linux-kernel - Re: [PATCH V2] mm/hugetlb: Make __NR_USED

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: <3733d1e8-6e9d-4337-aeda-a6d5366d4a4c@arm.com>
Date: Tue, 3 Dec 2024 13:16:39 +0530
From: Anshuman Khandual <anshuman.khandual@....com>
To: Muchun Song <muchun.song@...ux.dev>
Cc: linux-mm@...ck.org, Andrew Morton <akpm@...ux-foundation.org>,
 Oscar Salvador <osalvador@...e.de>, linux-kernel@...r.kernel.org
Subject: Re: [PATCH V2] mm/hugetlb: Make __NR_USED_SUBPAGE check conditional


On 12/3/24 08:59, Muchun Song wrote:
> 
> 
>> On Dec 3, 2024, at 10:32, Anshuman Khandual <anshuman.khandual@....com> wrote:
>>
>> The HugeTLB order check against __NR_USED_SUBPAGE is required only when
>> HUGETLB_PAGE_OPTIMIZE_VMEMMAP is enabled. Hence BUG_ON() trigger should
>> happen only when applicable.
>>
>> Cc: Muchun Song <muchun.song@...ux.dev>
>> Cc: Andrew Morton <akpm@...ux-foundation.org>
>> Cc: Oscar Salvador <osalvador@...e.de>
>> Cc: linux-mm@...ck.org
>> Cc: linux-kernel@...r.kernel.org
>> Signed-off-by: Anshuman Khandual <anshuman.khandual@....com>
>> ---
>> This patch applies on v6.13-rc1
>>
>> Changes in V2:
>>
>> - Fixed #ifdef with CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP per Oscar
>>
>> Changes in V1:
>>
>> https://lore.kernel.org/all/20241202090728.78935-1-anshuman.khandual@arm.com/
>>
>> mm/hugetlb.c | 6 ++++--
>> 1 file changed, 4 insertions(+), 2 deletions(-)
>>
>> diff --git a/mm/hugetlb.c b/mm/hugetlb.c
>> index ea2ed8e301ef..e6a5b21e3578 100644
>> --- a/mm/hugetlb.c
>> +++ b/mm/hugetlb.c
>> @@ -4513,11 +4513,13 @@ void __init hugetlb_add_hstate(unsigned int order)
>> 	struct hstate *h;
>> 	unsigned long i;
>>
>> - 	if (size_to_hstate(PAGE_SIZE << order)) {
>> + 	if (size_to_hstate(PAGE_SIZE << order))
>> 		return;
>> - 	}
>> +
>> 	BUG_ON(hugetlb_max_hstate >= HUGE_MAX_HSTATE);
>> +#ifdef CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP
>> 	BUG_ON(order < order_base_2(__NR_USED_SUBPAGE));
> 
> Hi Anshuman,
> 
> __NR_USED_SUBPAGE indicates how many struct pages are used to store
> extra metadata for a HugeTLB page. So we need to make sure the order
> of a HugeTLB page should be bigger than or equal to order_base_2(__NR_USED_SUBPAGE).
> So It is not related to HVO. I don't think the changes make sense.

I did think about that but order_base_2(__NR_USED_SUBPAGE) actually
turns out to be just 2 as __NR_USED_SUBPAGE is 3. Would not HugeTLB
size be always greater than four base pages in reality, thus making
this BUG_ON() check just redundant ?

> 
> Thanks.
> 
>> +#endif
>> 	h = &hstates[hugetlb_max_hstate++];
>> 	__mutex_init(&h->resize_lock, "resize mutex", &h->resize_key);
>> 	h->order = order;
>> -- 
>> 2.30.2
>>
>