qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Eduardo Habkost <ehabkost@redhat.com>
To: qemu-devel@nongnu.org
Cc: Paolo Bonzini <pbonzini@redhat.com>,
	Igor Mammedov <imammedo@redhat.com>,
	Richard Henderson <richard.henderson@linaro.org>,
	Eduardo Habkost <ehabkost@redhat.com>,
	Babu Moger <babu.moger@amd.com>
Subject: [PATCH 3/3] i386: Sanity check CPU model feature sets
Date: Mon,  1 Feb 2021 17:54:04 -0500	[thread overview]
Message-ID: <20210201225404.3941395-4-ehabkost@redhat.com> (raw)
In-Reply-To: <20210201225404.3941395-1-ehabkost@redhat.com>

All CPU models must refer only to features that have their names
defined in feature_word_info[].feat_names, otherwise error
reporting and query-cpu-model-expansion will break.

Validate CPU feature flags in x86_cpudef_validate(), we can catch
mistakes more easily.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
---
 target/i386/cpu.c | 15 +++++++++++++++
 1 file changed, 15 insertions(+)

diff --git a/target/i386/cpu.c b/target/i386/cpu.c
index 6285fb00eb8..3c066738e82 100644
--- a/target/i386/cpu.c
+++ b/target/i386/cpu.c
@@ -5435,12 +5435,27 @@ static void x86_register_cpu_model_type(const char *name, X86CPUModel *model)
 static void x86_cpudef_validate(X86CPUDefinition *def)
 {
 #ifndef NDEBUG
+    FeatureWord w;
+    int bitnr;
+
     /* AMD aliases are handled at runtime based on CPUID vendor, so
      * they shouldn't be set on the CPU model table.
      */
     assert(!(def->features[FEAT_8000_0001_EDX] & CPUID_EXT2_AMD_ALIASES));
     /* catch mistakes instead of silently truncating model_id when too long */
     assert(def->model_id && strlen(def->model_id) <= 48);
+
+    /*
+     * CPU models must enable only features with valid names, otherwise
+     * error reporting and query-cpu-model-expansion can't work correctly.
+     */
+    for (w = 0; w < FEATURE_WORDS; w++) {
+        for (bitnr = 0; bitnr < 64; bitnr++) {
+            uint64_t mask = (1ULL << bitnr);
+            assert(!(def->features[w] & mask) ||
+                   feature_word_info[w].feat_names[bitnr]);
+        }
+    }
 #endif
 }
 
-- 
2.28.0



      parent reply	other threads:[~2021-02-01 22:57 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-02-01 22:54 [PATCH 0/3] i386: Ensure feature names are always defined Eduardo Habkost
2021-02-01 22:54 ` [PATCH 1/3] i386: Add missing "vmx-ept-wb" feature name Eduardo Habkost
2021-02-01 22:59   ` Paolo Bonzini
2021-02-01 23:05     ` Eduardo Habkost
2021-02-01 23:28       ` Paolo Bonzini
2021-02-02  0:18         ` Eduardo Habkost
2021-02-02  7:54           ` Paolo Bonzini
2021-02-02 15:25             ` Eduardo Habkost
2021-02-01 22:54 ` [PATCH 2/3] i386: Move asserts to separate x86_cpudef_validate() function Eduardo Habkost
2021-02-02 16:02   ` Philippe Mathieu-Daudé
2021-02-01 22:54 ` Eduardo Habkost [this message]

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=20210201225404.3941395-4-ehabkost@redhat.com \
    --to=ehabkost@redhat.com \
    --cc=babu.moger@amd.com \
    --cc=imammedo@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=richard.henderson@linaro.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).