linux-kernel - Re: [PATCH 3/9] mm/hugetlb_cgroup: convert set_hugetlb

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] [day] [month] [year] [list]

Date:   Tue, 1 Nov 2022 09:43:11 -0700
From:   Sidhartha Kumar <sidhartha.kumar@...cle.com>
To:     Mike Kravetz <mike.kravetz@...cle.com>
Cc:     linux-kernel@...r.kernel.org, linux-mm@...ck.org,
        akpm@...ux-foundation.org, songmuchun@...edance.com,
        willy@...radead.org, almasrymina@...gle.com, linmiaohe@...wei.com,
        minhquangbui99@...il.com, aneesh.kumar@...ux.ibm.com
Subject: Re: [PATCH 3/9] mm/hugetlb_cgroup: convert set_hugetlb_cgroup*() to
 folios



On 10/31/22 9:38 AM, Mike Kravetz wrote:
> On 10/13/22 20:12, Sidhartha Kumar wrote:
>> Allows __prep_new_huge_page() to operate on a folio by converting
>> set_hugetlb_cgroup*() to take in a folio.
>>
>> Signed-off-by: Sidhartha Kumar <sidhartha.kumar@...cle.com>
>> --- a/mm/hugetlb.c
>> +++ b/mm/hugetlb.c
>> @@ -1758,19 +1758,21 @@ static void __prep_account_new_huge_page(struct hstate *h, int nid)
>>   	h->nr_huge_pages_node[nid]++;
>>   }
>>   
>> -static void __prep_new_huge_page(struct hstate *h, struct page *page)
>> +static void __prep_new_hugetlb_folio(struct hstate *h, struct folio *folio)
>>   {
>> -	hugetlb_vmemmap_optimize(h, page);
>> -	INIT_LIST_HEAD(&page->lru);
>> -	set_compound_page_dtor(page, HUGETLB_PAGE_DTOR);
>> -	hugetlb_set_page_subpool(page, NULL);
>> -	set_hugetlb_cgroup(page, NULL);
>> -	set_hugetlb_cgroup_rsvd(page, NULL);
>> +	hugetlb_vmemmap_optimize(h, &folio->page);
>> +	INIT_LIST_HEAD(&folio->lru);
>> +	folio->_folio_dtor = HUGETLB_PAGE_DTOR;
> Seems like we should have a routine 'set_folio_dtor' that has the same
> functionality as set_compound_page_dtor.  Here, we loose the check for a
> valid DTOR value (although not terribly valuable).

I agree with the need for a 'set_folio_dtor' routine, I'll send out a 
patch for that as well.

> Not required for this patch, but something to note.
>
>> +	hugetlb_set_folio_subpool(folio, NULL);
>> +	set_hugetlb_cgroup(folio, NULL);
>> +	set_hugetlb_cgroup_rsvd(folio, NULL);
>>   }
>>   
>>   static void prep_new_huge_page(struct hstate *h, struct page *page, int nid)
>>   {
>> -	__prep_new_huge_page(h, page);
>> +	struct folio *folio = page_folio(page);
>> +
>> +	__prep_new_hugetlb_folio(h, folio);
>>   	spin_lock_irq(&hugetlb_lock);
>>   	__prep_account_new_huge_page(h, nid);
>>   	spin_unlock_irq(&hugetlb_lock);
>> @@ -2731,8 +2733,10 @@ static int alloc_and_dissolve_huge_page(struct hstate *h, struct page *old_page,
>>   					struct list_head *list)
>>   {
>>   	gfp_t gfp_mask = htlb_alloc_mask(h) | __GFP_THISNODE;
>> -	int nid = page_to_nid(old_page);
>> +	struct folio *old_folio = page_folio(old_page);
>> +	int nid = folio_nid(old_folio);
>>   	struct page *new_page;
>> +	struct folio *new_folio;
>>   	int ret = 0;
>>   
>>   	/*
>> @@ -2745,16 +2749,17 @@ static int alloc_and_dissolve_huge_page(struct hstate *h, struct page *old_page,
>>   	new_page = alloc_buddy_huge_page(h, gfp_mask, nid, NULL, NULL);
>>   	if (!new_page)
>>   		return -ENOMEM;
>> -	__prep_new_huge_page(h, new_page);
>> +	new_folio = page_folio(new_page);
>> +	__prep_new_hugetlb_folio(h, new_folio);
>>   
>>   retry:
>>   	spin_lock_irq(&hugetlb_lock);
>> -	if (!PageHuge(old_page)) {
>> +	if (!folio_test_hugetlb(old_folio)) {
>>   		/*
>>   		 * Freed from under us. Drop new_page too.
>>   		 */
>>   		goto free_new;
>> -	} else if (page_count(old_page)) {
>> +	} else if (folio_ref_count(old_folio)) {
>>   		/*
>>   		 * Someone has grabbed the page, try to isolate it here.
>>   		 * Fail with -EBUSY if not possible.
>> @@ -2763,7 +2768,7 @@ static int alloc_and_dissolve_huge_page(struct hstate *h, struct page *old_page,
>>   		ret = isolate_hugetlb(old_page, list);
>>   		spin_lock_irq(&hugetlb_lock);
>>   		goto free_new;
>> -	} else if (!HPageFreed(old_page)) {
>> +	} else if (!folio_test_hugetlb(old_folio)) {
> Should that be?
> 	} else if (!folio_test_hugetlb_freed(old_folio)) {

Yes good catch, I will fix in a v2.

Thanks,
Sidhartha Kumar

>