linux-kernel - Re: [PATCH 5/8] x86/mmu: Add mm-based PASID refcounting

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 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <YVSw/6BAFvh9C+ct@agluck-desk2.amr.corp.intel.com>
Date:   Wed, 29 Sep 2021 11:31:27 -0700
From:   "Luck, Tony" <tony.luck@...el.com>
To:     Fenghua Yu <fenghua.yu@...el.com>
Cc:     Thomas Gleixner <tglx@...utronix.de>,
        Andy Lutomirski <luto@...nel.org>,
        Peter Zijlstra <peterz@...radead.org>,
        Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
        Dave Hansen <dave.hansen@...el.com>,
        Lu Baolu <baolu.lu@...ux.intel.com>,
        Joerg Roedel <joro@...tes.org>,
        Josh Poimboeuf <jpoimboe@...hat.com>,
        Dave Jiang <dave.jiang@...el.com>,
        Jacob Jun Pan <jacob.jun.pan@...el.com>,
        Raj Ashok <ashok.raj@...el.com>,
        "Shankar, Ravi V" <ravi.v.shankar@...el.com>,
        iommu@...ts.linux-foundation.org,
        the arch/x86 maintainers <x86@...nel.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 5/8] x86/mmu: Add mm-based PASID refcounting

On Wed, Sep 29, 2021 at 06:07:22PM +0000, Fenghua Yu wrote:
> Add
> +#ifdef CONFIG_IOMMU_SUPPORT
> because mm->pasid and current-pasid_activated are defined only if
> CONFIG_IOMMU_SUPPORT is defined.
> 
> > +	if (user_mode(regs) && current->mm->pasid && !current->pasid_activated) {
> 
> Maybe need to add "&& cpu_feature_enabled(X86_FEATURE_ENQCMD)" because
> the IA32_PASID MSR is only used when ENQCMD is enabled?
> 
> > +		current->pasid_activated = 1;
> > +		wrmsrl(MSR_IA32_PASID, current->mm->pasid | MSR_IA32_PASID_VALID);
> > +		return;
> > +	}
> > +
> 
> +endif

New version that addresses those issues.  Has ugly #ifdef in C
code :-(  If that's unacceptable, then could create some stub
functions, or add a call to __try_fixup_pasid() that's in a
file in the iommu code that is only built when CONFIG_IOMMU_SUPPORT
is set.  But either of those move the details far away from the
#GP handler so make extra work for anyone trying to follow along
with what is happening here.

-Tony

---

diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c
index a58800973aed..5a3c87fd65de 100644
--- a/arch/x86/kernel/traps.c
+++ b/arch/x86/kernel/traps.c
@@ -528,6 +528,32 @@ static enum kernel_gp_hint get_kernel_gp_address(struct pt_regs *regs,
 
 #define GPFSTR "general protection fault"
 
+/*
+ * When a user executes the ENQCMD instruction it will #GP
+ * fault if the IA32_PASID MSR has not been set up with a
+ * valid PASID.
+ * So if the process has been allocated a PASID (mm->pasid)
+ * AND the IA32_PASID MSR has not been initialized, try to
+ * fix this #GP by initializing the IA32_PASID MSR.
+ * If the #GP was for some other reason, it will trigger
+ * again, but this routine will return false and the #GP
+ * will be processed.
+ */
+static void try_fixup_pasid(void)
+{
+	if (!cpu_feature_enabled(X86_FEATURE_ENQCMD))
+		return false;
+
+#ifdef CONFIG_IOMMU_SUPPORT
+	if (current->mm->pasid && !current->pasid_activated) {
+		current->pasid_activated = 1;
+		wrmsrl(MSR_IA32_PASID, current->mm->pasid);
+		return true;
+	}
+#endif
+	return false;
+}
+
 DEFINE_IDTENTRY_ERRORCODE(exc_general_protection)
 {
 	char desc[sizeof(GPFSTR) + 50 + 2*sizeof(unsigned long) + 1] = GPFSTR;
@@ -536,6 +562,9 @@ DEFINE_IDTENTRY_ERRORCODE(exc_general_protection)
 	unsigned long gp_addr;
 	int ret;
 
+	if (user_mode(regs) && try_fixup_pasid())
+		return;
+
 	cond_local_irq_enable(regs);
 
 	if (static_cpu_has(X86_FEATURE_UMIP)) {