[PATCH 1/2] x86: use memblock_alloc_range() or memblock_alloc_base()

[PATCH 1/2] x86: use memblock_alloc_range() or memblock_alloc_base()

[Akinobu Mita]

Replace memblock_find_in_range() and memblock_reserve() with
memblock_alloc_range() or memblock_alloc_base().

Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: x86@kernel.org
Cc: linux-mm@kvack.org
---
 arch/x86/kernel/aperture_64.c |  6 +++---
 arch/x86/kernel/setup.c       | 22 ++++++++++------------
 arch/x86/mm/init.c            |  7 +++----
 arch/x86/mm/numa.c            |  4 +---
 arch/x86/mm/numa_emulation.c  |  5 ++---
 arch/x86/realmode/init.c      |  3 +--
 6 files changed, 20 insertions(+), 27 deletions(-)

diff --git a/arch/x86/kernel/aperture_64.c b/arch/x86/kernel/aperture_64.c
index 76164e1..baaa7c9 100644
--- a/arch/x86/kernel/aperture_64.c
+++ b/arch/x86/kernel/aperture_64.c
@@ -74,14 +74,14 @@ static u32 __init allocate_aperture(void)
 	 * memory. Unfortunately we cannot move it up because that would
 	 * make the IOMMU useless.
 	 */
-	addr = memblock_find_in_range(GART_MIN_ADDR, GART_MAX_ADDR,
-				      aper_size, aper_size);
+	addr = memblock_alloc_range(aper_size, aper_size,
+				    GART_MIN_ADDR, GART_MAX_ADDR);
+
 	if (!addr) {
 		pr_err("Cannot allocate aperture memory hole [mem %#010lx-%#010lx] (%uKB)\n",
 		       addr, addr + aper_size - 1, aper_size >> 10);
 		return 0;
 	}
-	memblock_reserve(addr, aper_size);
 	pr_info("Mapping aperture over RAM [mem %#010lx-%#010lx] (%uKB)\n",
 		addr, addr + aper_size - 1, aper_size >> 10);
 	register_nosave_region(addr >> PAGE_SHIFT,
diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index 41ead8d..fa609c9 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -327,8 +327,8 @@ static void __init relocate_initrd(void)
 	char *p, *q;
 
 	/* We need to move the initrd down into directly mapped mem */
-	relocated_ramdisk = memblock_find_in_range(0, PFN_PHYS(max_pfn_mapped),
-						   area_size, PAGE_SIZE);
+	relocated_ramdisk = memblock_alloc_base(area_size, PAGE_SIZE,
+						PFN_PHYS(max_pfn_mapped));
 
 	if (!relocated_ramdisk)
 		panic("Cannot find place for new RAMDISK of size %lld\n",
@@ -336,7 +336,6 @@ static void __init relocate_initrd(void)
 
 	/* Note: this includes all the mem currently occupied by
 	   the initrd, we rely on that fact to keep the data intact. */
-	memblock_reserve(relocated_ramdisk, area_size);
 	initrd_start = relocated_ramdisk + PAGE_OFFSET;
 	initrd_end   = initrd_start + ramdisk_size;
 	printk(KERN_INFO "Allocated new RAMDISK: [mem %#010llx-%#010llx]\n",
@@ -545,8 +544,8 @@ static void __init reserve_crashkernel_low(void)
 			return;
 	}
 
-	low_base = memblock_find_in_range(low_size, (1ULL<<32),
-					low_size, alignment);
+	low_base = memblock_alloc_range(low_size, alignment,
+					low_size, 1ULL << 32);
 
 	if (!low_base) {
 		if (!auto_set)
@@ -555,7 +554,6 @@ static void __init reserve_crashkernel_low(void)
 		return;
 	}
 
-	memblock_reserve(low_base, low_size);
 	pr_info("Reserving %ldMB of low memory at %ldMB for crashkernel (System low RAM: %ldMB)\n",
 			(unsigned long)(low_size >> 20),
 			(unsigned long)(low_base >> 20),
@@ -593,10 +591,10 @@ static void __init reserve_crashkernel(void)
 		/*
 		 *  kexec want bzImage is below CRASH_KERNEL_ADDR_MAX
 		 */
-		crash_base = memblock_find_in_range(alignment,
+		crash_base = memblock_alloc_range(crash_size, alignment,
+					alignment,
 					high ? CRASH_KERNEL_ADDR_HIGH_MAX :
-					       CRASH_KERNEL_ADDR_LOW_MAX,
-					crash_size, alignment);
+					       CRASH_KERNEL_ADDR_LOW_MAX);
 
 		if (!crash_base) {
 			pr_info("crashkernel reservation failed - No suitable area found.\n");
@@ -606,14 +604,14 @@ static void __init reserve_crashkernel(void)
 	} else {
 		unsigned long long start;
 
-		start = memblock_find_in_range(crash_base,
-				 crash_base + crash_size, crash_size, 1<<20);
+		start = memblock_alloc_range(crash_size, 1 << 20,
+					     crash_base,
+					     crash_base + crash_size);
 		if (start != crash_base) {
 			pr_info("crashkernel reservation failed - memory is in use.\n");
 			return;
 		}
 	}
-	memblock_reserve(crash_base, crash_size);
 
 	printk(KERN_INFO "Reserving %ldMB of memory at %ldMB "
 			"for crashkernel (System RAM: %ldMB)\n",
diff --git a/arch/x86/mm/init.c b/arch/x86/mm/init.c
index 66dba36..76f67be 100644
--- a/arch/x86/mm/init.c
+++ b/arch/x86/mm/init.c
@@ -61,12 +61,11 @@ __ref void *alloc_low_pages(unsigned int num)
 		unsigned long ret;
 		if (min_pfn_mapped >= max_pfn_mapped)
 			panic("alloc_low_pages: ran out of memory");
-		ret = memblock_find_in_range(min_pfn_mapped << PAGE_SHIFT,
-					max_pfn_mapped << PAGE_SHIFT,
-					PAGE_SIZE * num , PAGE_SIZE);
+		ret = memblock_alloc_range(PAGE_SIZE * num, PAGE_SIZE,
+					   min_pfn_mapped << PAGE_SHIFT,
+					   max_pfn_mapped << PAGE_SHIFT);
 		if (!ret)
 			panic("alloc_low_pages: can not alloc memory");
-		memblock_reserve(ret, PAGE_SIZE * num);
 		pfn = ret >> PAGE_SHIFT;
 	} else {
 		pfn = pgt_buf_end;
diff --git a/arch/x86/mm/numa.c b/arch/x86/mm/numa.c
index a32b706..f7c0718 100644
--- a/arch/x86/mm/numa.c
+++ b/arch/x86/mm/numa.c
@@ -375,15 +375,13 @@ static int __init numa_alloc_distance(void)
 	cnt++;
 	size = cnt * cnt * sizeof(numa_distance[0]);
 
-	phys = memblock_find_in_range(0, PFN_PHYS(max_pfn_mapped),
-				      size, PAGE_SIZE);
+	phys = memblock_alloc_base(size, PAGE_SIZE, PFN_PHYS(max_pfn_mapped));
 	if (!phys) {
 		pr_warning("NUMA: Warning: can't allocate distance table!\n");
 		/* don't retry until explicitly reset */
 		numa_distance = (void *)1LU;
 		return -ENOMEM;
 	}
-	memblock_reserve(phys, size);
 
 	numa_distance = __va(phys);
 	numa_distance_cnt = cnt;
diff --git a/arch/x86/mm/numa_emulation.c b/arch/x86/mm/numa_emulation.c
index a8f90ce..341d5ae 100644
--- a/arch/x86/mm/numa_emulation.c
+++ b/arch/x86/mm/numa_emulation.c
@@ -357,13 +357,12 @@ void __init numa_emulation(struct numa_meminfo *numa_meminfo, int numa_dist_cnt)
 	if (numa_dist_cnt) {
 		u64 phys;
 
-		phys = memblock_find_in_range(0, PFN_PHYS(max_pfn_mapped),
-					      phys_size, PAGE_SIZE);
+		phys = memblock_alloc_base(phys_size, PAGE_SIZE,
+					   PFN_PHYS(max_pfn_mapped));
 		if (!phys) {
 			pr_warning("NUMA: Warning: can't allocate copy of distance table, disabling emulation\n");
 			goto no_emu;
 		}
-		memblock_reserve(phys, phys_size);
 		phys_dist = __va(phys);
 
 		for (i = 0; i < numa_dist_cnt; i++)
diff --git a/arch/x86/realmode/init.c b/arch/x86/realmode/init.c
index bad628a..7867ae6 100644
--- a/arch/x86/realmode/init.c
+++ b/arch/x86/realmode/init.c
@@ -15,12 +15,11 @@ void __init reserve_real_mode(void)
 	size_t size = PAGE_ALIGN(real_mode_blob_end - real_mode_blob);
 
 	/* Has to be under 1M so we can execute real-mode AP code. */
-	mem = memblock_find_in_range(0, 1<<20, size, PAGE_SIZE);
+	mem = memblock_alloc_base(size, PAGE_SIZE, 1 << 20);
 	if (!mem)
 		panic("Cannot allocate trampoline\n");
 
 	base = __va(mem);
-	memblock_reserve(mem, size);
 	real_mode_header = (struct real_mode_header *) base;
 	printk(KERN_DEBUG "Base memory trampoline at [%p] %llx size %zu\n",
 	       base, (unsigned long long)mem, size);
-- 
1.9.1

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

[PATCH 2/2] mm: use memblock_alloc_range()

[Akinobu Mita]

Replace memblock_find_in_range() and memblock_reserve() with
memblock_alloc_range().

Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Cc: linux-mm@kvack.org
---
 mm/memblock.c | 10 ++--------
 1 file changed, 2 insertions(+), 8 deletions(-)

diff --git a/mm/memblock.c b/mm/memblock.c
index 6d2f219..4d98d93 100644
--- a/mm/memblock.c
+++ b/mm/memblock.c
@@ -1151,21 +1151,16 @@ static void * __init memblock_virt_alloc_internal(
 	if (WARN_ON_ONCE(slab_is_available()))
 		return kzalloc_node(size, GFP_NOWAIT, nid);
 
-	if (!align)
-		align = SMP_CACHE_BYTES;
-
 	if (max_addr > memblock.current_limit)
 		max_addr = memblock.current_limit;
 
 again:
-	alloc = memblock_find_in_range_node(size, align, min_addr, max_addr,
-					    nid);
+	alloc = memblock_alloc_range_nid(size, align, min_addr, max_addr, nid);
 	if (alloc)
 		goto done;
 
 	if (nid != NUMA_NO_NODE) {
-		alloc = memblock_find_in_range_node(size, align, min_addr,
-						    max_addr,  NUMA_NO_NODE);
+		alloc = memblock_alloc_range(size, align, min_addr, max_addr);
 		if (alloc)
 			goto done;
 	}
@@ -1178,7 +1173,6 @@ again:
 	}
 
 done:
-	memblock_reserve(alloc, size);
 	ptr = phys_to_virt(alloc);
 	memset(ptr, 0, size);
 
-- 
1.9.1

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH 1/2] x86: use memblock_alloc_range() or memblock_alloc_base()

[Andrew Morton]

On Sun, 24 Aug 2014 23:56:02 +0900 Akinobu Mita <akinobu.mita@gmail.com> wrote:

> Replace memblock_find_in_range() and memblock_reserve() with
> memblock_alloc_range() or memblock_alloc_base().

Please spend a little more time preparing the changelogs?

Why are we making this change?  Because memblock_alloc_range() is
equivalent to memblock_find_in_range()+memblock_reserve() and it's just
a cleanup?  Or is there some deeper functional reason?

Does memblock_find_in_range() need to exist?  Can we convert all
callers to memblock_alloc_range()?

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH 1/2] x86: use memblock_alloc_range() or memblock_alloc_base()

[Akinobu Mita]

2014-08-28 5:53 GMT+09:00 Andrew Morton <akpm@linux-foundation.org>:
> On Sun, 24 Aug 2014 23:56:02 +0900 Akinobu Mita <akinobu.mita@gmail.com> wrote:
>
>> Replace memblock_find_in_range() and memblock_reserve() with
>> memblock_alloc_range() or memblock_alloc_base().
>
> Please spend a little more time preparing the changelogs?

OK, I'll be careful next time.

> Why are we making this change?  Because memblock_alloc_range() is
> equivalent to memblock_find_in_range()+memblock_reserve() and it's just
> a cleanup?  Or is there some deeper functional reason?

This is just a cleanup and I thought there are no functional change.
But I've just realized that the conversion to memblock_alloc_base() in
this patch changes the behaviour in the error case.
Because memblock_alloc_base calls panic if it can't allocate.

So please drop this patch from -mm tree for now.

> Does memblock_find_in_range() need to exist?  Can we convert all
> callers to memblock_alloc_range()?

There are two callsites where we can't simply convert with
memblock_alloc_range (arch/s390/kernel/setup.c, arch/x86/mm/init.c).

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

kmemleak: Cannot insert [...] into the object search tree (overlaps existing) (mm: use memblock_alloc_range())

[Sabrina Dubroca]

Hello,

2014-08-24, 23:56:03 +0900, Akinobu Mita wrote:
> Replace memblock_find_in_range() and memblock_reserve() with
> memblock_alloc_range().
> 
> Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
> Cc: linux-mm@kvack.org

This patch is included in linux-next, and when I boot next-20140901,
on a 32-bit build, I get this message:


kmemleak: Cannot insert 0xf6556000 into the object search tree (overlaps existing)
CPU: 0 PID: 0 Comm: swapper/0 Not tainted 3.17.0-rc3-next-20140901 #126
Hardware name: Dell Inc. Latitude D830                   /0UY141, BIOS A02 06/07/2007
 f6556000 00000000 c1891f64 c16226e8 f5c090c0 c1891f98 c11a934c c17f8768
 f6556000 00200282 f5c090e4 00000005 00000010 f5c09124 00000000 c19984e8
 00000002 c19f5800 c1891fb0 c162216d 00000020 c19984e8 00000002 c19f5800
Call Trace:
 [<c16226e8>] dump_stack+0x48/0x69
 [<c11a934c>] create_object+0x23c/0x290
 [<c162216d>] early_alloc+0x98/0x120
 [<c195b10d>] kmemleak_init+0x129/0x226
 [<c19399f7>] start_kernel+0x2d5/0x38d
 [<c19392ab>] i386_start_kernel+0x79/0x7d
kmemleak: Kernel memory leak detector disabled
kmemleak: Object 0xf6556000 (size 16777216):
kmemleak:   comm "swapper/0", pid 0, jiffies 4294877296
kmemleak:   min_count = 0
kmemleak:   count = 0
kmemleak:   flags = 0x1
kmemleak:   checksum = 0
kmemleak:   backtrace:
     [<c1620048>] kmemleak_alloc+0xa8/0xb0
     [<c19595bd>] memblock_alloc_range_nid+0x46/0x50
     [<c195965f>] memblock_virt_alloc_internal+0x89/0xe7
     [<c195978b>] memblock_virt_alloc_try_nid_nopanic+0x58/0x60
     [<c161fc2f>] alloc_node_mem_map.constprop.72+0x4b/0x8c
     [<c195623f>] free_area_init_node+0xee/0x3a1
     [<c1956860>] free_area_init_nodes+0x36e/0x380
     [<c194b8a5>] zone_sizes_init+0x33/0x39
     [<c194c112>] paging_init+0xaa/0xad
     [<c194c169>] native_pagetable_init+0x54/0xe7
     [<c193c9d9>] setup_arch+0xb21/0xc07
     [<c193979b>] start_kernel+0x79/0x38d
     [<c19392ab>] i386_start_kernel+0x79/0x7d
     [<ffffffff>] 0xffffffff
kmemleak: Early log backtrace:
   [<c1620048>] kmemleak_alloc+0xa8/0xb0
   [<c19596a2>] memblock_virt_alloc_internal+0xcc/0xe7
   [<c195978b>] memblock_virt_alloc_try_nid_nopanic+0x58/0x60
   [<c161fc2f>] alloc_node_mem_map.constprop.72+0x4b/0x8c
   [<c195623f>] free_area_init_node+0xee/0x3a1
   [<c1956860>] free_area_init_nodes+0x36e/0x380
   [<c194b8a5>] zone_sizes_init+0x33/0x39
   [<c194c112>] paging_init+0xaa/0xad
   [<c194c169>] native_pagetable_init+0x54/0xe7
   [<c193c9d9>] setup_arch+0xb21/0xc07
   [<c193979b>] start_kernel+0x79/0x38d
   [<c19392ab>] i386_start_kernel+0x79/0x7d
   [<ffffffff>] 0xffffffff



git bisect pointed to this patch:

abc65ff21e61d49269bf8fafd486fff2e3679c21 is the first bad commit
commit abc65ff21e61d49269bf8fafd486fff2e3679c21
Author: Akinobu Mita <akinobu.mita@gmail.com>
Date:   Mon Sep 1 23:48:54 2014 +0100

    mm: use memblock_alloc_range()
    
    Replace memblock_find_in_range() and memblock_reserve() with
    the equivalent memblock_alloc_range().
    
    Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
    Cc: Thomas Gleixner <tglx@linutronix.de>
    Cc: Ingo Molnar <mingo@redhat.com>
    Cc: "H. Peter Anvin" <hpa@zytor.com>
    Cc: Yinghai Lu <yinghai@kernel.org>
    Signed-off-by: Andrew Morton <akpm@linux-foundation.org>


## git bisect log

git bisect start
# good: [a3793b0cedfc0bc3212e5ebf5b79710c910687c4] Merge remote-tracking branch 'llvmlinux/for-next'
git bisect good a3793b0cedfc0bc3212e5ebf5b79710c910687c4
# bad: [03af78748485f63e8ed21d2e2585b5d1ec862ba6] Add linux-next specific files for 20140901
git bisect bad 03af78748485f63e8ed21d2e2585b5d1ec862ba6
# good: [42330182961e380c6ac85c1482ff8115ddc487dd] mempolicy: unexport get_vma_policy() and remove its "task" arg
git bisect good 42330182961e380c6ac85c1482ff8115ddc487dd
# bad: [2437e4e8841cafe9c086a98d4b0186196c7e10af] MAINTAINERS: remove non existent files
git bisect bad 2437e4e8841cafe9c086a98d4b0186196c7e10af
# bad: [36fb2fa2c928947f728dc8119a7143fe9f61033c] zsmalloc: change return value unit of zs_get_total_size_bytes
git bisect bad 36fb2fa2c928947f728dc8119a7143fe9f61033c
# bad: [658f7da49d34bc6187e6cd1ec57933d1a2a76035] mm: introduce dump_vma
git bisect bad 658f7da49d34bc6187e6cd1ec57933d1a2a76035
# skip: [2090938a202a34e6ea28a40a9b98214795546882] mm: introduce common page state for ballooned memory
git bisect skip 2090938a202a34e6ea28a40a9b98214795546882
# skip: [843cbba246f248585f88dffec89304545d2f3bde] mm-introduce-common-page-state-for-ballooned-memory-fix
git bisect skip 843cbba246f248585f88dffec89304545d2f3bde
# bad: [92a1357eacd714671071871e95bdaf9144aa622a] mm-balloon_compaction-general-cleanup-checkpatch-fixes
git bisect bad 92a1357eacd714671071871e95bdaf9144aa622a
# skip: [b19a479c8ae91329771288310701f996bc100947] selftests/vm/transhuge-stress: stress test for memory compaction
git bisect skip b19a479c8ae91329771288310701f996bc100947
# bad: [e0e398dffe88d10dcda4e41941d677aa337410e5] mm/balloon_compaction: ignore anonymous pages
git bisect bad e0e398dffe88d10dcda4e41941d677aa337410e5
# bad: [15ccb0a452bb0e2f0edb25747110aae73fd9a962] include/linux/migrate.h: remove migrate_page #define
git bisect bad 15ccb0a452bb0e2f0edb25747110aae73fd9a962
# bad: [abc65ff21e61d49269bf8fafd486fff2e3679c21] mm: use memblock_alloc_range()
git bisect bad abc65ff21e61d49269bf8fafd486fff2e3679c21
# first bad commit: [abc65ff21e61d49269bf8fafd486fff2e3679c21] mm: use memblock_alloc_range()



Thanks,

-- 
Sabrina

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: kmemleak: Cannot insert [...] into the object search tree (overlaps existing) (mm: use memblock_alloc_range())

[Akinobu Mita]

2014-09-03 2:10 GMT+09:00 Sabrina Dubroca <sd@queasysnail.net>:
> Hello,
>
> 2014-08-24, 23:56:03 +0900, Akinobu Mita wrote:
>> Replace memblock_find_in_range() and memblock_reserve() with
>> memblock_alloc_range().
>>
>> Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
>> Cc: linux-mm@kvack.org
>
> This patch is included in linux-next, and when I boot next-20140901,
> on a 32-bit build, I get this message:
>
>
> kmemleak: Cannot insert 0xf6556000 into the object search tree (overlaps existing)

kmemleak_alloc() in memblock_virt_alloc_internal() should have been removed by
the conversion in this patch.  Otherwise kmemleak() is called twice because
memblock_alloc_range() also calls it.

Andrew, could you drop this patch for now.  I'll send the patch with this fix.

Thanks for the report.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>