From: "Aneesh Kumar K.V" <aneesh.kumar@linux.ibm.com> To: dan.j.williams@intel.com, mpe@ellerman.id.au Cc: linux-nvdimm@lists.01.org, linuxppc-dev@lists.ozlabs.org, "Aneesh Kumar K.V" <aneesh.kumar@linux.ibm.com> Subject: [RFC PATCH 2/4] libnvdimm/namespace: Disable the region if the namespace size is not aligned correctly Date: Mon, 28 Oct 2019 15:18:23 +0530 [thread overview] Message-ID: <20191028094825.21448-2-aneesh.kumar@linux.ibm.com> (raw) In-Reply-To: <20191028094825.21448-1-aneesh.kumar@linux.ibm.com> During namespace initialization, if kernel finds the namespace size not aligned as per arch-specific restriction, disable the region completely. Even though kernel validates the namespace size while creating the namespace nvdimm core still needs to make sure namespaces with wrong size are not initialized. This can happen when users move SCM across different architectures and with architectures like ppc64 when we use different MMU translation modes. ppc64 allow booting systems with different translation mode based on kernel parameter and the two translation mode (hash and radix) have different direct-map page size restrictions. Marking the region disabled enables the user to either re-init the namespace using the dimm interface or boot back to the right kernel supporting the alignment. Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> --- drivers/nvdimm/namespace_devs.c | 15 +++++++++++++++ 1 file changed, 15 insertions(+) diff --git a/drivers/nvdimm/namespace_devs.c b/drivers/nvdimm/namespace_devs.c index cca0a3ba1d2c..fcde9f2bf2ea 100644 --- a/drivers/nvdimm/namespace_devs.c +++ b/drivers/nvdimm/namespace_devs.c @@ -1912,6 +1912,7 @@ static struct device *create_namespace_pmem(struct nd_region *nd_region, struct nd_label_ent *label_ent; struct nd_namespace_pmem *nspm; struct nd_mapping *nd_mapping; + unsigned long map_align_size; resource_size_t size = 0; struct resource *res; struct device *dev; @@ -2010,6 +2011,20 @@ static struct device *create_namespace_pmem(struct nd_region *nd_region, } + map_align_size = arch_validate_namespace_size(nd_region->ndr_mappings, size); + if (map_align_size) { + dev_err(&nd_region->dev, "%llu is not %ldK aligned\n", size, + map_align_size/ SZ_1K); + /* + * Disable this region completely. A wrongly sized namespace + * size implies the start address of other namespace would also + * be wrong and we would find confusing crashes w.r.t + * direct-map address. + */ + rc = -EINVAL; + goto err; + } + if (!nspm->alt_name || !nspm->uuid) { rc = -ENOMEM; goto err; -- 2.21.0 _______________________________________________ Linux-nvdimm mailing list -- linux-nvdimm@lists.01.org To unsubscribe send an email to linux-nvdimm-leave@lists.01.org
WARNING: multiple messages have this Message-ID (diff)
From: "Aneesh Kumar K.V" <aneesh.kumar@linux.ibm.com> To: dan.j.williams@intel.com, mpe@ellerman.id.au Cc: "Aneesh Kumar K.V" <aneesh.kumar@linux.ibm.com>, linuxppc-dev@lists.ozlabs.org, linux-nvdimm@lists.01.org Subject: [RFC PATCH 2/4] libnvdimm/namespace: Disable the region if the namespace size is not aligned correctly Date: Mon, 28 Oct 2019 15:18:23 +0530 [thread overview] Message-ID: <20191028094825.21448-2-aneesh.kumar@linux.ibm.com> (raw) In-Reply-To: <20191028094825.21448-1-aneesh.kumar@linux.ibm.com> During namespace initialization, if kernel finds the namespace size not aligned as per arch-specific restriction, disable the region completely. Even though kernel validates the namespace size while creating the namespace nvdimm core still needs to make sure namespaces with wrong size are not initialized. This can happen when users move SCM across different architectures and with architectures like ppc64 when we use different MMU translation modes. ppc64 allow booting systems with different translation mode based on kernel parameter and the two translation mode (hash and radix) have different direct-map page size restrictions. Marking the region disabled enables the user to either re-init the namespace using the dimm interface or boot back to the right kernel supporting the alignment. Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> --- drivers/nvdimm/namespace_devs.c | 15 +++++++++++++++ 1 file changed, 15 insertions(+) diff --git a/drivers/nvdimm/namespace_devs.c b/drivers/nvdimm/namespace_devs.c index cca0a3ba1d2c..fcde9f2bf2ea 100644 --- a/drivers/nvdimm/namespace_devs.c +++ b/drivers/nvdimm/namespace_devs.c @@ -1912,6 +1912,7 @@ static struct device *create_namespace_pmem(struct nd_region *nd_region, struct nd_label_ent *label_ent; struct nd_namespace_pmem *nspm; struct nd_mapping *nd_mapping; + unsigned long map_align_size; resource_size_t size = 0; struct resource *res; struct device *dev; @@ -2010,6 +2011,20 @@ static struct device *create_namespace_pmem(struct nd_region *nd_region, } + map_align_size = arch_validate_namespace_size(nd_region->ndr_mappings, size); + if (map_align_size) { + dev_err(&nd_region->dev, "%llu is not %ldK aligned\n", size, + map_align_size/ SZ_1K); + /* + * Disable this region completely. A wrongly sized namespace + * size implies the start address of other namespace would also + * be wrong and we would find confusing crashes w.r.t + * direct-map address. + */ + rc = -EINVAL; + goto err; + } + if (!nspm->alt_name || !nspm->uuid) { rc = -ENOMEM; goto err; -- 2.21.0
next prev parent reply other threads:[~2019-10-28 9:48 UTC|newest] Thread overview: 41+ messages / expand[flat|nested] mbox.gz Atom feed top 2019-10-28 9:48 [RFC PATCH 1/4] libnvdimm/namespace: Make namespace size validation arch dependent Aneesh Kumar K.V 2019-10-28 9:48 ` Aneesh Kumar K.V 2019-10-28 9:48 ` Aneesh Kumar K.V [this message] 2019-10-28 9:48 ` [RFC PATCH 2/4] libnvdimm/namespace: Disable the region if the namespace size is not aligned correctly Aneesh Kumar K.V 2019-10-30 9:12 ` kbuild test robot 2019-10-28 9:48 ` [RFC PATCH 3/4] libnvdimm/namespace: Use direct-mapping page size to validate namespace size Aneesh Kumar K.V 2019-10-28 9:48 ` Aneesh Kumar K.V 2019-10-28 9:48 ` [RFC PATCH 4/4] libnvdimm/namespace: Add debug check while initializing namespace resource size Aneesh Kumar K.V 2019-10-28 9:48 ` Aneesh Kumar K.V 2019-10-28 21:21 ` [RFC PATCH 1/4] libnvdimm/namespace: Make namespace size validation arch dependent Ira Weiny 2019-10-28 21:21 ` Ira Weiny 2019-10-29 4:34 ` Aneesh Kumar K.V 2019-10-29 4:34 ` Aneesh Kumar K.V 2019-10-28 23:08 ` Dan Williams 2019-10-28 23:08 ` Dan Williams 2019-10-29 4:34 ` Aneesh Kumar K.V 2019-10-29 4:34 ` Aneesh Kumar K.V 2019-10-29 5:30 ` Dan Williams 2019-10-29 5:30 ` Dan Williams 2019-10-31 5:35 ` Aneesh Kumar K.V 2019-10-31 5:35 ` Aneesh Kumar K.V 2019-10-31 6:30 ` Dan Williams 2019-10-31 6:30 ` Dan Williams 2019-10-31 8:37 ` Aneesh Kumar K.V 2019-10-31 8:37 ` Aneesh Kumar K.V 2019-10-31 15:53 ` Dan Williams 2019-10-31 15:53 ` Dan Williams 2019-11-06 10:44 ` Aneesh Kumar K.V 2019-11-06 10:44 ` Aneesh Kumar K.V 2019-11-16 12:15 ` Aneesh Kumar K.V 2019-11-16 12:15 ` Aneesh Kumar K.V 2019-11-16 18:50 ` Dan Williams 2019-11-16 18:50 ` Dan Williams 2019-11-18 9:51 ` Aneesh Kumar K.V 2019-11-18 9:51 ` Aneesh Kumar K.V 2019-11-19 17:58 ` Dan Williams 2019-11-19 17:58 ` Dan Williams 2019-11-20 3:19 ` Aneesh Kumar K.V 2019-11-20 3:19 ` Aneesh Kumar K.V 2019-11-20 3:46 ` Dan Williams 2019-11-20 3:46 ` Dan Williams
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=20191028094825.21448-2-aneesh.kumar@linux.ibm.com \ --to=aneesh.kumar@linux.ibm.com \ --cc=dan.j.williams@intel.com \ --cc=linux-nvdimm@lists.01.org \ --cc=linuxppc-dev@lists.ozlabs.org \ --cc=mpe@ellerman.id.au \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
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.