linux-kernel - Re: [PATCH] mm/page_alloc: Make alloc_gigantic

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: <20191016085123.GO317@dhcp22.suse.cz>
Date:   Wed, 16 Oct 2019 10:51:23 +0200
From:   Michal Hocko <mhocko@...nel.org>
To:     David Hildenbrand <david@...hat.com>
Cc:     Anshuman Khandual <anshuman.khandual@....com>, linux-mm@...ck.org,
        Mike Kravetz <mike.kravetz@...cle.com>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Vlastimil Babka <vbabka@...e.cz>,
        David Rientjes <rientjes@...gle.com>,
        Andrea Arcangeli <aarcange@...hat.com>,
        Oscar Salvador <osalvador@...e.de>,
        Mel Gorman <mgorman@...hsingularity.net>,
        Mike Rapoport <rppt@...ux.ibm.com>,
        Dan Williams <dan.j.williams@...el.com>,
        Pavel Tatashin <pavel.tatashin@...rosoft.com>,
        Matthew Wilcox <willy@...radead.org>,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH] mm/page_alloc: Make alloc_gigantic_page() available for
 general use

On Wed 16-10-19 10:08:21, David Hildenbrand wrote:
> On 16.10.19 09:34, Anshuman Khandual wrote:
[...]
> > +static bool pfn_range_valid_contig(struct zone *z, unsigned long start_pfn,
> > +				   unsigned long nr_pages)
> > +{
> > +	unsigned long i, end_pfn = start_pfn + nr_pages;
> > +	struct page *page;
> > +
> > +	for (i = start_pfn; i < end_pfn; i++) {
> > +		page = pfn_to_online_page(i);
> > +		if (!page)
> > +			return false;
> > +
> > +		if (page_zone(page) != z)
> > +			return false;
> > +
> > +		if (PageReserved(page))
> > +			return false;
> > +
> > +		if (page_count(page) > 0)
> > +			return false;
> > +
> > +		if (PageHuge(page))
> > +			return false;
> > +	}
> 
> We might still try to allocate a lot of ranges that contain unmovable data
> (we could avoid isolating a lot of page blocks in the first place). I'd love
> to see something like pfn_range_movable() (similar, but different to
> is_mem_section_removable(), which uses has_unmovable_pages()).

Just to make sure I understand. Do you want has_unmovable_pages to be
called inside pfn_range_valid_contig?
[...]
> > +struct page *alloc_contig_pages(unsigned long nr_pages, gfp_t gfp_mask,
> > +				int nid, nodemask_t *nodemask)
> > +{
> > +	unsigned long ret, pfn, flags;
> > +	struct zonelist *zonelist;
> > +	struct zone *zone;
> > +	struct zoneref *z;
> > +
> > +	zonelist = node_zonelist(nid, gfp_mask);
> > +	for_each_zone_zonelist_nodemask(zone, z, zonelist,
> > +					gfp_zone(gfp_mask), nodemask) {
> 
> One important part is to never use the MOVABLE zone here (otherwise
> unmovable data would end up on the movable zone). But I guess the caller is
> responsible for that (not pass GFP_MOVABLE) like gigantic pages do.

Well, if the caller uses GFP_MOVABLE then the movability should be
implemented in some form. If that is not the case then it is a bug on
the caller behalf.

> > +		spin_lock_irqsave(&zone->lock, flags);
> > +
> > +		pfn = ALIGN(zone->zone_start_pfn, nr_pages);
> 
> This alignment does not make too much sense when allowing passing in !power
> of two orders. Maybe the caller should specify the requested alignment
> instead? Or should we enforce this to be aligned to make our life easier for
> now?

Are there any usecases that would require than the page alignment?
-- 
Michal Hocko
SUSE Labs