linux-kernel - RE: [PATCH 2/6] x86/tdx: Retry TDVMCALL_MAP

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: <SA1PR21MB13352BBC75217A9D019351DEBF139@SA1PR21MB1335.namprd21.prod.outlook.com>
Date:   Mon, 28 Nov 2022 00:07:27 +0000
From:   Dexuan Cui <decui@...rosoft.com>
To:     "Michael Kelley (LINUX)" <mikelley@...rosoft.com>,
        "Kirill A. Shutemov" <kirill@...temov.name>
CC:     "ak@...ux.intel.com" <ak@...ux.intel.com>,
        "arnd@...db.de" <arnd@...db.de>, "bp@...en8.de" <bp@...en8.de>,
        "brijesh.singh@....com" <brijesh.singh@....com>,
        "Williams, Dan J" <dan.j.williams@...el.com>,
        "dave.hansen@...ux.intel.com" <dave.hansen@...ux.intel.com>,
        Haiyang Zhang <haiyangz@...rosoft.com>,
        "hpa@...or.com" <hpa@...or.com>,
        "jane.chu@...cle.com" <jane.chu@...cle.com>,
        "kirill.shutemov@...ux.intel.com" <kirill.shutemov@...ux.intel.com>,
        KY Srinivasan <kys@...rosoft.com>,
        "linux-arch@...r.kernel.org" <linux-arch@...r.kernel.org>,
        "linux-hyperv@...r.kernel.org" <linux-hyperv@...r.kernel.org>,
        "luto@...nel.org" <luto@...nel.org>,
        "mingo@...hat.com" <mingo@...hat.com>,
        "peterz@...radead.org" <peterz@...radead.org>,
        "rostedt@...dmis.org" <rostedt@...dmis.org>,
        "sathyanarayanan.kuppuswamy@...ux.intel.com" 
        <sathyanarayanan.kuppuswamy@...ux.intel.com>,
        "seanjc@...gle.com" <seanjc@...gle.com>,
        "tglx@...utronix.de" <tglx@...utronix.de>,
        "tony.luck@...el.com" <tony.luck@...el.com>,
        "wei.liu@...nel.org" <wei.liu@...nel.org>,
        "x86@...nel.org" <x86@...nel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: RE: [PATCH 2/6] x86/tdx: Retry TDVMCALL_MAP_GPA() when needed

> From: Michael Kelley (LINUX) <mikelley@...rosoft.com>
> Sent: Wednesday, November 23, 2022 5:30 AM
> > [...]
> > static bool tdx_map_gpa(phys_addr_t start, phys_addr_t end, bool enc)
> > {
> >         int max_retry_cnt = 1000, retry_cnt = 0;
> >         struct tdx_hypercall_args args;
> >         u64 map_fail_paddr, ret;
> >
> >         while (1) {
> >                 args.r10 = TDX_HYPERCALL_STANDARD;
> >                 args.r11 = TDVMCALL_MAP_GPA;
> >                 args.r12 = start;
> >                 args.r13 = end - start;
> >                 args.r14 = 0;
> >                 args.r15 = 0;
> >
> >                 ret = __tdx_hypercall(&args, TDX_HCALL_HAS_OUTPUT);
> >                 if (!ret)
> >                         break;
> 
> The above test is redundant and can be removed.  The "success" case is
> implicitly handled by the test below for != TDVMCALL_STATUS_RETRY.

Good point. Will remove the redundant test.

> >                 if (ret != TDVMCALL_STATUS_RETRY)
> >                         break;
> >                 /*
> >                  * The guest must retry the operation for the pages in
> the
> >                  * region starting at the GPA specified in R11. Make sure
> R11
> >                  * contains a sane value.
> >                  */
> >                 map_fail_paddr = args.r11 ;
> >                 if (map_fail_paddr < start || map_fail_paddr >= end)
> >                         return false;
> >
> >                 if (map_fail_paddr == start) {
> >                         retry_cnt++;
> >                         if (retry_cnt > max_retry_cnt)
> >                                 return false;
> >                 } else {
> >                         retry_cnt = 0;;
> >                         start = map_fail_paddr;
> 
> Just summarizing the code, we increment the retry count if the hypercall
> returns STATUS_RETRY but did nothing (i.e., map_fail_paddr == start).  But
> if the hypercall returns STATUS_RETRY after making at least some progress,
> then we reset the retry count.   So in the worst case, for example, if the
> hypercall processed only one page on each invocation, the loop will continue
> until completion, without hitting any retry limits.  That scenario seems
> plausible and within the spec.

Exactly.

> Do we have any indication about the likelihood of the "RETRY but did
> nothing" case? The spec doesn't appear to disallow this case, but does
> Hyper-V actually do this?  It seems like a weird case.
> 
> Michael

Yes, Hyper-V does do this, according to my test. It looks like this is not
because the operation is too time-consuming -- it looks like there is some
Hyper-V specific activity going on.