All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH 0/2] Memblock cleanup plus memory removal fix
@ 2017-01-03 20:56 Laura Abbott
  2017-01-03 20:56 ` [PATCH 1/2] arm: Cleanup sanity_check_meminfo Laura Abbott
  2017-01-03 20:56 ` [PATCH 2/2] arm: Adjust memory boundaries after reservations Laura Abbott
  0 siblings, 2 replies; 5+ messages in thread
From: Laura Abbott @ 2017-01-03 20:56 UTC (permalink / raw)
  To: linux-arm-kernel

Hi,

During the course of https://marc.info/?l=linux-arm-kernel&m=148145259511248,
Grygorii Strashko reminded me of another issue where I proposed a patch but
never followed up on it. The patch in
http://lists.infradead.org/pipermail/linux-arm-kernel/2014-October/296978.html
did some cleanup and renaming of sanity_check_meminfo. I think this makes the
code more readable so I'd like to resurect it and rebase my fix
http://lists.infradead.org/pipermail/linux-arm-kernel/2016-December/474060.html
on top of it.

Thanks,
Laura

Laura Abbott (2):
  arm: Cleanup sanity_check_meminfo
  arm: Adjust memory boundaries after reservations

 arch/arm/kernel/setup.c | 12 ++++++++--
 arch/arm/mm/mmu.c       | 62 +++++++++++++++++--------------------------------
 arch/arm/mm/nommu.c     |  8 +++----
 3 files changed, 35 insertions(+), 47 deletions(-)

-- 
2.7.4

^ permalink raw reply	[flat|nested] 5+ messages in thread

* [PATCH 1/2] arm: Cleanup sanity_check_meminfo
  2017-01-03 20:56 [PATCH 0/2] Memblock cleanup plus memory removal fix Laura Abbott
@ 2017-01-03 20:56 ` Laura Abbott
  2017-01-04  1:12   ` Nicolas Pitre
  2017-01-03 20:56 ` [PATCH 2/2] arm: Adjust memory boundaries after reservations Laura Abbott
  1 sibling, 1 reply; 5+ messages in thread
From: Laura Abbott @ 2017-01-03 20:56 UTC (permalink / raw)
  To: linux-arm-kernel


The logic for sanity_check_meminfo has become difficult to
follow. Clean up the code so it's more obvious what the code
is actually trying to do. Additionally, meminfo is now removed
so rename the function to better describe it's purpose.

Originally proposed in http://lists.infradead.org/pipermail/linux-arm-kernel/2014-October/296978.html

Signed-off-by: Laura Abbott <lauraa@codeaurora.org>
Signed-off-by: Laura Abbott <labbott@redhat.com>
---
 arch/arm/kernel/setup.c |  4 ++--
 arch/arm/mm/mmu.c       | 59 +++++++++++++++----------------------------------
 arch/arm/mm/nommu.c     |  8 +++----
 3 files changed, 24 insertions(+), 47 deletions(-)

diff --git a/arch/arm/kernel/setup.c b/arch/arm/kernel/setup.c
index 34e3f3c..8a8051c 100644
--- a/arch/arm/kernel/setup.c
+++ b/arch/arm/kernel/setup.c
@@ -81,7 +81,7 @@ __setup("fpe=", fpe_setup);
 extern void init_default_cache_policy(unsigned long);
 extern void paging_init(const struct machine_desc *desc);
 extern void early_paging_init(const struct machine_desc *);
-extern void sanity_check_meminfo(void);
+extern void adjust_lowmem_bounds(void);
 extern enum reboot_mode reboot_mode;
 extern void setup_dma_zone(const struct machine_desc *desc);
 
@@ -1093,7 +1093,7 @@ void __init setup_arch(char **cmdline_p)
 	setup_dma_zone(mdesc);
 	xen_early_init();
 	efi_init();
-	sanity_check_meminfo();
+	adjust_lowmem_bounds();
 	arm_memblock_init(mdesc);
 
 	early_ioremap_reset();
diff --git a/arch/arm/mm/mmu.c b/arch/arm/mm/mmu.c
index 4001dd1..32ecdfd 100644
--- a/arch/arm/mm/mmu.c
+++ b/arch/arm/mm/mmu.c
@@ -1152,13 +1152,11 @@ early_param("vmalloc", early_vmalloc);
 
 phys_addr_t arm_lowmem_limit __initdata = 0;
 
-void __init sanity_check_meminfo(void)
+void __init adjust_lowmem_bounds(void)
 {
 	phys_addr_t memblock_limit = 0;
-	int highmem = 0;
 	u64 vmalloc_limit;
 	struct memblock_region *reg;
-	bool should_use_highmem = false;
 
 	/*
 	 * Let's use our own (unoptimized) equivalent of __pa() that is
@@ -1172,43 +1170,13 @@ void __init sanity_check_meminfo(void)
 	for_each_memblock(memory, reg) {
 		phys_addr_t block_start = reg->base;
 		phys_addr_t block_end = reg->base + reg->size;
-		phys_addr_t size_limit = reg->size;
 
-		if (reg->base >= vmalloc_limit)
-			highmem = 1;
-		else
-			size_limit = vmalloc_limit - reg->base;
 
-
-		if (!IS_ENABLED(CONFIG_HIGHMEM) || cache_is_vipt_aliasing()) {
-
-			if (highmem) {
-				pr_notice("Ignoring RAM at %pa-%pa (!CONFIG_HIGHMEM)\n",
-					  &block_start, &block_end);
-				memblock_remove(reg->base, reg->size);
-				should_use_highmem = true;
-				continue;
-			}
-
-			if (reg->size > size_limit) {
-				phys_addr_t overlap_size = reg->size - size_limit;
-
-				pr_notice("Truncating RAM at %pa-%pa",
-					  &block_start, &block_end);
-				block_end = vmalloc_limit;
-				pr_cont(" to -%pa", &block_end);
-				memblock_remove(vmalloc_limit, overlap_size);
-				should_use_highmem = true;
-			}
-		}
-
-		if (!highmem) {
-			if (block_end > arm_lowmem_limit) {
-				if (reg->size > size_limit)
-					arm_lowmem_limit = vmalloc_limit;
-				else
-					arm_lowmem_limit = block_end;
-			}
+		if (reg->base < vmalloc_limit) {
+			if (block_end > arm_lowmem_limit)
+				arm_lowmem_limit = min(
+						(phys_addr_t)vmalloc_limit,
+						block_end);
 
 			/*
 			 * Find the first non-pmd-aligned page, and point
@@ -1233,9 +1201,6 @@ void __init sanity_check_meminfo(void)
 		}
 	}
 
-	if (should_use_highmem)
-		pr_notice("Consider using a HIGHMEM enabled kernel.\n");
-
 	high_memory = __va(arm_lowmem_limit - 1) + 1;
 
 	/*
@@ -1248,6 +1213,18 @@ void __init sanity_check_meminfo(void)
 	if (!memblock_limit)
 		memblock_limit = arm_lowmem_limit;
 
+	if (!IS_ENABLED(CONFIG_HIGHMEM) || cache_is_vipt_aliasing()) {
+		if (memblock_end_of_DRAM() > arm_lowmem_limit) {
+			phys_addr_t end = memblock_end_of_DRAM();
+
+			pr_notice("Ignoring RAM at %pa-%pa\n",
+				  &memblock_limit, &end);
+			pr_notice("Consider using a HIGHMEM enabled kernel.\n");
+
+			memblock_remove(memblock_limit, end);
+		}
+	}
+
 	memblock_set_current_limit(memblock_limit);
 }
 
diff --git a/arch/arm/mm/nommu.c b/arch/arm/mm/nommu.c
index 2740967..13a25d6 100644
--- a/arch/arm/mm/nommu.c
+++ b/arch/arm/mm/nommu.c
@@ -85,7 +85,7 @@ static unsigned long irbar_read(void)
 }
 
 /* MPU initialisation functions */
-void __init sanity_check_meminfo_mpu(void)
+void __init adjust_lowmem_bounds_mpu(void)
 {
 	phys_addr_t phys_offset = PHYS_OFFSET;
 	phys_addr_t aligned_region_size, specified_mem_size, rounded_mem_size;
@@ -274,7 +274,7 @@ void __init mpu_setup(void)
 	}
 }
 #else
-static void sanity_check_meminfo_mpu(void) {}
+static void adjust_lowmem_bounds_mpu(void) {}
 static void __init mpu_setup(void) {}
 #endif /* CONFIG_ARM_MPU */
 
@@ -295,10 +295,10 @@ void __init arm_mm_memblock_reserve(void)
 #endif
 }
 
-void __init sanity_check_meminfo(void)
+void __init adjust_lowmem_bounds(void)
 {
 	phys_addr_t end;
-	sanity_check_meminfo_mpu();
+	adjust_lowmem_bounds_mpu();
 	end = memblock_end_of_DRAM();
 	high_memory = __va(end - 1) + 1;
 	memblock_set_current_limit(end);
-- 
2.7.4

^ permalink raw reply related	[flat|nested] 5+ messages in thread

* [PATCH 2/2] arm: Adjust memory boundaries after reservations
  2017-01-03 20:56 [PATCH 0/2] Memblock cleanup plus memory removal fix Laura Abbott
  2017-01-03 20:56 ` [PATCH 1/2] arm: Cleanup sanity_check_meminfo Laura Abbott
@ 2017-01-03 20:56 ` Laura Abbott
  1 sibling, 0 replies; 5+ messages in thread
From: Laura Abbott @ 2017-01-03 20:56 UTC (permalink / raw)
  To: linux-arm-kernel


adjust_lowmem_bounds is responsible for setting up the boundary for
lowmem/hihgmme. This needs to be setup before memblock reservations can
occur. At the time memblock reservations can occur, memory can also be
removed from the system. The lowmem/highmem boundary and end of memory
may be affected by this but it is currently not recalculated. On some
systems this may be harmless, on o thers this may result in incorrect
ranges being passed to the main memory allocator. Correct this by
recalculating the lowmem/highmem boundary after all reservations have
been made.

Signed-off-by: Laura Abbott <labbott@redhat.com>
---
 arch/arm/kernel/setup.c | 8 ++++++++
 arch/arm/mm/mmu.c       | 9 ++++++---
 2 files changed, 14 insertions(+), 3 deletions(-)

diff --git a/arch/arm/kernel/setup.c b/arch/arm/kernel/setup.c
index 8a8051c..4625115 100644
--- a/arch/arm/kernel/setup.c
+++ b/arch/arm/kernel/setup.c
@@ -1093,8 +1093,16 @@ void __init setup_arch(char **cmdline_p)
 	setup_dma_zone(mdesc);
 	xen_early_init();
 	efi_init();
+	/*
+	 * Make sure the calcualtion for lowmem/highmem is set appropriately
+	 * before reserving/allocating any mmeory
+	 */
 	adjust_lowmem_bounds();
 	arm_memblock_init(mdesc);
+	/*
+	 * Memory may have been removed so recalculate the bounds.
+	 */
+	adjust_lowmem_bounds();
 
 	early_ioremap_reset();
 
diff --git a/arch/arm/mm/mmu.c b/arch/arm/mm/mmu.c
index 32ecdfd..3f12a64 100644
--- a/arch/arm/mm/mmu.c
+++ b/arch/arm/mm/mmu.c
@@ -1157,6 +1157,7 @@ void __init adjust_lowmem_bounds(void)
 	phys_addr_t memblock_limit = 0;
 	u64 vmalloc_limit;
 	struct memblock_region *reg;
+	phys_addr_t lowmem_limit = 0;
 
 	/*
 	 * Let's use our own (unoptimized) equivalent of __pa() that is
@@ -1173,8 +1174,8 @@ void __init adjust_lowmem_bounds(void)
 
 
 		if (reg->base < vmalloc_limit) {
-			if (block_end > arm_lowmem_limit)
-				arm_lowmem_limit = min(
+			if (block_end > lowmem_limit)
+				lowmem_limit = min(
 						(phys_addr_t)vmalloc_limit,
 						block_end);
 
@@ -1195,12 +1196,14 @@ void __init adjust_lowmem_bounds(void)
 				if (!IS_ALIGNED(block_start, PMD_SIZE))
 					memblock_limit = block_start;
 				else if (!IS_ALIGNED(block_end, PMD_SIZE))
-					memblock_limit = arm_lowmem_limit;
+					memblock_limit = lowmem_limit;
 			}
 
 		}
 	}
 
+	arm_lowmem_limit = lowmem_limit;
+
 	high_memory = __va(arm_lowmem_limit - 1) + 1;
 
 	/*
-- 
2.7.4

^ permalink raw reply related	[flat|nested] 5+ messages in thread

* [PATCH 1/2] arm: Cleanup sanity_check_meminfo
  2017-01-03 20:56 ` [PATCH 1/2] arm: Cleanup sanity_check_meminfo Laura Abbott
@ 2017-01-04  1:12   ` Nicolas Pitre
  2017-01-04 16:50     ` Laura Abbott
  0 siblings, 1 reply; 5+ messages in thread
From: Nicolas Pitre @ 2017-01-04  1:12 UTC (permalink / raw)
  To: linux-arm-kernel

On Tue, 3 Jan 2017, Laura Abbott wrote:

> 
> The logic for sanity_check_meminfo has become difficult to
> follow. Clean up the code so it's more obvious what the code
> is actually trying to do.

Absolutely. This function has a long history of subtle bugs ... and 
broken fixes.

> +		if (reg->base < vmalloc_limit) {
> +			if (block_end > arm_lowmem_limit)
> +				arm_lowmem_limit = min(
> +						(phys_addr_t)vmalloc_limit,
> +						block_end);

Just to illustrate how subtle this can be, the above would reintroduce a 
bug that was fixed in commit b9a019899f.


Nicolas

^ permalink raw reply	[flat|nested] 5+ messages in thread

* [PATCH 1/2] arm: Cleanup sanity_check_meminfo
  2017-01-04  1:12   ` Nicolas Pitre
@ 2017-01-04 16:50     ` Laura Abbott
  0 siblings, 0 replies; 5+ messages in thread
From: Laura Abbott @ 2017-01-04 16:50 UTC (permalink / raw)
  To: linux-arm-kernel

On 01/03/2017 05:12 PM, Nicolas Pitre wrote:
> On Tue, 3 Jan 2017, Laura Abbott wrote:
> 
>>
>> The logic for sanity_check_meminfo has become difficult to
>> follow. Clean up the code so it's more obvious what the code
>> is actually trying to do.
> 
> Absolutely. This function has a long history of subtle bugs ... and 
> broken fixes.
> 
>> +		if (reg->base < vmalloc_limit) {
>> +			if (block_end > arm_lowmem_limit)
>> +				arm_lowmem_limit = min(
>> +						(phys_addr_t)vmalloc_limit,
>> +						block_end);
> 
> Just to illustrate how subtle this can be, the above would reintroduce a 
> bug that was fixed in commit b9a019899f.
> 
> 
> Nicolas
> 

Yup, I saw that commit and thought "I need to keep the fix". And yet I
still managed to get it wrong. I'll fix in a v2.

Thanks,
Laura

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2017-01-04 16:50 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-01-03 20:56 [PATCH 0/2] Memblock cleanup plus memory removal fix Laura Abbott
2017-01-03 20:56 ` [PATCH 1/2] arm: Cleanup sanity_check_meminfo Laura Abbott
2017-01-04  1:12   ` Nicolas Pitre
2017-01-04 16:50     ` Laura Abbott
2017-01-03 20:56 ` [PATCH 2/2] arm: Adjust memory boundaries after reservations Laura Abbott

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.