Re: [PATCH 2/2] x86/MCE: Add command line option to extend MCE Records pool

From: Sohil Mehta <sohil.mehta@intel.com>
To: "Luck, Tony" <tony.luck@intel.com>,
	"Naik, Avadhut" <avadnaik@amd.com>,
	"x86@kernel.org" <x86@kernel.org>,
	"linux-edac@vger.kernel.org" <linux-edac@vger.kernel.org>
Cc: "bp@alien8.de" <bp@alien8.de>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	"yazen.ghannam@amd.com" <yazen.ghannam@amd.com>,
	Avadhut Naik <avadhut.naik@amd.com>
Subject: Re: [PATCH 2/2] x86/MCE: Add command line option to extend MCE Records pool
Date: Fri, 9 Feb 2024 13:02:30 -0800	[thread overview]
Message-ID: <4850ef1f-db20-4fb2-8658-51f27330c343@intel.com> (raw)
In-Reply-To: <SJ1PR11MB6083D8390032B7072443F83CFC4B2@SJ1PR11MB6083.namprd11.prod.outlook.com>

On 2/9/2024 12:28 PM, Luck, Tony wrote:
>> How about being more conservative with the allocations in the previous
>> patch so that we don't need to introduce this additional mechanism right
>> now? Later, if there is really a need for some specific usage, the patch
>> can be re-submitted then with the supporting data.
> 
> There used to be a rule-of-thumb when configuring systems to have at least
> one GByte of memory per CPU. Anyone following that rule shouldn't be
> worried about sub-kilobyte allocations per CPU.
> 

I meant, to avoid the need for this second patch we can always start
lower and increase it later.

256 bytes per cpu seems fine to me as done in the previous patch. But,
if that seems too high as described by Avadhut below then maybe we can
start with 200 bytes or any other number. It's just heuristic IIUC.

https://lore.kernel.org/lkml/8d2d0dac-b188-4826-a43a-bb5fc0528f0d@amd.com/

Sohil