same command, due to the maze of #ifdefs.

The QAPI schema language has support for many years for expressing conditions against commands when declaring them. This results in the QAPI code generator omitting their implementation entirely at build time. This has mutliple benefits

* The unsupported commands are guaranteed to not exist at runtime
* No stubs need ever be defined in the code
* The generated QAPI reference manual documents the build conditions

This series is broadly split into three parts

* Moving tonnes of Linux only commands out of commands-posix.c into commands-linux.c to remove many #ifdefs.
* Adding 'if' conditions in the QAPI schema to reflect the build conditions, removing many more #ifdefs
* Sanitizing the logic for disabling/enabling commands at runtime to guarantee consistency

Changed in v2:

- Make FSFreeze error reporting distinguish inability to enable VSS from user config choice

- Fully remove ga_command_init_blockedrpcs() methods. No more special case disabling of commands. Either they're disabled at build time, or disabled by user config, or by well defined rule ie not permitted during FS freeze.

- Apply rules later in startup to avoid crash from NULL config pointer

- Document changed error messages in commit messages

- Add -c / --config command line parameter

- Fix mistaken enabling of fsfreeze hooks on win32

- Remove pointless 'blockrpcs_key' variable

- Allow concurrent setting of allow and block lists for RPC commands

Daniel P. Berrangé (22):
qga: drop blocking of guest-get-memory-block-size command
qga: move linux vcpu command impls to commands-linux.c
qga: move linux suspend command impls to commands-linux.c
qga: move linux fs/disk command impls to commands-linux.c
qga: move linux disk/cpu stats command impls to commands-linux.c
qga: move linux memory block command impls to commands-linux.c
qga: move CONFIG_FSFREEZE/TRIM to be meson defined options
qga: conditionalize schema for commands unsupported on Windows
qga: conditionalize schema for commands unsupported on non-Linux POSIX
qga: conditionalize schema for commands requiring getifaddrs
qga: conditionalize schema for commands requiring linux/win32
qga: conditionalize schema for commands only supported on Windows
qga: conditionalize schema for commands requiring fsfreeze
qga: conditionalize schema for commands requiring fstrim
qga: conditionalize schema for commands requiring libudev
qga: conditionalize schema for commands requiring utmpx
qga: conditionalize schema for commands not supported on other UNIX
qga: don't disable fsfreeze commands if vss_init fails
qga: move declare of QGAConfig struct to top of file
qga: remove pointless 'blockrpcs_key' variable
qga: allow configuration file path via the cli
qga: centralize logic for disabling/enabling commands

Quoted 14 lines
 docs/interop/qemu-ga.rst |   19 +
 meson.build              |   16 +
 qga/commands-bsd.c       |   24 -
 qga/commands-common.h    |    9 -
 qga/commands-linux.c     | 1805 +++++++++++++++++++++++++++++
 qga/commands-posix.c     | 2373 +++-----------------------------------
 qga/commands-win32.c     |   78 +-
 qga/main.c               |  216 ++--
 qga/qapi-schema.json     |  153 ++-
 9 files changed, 2234 insertions(+), 2459 deletions(-)
Signature · 1 line
2.45.1

With regards,
Daniel

Signature · 3 lines
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|
#67 lore raw

Hi Daniel

Quoted 2 lines · Daniel P. Berrangé

Ping: for any review comments from QGA maintainers ?

Maybe you could resend for patchew to correctly handle the series.

thanks

Quoted 107 lines · Daniel P. Berrangé
Quoted 94 lines

This series is a side effect of other work I started, to attempt to make the QGA safe to use in confidential VMs by automatically restricting the permitted commands. Since this cleanup stands on its own, I'm sending it now.

The QGA codebase has a very complicated maze of #ifdefs to create stubs for the various commands that cannot be implemented on certain platforms. It then has further logic to dynamically disable the stub commands at runtime, except this is not consistently applied, so some commands remain enabled despite being merely stubs.

The resulting code is hard to follow, when trying to understand exactly what commands are available under what circumstances, and when changing impls it is easy to get the #ifdefs wrong, resulting in stubs getting missed on platforms without a real impl. In some cases, we have multiple stubs for the same command, due to the maze of #ifdefs.

The QAPI schema language has support for many years for expressing conditions against commands when declaring them. This results in the QAPI code generator omitting their implementation entirely at build time. This has mutliple benefits

* The unsupported commands are guaranteed to not exist at runtime
* No stubs need ever be defined in the code
* The generated QAPI reference manual documents the build conditions

This series is broadly split into three parts

* Moving tonnes of Linux only commands out of commands-posix.c into commands-linux.c to remove many #ifdefs.
* Adding 'if' conditions in the QAPI schema to reflect the build conditions, removing many more #ifdefs
* Sanitizing the logic for disabling/enabling commands at runtime to guarantee consistency

Changed in v2:

- Make FSFreeze error reporting distinguish inability to enable VSS from user config choice

- Fully remove ga_command_init_blockedrpcs() methods. No more special case disabling of commands. Either they're disabled at build time, or disabled by user config, or by well defined rule ie not permitted during FS freeze.

- Apply rules later in startup to avoid crash from NULL config pointer

- Document changed error messages in commit messages

- Add -c / --config command line parameter

- Fix mistaken enabling of fsfreeze hooks on win32

- Remove pointless 'blockrpcs_key' variable

- Allow concurrent setting of allow and block lists for RPC commands

Daniel P. Berrangé (22):
qga: drop blocking of guest-get-memory-block-size command
qga: move linux vcpu command impls to commands-linux.c
qga: move linux suspend command impls to commands-linux.c
qga: move linux fs/disk command impls to commands-linux.c
qga: move linux disk/cpu stats command impls to commands-linux.c
qga: move linux memory block command impls to commands-linux.c
qga: move CONFIG_FSFREEZE/TRIM to be meson defined options
qga: conditionalize schema for commands unsupported on Windows
qga: conditionalize schema for commands unsupported on non-Linux POSIX
qga: conditionalize schema for commands requiring getifaddrs
qga: conditionalize schema for commands requiring linux/win32
qga: conditionalize schema for commands only supported on Windows
qga: conditionalize schema for commands requiring fsfreeze
qga: conditionalize schema for commands requiring fstrim
qga: conditionalize schema for commands requiring libudev
qga: conditionalize schema for commands requiring utmpx
qga: conditionalize schema for commands not supported on other UNIX
qga: don't disable fsfreeze commands if vss_init fails
qga: move declare of QGAConfig struct to top of file
qga: remove pointless 'blockrpcs_key' variable
qga: allow configuration file path via the cli
qga: centralize logic for disabling/enabling commands

 docs/interop/qemu-ga.rst |   19 +
 meson.build              |   16 +
 qga/commands-bsd.c       |   24 -
 qga/commands-common.h    |    9 -
 qga/commands-linux.c     | 1805 +++++++++++++++++++++++++++++
 qga/commands-posix.c     | 2373 +++-----------------------------------
 qga/commands-win32.c     |   78 +-
 qga/main.c               |  216 ++--
 qga/qapi-schema.json     |  153 ++-
 9 files changed, 2234 insertions(+), 2459 deletions(-)
Quoted 3 lines

--
2.45.1

With regards,
Daniel
--

|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|
#68 lore raw
Quoted 7 lines · Marc-André Lureau

Hi Daniel

Quoted 2 lines · Daniel P. Berrangé

Ping: for any review comments from QGA maintainers ?

Maybe you could resend for patchew to correctly handle the series.

I don't want to spam the list again just because patchew didn't handle it. If anyone can't just pull the messages from mail
though, they are also at https://gitlab.com/berrange/qemu/-/tags/qga-conditions-v2

With regards,
Daniel

Signature · 3 lines
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|
#69 lore raw

Hi

Quoted 6 lines · Daniel P. Berrangé
Quoted 4 lines · Marc-André Lureau

Hi Daniel

Quoted 1 line · Daniel P. Berrangé
Quoted 7 lines
Quoted 3 lines
Quoted 1 line

Ping: for any review comments from QGA maintainers ?

Maybe you could resend for patchew to correctly handle the series.

I don't want to spam the list again just because patchew didn't handle it. If anyone can't just pull the messages from mail
though, they are also at https://gitlab.com/berrange/qemu/-/tags/qga-conditions-v2

patchew also handles R-b/A-b/T-b tags, mail ref, versioning etc..