From d9cbe09d39aa13f6924dc5fb88325de7ef01a72e Mon Sep 17 00:00:00 2001 From: Dan Williams Date: Thu, 3 Mar 2016 09:14:36 -0800 Subject: libnvdimm, pmem: fix 'pfn' support for section-misaligned namespaces The altmap for a section-misaligned namespace needs to arrange for the base_pfn to be section-aligned. As a result the 'reserve' region (pfns from base that do not have a struct page) must be increased. Otherwise we trip the altmap validation check in __add_pages: if (altmap->base_pfn != phys_start_pfn || vmem_altmap_offset(altmap) > nr_pages) { pr_warn_once("memory add fail, invalid altmap\n"); return -EINVAL; } Signed-off-by: Dan Williams --- drivers/nvdimm/pmem.c | 24 ++++++++++++++++++++++-- 1 file changed, 22 insertions(+), 2 deletions(-) (limited to 'drivers/nvdimm/pmem.c') diff --git a/drivers/nvdimm/pmem.c b/drivers/nvdimm/pmem.c index 8d0b54670184..59d568ab7556 100644 --- a/drivers/nvdimm/pmem.c +++ b/drivers/nvdimm/pmem.c @@ -356,6 +356,26 @@ static int nvdimm_namespace_detach_pfn(struct nd_namespace_common *ndns) return 0; } +/* + * We hotplug memory at section granularity, pad the reserved area from + * the previous section base to the namespace base address. + */ +static unsigned long init_altmap_base(resource_size_t base) +{ + unsigned long base_pfn = __phys_to_pfn(base); + + return PFN_SECTION_ALIGN_DOWN(base_pfn); +} + +static unsigned long init_altmap_reserve(resource_size_t base) +{ + unsigned long reserve = __phys_to_pfn(SZ_8K); + unsigned long base_pfn = __phys_to_pfn(base); + + reserve += base_pfn - PFN_SECTION_ALIGN_DOWN(base_pfn); + return reserve; +} + static int nvdimm_namespace_attach_pfn(struct nd_namespace_common *ndns) { struct nd_namespace_io *nsio = to_nd_namespace_io(&ndns->dev); @@ -369,8 +389,8 @@ static int nvdimm_namespace_attach_pfn(struct nd_namespace_common *ndns) phys_addr_t offset; int rc; struct vmem_altmap __altmap = { - .base_pfn = __phys_to_pfn(nsio->res.start), - .reserve = __phys_to_pfn(SZ_8K), + .base_pfn = init_altmap_base(nsio->res.start), + .reserve = init_altmap_reserve(nsio->res.start), }; if (!nd_pfn->uuid || !nd_pfn->ndns) -- cgit v1.2.1 From cfe30b872058f211630eda7f65fb19d83beaaa3c Mon Sep 17 00:00:00 2001 From: Dan Williams Date: Thu, 3 Mar 2016 09:38:00 -0800 Subject: libnvdimm, pmem: adjust for section collisions with 'System RAM' On a platform where 'Persistent Memory' and 'System RAM' are mixed within a given sparsemem section, trim the namespace and notify about the sub-optimal alignment. Cc: Toshi Kani Cc: Ross Zwisler Signed-off-by: Dan Williams --- drivers/nvdimm/pmem.c | 125 ++++++++++++++++++++++++++++++++++++-------------- 1 file changed, 91 insertions(+), 34 deletions(-) (limited to 'drivers/nvdimm/pmem.c') diff --git a/drivers/nvdimm/pmem.c b/drivers/nvdimm/pmem.c index 59d568ab7556..0cb450e1b400 100644 --- a/drivers/nvdimm/pmem.c +++ b/drivers/nvdimm/pmem.c @@ -43,7 +43,10 @@ struct pmem_device { phys_addr_t data_offset; u64 pfn_flags; void __pmem *virt_addr; + /* immutable base size of the namespace */ size_t size; + /* trim size when namespace capacity has been section aligned */ + u32 pfn_pad; struct badblocks bb; }; @@ -145,7 +148,7 @@ static long pmem_direct_access(struct block_device *bdev, sector_t sector, *kaddr = pmem->virt_addr + offset; *pfn = phys_to_pfn_t(pmem->phys_addr + offset, pmem->pfn_flags); - return pmem->size - offset; + return pmem->size - pmem->pfn_pad - offset; } static const struct block_device_operations pmem_fops = { @@ -236,7 +239,8 @@ static int pmem_attach_disk(struct device *dev, disk->flags = GENHD_FL_EXT_DEVT; nvdimm_namespace_disk_name(ndns, disk->disk_name); disk->driverfs_dev = dev; - set_capacity(disk, (pmem->size - pmem->data_offset) / 512); + set_capacity(disk, (pmem->size - pmem->pfn_pad - pmem->data_offset) + / 512); pmem->pmem_disk = disk; devm_exit_badblocks(dev, &pmem->bb); if (devm_init_badblocks(dev, &pmem->bb)) @@ -279,6 +283,9 @@ static int nd_pfn_init(struct nd_pfn *nd_pfn) struct nd_pfn_sb *pfn_sb = kzalloc(sizeof(*pfn_sb), GFP_KERNEL); struct pmem_device *pmem = dev_get_drvdata(&nd_pfn->dev); struct nd_namespace_common *ndns = nd_pfn->ndns; + u32 start_pad = 0, end_trunc = 0; + resource_size_t start, size; + struct nd_namespace_io *nsio; struct nd_region *nd_region; unsigned long npfns; phys_addr_t offset; @@ -304,21 +311,56 @@ static int nd_pfn_init(struct nd_pfn *nd_pfn) } memset(pfn_sb, 0, sizeof(*pfn_sb)); - npfns = (pmem->size - SZ_8K) / SZ_4K; + + /* + * Check if pmem collides with 'System RAM' when section aligned and + * trim it accordingly + */ + nsio = to_nd_namespace_io(&ndns->dev); + start = PHYS_SECTION_ALIGN_DOWN(nsio->res.start); + size = resource_size(&nsio->res); + if (region_intersects(start, size, IORESOURCE_SYSTEM_RAM, + IORES_DESC_NONE) == REGION_MIXED) { + + start = nsio->res.start; + start_pad = PHYS_SECTION_ALIGN_UP(start) - start; + } + + start = nsio->res.start; + size = PHYS_SECTION_ALIGN_UP(start + size) - start; + if (region_intersects(start, size, IORESOURCE_SYSTEM_RAM, + IORES_DESC_NONE) == REGION_MIXED) { + size = resource_size(&nsio->res); + end_trunc = start + size - PHYS_SECTION_ALIGN_DOWN(start + size); + } + + if (start_pad + end_trunc) + dev_info(&nd_pfn->dev, "%s section collision, truncate %d bytes\n", + dev_name(&ndns->dev), start_pad + end_trunc); + /* * Note, we use 64 here for the standard size of struct page, * debugging options may cause it to be larger in which case the * implementation will limit the pfns advertised through * ->direct_access() to those that are included in the memmap. */ + start += start_pad; + npfns = (pmem->size - start_pad - end_trunc - SZ_8K) / SZ_4K; if (nd_pfn->mode == PFN_MODE_PMEM) - offset = ALIGN(SZ_8K + 64 * npfns, nd_pfn->align); + offset = ALIGN(start + SZ_8K + 64 * npfns, nd_pfn->align) + - start; else if (nd_pfn->mode == PFN_MODE_RAM) - offset = ALIGN(SZ_8K, nd_pfn->align); + offset = ALIGN(start + SZ_8K, nd_pfn->align) - start; else goto err; - npfns = (pmem->size - offset) / SZ_4K; + if (offset + start_pad + end_trunc >= pmem->size) { + dev_err(&nd_pfn->dev, "%s unable to satisfy requested alignment\n", + dev_name(&ndns->dev)); + goto err; + } + + npfns = (pmem->size - offset - start_pad - end_trunc) / SZ_4K; pfn_sb->mode = cpu_to_le32(nd_pfn->mode); pfn_sb->dataoff = cpu_to_le64(offset); pfn_sb->npfns = cpu_to_le64(npfns); @@ -326,6 +368,9 @@ static int nd_pfn_init(struct nd_pfn *nd_pfn) memcpy(pfn_sb->uuid, nd_pfn->uuid, 16); memcpy(pfn_sb->parent_uuid, nd_dev_to_uuid(&ndns->dev), 16); pfn_sb->version_major = cpu_to_le16(1); + pfn_sb->version_minor = cpu_to_le16(1); + pfn_sb->start_pad = cpu_to_le32(start_pad); + pfn_sb->end_trunc = cpu_to_le32(end_trunc); checksum = nd_sb_checksum((struct nd_gen_sb *) pfn_sb); pfn_sb->checksum = cpu_to_le64(checksum); @@ -376,41 +421,36 @@ static unsigned long init_altmap_reserve(resource_size_t base) return reserve; } -static int nvdimm_namespace_attach_pfn(struct nd_namespace_common *ndns) +static int __nvdimm_namespace_attach_pfn(struct nd_pfn *nd_pfn) { - struct nd_namespace_io *nsio = to_nd_namespace_io(&ndns->dev); - struct nd_pfn *nd_pfn = to_nd_pfn(ndns->claim); - struct device *dev = &nd_pfn->dev; - struct nd_region *nd_region; - struct vmem_altmap *altmap; - struct nd_pfn_sb *pfn_sb; - struct pmem_device *pmem; - struct request_queue *q; - phys_addr_t offset; int rc; + struct resource res; + struct request_queue *q; + struct pmem_device *pmem; + struct vmem_altmap *altmap; + struct device *dev = &nd_pfn->dev; + struct nd_pfn_sb *pfn_sb = nd_pfn->pfn_sb; + struct nd_namespace_common *ndns = nd_pfn->ndns; + u32 start_pad = __le32_to_cpu(pfn_sb->start_pad); + u32 end_trunc = __le32_to_cpu(pfn_sb->end_trunc); + struct nd_namespace_io *nsio = to_nd_namespace_io(&ndns->dev); + resource_size_t base = nsio->res.start + start_pad; struct vmem_altmap __altmap = { - .base_pfn = init_altmap_base(nsio->res.start), - .reserve = init_altmap_reserve(nsio->res.start), + .base_pfn = init_altmap_base(base), + .reserve = init_altmap_reserve(base), }; - if (!nd_pfn->uuid || !nd_pfn->ndns) - return -ENODEV; - - nd_region = to_nd_region(dev->parent); - rc = nd_pfn_init(nd_pfn); - if (rc) - return rc; - - pfn_sb = nd_pfn->pfn_sb; - offset = le64_to_cpu(pfn_sb->dataoff); + pmem = dev_get_drvdata(dev); + pmem->data_offset = le64_to_cpu(pfn_sb->dataoff); + pmem->pfn_pad = start_pad + end_trunc; nd_pfn->mode = le32_to_cpu(nd_pfn->pfn_sb->mode); if (nd_pfn->mode == PFN_MODE_RAM) { - if (offset < SZ_8K) + if (pmem->data_offset < SZ_8K) return -EINVAL; nd_pfn->npfns = le64_to_cpu(pfn_sb->npfns); altmap = NULL; } else if (nd_pfn->mode == PFN_MODE_PMEM) { - nd_pfn->npfns = (resource_size(&nsio->res) - offset) + nd_pfn->npfns = (pmem->size - pmem->pfn_pad - pmem->data_offset) / PAGE_SIZE; if (le64_to_cpu(nd_pfn->pfn_sb->npfns) > nd_pfn->npfns) dev_info(&nd_pfn->dev, @@ -418,7 +458,7 @@ static int nvdimm_namespace_attach_pfn(struct nd_namespace_common *ndns) le64_to_cpu(nd_pfn->pfn_sb->npfns), nd_pfn->npfns); altmap = & __altmap; - altmap->free = __phys_to_pfn(offset - SZ_8K); + altmap->free = __phys_to_pfn(pmem->data_offset - SZ_8K); altmap->alloc = 0; } else { rc = -ENXIO; @@ -426,10 +466,12 @@ static int nvdimm_namespace_attach_pfn(struct nd_namespace_common *ndns) } /* establish pfn range for lookup, and switch to direct map */ - pmem = dev_get_drvdata(dev); q = pmem->pmem_queue; + memcpy(&res, &nsio->res, sizeof(res)); + res.start += start_pad; + res.end -= end_trunc; devm_memunmap(dev, (void __force *) pmem->virt_addr); - pmem->virt_addr = (void __pmem *) devm_memremap_pages(dev, &nsio->res, + pmem->virt_addr = (void __pmem *) devm_memremap_pages(dev, &res, &q->q_usage_counter, altmap); pmem->pfn_flags |= PFN_MAP; if (IS_ERR(pmem->virt_addr)) { @@ -438,7 +480,6 @@ static int nvdimm_namespace_attach_pfn(struct nd_namespace_common *ndns) } /* attach pmem disk in "pfn-mode" */ - pmem->data_offset = offset; rc = pmem_attach_disk(dev, ndns, pmem); if (rc) goto err; @@ -447,6 +488,22 @@ static int nvdimm_namespace_attach_pfn(struct nd_namespace_common *ndns) err: nvdimm_namespace_detach_pfn(ndns); return rc; + +} + +static int nvdimm_namespace_attach_pfn(struct nd_namespace_common *ndns) +{ + struct nd_pfn *nd_pfn = to_nd_pfn(ndns->claim); + int rc; + + if (!nd_pfn->uuid || !nd_pfn->ndns) + return -ENODEV; + + rc = nd_pfn_init(nd_pfn); + if (rc) + return rc; + /* we need a valid pfn_sb before we can init a vmem_altmap */ + return __nvdimm_namespace_attach_pfn(nd_pfn); } static int nd_pmem_probe(struct device *dev) -- cgit v1.2.1 From 45f68802f2542a6ad1550dab9c07004de6e0df40 Mon Sep 17 00:00:00 2001 From: Dan Williams Date: Sun, 6 Mar 2016 08:04:12 -0800 Subject: libnvdimm, pmem: fix ia64 build, use PHYS_PFN drivers/nvdimm/pmem.c: In function 'nvdimm_namespace_attach_pfn': drivers/nvdimm/pmem.c:367:3: error: implicit declaration of function '__phys_to_pfn' [-Werror=implicit-function-declaration] .base_pfn = __phys_to_pfn(nsio->res.start), ia64 does not provide __phys_to_pfn(), just use the PHYS_PFN() alias. Cc: Guenter Roeck Reported-by: kbuild test robot Signed-off-by: Dan Williams --- drivers/nvdimm/pmem.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'drivers/nvdimm/pmem.c') diff --git a/drivers/nvdimm/pmem.c b/drivers/nvdimm/pmem.c index 0cb450e1b400..74e2569910d8 100644 --- a/drivers/nvdimm/pmem.c +++ b/drivers/nvdimm/pmem.c @@ -407,15 +407,15 @@ static int nvdimm_namespace_detach_pfn(struct nd_namespace_common *ndns) */ static unsigned long init_altmap_base(resource_size_t base) { - unsigned long base_pfn = __phys_to_pfn(base); + unsigned long base_pfn = PHYS_PFN(base); return PFN_SECTION_ALIGN_DOWN(base_pfn); } static unsigned long init_altmap_reserve(resource_size_t base) { - unsigned long reserve = __phys_to_pfn(SZ_8K); - unsigned long base_pfn = __phys_to_pfn(base); + unsigned long reserve = PHYS_PFN(SZ_8K); + unsigned long base_pfn = PHYS_PFN(base); reserve += base_pfn - PFN_SECTION_ALIGN_DOWN(base_pfn); return reserve; @@ -458,7 +458,7 @@ static int __nvdimm_namespace_attach_pfn(struct nd_pfn *nd_pfn) le64_to_cpu(nd_pfn->pfn_sb->npfns), nd_pfn->npfns); altmap = & __altmap; - altmap->free = __phys_to_pfn(pmem->data_offset - SZ_8K); + altmap->free = PHYS_PFN(pmem->data_offset - SZ_8K); altmap->alloc = 0; } else { rc = -ENXIO; -- cgit v1.2.1 From 55155291b32d24371256adbcc67f9f53cf3f314f Mon Sep 17 00:00:00 2001 From: NeilBrown Date: Wed, 9 Mar 2016 09:21:54 +1100 Subject: pmem: don't allocate unused major device number When alloc_disk(0) or alloc_disk-node(0, XX) is used, the ->major number is completely ignored: all devices are allocated with a major of BLOCK_EXT_MAJOR. So there is no point allocating pmem_major. Signed-off-by: NeilBrown Signed-off-by: Dan Williams --- drivers/nvdimm/pmem.c | 19 +------------------ 1 file changed, 1 insertion(+), 18 deletions(-) (limited to 'drivers/nvdimm/pmem.c') diff --git a/drivers/nvdimm/pmem.c b/drivers/nvdimm/pmem.c index 74e2569910d8..ba8d5b6bfad0 100644 --- a/drivers/nvdimm/pmem.c +++ b/drivers/nvdimm/pmem.c @@ -50,8 +50,6 @@ struct pmem_device { struct badblocks bb; }; -static int pmem_major; - static bool is_bad_pmem(struct badblocks *bb, sector_t sector, unsigned int len) { if (bb->count) { @@ -231,8 +229,6 @@ static int pmem_attach_disk(struct device *dev, return -ENOMEM; } - disk->major = pmem_major; - disk->first_minor = 0; disk->fops = &pmem_fops; disk->private_data = pmem; disk->queue = pmem->pmem_queue; @@ -579,26 +575,13 @@ static struct nd_device_driver nd_pmem_driver = { static int __init pmem_init(void) { - int error; - - pmem_major = register_blkdev(0, "pmem"); - if (pmem_major < 0) - return pmem_major; - - error = nd_driver_register(&nd_pmem_driver); - if (error) { - unregister_blkdev(pmem_major, "pmem"); - return error; - } - - return 0; + return nd_driver_register(&nd_pmem_driver); } module_init(pmem_init); static void pmem_exit(void) { driver_unregister(&nd_pmem_driver.drv); - unregister_blkdev(pmem_major, "pmem"); } module_exit(pmem_exit); -- cgit v1.2.1