From: David Hildenbrand <david@redhat.com>
To: Zi Yan <ziy@nvidia.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>,
Vlastimil Babka <vbabka@suse.cz>,
linux-mm@kvack.org, Andrew Morton <akpm@linux-foundation.org>,
Oscar Salvador <osalvador@suse.de>,
Baolin Wang <baolin.wang@linux.alibaba.com>,
"Kirill A . Shutemov" <kirill.shutemov@linux.intel.com>,
Mel Gorman <mgorman@techsingularity.net>,
Suren Baghdasaryan <surenb@google.com>,
Michal Hocko <mhocko@suse.com>,
Brendan Jackman <jackmanb@google.com>,
Richard Chang <richardycc@google.com>,
linux-kernel@vger.kernel.org
Subject: Re: [PATCH v6 6/6] mm/page_isolation: remove migratetype parameter from more functions.
Date: Fri, 30 May 2025 22:08:58 +0200 [thread overview]
Message-ID: <50da1626-5894-42ef-b3fd-935a331f14c4@redhat.com> (raw)
In-Reply-To: <8C1EFB3C-0F1C-46DE-878A-3546A86A141F@nvidia.com>
On 30.05.25 21:58, Zi Yan wrote:
> On 30 May 2025, at 15:56, David Hildenbrand wrote:
>
>> On 30.05.25 18:22, Zi Yan wrote:
>>> migratetype is no longer overwritten during pageblock isolation,
>>> start_isolate_page_range(), has_unmovable_pages(), and
>>> set_migratetype_isolate() no longer need which migratetype to restore
>>> during isolation failure.
>>>
>>> For has_unmoable_pages(), it needs to know if the isolation is for CMA
>>> allocation, so adding CMA_ALLOCATION to provide the information. At the
>>> same time change isolation flags to enum pb_isolate_mode
>>> (PB_ISOLATE_MODE_MEM_OFFLINE, PB_ISOLATE_MODE_CMA_ALLOC,
>>> PB_ISOLATE_MODE_OTHER). Remove REPORT_FAILURE and check
>>> MEMORY_OFFLINE instead, since only PB_ISOLATE_MODE_MEM_OFFLINE reports
>>> isolation failures.
>>>
>>> alloc_contig_range() no longer needs migratetype. Replace it with
>>> enum acr_flags_t to tell if an allocation is for CMA. So does
>>> __alloc_contig_migrate_range().
>>>
>>> Signed-off-by: Zi Yan <ziy@nvidia.com>
>>> ---
>>> drivers/virtio/virtio_mem.c | 2 +-
>>> include/linux/gfp.h | 9 ++++-
>>> include/linux/page-isolation.h | 20 ++++++++--
>>> include/trace/events/kmem.h | 14 ++++---
>>> mm/cma.c | 2 +-
>>> mm/memory_hotplug.c | 6 +--
>>> mm/page_alloc.c | 27 ++++++-------
>>> mm/page_isolation.c | 70 +++++++++++++++-------------------
>>> 8 files changed, 82 insertions(+), 68 deletions(-)
>>>
>>> diff --git a/drivers/virtio/virtio_mem.c b/drivers/virtio/virtio_mem.c
>>> index 56d0dbe62163..6bce70b139b2 100644
>>> --- a/drivers/virtio/virtio_mem.c
>>> +++ b/drivers/virtio/virtio_mem.c
>>> @@ -1243,7 +1243,7 @@ static int virtio_mem_fake_offline(struct virtio_mem *vm, unsigned long pfn,
>>> if (atomic_read(&vm->config_changed))
>>> return -EAGAIN;
>>> - rc = alloc_contig_range(pfn, pfn + nr_pages, MIGRATE_MOVABLE,
>>> + rc = alloc_contig_range(pfn, pfn + nr_pages, ACR_OTHER,
>>> GFP_KERNEL);
>>> if (rc == -ENOMEM)
>>> /* whoops, out of memory */
>>> diff --git a/include/linux/gfp.h b/include/linux/gfp.h
>>> index be160e8d8bcb..51990d571e3e 100644
>>> --- a/include/linux/gfp.h
>>> +++ b/include/linux/gfp.h
>>> @@ -423,9 +423,16 @@ static inline bool gfp_compaction_allowed(gfp_t gfp_mask)
>>> extern gfp_t vma_thp_gfp_mask(struct vm_area_struct *vma);
>>> #ifdef CONFIG_CONTIG_ALLOC
>>> +
>>> +enum acr_flags_t {
>>> + ACR_CMA, // CMA allocation
>>> + ACR_OTHER, // other allocation
>>> +};
>>
>> Hm, enum != flags.
>>
>> If you want to use flags, then just have ACR_CMA. ACR_OTHER is implied if not set.
>>
>> And ACR_CMA would then have to be "1" etc.
>
> I have a fixup to change acr_flags_t to acr_mode.
>
>>
>>> +
>>> /* The below functions must be run on a range from a single zone. */
>>> extern int alloc_contig_range_noprof(unsigned long start, unsigned long end,
>>> - unsigned migratetype, gfp_t gfp_mask);
>>> + enum acr_flags_t alloc_flags,
>>> + gfp_t gfp_mask);
>>> #define alloc_contig_range(...) alloc_hooks(alloc_contig_range_noprof(__VA_ARGS__))
>>> extern struct page *alloc_contig_pages_noprof(unsigned long nr_pages, gfp_t gfp_mask,
>>> diff --git a/include/linux/page-isolation.h b/include/linux/page-isolation.h
>>> index 7a681a49e73c..3e2f960e166c 100644
>>> --- a/include/linux/page-isolation.h
>>> +++ b/include/linux/page-isolation.h
>>> @@ -38,8 +38,20 @@ static inline void set_pageblock_isolate(struct page *page)
>>> }
>>> #endif
>>> -#define MEMORY_OFFLINE 0x1
>>> -#define REPORT_FAILURE 0x2
>>> +/*
>>> + * Pageblock isolation modes:
>>> + * PB_ISOLATE_MODE_MEM_OFFLINE - isolate to offline (!allocate) memory
>>> + * e.g., skip over PageHWPoison() pages and
>>> + * PageOffline() pages. Unmovable pages will be
>>> + * reported in this mode.
>>> + * PB_ISOLATE_MODE_CMA_ALLOC - isolate for CMA allocations
>>> + * PB_ISOLATE_MODE_OTHER - isolate for other purposes
>>> + */
>>> +enum pb_isolate_mode {
>>> + PB_ISOLATE_MODE_MEM_OFFLINE,
>>> + PB_ISOLATE_MODE_CMA_ALLOC,
>>> + PB_ISOLATE_MODE_OTHER,
>>> +};
>>
>> It's late on friady, but it looks like we are duplicating things here.
>>
>> Let me think about that once my brain is recharged :)
>
> Sure. Take your time.
Could we abstract both settings and use a single one? Then, we could
simply reject if MEM_OFFLINE is passed into alloc_contig_range().
alloc_contig_pages and page isolation, hmmmm, MEM_OFFLINE is kind-of an
allocation. CMA is an allocation.
Just an idea, not sure ...
--
Cheers,
David / dhildenb
next prev parent reply other threads:[~2025-05-30 20:09 UTC|newest]
Thread overview: 31+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-05-30 16:22 [PATCH v6 0/6] Make MIGRATE_ISOLATE a standalone bit Zi Yan
2025-05-30 16:22 ` [PATCH v6 1/6] mm/page_alloc: pageblock flags functions clean up Zi Yan
2025-05-30 17:01 ` Vlastimil Babka
2025-05-30 19:39 ` David Hildenbrand
2025-05-30 16:22 ` [PATCH v6 2/6] mm/page_isolation: make page isolation a standalone bit Zi Yan
2025-05-30 17:04 ` Vlastimil Babka
2025-05-30 19:41 ` David Hildenbrand
2025-05-30 16:22 ` [PATCH v6 3/6] mm/page_alloc: add support for initializing pageblock as isolated Zi Yan
2025-05-30 17:06 ` Vlastimil Babka
2025-05-30 17:10 ` Zi Yan
2025-05-30 17:59 ` Zi Yan
2025-05-30 19:44 ` David Hildenbrand
2025-05-30 19:48 ` Zi Yan
2025-05-30 16:22 ` [PATCH v6 4/6] mm/page_isolation: remove migratetype from move_freepages_block_isolate() Zi Yan
2025-05-30 17:10 ` Vlastimil Babka
2025-05-30 19:52 ` David Hildenbrand
2025-05-30 19:54 ` Zi Yan
2025-05-30 16:22 ` [PATCH v6 5/6] mm/page_isolation: remove migratetype from undo_isolate_page_range() Zi Yan
2025-05-30 16:22 ` [PATCH v6 6/6] mm/page_isolation: remove migratetype parameter from more functions Zi Yan
2025-05-30 17:15 ` Vlastimil Babka
2025-05-30 17:21 ` Zi Yan
2025-05-30 18:01 ` Zi Yan
2025-05-30 19:56 ` David Hildenbrand
2025-05-30 19:58 ` Zi Yan
2025-05-30 20:08 ` David Hildenbrand [this message]
2025-05-30 20:46 ` Zi Yan
2025-05-30 20:55 ` David Hildenbrand
2025-06-02 14:18 ` Zi Yan
2025-06-02 16:25 ` David Hildenbrand
2025-06-02 16:27 ` Zi Yan
2025-06-02 16:29 ` David Hildenbrand
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=50da1626-5894-42ef-b3fd-935a331f14c4@redhat.com \
--to=david@redhat.com \
--cc=akpm@linux-foundation.org \
--cc=baolin.wang@linux.alibaba.com \
--cc=hannes@cmpxchg.org \
--cc=jackmanb@google.com \
--cc=kirill.shutemov@linux.intel.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mgorman@techsingularity.net \
--cc=mhocko@suse.com \
--cc=osalvador@suse.de \
--cc=richardycc@google.com \
--cc=surenb@google.com \
--cc=vbabka@suse.cz \
--cc=ziy@nvidia.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).