lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Date:   Thu, 23 Nov 2017 14:13:31 +0100
From:   Borislav Petkov <bp@...en8.de>
To:     qemu-devel@...gnu.org
Cc:     x86-ml <x86@...nel.org>, lkml <linux-kernel@...r.kernel.org>
Subject: qemu x86 CPUID leafs override

Hi guys,

I'm using the hack below to do some quick kernel testing by setting
arbitrary feature bits and then make it execute the code for that
feature.

For example, boot with:

-cpu EPYC,cpuid-leaf=0x80000007,ebx=0xf

to set some RAS feature bits and test newer RAS code.

Would something like that be of interest to a wider audience?

It is rough and ugly but if deemed useful, I could try to clean it up.

Thx.

---
diff --git a/target/i386/cpu.c b/target/i386/cpu.c
index 045d66191f28..249fb23be696 100644
--- a/target/i386/cpu.c
+++ b/target/i386/cpu.c
@@ -2732,6 +2732,13 @@ void cpu_x86_cpuid(CPUX86State *env, uint32_t index, uint32_t count,
     uint32_t limit;
     uint32_t signature[3];
 
+    /*
+     * Pull up max xlevel in case the one we've specified on the cmdline is
+     * higher.
+     */
+    if (cpu->cpuid_leaf && env->cpuid_xlevel < cpu->cpuid_leaf)
+	    env->cpuid_xlevel = cpu->cpuid_leaf;
+
     /* Calculate & apply limits for different index ranges */
     if (index >= 0xC0000000) {
         limit = env->cpuid_xlevel2;
@@ -3140,6 +3147,22 @@ void cpu_x86_cpuid(CPUX86State *env, uint32_t index, uint32_t count,
         *edx = 0;
         break;
     }
+
+    /* Do CPUID overrides: */
+    if (cpu->cpuid_leaf && cpu->cpuid_leaf == index) {
+
+	    if (cpu->eax)
+		    *eax = cpu->eax;
+
+	    if (cpu->ebx)
+		    *ebx = cpu->ebx;
+
+	    if (cpu->ecx)
+		    *ecx = cpu->ecx;
+
+	    if (cpu->edx)
+		    *edx = cpu->edx;
+    }
 }
 
 /* CPUClass::reset() */
@@ -4173,6 +4196,11 @@ static Property x86_cpu_properties[] = {
      * to the specific Windows version being used."
      */
     DEFINE_PROP_INT32("x-hv-max-vps", X86CPU, hv_max_vps, -1),
+    DEFINE_PROP_UINT32("cpuid-leaf", X86CPU, cpuid_leaf, UINT32_MAX),
+    DEFINE_PROP_UINT32("eax", X86CPU, eax, UINT32_MAX),
+    DEFINE_PROP_UINT32("ebx", X86CPU, ebx, UINT32_MAX),
+    DEFINE_PROP_UINT32("ecx", X86CPU, ecx, UINT32_MAX),
+    DEFINE_PROP_UINT32("edx", X86CPU, edx, UINT32_MAX),
     DEFINE_PROP_END_OF_LIST()
 };
 
diff --git a/target/i386/cpu.h b/target/i386/cpu.h
index b086b1528b89..b336b0849456 100644
--- a/target/i386/cpu.h
+++ b/target/i386/cpu.h
@@ -1284,6 +1284,13 @@ struct X86CPU {
     int32_t thread_id;
 
     int32_t hv_max_vps;
+
+    /*
+     * CPUID overrides:
+     */
+    uint32_t cpuid_leaf;
+    uint32_t eax, ebx, ecx, edx;
+
 };
 
 static inline X86CPU *x86_env_get_cpu(CPUX86State *env)

-- 
Regards/Gruss,
    Boris.

Good mailing practices for 400: avoid top-posting and trim the reply.

Powered by blists - more mailing lists