[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <17c958bc-d421-f48b-a07a-bf22afc32d33@redhat.com>
Date: Thu, 4 Mar 2021 18:21:54 +0100
From: David Hildenbrand <david@...hat.com>
To: Mike Kravetz <mike.kravetz@...cle.com>,
Chen Wandun <chenwandun@...wei.com>, akpm@...ux-foundation.org,
linux-mm@...ck.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] mm/hugetlb: suppress wrong warning info when alloc
gigantic page
On 04.03.21 18:20, Mike Kravetz wrote:
> On 3/4/21 1:35 AM, David Hildenbrand wrote:
>> On 19.02.21 20:14, Mike Kravetz wrote:
>>> On 2/19/21 4:39 AM, Chen Wandun wrote:
>>>> If hugetlb_cma is enabled, it will skip boot time allocation
>>>> when allocating gigantic page, that doesn't means allocation
>>>> failure, so suppress this warning info.
>>>>
>>>
>>> Normally the addition of warning messages is discouraged. However, in
>>> this case the additional message provides value. Why?
>>>
>>> Prior to the commit cf11e85fc08c, one could have a kernel command line
>>> that contains:
>>>
>>> hugepagesz=1G hugepages=16
>>>
>>> This would allocate 16 1G pages at boot time.
>>>
>>> After the commit, someone could specify a command line containing:
>>>
>>> hugepagesz=1G hugepages=16 hugetlb_cma=16G
>>>
>>> In this case, 16G of CMA will be reserved for 1G huge page allocations
>>> after boot time. The parameter 'hugepages=16' is ignored, and the warning
>>> message is logged. The warning message should only be logged when the
>>> kernel parameter 'hugepages=' is ignored.
>>>
>>> IMO, it make sense to log a warning if ignoring a user specified parameter.
>>> The user should not be attempting boot time allocation and CMA reservation
>>> for 1G pages.
>>>
>>> I do not think we should drop the warning as the it tells the user thay
>>> have specified two incompatible allocation options.
>>>
>>
>> I agree. It has value.
>>
>
> Hi David,
>
> Sorry my above reply was too quick as I did not take a close look at
> the code/patch. See,
>
> https://lore.kernel.org/mm-commits/YDAbeDsG7GhV6s6B@carbon.dhcp.thefacebook.com/
>
> This patch is actually in Andrew's tree.
>
Oh, I missed that discussion - thanks for the pointer!
--
Thanks,
David / dhildenb
Powered by blists - more mailing lists