Merge branch 'akpm' (patches from Andrew)

b5d903c2 · Linus Torvalds <[email protected]> · 2018-06-15 08:51:42 +09:00

Merge more updates from Andrew Morton:

 - MM remainders

 - various misc things

 - kcov updates

* emailed patches from Andrew Morton <[email protected]>: (27 commits)
  lib/test_printf.c: call wait_for_random_bytes() before plain %p tests
  hexagon: drop the unused variable zero_page_mask
  hexagon: fix printk format warning in setup.c
  mm: fix oom_kill event handling
  treewide: use PHYS_ADDR_MAX to avoid type casting ULLONG_MAX
  mm: use octal not symbolic permissions
  ipc: use new return type vm_fault_t
  sysvipc/sem: mitigate semnum index against spectre v1
  fault-injection: reorder config entries
  arm: port KCOV to arm
  sched/core / kcov: avoid kcov_area during task switch
  kcov: prefault the kcov_area
  kcov: ensure irq code sees a valid area
  kernel/relay.c: change return type to vm_fault_t
  exofs: avoid VLA in structures
  coredump: fix spam with zero VMA process
  fat: use fat_fs_error() instead of BUG_ON() in __fat_get_block()
  proc: skip branch in /proc/*/* lookup
  mremap: remove LATENCY_LIMIT from mremap to reduce the number of TLB shootdowns
  mm/memblock: add missing include <linux/bootmem.h>
  ...

Commit: b5d903c2d656e9bc54bc76554a477d796a63120d
Parents: 7a932516 · ee410f15

Merge commit. Changes below are compared with the first parent: 7a932516.

Changes

56 files changed · +397 −212

File	Status	Add	Del	Content
arch/arm/Kconfig	modified	+2	−1	View file
arch/arm/boot/compressed/Makefile	modified	+3	−0	View file
arch/arm/kvm/hyp/Makefile	modified	+8	−0	View file
arch/arm/vdso/Makefile	modified	+3	−0	View file
arch/arm64/mm/init.c	modified	+3	−3	View file
arch/hexagon/include/asm/pgtable.h	modified	+0	−1	View file
arch/hexagon/kernel/setup.c	modified	+1	−1	View file
arch/hexagon/mm/init.c	modified	+0	−3	View file
arch/mips/kernel/setup.c	modified	+2	−2	View file
arch/powerpc/mm/mem.c	modified	+1	−1	View file
arch/sparc/mm/init_64.c	modified	+1	−1	View file
arch/x86/mm/init.c	modified	+3	−1	View file
arch/x86/mm/init_32.c	modified	+1	−1	View file
arch/x86/mm/init_64.c	modified	+1	−1	View file
drivers/firmware/efi/arm-init.c	modified	+1	−1	View file
drivers/remoteproc/qcom_q6v5_pil.c	modified	+1	−1	View file
drivers/soc/qcom/mdt_loader.c	modified	+2	−2	View file
fs/binfmt_elf.c	modified	+9	−8	View file
fs/exofs/ore.c	modified	+49	−35	View file
fs/exofs/ore_raid.c	modified	+55	−20	View file
fs/exofs/super.c	modified	+11	−12	View file
fs/fat/inode.c	modified	+7	−1	View file
fs/proc/base.c	modified	+3	−6	View file
include/linux/kcov.h	modified	+14	−0	View file
include/linux/memcontrol.h	modified	+22	−4	View file
include/linux/sched.h	modified	+1	−1	View file
include/linux/slab.h	modified	+1	−0	View file
ipc/sem.c	modified	+14	−4	View file
ipc/shm.c	modified	+1	−1	View file
kernel/fork.c	modified	+8	−0	View file
kernel/kcov.c	modified	+19	−2	View file
kernel/kexec_core.c	modified	+4	−0	View file
kernel/relay.c	modified	+1	−1	View file
kernel/sched/core.c	modified	+4	−0	View file
lib/Kconfig.debug	modified	+18	−18	View file
lib/test_printf.c	modified	+7	−0	View file
mm/cleancache.c	modified	+4	−6	View file
mm/cma_debug.c	modified	+10	−15	View file
mm/compaction.c	modified	+1	−1	View file
mm/dmapool.c	modified	+1	−1	View file
mm/failslab.c	modified	+1	−1	View file
mm/frontswap.c	modified	+5	−6	View file
mm/ksm.c	modified	+10	−4	View file
mm/memblock.c	modified	+7	−3	View file
mm/memcontrol.c	modified	+8	−2	View file
mm/mremap.c	modified	+0	−4	View file
mm/oom_kill.c	modified	+1	−1	View file
mm/page_alloc.c	modified	+1	−1	View file
mm/page_idle.c	modified	+1	−1	View file
mm/page_owner.c	modified	+2	−2	View file
mm/shmem.c	modified	+5	−4	View file
mm/slab_common.c	modified	+34	−3	View file
mm/swapfile.c	modified	+1	−1	View file
mm/vmalloc.c	modified	+2	−2	View file
mm/zsmalloc.c	modified	+3	−2	View file
mm/zswap.c	modified	+19	−19	View file

Only the first 600 diff lines are shown. Download the complete patch.

arch/arm/Kconfig

View file

diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig
index 0be4397..483d285 100644
--- a/arch/arm/Kconfig
+++ b/arch/arm/Kconfig
@@ -8,9 +8,10 @@ config ARM
 	select ARCH_HAS_DEVMEM_IS_ALLOWED
 	select ARCH_HAS_ELF_RANDOMIZE
 	select ARCH_HAS_FORTIFY_SOURCE
+	select ARCH_HAS_KCOV
 	select ARCH_HAS_PTE_SPECIAL if ARM_LPAE
-	select ARCH_HAS_SET_MEMORY
 	select ARCH_HAS_PHYS_TO_DMA
+	select ARCH_HAS_SET_MEMORY
 	select ARCH_HAS_STRICT_KERNEL_RWX if MMU && !XIP_KERNEL
 	select ARCH_HAS_STRICT_MODULE_RWX if MMU
 	select ARCH_HAS_TICK_BROADCAST if GENERIC_CLOCKEVENTS_BROADCAST

arch/arm/boot/compressed/Makefile

View file

diff --git a/arch/arm/boot/compressed/Makefile b/arch/arm/boot/compressed/Makefile
index a3c5fbc..1f5a5ff 100644
--- a/arch/arm/boot/compressed/Makefile
+++ b/arch/arm/boot/compressed/Makefile
@@ -25,6 +25,9 @@ endif
 
 GCOV_PROFILE		:= n
 
+# Prevents link failures: __sanitizer_cov_trace_pc() is not linked in.
+KCOV_INSTRUMENT		:= n
+
 #
 # Architecture dependencies
 #

arch/arm/kvm/hyp/Makefile

View file

diff --git a/arch/arm/kvm/hyp/Makefile b/arch/arm/kvm/hyp/Makefile
index 7fc0638..d2b5ec9 100644
--- a/arch/arm/kvm/hyp/Makefile
+++ b/arch/arm/kvm/hyp/Makefile
@@ -23,3 +23,11 @@ obj-$(CONFIG_KVM_ARM_HOST) += hyp-entry.o
 obj-$(CONFIG_KVM_ARM_HOST) += switch.o
 CFLAGS_switch.o		   += $(CFLAGS_ARMV7VE)
 obj-$(CONFIG_KVM_ARM_HOST) += s2-setup.o
+
+# KVM code is run at a different exception code with a different map, so
+# compiler instrumentation that inserts callbacks or checks into the code may
+# cause crashes. Just disable it.
+GCOV_PROFILE	:= n
+KASAN_SANITIZE	:= n
+UBSAN_SANITIZE	:= n
+KCOV_INSTRUMENT	:= n

arch/arm/vdso/Makefile

View file

diff --git a/arch/arm/vdso/Makefile b/arch/arm/vdso/Makefile
index bb41182..f4efff9 100644
--- a/arch/arm/vdso/Makefile
+++ b/arch/arm/vdso/Makefile
@@ -30,6 +30,9 @@ CFLAGS_vgettimeofday.o = -O2
 # Disable gcov profiling for VDSO code
 GCOV_PROFILE := n
 
+# Prevents link failures: __sanitizer_cov_trace_pc() is not linked in.
+KCOV_INSTRUMENT := n
+
 # Force dependency
 $(obj)/vdso.o : $(obj)/vdso.so
 

arch/arm64/mm/init.c

View file

diff --git a/arch/arm64/mm/init.c b/arch/arm64/mm/init.c
index 1b18b47..325cfb3 100644
--- a/arch/arm64/mm/init.c
+++ b/arch/arm64/mm/init.c
@@ -310,7 +310,7 @@ static void __init arm64_memory_present(void)
 }
 #endif
 
-static phys_addr_t memory_limit = (phys_addr_t)ULLONG_MAX;
+static phys_addr_t memory_limit = PHYS_ADDR_MAX;
 
 /*
  * Limit the memory size that was specified via FDT.
@@ -401,7 +401,7 @@ void __init arm64_memblock_init(void)
 	 * high up in memory, add back the kernel region that must be accessible
 	 * via the linear mapping.
 	 */
-	if (memory_limit != (phys_addr_t)ULLONG_MAX) {
+	if (memory_limit != PHYS_ADDR_MAX) {
 		memblock_mem_limit_remove_map(memory_limit);
 		memblock_add(__pa_symbol(_text), (u64)(_end - _text));
 	}
@@ -666,7 +666,7 @@ __setup("keepinitrd", keepinitrd_setup);
  */
 static int dump_mem_limit(struct notifier_block *self, unsigned long v, void *p)
 {
-	if (memory_limit != (phys_addr_t)ULLONG_MAX) {
+	if (memory_limit != PHYS_ADDR_MAX) {
 		pr_emerg("Memory Limit: %llu MB\n", memory_limit >> 20);
 	} else {
 		pr_emerg("Memory Limit: none\n");

arch/hexagon/include/asm/pgtable.h

View file

diff --git a/arch/hexagon/include/asm/pgtable.h b/arch/hexagon/include/asm/pgtable.h
index aef02f7..65125d0 100644
--- a/arch/hexagon/include/asm/pgtable.h
+++ b/arch/hexagon/include/asm/pgtable.h
@@ -30,7 +30,6 @@
 
 /* A handy thing to have if one has the RAM. Declared in head.S */
 extern unsigned long empty_zero_page;
-extern unsigned long zero_page_mask;
 
 /*
  * The PTE model described here is that of the Hexagon Virtual Machine,

arch/hexagon/kernel/setup.c

View file

diff --git a/arch/hexagon/kernel/setup.c b/arch/hexagon/kernel/setup.c
index 6981949..dc8c7e7 100644
--- a/arch/hexagon/kernel/setup.c
+++ b/arch/hexagon/kernel/setup.c
@@ -66,7 +66,7 @@ void __init setup_arch(char **cmdline_p)
 	 */
 	__vmsetvec(_K_VM_event_vector);
 
-	printk(KERN_INFO "PHYS_OFFSET=0x%08x\n", PHYS_OFFSET);
+	printk(KERN_INFO "PHYS_OFFSET=0x%08lx\n", PHYS_OFFSET);
 
 	/*
 	 * Simulator has a few differences from the hardware.

arch/hexagon/mm/init.c

View file

diff --git a/arch/hexagon/mm/init.c b/arch/hexagon/mm/init.c
index 192584d..1495d45 100644
--- a/arch/hexagon/mm/init.c
+++ b/arch/hexagon/mm/init.c
@@ -39,9 +39,6 @@ unsigned long __phys_offset;	/*  physical kernel offset >> 12  */
 /*  Set as variable to limit PMD copies  */
 int max_kernel_seg = 0x303;
 
-/*  think this should be (page_size-1) the way it's used...*/
-unsigned long zero_page_mask;
-
 /*  indicate pfn's of high memory  */
 unsigned long highstart_pfn, highend_pfn;
 

arch/mips/kernel/setup.c

View file

diff --git a/arch/mips/kernel/setup.c b/arch/mips/kernel/setup.c
index 563188a..2c96c0c 100644
--- a/arch/mips/kernel/setup.c
+++ b/arch/mips/kernel/setup.c
@@ -93,7 +93,7 @@ void __init add_memory_region(phys_addr_t start, phys_addr_t size, long type)
 	 * If the region reaches the top of the physical address space, adjust
 	 * the size slightly so that (start + size) doesn't overflow
 	 */
-	if (start + size - 1 == (phys_addr_t)ULLONG_MAX)
+	if (start + size - 1 == PHYS_ADDR_MAX)
 		--size;
 
 	/* Sanity check */
@@ -376,7 +376,7 @@ static void __init bootmem_init(void)
 	unsigned long reserved_end;
 	unsigned long mapstart = ~0UL;
 	unsigned long bootmap_size;
-	phys_addr_t ramstart = (phys_addr_t)ULLONG_MAX;
+	phys_addr_t ramstart = PHYS_ADDR_MAX;
 	bool bootmap_valid = false;
 	int i;
 

arch/powerpc/mm/mem.c

View file

diff --git a/arch/powerpc/mm/mem.c b/arch/powerpc/mm/mem.c
index 8cecda4..5c8530d 100644
--- a/arch/powerpc/mm/mem.c
+++ b/arch/powerpc/mm/mem.c
@@ -215,7 +215,7 @@ void __init mem_topology_setup(void)
 	/* Place all memblock_regions in the same node and merge contiguous
 	 * memblock_regions
 	 */
-	memblock_set_node(0, (phys_addr_t)ULLONG_MAX, &memblock.memory, 0);
+	memblock_set_node(0, PHYS_ADDR_MAX, &memblock.memory, 0);
 }
 
 void __init initmem_init(void)

arch/sparc/mm/init_64.c

View file

diff --git a/arch/sparc/mm/init_64.c b/arch/sparc/mm/init_64.c
index 8aeb1aa..f396048 100644
--- a/arch/sparc/mm/init_64.c
+++ b/arch/sparc/mm/init_64.c
@@ -1620,7 +1620,7 @@ static void __init bootmem_init_nonnuma(void)
 	       (top_of_ram - total_ram) >> 20);
 
 	init_node_masks_nonnuma();
-	memblock_set_node(0, (phys_addr_t)ULLONG_MAX, &memblock.memory, 0);
+	memblock_set_node(0, PHYS_ADDR_MAX, &memblock.memory, 0);
 	allocate_node_data(0);
 	node_set_online(0);
 }

arch/x86/mm/init.c

View file

diff --git a/arch/x86/mm/init.c b/arch/x86/mm/init.c
index fec82b5..cee58a9 100644
--- a/arch/x86/mm/init.c
+++ b/arch/x86/mm/init.c
@@ -706,7 +706,9 @@ void __init init_mem_mapping(void)
  */
 int devmem_is_allowed(unsigned long pagenr)
 {
-	if (page_is_ram(pagenr)) {
+	if (region_intersects(PFN_PHYS(pagenr), PAGE_SIZE,
+				IORESOURCE_SYSTEM_RAM, IORES_DESC_NONE)
+			!= REGION_DISJOINT) {
 		/*
 		 * For disallowed memory regions in the low 1MB range,
 		 * request that the page be shown as all zeros.

arch/x86/mm/init_32.c

View file

diff --git a/arch/x86/mm/init_32.c b/arch/x86/mm/init_32.c
index c893c6a..979e0a0 100644
--- a/arch/x86/mm/init_32.c
+++ b/arch/x86/mm/init_32.c
@@ -692,7 +692,7 @@ void __init initmem_init(void)
 	high_memory = (void *) __va(max_low_pfn * PAGE_SIZE - 1) + 1;
 #endif
 
-	memblock_set_node(0, (phys_addr_t)ULLONG_MAX, &memblock.memory, 0);
+	memblock_set_node(0, PHYS_ADDR_MAX, &memblock.memory, 0);
 	sparse_memory_present_with_active_regions(0);
 
 #ifdef CONFIG_FLATMEM

arch/x86/mm/init_64.c

View file

diff --git a/arch/x86/mm/init_64.c b/arch/x86/mm/init_64.c
index 17383f9..045f492 100644
--- a/arch/x86/mm/init_64.c
+++ b/arch/x86/mm/init_64.c
@@ -742,7 +742,7 @@ kernel_physical_mapping_init(unsigned long paddr_start,
 #ifndef CONFIG_NUMA
 void __init initmem_init(void)
 {
-	memblock_set_node(0, (phys_addr_t)ULLONG_MAX, &memblock.memory, 0);
+	memblock_set_node(0, PHYS_ADDR_MAX, &memblock.memory, 0);
 }
 #endif
 

drivers/firmware/efi/arm-init.c

View file

diff --git a/drivers/firmware/efi/arm-init.c b/drivers/firmware/efi/arm-init.c
index 80d1a88..b5214c1 100644
--- a/drivers/firmware/efi/arm-init.c
+++ b/drivers/firmware/efi/arm-init.c
@@ -193,7 +193,7 @@ static __init void reserve_regions(void)
 	 * uses its own memory map instead.
 	 */
 	memblock_dump_all();
-	memblock_remove(0, (phys_addr_t)ULLONG_MAX);
+	memblock_remove(0, PHYS_ADDR_MAX);
 
 	for_each_efi_memory_desc(md) {
 		paddr = md->phys_addr;

drivers/remoteproc/qcom_q6v5_pil.c

View file

diff --git a/drivers/remoteproc/qcom_q6v5_pil.c b/drivers/remoteproc/qcom_q6v5_pil.c
index 2190deb..2bf8e7c 100644
--- a/drivers/remoteproc/qcom_q6v5_pil.c
+++ b/drivers/remoteproc/qcom_q6v5_pil.c
@@ -686,7 +686,7 @@ static int q6v5_mpss_load(struct q6v5 *qproc)
 	struct elf32_hdr *ehdr;
 	phys_addr_t mpss_reloc;
 	phys_addr_t boot_addr;
-	phys_addr_t min_addr = (phys_addr_t)ULLONG_MAX;
+	phys_addr_t min_addr = PHYS_ADDR_MAX;
 	phys_addr_t max_addr = 0;
 	bool relocate = false;
 	char seg_name[10];

drivers/soc/qcom/mdt_loader.c

View file

diff --git a/drivers/soc/qcom/mdt_loader.c b/drivers/soc/qcom/mdt_loader.c
index 17b314d..dc09d7a 100644
--- a/drivers/soc/qcom/mdt_loader.c
+++ b/drivers/soc/qcom/mdt_loader.c
@@ -50,7 +50,7 @@ ssize_t qcom_mdt_get_size(const struct firmware *fw)
 	const struct elf32_phdr *phdrs;
 	const struct elf32_phdr *phdr;
 	const struct elf32_hdr *ehdr;
-	phys_addr_t min_addr = (phys_addr_t)ULLONG_MAX;
+	phys_addr_t min_addr = PHYS_ADDR_MAX;
 	phys_addr_t max_addr = 0;
 	int i;
 
@@ -97,7 +97,7 @@ int qcom_mdt_load(struct device *dev, const struct firmware *fw,
 	const struct elf32_hdr *ehdr;
 	const struct firmware *seg_fw;
 	phys_addr_t mem_reloc;
-	phys_addr_t min_addr = (phys_addr_t)ULLONG_MAX;
+	phys_addr_t min_addr = PHYS_ADDR_MAX;
 	phys_addr_t max_addr = 0;
 	size_t fw_name_len;
 	ssize_t offset;

fs/binfmt_elf.c

View file

diff --git a/fs/binfmt_elf.c b/fs/binfmt_elf.c
index 070b618..0ac456b 100644
--- a/fs/binfmt_elf.c
+++ b/fs/binfmt_elf.c
@@ -1621,8 +1621,8 @@ static int fill_files_note(struct memelfnote *note)
 	if (size >= MAX_FILE_NOTE_SIZE) /* paranoia check */
 		return -EINVAL;
 	size = round_up(size, PAGE_SIZE);
-	data = vmalloc(size);
-	if (!data)
+	data = kvmalloc(size, GFP_KERNEL);
+	if (ZERO_OR_NULL_PTR(data))
 		return -ENOMEM;
 
 	start_end_ofs = data + 2;
@@ -1639,7 +1639,7 @@ static int fill_files_note(struct memelfnote *note)
 		filename = file_path(file, name_curpos, remaining);
 		if (IS_ERR(filename)) {
 			if (PTR_ERR(filename) == -ENAMETOOLONG) {
-				vfree(data);
+				kvfree(data);
 				size = size * 5 / 4;
 				goto alloc;
 			}
@@ -1932,7 +1932,7 @@ static void free_note_info(struct elf_note_info *info)
 		kfree(t);
 	}
 	kfree(info->psinfo.data);
-	vfree(info->files.data);
+	kvfree(info->files.data);
 }
 
 #else
@@ -2148,7 +2148,7 @@ static void free_note_info(struct elf_note_info *info)
 
 	/* Free data possibly allocated by fill_files_note(): */
 	if (info->notes_files)
-		vfree(info->notes_files->data);
+		kvfree(info->notes_files->data);
 
 	kfree(info->prstatus);
 	kfree(info->psinfo);
@@ -2294,8 +2294,9 @@ static int elf_core_dump(struct coredump_params *cprm)
 
 	if (segs - 1 > ULONG_MAX / sizeof(*vma_filesz))
 		goto end_coredump;
-	vma_filesz = vmalloc(array_size(sizeof(*vma_filesz), (segs - 1)));
-	if (!vma_filesz)
+	vma_filesz = kvmalloc(array_size(sizeof(*vma_filesz), (segs - 1)),
+			      GFP_KERNEL);
+	if (ZERO_OR_NULL_PTR(vma_filesz))
 		goto end_coredump;
 
 	for (i = 0, vma = first_vma(current, gate_vma); vma != NULL;
@@ -2402,7 +2403,7 @@ end_coredump:
 cleanup:
 	free_note_info(&info);
 	kfree(shdr4extnum);
-	vfree(vma_filesz);
+	kvfree(vma_filesz);
 	kfree(phdr4note);
 	kfree(elf);
 out:

fs/exofs/ore.c

View file

diff --git a/fs/exofs/ore.c b/fs/exofs/ore.c
index ddbf872..1b8b446 100644
--- a/fs/exofs/ore.c
+++ b/fs/exofs/ore.c
@@ -146,68 +146,82 @@ int  _ore_get_io_state(struct ore_layout *layout,
 			struct ore_io_state **pios)
 {
 	struct ore_io_state *ios;
-	struct page **pages;
-	struct osd_sg_entry *sgilist;
+	size_t size_ios, size_extra, size_total;
+	void *ios_extra;
+
+	/*
+	 * The desired layout looks like this, with the extra_allocation
+	 * items pointed at from fields within ios or per_dev:
+
 	struct __alloc_all_io_state {
 		struct ore_io_state ios;
 		struct ore_per_dev_state per_dev[numdevs];
 		union {
 			struct osd_sg_entry sglist[sgs_per_dev * numdevs];
 			struct page *pages[num_par_pages];
-		};
-	} *_aios;
-
-	if (likely(sizeof(*_aios) <= PAGE_SIZE)) {
-		_aios = kzalloc(sizeof(*_aios), GFP_KERNEL);
-		if (unlikely(!_aios)) {
-			ORE_DBGMSG("Failed kzalloc bytes=%zd\n",
-				   sizeof(*_aios));
+		} extra_allocation;
+	} whole_allocation;
+
+	*/
+
+	/* This should never happen, so abort early if it ever does. */
+	if (sgs_per_dev && num_par_pages) {
+		ORE_DBGMSG("Tried to use both pages and sglist\n");
+		*pios = NULL;
+		return -EINVAL;
+	}
+
+	if (numdevs > (INT_MAX - sizeof(*ios)) /
+		       sizeof(struct ore_per_dev_state))
+		return -ENOMEM;
+	size_ios = sizeof(*ios) + sizeof(struct ore_per_dev_state) * numdevs;
+
+	if (sgs_per_dev * numdevs > INT_MAX / sizeof(struct osd_sg_entry))
+		return -ENOMEM;
+	if (num_par_pages > INT_MAX / sizeof(struct page *))
+		return -ENOMEM;
+	size_extra = max(sizeof(struct osd_sg_entry) * (sgs_per_dev * numdevs),
+			 sizeof(struct page *) * num_par_pages);
+
+	size_total = size_ios + size_extra;
+
+	if (likely(size_total <= PAGE_SIZE)) {
+		ios = kzalloc(size_total, GFP_KERNEL);
+		if (unlikely(!ios)) {
+			ORE_DBGMSG("Failed kzalloc bytes=%zd\n", size_total);
 			*pios = NULL;
 			return -ENOMEM;
 		}
-		pages = num_par_pages ? _aios->pages : NULL;
-		sgilist = sgs_per_dev ? _aios->sglist : NULL;
-		ios = &_aios->ios;
+		ios_extra = (char *)ios + size_ios;
 	} else {
-		struct __alloc_small_io_state {
-			struct ore_io_state ios;
-			struct ore_per_dev_state per_dev[numdevs];
-		} *_aio_small;
-		union __extra_part {
-			struct osd_sg_entry sglist[sgs_per_dev * numdevs];
-			struct page *pages[num_par_pages];
-		} *extra_part;
-
-		_aio_small = kzalloc(sizeof(*_aio_small), GFP_KERNEL);
-		if (unlikely(!_aio_small)) {
+		ios = kzalloc(size_ios, GFP_KERNEL);
+		if (unlikely(!ios)) {
 			ORE_DBGMSG("Failed alloc first part bytes=%zd\n",
-				   sizeof(*_aio_small));
+				   size_ios);
 			*pios = NULL;
 			return -ENOMEM;
 		}
-		extra_part = kzalloc(sizeof(*extra_part), GFP_KERNEL);
-		if (unlikely(!extra_part)) {
+		ios_extra = kzalloc(size_extra, GFP_KERNEL);
+		if (unlikely(!ios_extra)) {
 			ORE_DBGMSG("Failed alloc second part bytes=%zd\n",
-				   sizeof(*extra_part));
-			kfree(_aio_small);
+				   size_extra);
+			kfree(ios);
 			*pios = NULL;
 			return -ENOMEM;
 		}
 
-		pages = num_par_pages ? extra_part->pages : NULL;
-		sgilist = sgs_per_dev ? extra_part->sglist : NULL;
 		/* In this case the per_dev[0].sgilist holds the pointer to
 		 * be freed
 		 */
-		ios = &_aio_small->ios;
 		ios->extra_part_alloc = true;
 	}
 
-	if (pages) {
-		ios->parity_pages = pages;
+	if (num_par_pages) {
+		ios->parity_pages = ios_extra;
 		ios->max_par_pages = num_par_pages;
 	}
-	if (sgilist) {
+	if (sgs_per_dev) {
+		struct osd_sg_entry *sgilist = ios_extra;
 		unsigned d;
 
 		for (d = 0; d < numdevs; ++d) {

fs/exofs/ore_raid.c

View file

diff --git a/fs/exofs/ore_raid.c b/fs/exofs/ore_raid.c
index 27cbdb6..199590f 100644
--- a/fs/exofs/ore_raid.c
+++ b/fs/exofs/ore_raid.c
@@ -71,6 +71,11 @@ static int _sp2d_alloc(unsigned pages_in_unit, unsigned group_width,
 {
 	struct __stripe_pages_2d *sp2d;
 	unsigned data_devs = group_width - parity;
+
+	/*
+	 * Desired allocation layout is, though when larger than PAGE_SIZE,
+	 * each struct __alloc_1p_arrays is separately allocated:
+
 	struct _alloc_all_bytes {
 		struct __alloc_stripe_pages_2d {
 			struct __stripe_pages_2d sp2d;
@@ -82,55 +87,85 @@ static int _sp2d_alloc(unsigned pages_in_unit, unsigned group_width,
 			char page_is_read[data_devs];
 		} __a1pa[pages_in_unit];
 	} *_aab;
+
 	struct __alloc_1p_arrays *__a1pa;
 	struct __alloc_1p_arrays *__a1pa_end;
-	const unsigned sizeof__a1pa = sizeof(_aab->__a1pa[0]);
+
+	*/
+
+	char *__a1pa;
+	char *__a1pa_end;
+
+	const size_t sizeof_stripe_pages_2d =
+		sizeof(struct __stripe_pages_2d) +
+		sizeof(struct __1_page_stripe) * pages_in_unit;
+	const size_t sizeof__a1pa =
+		ALIGN(sizeof(struct page *) * (2 * group_width) + data_devs,
+		      sizeof(void *));
+	const size_t sizeof__a1pa_arrays = sizeof__a1pa * pages_in_unit;
+	const size_t alloc_total = sizeof_stripe_pages_2d +
+				   sizeof__a1pa_arrays;
+
 	unsigned num_a1pa, alloc_size, i;
 
 	/* FIXME: check these numbers in ore_verify_layout */
-	BUG_ON(sizeof(_aab->__asp2d) > PAGE_SIZE);
+	BUG_ON(sizeof_stripe_pages_2d > PAGE_SIZE);
 	BUG_ON(sizeof__a1pa > PAGE_SIZE);
 
-	if (sizeof(*_aab) > PAGE_SIZE) {
-		num_a1pa = (PAGE_SIZE - sizeof(_aab->__asp2d)) / sizeof__a1pa;
-		alloc_size = sizeof(_aab->__asp2d) + sizeof__a1pa * num_a1pa;
+	/*
+	 * If alloc_total would be larger than PAGE_SIZE, only allocate
+	 * as many a1pa items as would fill the rest of the page, instead
+	 * of the full pages_in_unit count.
+	 */
+	if (alloc_total > PAGE_SIZE) {
+		num_a1pa = (PAGE_SIZE - sizeof_stripe_pages_2d) / sizeof__a1pa;
+		alloc_size = sizeof_stripe_pages_2d + sizeof__a1pa * num_a1pa;
 	} else {
 		num_a1pa = pages_in_unit;
-		alloc_size = sizeof(*_aab);
+		alloc_size = alloc_total;
 	}
 
-	_aab = kzalloc(alloc_size, GFP_KERNEL);
-	if (unlikely(!_aab)) {
+	*psp2d = sp2d = kzalloc(alloc_size, GFP_KERNEL);
+	if (unlikely(!sp2d)) {
 		ORE_DBGMSG("!! Failed to alloc sp2d size=%d\n", alloc_size);
 		return -ENOMEM;
 	}
+	/* From here Just call _sp2d_free */
 
-	sp2d = &_aab->__asp2d.sp2d;
-	*psp2d = sp2d; /* From here Just call _sp2d_free */
-
-	__a1pa = _aab->__a1pa;
-	__a1pa_end = __a1pa + num_a1pa;
+	/* Find start of a1pa area. */
+	__a1pa = (char *)sp2d + sizeof_stripe_pages_2d;
+	/* Find end of the _allocated_ a1pa area. */
+	__a1pa_end = __a1pa + alloc_size;
 
+	/* Allocate additionally needed a1pa items in PAGE_SIZE chunks. */
 	for (i = 0; i < pages_in_unit; ++i) {
+		struct __1_page_stripe *stripe = &sp2d->_1p_stripes[i];
+
 		if (unlikely(__a1pa >= __a1pa_end)) {
 			num_a1pa = min_t(unsigned, PAGE_SIZE / sizeof__a1pa,
 							pages_in_unit - i);
+			alloc_size = sizeof__a1pa * num_a1pa;
 
-			__a1pa = kcalloc(num_a1pa, sizeof__a1pa, GFP_KERNEL);
+			__a1pa = kzalloc(alloc_size, GFP_KERNEL);
 			if (unlikely(!__a1pa)) {
 				ORE_DBGMSG("!! Failed to _alloc_1p_arrays=%d\n",
 					   num_a1pa);
 				return -ENOMEM;
 			}
-			__a1pa_end = __a1pa + num_a1pa;
+			__a1pa_end = __a1pa + alloc_size;
 			/* First *pages is marked for kfree of the buffer */
-			sp2d->_1p_stripes[i].alloc = true;
+			stripe->alloc = true;
 		}
 
-		sp2d->_1p_stripes[i].pages = __a1pa->pages;
-		sp2d->_1p_stripes[i].scribble = __a1pa->scribble ;
-		sp2d->_1p_stripes[i].page_is_read = __a1pa->page_is_read;
-		++__a1pa;
+		/*
+		 * Attach all _lp_stripes pointers to the allocation for
+		 * it which was either part of the original PAGE_SIZE
+		 * allocation or the subsequent allocation in this loop.
+		 */
+		stripe->pages = (void *)__a1pa;
+		stripe->scribble = stripe->pages + group_width;
+		stripe->page_is_read = (char *)stripe->scribble + group_width;
+		__a1pa += sizeof__a1pa;
 	}
 
 	sp2d->parity = parity;

fs/exofs/super.c

View file

diff --git a/fs/exofs/super.c b/fs/exofs/super.c
index 719a315..41cf2fb 100644
--- a/fs/exofs/super.c
+++ b/fs/exofs/super.c
@@ -549,27 +549,26 @@ static int exofs_devs_2_odi(struct exofs_dt_device_info *dt_dev,
 static int __alloc_dev_table(struct exofs_sb_info *sbi, unsigned numdevs,
 		      struct exofs_dev **peds)
 {
-	struct __alloc_ore_devs_and_exofs_devs {
-		/* Twice bigger table: See exofs_init_comps() and comment at
-		 * exofs_read_lookup_dev_table()
-		 */
-		struct ore_dev *oreds[numdevs * 2 - 1];
-		struct exofs_dev eds[numdevs];
-	} *aoded;
+	/* Twice bigger table: See exofs_init_comps() and comment at
+	 * exofs_read_lookup_dev_table()
+	 */
+	const size_t numores = numdevs * 2 - 1;
 	struct exofs_dev *eds;
 	unsigned i;
 
-	aoded = kzalloc(sizeof(*aoded), GFP_KERNEL);
-	if (unlikely(!aoded)) {
+	sbi->oc.ods = kzalloc(numores * sizeof(struct ore_dev *) +
+			      numdevs * sizeof(struct exofs_dev), GFP_KERNEL);