Skip to contentSkip to conversation

powerpc/fadump: handle CMA activation failure appropriately

PATCH Superseded Hari Bathini 2 participants 6 messages Jan 12, 2022 – Jan 17, 2022

Subsystems (3)

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

v2 cover powerpc/fadump: handle CMA activation failure appropriately

While commit a4e92ce8e4c8 ("powerpc/fadump: Reservationless firmware assisted dump"), introduced Linux kernel's Contiguous Memory Allocator (CMA) based reservation for fadump, it came with the assumption that the memory remains reserved even if CMA activation fails. It ensures no kernel pages reside in the reserved memory region, which can't be mapped into the /proc/vmcore.

But commit 072355c1cf2d ("mm/cma: expose all pages to the buddy if activation of an area fails") started returning all pages to buddy allocator if CMA activation fails. This led to warning messages like below while running crash-utility on vmcore of a kernel having above two commits:

crash: seek error: kernel virtual address: <from reserved region>

as reserved memory region ended up having kernel pages crash-utility was looking for. Fix this by introducing an option in CMA, to opt out from exposing pages to buddy allocator, on CMA activation failure.

Changes in v2:
* Replaced cma->free_pages_on_error with cma->reserve_pages_on_error & cma_dont_free_pages_on_error() with cma_reserve_pages_on_error() to avoid confusion and make the expectation on failure clearer.

Hari Bathini (2):
mm/cma: provide option to opt out from exposing pages on activation
failure
powerpc/fadump: opt out from freeing pages on cma activation failure

 arch/powerpc/kernel/fadump.c |  6 ++++++
 include/linux/cma.h          |  2 ++
 mm/cma.c                     | 15 +++++++++++++--
 mm/cma.h                     |  1 +
 4 files changed, 22 insertions(+), 2 deletions(-)
Signature · 1 line
2.34.1
#1 lore raw

PATCH v2 1/2 mm/cma: provide option to opt out from exposing pages on activation failure

Commit 072355c1cf2d ("mm/cma: expose all pages to the buddy if activation of an area fails") started exposing all pages to buddy allocator on CMA activation failure. But there can be CMA users that want to handle the reserved memory differently on CMA allocation failure. Provide an option to opt out from exposing pages to buddy for such cases.

---

Changes in v2:
* Changed cma->free_pages_on_error to cma->reserve_pages_on_error and cma_dont_free_pages_on_error() to cma_reserve_pages_on_error() to avoid cofusion.

 include/linux/cma.h |  2 ++
 mm/cma.c            | 15 +++++++++++++--
 mm/cma.h            |  1 +
 3 files changed, 16 insertions(+), 2 deletions(-)
include/linux/cma.h +2 -0
index bd801023504b..51d540eee18a 100644
@@ -50,4 +50,6 @@ extern bool cma_pages_valid(struct cma *cma, const struct page *pages, unsigned
 extern bool cma_release(struct cma *cma, const struct page *pages, unsigned long count);
 
 extern int cma_for_each_area(int (*it)(struct cma *cma, void *data), void *data);
+
+extern void cma_reserve_pages_on_error(struct cma *cma);
 #endif
mm/cma.c +13 -2
index bc9ca8f3c487..20626321f87f 100644
@@ -131,8 +131,10 @@ static void __init cma_activate_area(struct cma *cma)
 	bitmap_free(cma->bitmap);
 out_error:
 	/* Expose all pages to the buddy, they are useless for CMA. */
-	for (pfn = base_pfn; pfn < base_pfn + cma->count; pfn++)
-		free_reserved_page(pfn_to_page(pfn));
+	if (!cma->reserve_pages_on_error) {
+		for (pfn = base_pfn; pfn < base_pfn + cma->count; pfn++)
+			free_reserved_page(pfn_to_page(pfn));
+	}
 	totalcma_pages -= cma->count;
 	cma->count = 0;
 	pr_err("CMA area %s could not be activated\n", cma->name);
@@ -150,6 +152,14 @@ static int __init cma_init_reserved_areas(void)
 }
 core_initcall(cma_init_reserved_areas);
 
+void __init cma_reserve_pages_on_error(struct cma *cma)
+{
+	if (!cma)
+		return;
+
+	cma->reserve_pages_on_error = true;
+}
+
 /**
  * cma_init_reserved_mem() - create custom contiguous area from reserved memory
  * @base: Base address of the reserved area
@@ -204,6 +214,7 @@ int __init cma_init_reserved_mem(phys_addr_t base, phys_addr_t size,
 	cma->base_pfn = PFN_DOWN(base);
 	cma->count = size >> PAGE_SHIFT;
 	cma->order_per_bit = order_per_bit;
+	cma->reserve_pages_on_error = false;
 	*res_cma = cma;
 	cma_area_count++;
 	totalcma_pages += (size / PAGE_SIZE);
mm/cma.h +1 -0
index 2c775877eae2..88a0595670b7 100644
@@ -30,6 +30,7 @@ struct cma {
 	/* kobject requires dynamic object */
 	struct cma_kobject *cma_kobj;
 #endif
+	bool reserve_pages_on_error;
 };
 
 extern struct cma cma_areas[MAX_CMA_AREAS];
Signature · 1 line
2.34.1
Reviewed-by on v2 1/2
#2 lore raw

Re: [PATCH v2 1/2] mm/cma: provide option to opt out from exposing pages on activation failure

mm/cma.c static int __init cma_init_reserved_areas(void) · line 158

Quoted 3 lines
mm/cma.c +3 -0
+{
+	if (!cma)
+		return;

Do we really need that check for NULL?

mm/cma.c int __init cma_init_reserved_mem(phys_addr_t base, phys_addr_t size, · line 217

Quoted 12 lines
mm/cma.c +5 -0
+
+	cma->reserve_pages_on_error = true;
+}
+
 /**
  * cma_init_reserved_mem() - create custom contiguous area from reserved memory
  * @base: Base address of the reserved area
@@ -204,6 +214,7 @@ int __init cma_init_reserved_mem(phys_addr_t base, phys_addr_t size,
 	cma->base_pfn = PFN_DOWN(base);
 	cma->count = size >> PAGE_SHIFT;
 	cma->order_per_bit = order_per_bit;
+	cma->reserve_pages_on_error = false;

I think you can drop that; should already be initialized to 0.

Apart from that

Signature · 3 lines
Thanks,

David / dhildenb
#3 lore raw

Re: [PATCH v2 1/2] mm/cma: provide option to opt out from exposing pages on activation failure

Quoted 5 lines · David Hildenbrand

mm/cma.c static int __init cma_init_reserved_areas(void) · line 158

Quoted 3 lines
mm/cma.c +3 -0
+{
+	if (!cma)
+		return;

Do we really need that check for NULL?

Probably not.

Quoted 17 lines

mm/cma.c int __init cma_init_reserved_mem(phys_addr_t base, phys_addr_t size, · line 217

Quoted 13 lines
mm/cma.c +5 -0
+
+	cma->reserve_pages_on_error = true;
+}
+
  /**
   * cma_init_reserved_mem() - create custom contiguous area from reserved memory
   * @base: Base address of the reserved area
@@ -204,6 +214,7 @@ int __init cma_init_reserved_mem(phys_addr_t base, phys_addr_t size,
  	cma->base_pfn = PFN_DOWN(base);
  	cma->count = size >> PAGE_SHIFT;
  	cma->order_per_bit = order_per_bit;
+	cma->reserve_pages_on_error = false;

I think you can drop that; should already be initialized to 0.

Quoted 4 lines

Apart from that

Thanks for the review, David.
Posted v3 with the changes suggested above.

- Hari

#4 lore raw

PATCH v2 2/2 powerpc/fadump: opt out from freeing pages on cma activation failure

In reply to Hari Bathini

With commit a4e92ce8e4c8 ("powerpc/fadump: Reservationless firmware assisted dump"), Linux kernel's Contiguous Memory Allocator (CMA) based reservation was introduced in fadump. That change was aimed at using CMA to let applications utilize the memory reserved for fadump while blocking it from being used for kernel pages. The assumption was, even if CMA activation fails for whatever reason, the memory still remains reserved to avoid it from being used for kernel pages. But commit 072355c1cf2d ("mm/cma: expose all pages to the buddy if activation of an area fails") breaks this assumption as it started exposing all pages to buddy allocator on CMA activation failure. It led to warning messages like below while running crash-utility on vmcore of a kernel having above two commits:

crash: seek error: kernel virtual address: <from reserved region>

To fix this problem, opt out from exposing pages to buddy allocator on CMA activation failure for fadump reserved memory.

---
 arch/powerpc/kernel/fadump.c | 6 ++++++
 1 file changed, 6 insertions(+)
arch/powerpc/kernel/fadump.c +6 -0
index b7ceb041743c..82058b52e34a 100644
@@ -112,6 +112,12 @@ static int __init fadump_cma_init(void)
 		return 1;
 	}
 
+	/*
+	 *  If CMA activation fails, keep the pages reserved, instead of
+	 *  exposing them to buddy allocator. Same as 'fadump=nocma' case.
+	 */
+	cma_reserve_pages_on_error(fadump_cma);
+
 	/*
 	 * So we now have successfully initialized cma area for fadump.
 	 */
Signature · 1 line
2.34.1
Acked-by on v2 2/2
#5 lore raw

Re: [PATCH v2 2/2] powerpc/fadump: opt out from freeing pages on cma activation failure

Quoted 30 lines · Hari Bathini

With commit a4e92ce8e4c8 ("powerpc/fadump: Reservationless firmware assisted dump"), Linux kernel's Contiguous Memory Allocator (CMA) based reservation was introduced in fadump. That change was aimed at using CMA to let applications utilize the memory reserved for fadump while blocking it from being used for kernel pages. The assumption was, even if CMA activation fails for whatever reason, the memory still remains reserved to avoid it from being used for kernel pages. But commit 072355c1cf2d ("mm/cma: expose all pages to the buddy if activation of an area fails") breaks this assumption as it started exposing all pages to buddy allocator on CMA activation failure. It led to warning messages like below while running crash-utility on vmcore of a kernel having above two commits:

crash: seek error: kernel virtual address: <from reserved region>

To fix this problem, opt out from exposing pages to buddy allocator on CMA activation failure for fadump reserved memory.

---
 arch/powerpc/kernel/fadump.c | 6 ++++++
 1 file changed, 6 insertions(+)
arch/powerpc/kernel/fadump.c +0 -0
index b7ceb041743c..82058b52e34a 100644
@@ -112,6 +112,12 @@ static int __init fadump_cma_init(void)
 		return 1;
 	}

arch/powerpc/kernel/fadump.c static int __init fadump_cma_init(void) · line 123

Quoted 10 lines
arch/powerpc/kernel/fadump.c +6 -0
 
+	/*
+	 *  If CMA activation fails, keep the pages reserved, instead of
+	 *  exposing them to buddy allocator. Same as 'fadump=nocma' case.
+	 */
+	cma_reserve_pages_on_error(fadump_cma);
+
 	/*
 	 * So we now have successfully initialized cma area for fadump.
 	 */
Signature · 3 lines
Thanks,

David / dhildenb