On Tue, Dec 17, 2024 at 4:33 PM <yangge1116@xxxxxxx> wrote:
From: yangge <yangge1116@xxxxxxx>
Since commit 984fdba6a32e ("mm, compaction: use proper alloc_flags
in __compaction_suitable()") allow compaction to proceed when free
pages required for compaction reside in the CMA pageblocks, it's
possible that __compaction_suitable() always returns true, and in
some cases, it's not acceptable.
There are 4 NUMA nodes on my machine, and each NUMA node has 32GB
of memory. I have configured 16GB of CMA memory on each NUMA node,
and starting a 32GB virtual machine with device passthrough is
extremely slow, taking almost an hour.
During the start-up of the virtual machine, it will call
pin_user_pages_remote(..., FOLL_LONGTERM, ...) to allocate memory.
Long term GUP cannot allocate memory from CMA area, so a maximum
of 16 GB of no-CMA memory on a NUMA node can be used as virtual
machine memory. Since there is 16G of free CMA memory on the NUMA
Other unmovable allocations, like dma_buf, which can be large in a
Linux system, are
also unable to allocate memory from CMA. My question is whether the issue you
described applies to these allocations as well.
node, watermark for order-0 always be met for compaction, so
__compaction_suitable() always returns true, even if the node is
unable to allocate non-CMA memory for the virtual machine.
For costly allocations, because __compaction_suitable() always
returns true, __alloc_pages_slowpath() can't exit at the appropriate
place, resulting in excessively long virtual machine startup times.
Call trace:
__alloc_pages_slowpath
if (compact_result == COMPACT_SKIPPED ||
compact_result == COMPACT_DEFERRED)
goto nopage; // should exit __alloc_pages_slowpath() from here
Do we face the same issue if we allocate dma-buf while CMA has plenty
of free memory, but non-CMA has none?
In order to quickly fall back to remote node, we should remove
ALLOC_CMA both in __compaction_suitable() and __isolate_free_page()
in long term GUP flow. After this fix, starting a 32GB virtual machine
with device passthrough takes only a few seconds.
Fixes: 984fdba6a32e ("mm, compaction: use proper alloc_flags in __compaction_suitable()")
Cc: <stable@xxxxxxxxxxxxxxx>
Signed-off-by: yangge <yangge1116@xxxxxxx>
Reviewed-by: Baolin Wang <baolin.wang@xxxxxxxxxxxxxxxxx>
---
V6:
-- update cc->alloc_flags to keep the original loginc
V5:
- add 'alloc_flags' parameter for __isolate_free_page()
- remove 'usa_cma' variable
V4:
- rich the commit log description
V3:
- fix build errors
- add ALLOC_CMA both in should_continue_reclaim() and compaction_ready()
V2:
- using the 'cc->alloc_flags' to determin if 'ALLOC_CMA' is needed
- rich the commit log description
include/linux/compaction.h | 6 ++++--
mm/compaction.c | 26 +++++++++++++++-----------
mm/internal.h | 3 ++-
mm/page_alloc.c | 7 +++++--
mm/page_isolation.c | 3 ++-
mm/page_reporting.c | 2 +-
mm/vmscan.c | 4 ++--
7 files changed, 31 insertions(+), 20 deletions(-)
diff --git a/include/linux/compaction.h b/include/linux/compaction.h
index e947764..b4c3ac3 100644
--- a/include/linux/compaction.h
+++ b/include/linux/compaction.h
@@ -90,7 +90,8 @@ extern enum compact_result try_to_compact_pages(gfp_t gfp_mask,
struct page **page);
extern void reset_isolation_suitable(pg_data_t *pgdat);
extern bool compaction_suitable(struct zone *zone, int order,
- int highest_zoneidx);
+ int highest_zoneidx,
+ unsigned int alloc_flags);
extern void compaction_defer_reset(struct zone *zone, int order,
bool alloc_success);
@@ -108,7 +109,8 @@ static inline void reset_isolation_suitable(pg_data_t *pgdat)
}
static inline bool compaction_suitable(struct zone *zone, int order,
- int highest_zoneidx)
+ int highest_zoneidx,
+ unsigned int alloc_flags)
{
return false;
}
diff --git a/mm/compaction.c b/mm/compaction.c
index 07bd227..d92ba6c 100644
--- a/mm/compaction.c
+++ b/mm/compaction.c
@@ -655,7 +655,7 @@ static unsigned long isolate_freepages_block(struct compact_control *cc,
/* Found a free page, will break it into order-0 pages */
order = buddy_order(page);
- isolated = __isolate_free_page(page, order);
+ isolated = __isolate_free_page(page, order, cc->alloc_flags);
if (!isolated)
break;
set_page_private(page, order);
@@ -1634,7 +1634,7 @@ static void fast_isolate_freepages(struct compact_control *cc)
/* Isolate the page if available */
if (page) {
- if (__isolate_free_page(page, order)) {
+ if (__isolate_free_page(page, order, cc->alloc_flags)) {
set_page_private(page, order);
nr_isolated = 1 << order;
nr_scanned += nr_isolated - 1;
@@ -2381,6 +2381,7 @@ static enum compact_result compact_finished(struct compact_control *cc)
static bool __compaction_suitable(struct zone *zone, int order,
int highest_zoneidx,
+ unsigned int alloc_flags,
unsigned long wmark_target)
{
unsigned long watermark;
@@ -2395,25 +2396,26 @@ static bool __compaction_suitable(struct zone *zone, int order,
* even if compaction succeeds.
* For costly orders, we require low watermark instead of min for
* compaction to proceed to increase its chances.
- * ALLOC_CMA is used, as pages in CMA pageblocks are considered
- * suitable migration targets
+ * In addition to long term GUP flow, ALLOC_CMA is used, as pages in
+ * CMA pageblocks are considered suitable migration targets
I'm not sure if this document is correct for cases other than GUP.
