qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: john cooper <john.cooper@third-harmonic.com>
To: qemu-devel@nongnu.org
Cc: john cooper <john.cooper@redhat.com>,
	Anthony Liguori <aliguori@us.ibm.com>
Subject: [Qemu-devel] [PATCH] Allow "-cpu [check|enforce]" to be used implicitly with default CPU model.
Date: Fri, 28 Jan 2011 14:10:11 -0500	[thread overview]
Message-ID: <4D431493.9060701@third-harmonic.com> (raw)

Launching qemu with "-cpu [check|enforce]" without explicitly
prefixing a valid model name doesn't do as intuitively expected.
Rather qemu exits with a CLI parse error.

Unfortunately due to qemu's CLI argument parsing structure and
the fact cpu models are initialized depending on build configuration,
supporting the above implicitly for the default model is far more
clunky than it ought to be.  So as a minor concession to a far more
simple solution, a pseudo model name of "default" is added which
expands internally to qemu's build configuration default and is then
interpreted conventionally.  The result is then the following example
usage:

    # x86_64-softmmu/qemu-system-x86_64 -cpu default,check
    Using CPU model "qemu64,check"
    warning: host cpuid 0000_0001 lacks requested flag 'popcnt' [0x00800000]
    warning: host cpuid 8000_0001 lacks requested flag 'abm' [0x00000020]
    warning: host cpuid 8000_0001 lacks requested flag 'sse4a' [0x00000040]

    # x86_64-softmmu/qemu-system-x86_64 -cpu default,enforce,+sse4.2,+popcnt
    Using CPU model "qemu64,enforce,+sse4.2,+popcnt"
    warning: host cpuid 0000_0001 lacks requested flag 'sse4.2|sse4_2' [0x00100000]
    warning: host cpuid 0000_0001 lacks requested flag 'popcnt' [0x00800000]
    warning: host cpuid 8000_0001 lacks requested flag 'abm' [0x00000020]
    warning: host cpuid 8000_0001 lacks requested flag 'sse4a' [0x00000040]
    Unable to find x86 CPU definition

etc..  allowing both the ability here to "check|enforce" the default
model, as well as accepting arbitrary feature flags due to fan into the
existing flag parsing.  The resulting patch, which also indicates the CPU
model in use, is fairly simple.

Please review and apply.

Signed-off-by: john cooper <john.cooper@redhat.com>
---

diff --git a/hw/pc.c b/hw/pc.c
index 119c110..27c87b5 100644
--- a/hw/pc.c
+++ b/hw/pc.c
@@ -861,18 +861,49 @@ static CPUState *pc_new_cpu(const char *cpu_model)
     return env;
 }
 
+/* CLI cpu model name which expands to the actual configuration default
+ */
+#define CMDEF_KEYWORD     "default"
+#define CMDEF_KEYWORD_LN (sizeof (CMDEF_KEYWORD) - 1)
+
+/* set configuration default cpu model if current model string is
+ * uninitialized, or if user explicitly requests use of the config'ed
+ * default by specifying a cpu model name of "default".
+ * Use of "default" as a cpu model pseudo-name exists primarily to
+ * ease treatment of qualifier flags requested by the user without
+ * requiring knowledge of all cpu model names in advance of full "-cpu"
+ * option parsing.
+ */
+static const char *setdef_cpu_model(const char *model_str,
+    const char *default_str)
+{
+    int default_str_ln = strlen(default_str);
+
+    if (!model_str || !*model_str) {
+        return default_str;
+    } else if (strncmp(model_str, CMDEF_KEYWORD, CMDEF_KEYWORD_LN)) {
+        return model_str;
+    } else {
+        char *new = qemu_malloc(strlen(model_str) - CMDEF_KEYWORD_LN +
+            default_str_ln + 1);
+
+        strcpy(new, default_str);
+        strcpy(new + default_str_ln, model_str + CMDEF_KEYWORD_LN);
+        return new;
+    }
+}
+
 void pc_cpus_init(const char *cpu_model)
 {
     int i;
 
     /* init CPUs */
-    if (cpu_model == NULL) {
+    cpu_model = setdef_cpu_model(cpu_model,
 #ifdef TARGET_X86_64
-        cpu_model = "qemu64";
+        "qemu64");
 #else
-        cpu_model = "qemu32";
+        "qemu32");
 #endif
-    }
 
     for(i = 0; i < smp_cpus; i++) {
         pc_new_cpu(cpu_model);
diff --git a/target-i386/helper.c b/target-i386/helper.c
index 26ea1e5..1b0b586 100644
--- a/target-i386/helper.c
+++ b/target-i386/helper.c
@@ -1129,6 +1129,7 @@ CPUX86State *cpu_x86_init(const char *cpu_model)
     env = qemu_mallocz(sizeof(CPUX86State));
     cpu_exec_init(env);
     env->cpu_model_str = cpu_model;
+    fprintf(stderr, "Using CPU model \"%s\"\n", cpu_model);
 
     /* init various static tables */
     if (!inited) {

-- 
john.cooper@third-harmonic.com

                 reply	other threads:[~2011-01-28 19:38 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4D431493.9060701@third-harmonic.com \
    --to=john.cooper@third-harmonic.com \
    --cc=aliguori@us.ibm.com \
    --cc=john.cooper@redhat.com \
    --cc=qemu-devel@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).