| 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
| ||
|
Message-ID: <6D760A9C-E821-4D0D-8E9B-B6B33FD55ACC@nvidia.com>
Date: Tue, 27 May 2025 10:56:20 -0400
From: Zi Yan <ziy@...dia.com>
To: Vlastimil Babka <vbabka@...e.cz>
Cc: David Hildenbrand <david@...hat.com>,
Johannes Weiner <hannes@...xchg.org>, linux-mm@...ck.org,
Andrew Morton <akpm@...ux-foundation.org>,
Oscar Salvador <osalvador@...e.de>,
Baolin Wang <baolin.wang@...ux.alibaba.com>,
"Kirill A . Shutemov" <kirill.shutemov@...ux.intel.com>,
Mel Gorman <mgorman@...hsingularity.net>,
Suren Baghdasaryan <surenb@...gle.com>, Michal Hocko <mhocko@...e.com>,
Brendan Jackman <jackmanb@...gle.com>, Richard Chang <richardycc@...gle.com>,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH v5 2/6] mm/page_isolation: make page isolation a
standalone bit.
On 27 May 2025, at 6:11, Vlastimil Babka wrote:
> On 5/23/25 21:12, Zi Yan wrote:
>> During page isolation, the original migratetype is overwritten, since
>> MIGRATE_* are enums and stored in pageblock bitmaps. Change
>> MIGRATE_ISOLATE to be stored a standalone bit, PB_migrate_isolate, like
>> PB_migrate_skip, so that migratetype is not lost during pageblock
>> isolation.
>>
>> Signed-off-by: Zi Yan <ziy@...dia.com>
>
> <snip>
>
>> #define MEMORY_OFFLINE 0x1
>> diff --git a/include/linux/pageblock-flags.h b/include/linux/pageblock-flags.h
>> index 3acbb271a29a..f2f8540b95ca 100644
>> --- a/include/linux/pageblock-flags.h
>> +++ b/include/linux/pageblock-flags.h
>> @@ -20,7 +20,13 @@ enum pageblock_bits {
>> PB_migrate_end = PB_migrate + PB_migratetype_bits - 1,
>> /* 3 bits required for migrate types */
>> PB_migrate_skip,/* If set the block is skipped by compaction */
>> -
>> +#ifdef CONFIG_MEMORY_ISOLATION
>> + /*
>> + * Pageblock isolation is represented with a separate bit, so that
>> + * the migratetype of a block is not overwritten by isolation.
>> + */
>> + PB_migrate_isolate, /* If set the block is isolated */
>> +#endif
>> /*
>> * Assume the bits will always align on a word. If this assumption
>> * changes then get/set pageblock needs updating.
>> @@ -32,6 +38,11 @@ enum pageblock_bits {
>>
>> #define MIGRATETYPE_MASK ((1UL << (PB_migrate_end + 1)) - 1)
>>
>> +#ifdef CONFIG_MEMORY_ISOLATION
>> +#define MIGRATETYPE_AND_ISO_MASK \
>> + (((1UL << (PB_migrate_end + 1)) - 1) | BIT(PB_migrate_isolate))
>> +#endif
>
> I think if there was:
>
> #else
> #define MIGRATETYPE_AND_ISO_MASK MIGRATETYPE_MASK
> #endif
>
> you could avoid some #ifdef code later.
Sure. Will do.
>
>> #if defined(CONFIG_HUGETLB_PAGE)
>>
>> #ifdef CONFIG_HUGETLB_PAGE_SIZE_VARIABLE
>> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
>> index 0207164fcaf6..b2c623699461 100644
>> --- a/mm/page_alloc.c
>> +++ b/mm/page_alloc.c
>> @@ -360,8 +360,14 @@ get_pfnblock_bitmap_bitidx(const struct page *page, unsigned long pfn,
>> unsigned long *bitmap;
>> unsigned long word_bitidx;
>>
>> +#ifdef CONFIG_MEMORY_ISOLATION
>> + BUILD_BUG_ON(NR_PAGEBLOCK_BITS != 8);
>> + /* extra one for MIGRATE_ISOLATE */
>> + BUILD_BUG_ON(MIGRATE_TYPES > (1 << PB_migratetype_bits) + 1);
>
> This implicitly assumes MIGRATE_ISOLATE is the last of migratetypes so we
> can actually need less PB_migratetype_bits if we stop encoding it within
> them anymore, but there's nothing enforcing that (not even as a comment)?’
You point is valid. How about adding __MIGRATE_TYPES right before
MIGRATE_ISOLATE in enum migratetype and using
BUILD_BUG_ON(__MIGRATE_TYPES > (1 << PB_migratetype_bits));
here outside of #ifdef.
The next step of cleanup, as discussed with David, is to remove
MIGRATE_ISOLATE, so that all MIGRATE_TYPES are stored in
PB_migratetype_bits. And {get,set}_pfnblock_migratetype()
gives
struct pageblock_info {
enum migratetype migratetype;
bool isolated;
};
>
>> +#else
>> BUILD_BUG_ON(NR_PAGEBLOCK_BITS != 4);
>> BUILD_BUG_ON(MIGRATE_TYPES > (1 << PB_migratetype_bits));
>> +#endif
>> VM_BUG_ON_PAGE(!zone_spans_pfn(page_zone(page), pfn), page);
>>
>> bitmap = get_pageblock_bitmap(page, pfn);
>> @@ -435,7 +441,20 @@ bool get_pfnblock_bit(const struct page *page, unsigned long pfn,
>> __always_inline enum migratetype
>> get_pfnblock_migratetype(const struct page *page, unsigned long pfn)
>> {
>> - return __get_pfnblock_flags_mask(page, pfn, MIGRATETYPE_MASK);
>> + unsigned long mask = MIGRATETYPE_MASK;
>
> E.g. with my suggestion above you could use MIGRATETYPE_AND_ISO_MASK here.
>
>> + unsigned long flags;
>> +
>> +#ifdef CONFIG_MEMORY_ISOLATION
>> + mask = MIGRATETYPE_AND_ISO_MASK;
>> +#endif
>
> And drop this.
Sure. Will do.
>
>> + flags = __get_pfnblock_flags_mask(page, pfn, mask);
>> +
>> +#ifdef CONFIG_MEMORY_ISOLATION
>> + if (flags & BIT(PB_migrate_isolate))
>> + return MIGRATE_ISOLATE;
>> +#endif
>> + return flags & MIGRATETYPE_MASK;
>> }
>>
>> /**
>> @@ -513,12 +532,22 @@ void clear_pfnblock_bit(const struct page *page, unsigned long pfn,
>> __always_inline void set_pageblock_migratetype(struct page *page,
>> enum migratetype migratetype)
>> {
>> + unsigned long mask = MIGRATETYPE_MASK;
>> +
>> if (unlikely(page_group_by_mobility_disabled &&
>> migratetype < MIGRATE_PCPTYPES))
>> migratetype = MIGRATE_UNMOVABLE;
>>
>> +#ifdef CONFIG_MEMORY_ISOLATION
>> + if (migratetype == MIGRATE_ISOLATE) {
>> + set_pfnblock_bit(page, page_to_pfn(page), PB_migrate_isolate);
>> + return;
>> + }
>> + /* change mask to clear PB_migrate_isolate if it is set */
>> + mask = MIGRATETYPE_AND_ISO_MASK;
>> +#endif
>> __set_pfnblock_flags_mask(page, page_to_pfn(page),
>> - (unsigned long)migratetype, MIGRATETYPE_MASK);
>> + (unsigned long)migratetype, mask);
>
> This could just pass MIGRATETYPE_AND_ISO_MASK here.
Yep.
>
>> }
>>
>> #ifdef CONFIG_DEBUG_VM
Best Regards,
Yan, Zi
Powered by blists - more mailing lists