linux-kernel - Re: [RFC PATCH] x86/fpu/xstate: Add more diagnostic information on inconsistent xstate sizes

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: <388b09ea-7eb4-2878-5292-97475fde8a5a@intel.com>
Date:   Tue, 11 Apr 2023 21:38:50 -0700
From:   "Chang S. Bae" <chang.seok.bae@...el.com>
To:     Fenghua Yu <fenghua.yu@...el.com>,
        Dave Hansen <dave.hansen@...ux.intel.com>,
        Thomas Gleixner <tglx@...utronix.de>,
        Borislav Petkov <bp@...en8.de>, Ingo Molnar <mingo@...hat.com>
CC:     linux-kernel <linux-kernel@...r.kernel.org>, x86 <x86@...nel.org>,
        Chintan M Patel <chintan.m.patel@...el.com>,
        Thiago Macieira <thiago.macieira@...el.com>
Subject: Re: [RFC PATCH] x86/fpu/xstate: Add more diagnostic information on
 inconsistent xstate sizes

On 4/11/2023 6:21 PM, Fenghua Yu wrote:
> 
> First of all, max_features is shown already.

Yes.

> Kernel_size from CPUID.0xd.0x1:EBX takes XCR0 | IA32_XSS as input. 
> Platform may take wrong XCR0 or IA32_XSS and get wrong kernel_size. The  > purpose of this patch is to provide more debug info to help debug
> platform/kernel issue. So instead of a whole max_features, xgetbv() to 
> get XCR0 and xfeatures_mask_supervisor() to get IA32_XSS provides more 
> debug info in case platform may have issue in XCR0 or IA32_XSS.
> 
> In other words, splitting max_features into XCR0 and IA32_XSS and 
> showing them individually provide more useful debug info than one single 
> max_features value. >
> Does it make sense?

Hmm, I don't get it. I don't think whether the microcode takes those 
register values wrong or miscalculates the size does matter here.

print_xstate_offset_size() or something can decode the mask and readily 
shows off how it was calculated here. Then, probably that's it.

>> I still expect some acknowledgment of what is coded here for the 
>> kernel calculation details.
> 
> The kernel calculation is shown in
> +        print_xstate_offset_size();
> +        pr_info("x86/fpu: total size: %u bytes\n", size);
> 
> Isn't that detailed enough to show offset and size of each xstate and 
> sum of sizes?
> 
> After that,
> +    pr_info("x86/fpu: kernel_size from CPUID.0xd.0x%x:EBX: %u bytes\n",
> +               compacted ? 1 : 0, kernel_size);
> shows how kernel_size is calculated from CPUID?
> 
> Using the above debug info, a real platform CPUID issue is shown clearly.
> 
> What other details are needed?

I recall it was also asked to show which features are off or mismatched 
as compared to the CPU calculation. I'm not so sure about it.

Thanks,
Chang