linux-kernel - Re: [PATCH] x86/mce: Schedule mce_setup() on correct CPU for CPER decoding

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: <c9f885df-4650-248c-3708-9250a0b530dc@amd.com>
Date:   Thu, 15 Jun 2023 11:34:21 -0400
From:   Yazen Ghannam <yazen.ghannam@....com>
To:     Borislav Petkov <bp@...en8.de>
Cc:     yazen.ghannam@....com, linux-edac@...r.kernel.org,
        linux-kernel@...r.kernel.org, tony.luck@...el.com, x86@...nel.org
Subject: Re: [PATCH] x86/mce: Schedule mce_setup() on correct CPU for CPER
 decoding

On 6/15/2023 11:20 AM, Borislav Petkov wrote:
> On Mon, Apr 17, 2023 at 04:20:06PM +0000, Yazen Ghannam wrote:
>> @@ -97,20 +102,13 @@ int apei_smca_report_x86_error(struct cper_ia_proc_ctx *ctx_info, u64 lapic_id)
>>   	if (ctx_info->reg_arr_size < 48)
>>   		return -EINVAL;
>>   
>> -	mce_setup(&m);
>> -
>> -	m.extcpu = -1;
>> -	m.socketid = -1;
>> -
>> -	for_each_possible_cpu(cpu) {
>> -		if (cpu_data(cpu).initial_apicid == lapic_id) {
>> -			m.extcpu = cpu;
>> -			m.socketid = cpu_data(m.extcpu).phys_proc_id;
>> +	for_each_possible_cpu(cpu)
>> +		if (cpu_data(cpu).initial_apicid == lapic_id)
>>   			break;
>> -		}
>> -	}
>>   
>> -	m.apicid = lapic_id;
>> +	if (smp_call_function_single(cpu, __mce_setup, &m, 1))
> 
> I can see the following call-chain from NMI context which is a no-no:
> 
> ghes_notify_nmi
> |-> ghes_in_nmi_spool_from_list
>      |-> ghes_in_nmi_queue_one_entry
>          |-> __ghes_panic
> 	    |-> __ghes_print_estatus
> 	        |-> cper_estatus_print
> 		    |-> cper_estatus_print_section
> 		        |-> cper_print_proc_ia
> 			    |-> arch_apei_report_x86_error
> 			        |-> apei_smca_report_x86_error
> 				    |-> smp_call_function_single
> 
> 

Right, but in practice SMCA errors are not reported through GHES at 
runtime. They will only come in through BERT at boot time. There aren't 
any plans to change this, so the NMI issue won't be encountered.

I can include this info in the commit message and/or code comments. Is 
this okay?

We can solve the NMI issue if it ever comes up in the future. Unless 
there's an obvious change to avoid this now. Any suggestions?

Thanks,
Yazen