From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-ot1-x342.google.com (mail-ot1-x342.google.com [IPv6:2607:f8b0:4864:20::342]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by ml01.01.org (Postfix) with ESMTPS id 635A52114ACEA for ; Sat, 22 Sep 2018 17:11:06 -0700 (PDT) Received: by mail-ot1-x342.google.com with SMTP id e21-v6so3852900otk.10 for ; Sat, 22 Sep 2018 17:11:06 -0700 (PDT) MIME-Version: 1.0 References: <153549632073.4089.3609134467249378610.stgit@djiang5-desk3.ch.intel.com> <153549646600.4089.2626014553500613547.stgit@djiang5-desk3.ch.intel.com> In-Reply-To: <153549646600.4089.2626014553500613547.stgit@djiang5-desk3.ch.intel.com> From: Dan Williams Date: Sat, 22 Sep 2018 17:10:54 -0700 Message-ID: Subject: Re: [PATCH v8 05/12] nfit/libnvdimm: add unlock of nvdimm support for Intel DIMMs List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Errors-To: linux-nvdimm-bounces@lists.01.org Sender: "Linux-nvdimm" To: Dave Jiang Cc: "Schofield, Alison" , Kees Cook , linux-nvdimm , Eric Biggers , David Howells , keyrings@vger.kernel.org List-ID: On Tue, Aug 28, 2018 at 3:47 PM Dave Jiang wrote: > > Add support to allow query the security status of the Intel nvdimms and > also unlock the dimm via the kernel key management APIs. The passphrase is > expected to be pulled from userspace through keyutils. Moving the Intel > related bits to its own source file as well. > > Signed-off-by: Dave Jiang > Reviewed-by: Dan Williams > --- > drivers/acpi/nfit/Makefile | 1 > drivers/acpi/nfit/core.c | 3 + > drivers/acpi/nfit/intel.c | 152 ++++++++++++++++++++++++++++++++++++++++++++ > drivers/acpi/nfit/intel.h | 15 ++++ > drivers/nvdimm/Kconfig | 1 > drivers/nvdimm/dimm.c | 7 ++ > drivers/nvdimm/dimm_devs.c | 128 +++++++++++++++++++++++++++++++++++++ > drivers/nvdimm/nd-core.h | 4 + > drivers/nvdimm/nd.h | 2 + > include/linux/libnvdimm.h | 24 +++++++ > 10 files changed, 334 insertions(+), 3 deletions(-) > create mode 100644 drivers/acpi/nfit/intel.c > > diff --git a/drivers/acpi/nfit/Makefile b/drivers/acpi/nfit/Makefile > index a407e769f103..443c7ef4e6a6 100644 > --- a/drivers/acpi/nfit/Makefile > +++ b/drivers/acpi/nfit/Makefile > @@ -1,3 +1,4 @@ > obj-$(CONFIG_ACPI_NFIT) := nfit.o > nfit-y := core.o > +nfit-$(CONFIG_X86) += intel.o > nfit-$(CONFIG_X86_MCE) += mce.o > diff --git a/drivers/acpi/nfit/core.c b/drivers/acpi/nfit/core.c > index 21235d555b5f..9cb6a108ecba 100644 > --- a/drivers/acpi/nfit/core.c > +++ b/drivers/acpi/nfit/core.c > @@ -1904,7 +1904,8 @@ static int acpi_nfit_register_dimms(struct acpi_nfit_desc *acpi_desc) > nvdimm = nvdimm_create(acpi_desc->nvdimm_bus, nfit_mem, > acpi_nfit_dimm_attribute_groups, > flags, cmd_mask, flush ? flush->hint_count : 0, > - nfit_mem->flush_wpq, &nfit_mem->id[0]); > + nfit_mem->flush_wpq, &nfit_mem->id[0], > + acpi_nfit_get_security_ops(nfit_mem->family)); > if (!nvdimm) > return -ENOMEM; > > diff --git a/drivers/acpi/nfit/intel.c b/drivers/acpi/nfit/intel.c > new file mode 100644 > index 000000000000..4bfc1c1da339 > --- /dev/null > +++ b/drivers/acpi/nfit/intel.c > @@ -0,0 +1,152 @@ > +// SPDX-License-Identifier: GPL-2.0 > +/* Copyright(c) 2018 Intel Corporation. All rights reserved. */ > +/* > + * Intel specific NFIT ops > + */ > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include "intel.h" > +#include "nfit.h" > + > +static int intel_dimm_security_unlock(struct nvdimm_bus *nvdimm_bus, > + struct nvdimm *nvdimm, const struct nvdimm_key_data *nkey) > +{ > + struct nvdimm_bus_descriptor *nd_desc = to_nd_desc(nvdimm_bus); > + int cmd_rc, rc = 0; > + struct nfit_mem *nfit_mem = nvdimm_provider_data(nvdimm); > + struct { > + struct nd_cmd_pkg pkg; > + struct nd_intel_unlock_unit cmd; > + } nd_cmd = { > + .pkg = { > + .nd_command = NVDIMM_INTEL_UNLOCK_UNIT, > + .nd_family = NVDIMM_FAMILY_INTEL, > + .nd_size_in = ND_INTEL_PASSPHRASE_SIZE, > + .nd_size_out = ND_INTEL_STATUS_SIZE, > + .nd_fw_size = ND_INTEL_STATUS_SIZE, > + }, > + .cmd = { > + .status = 0, > + }, > + }; > + > + if (!test_bit(NVDIMM_INTEL_UNLOCK_UNIT, &nfit_mem->dsm_mask)) > + return -ENOTTY; > + > + memcpy(nd_cmd.cmd.passphrase, nkey->data, > + sizeof(nd_cmd.cmd.passphrase)); > + rc = nd_desc->ndctl(nd_desc, nvdimm, ND_CMD_CALL, &nd_cmd, > + sizeof(nd_cmd), &cmd_rc); > + if (rc < 0) > + goto out; > + if (cmd_rc < 0) { > + rc = cmd_rc; > + goto out; > + } > + > + switch (nd_cmd.cmd.status) { > + case 0: > + break; > + case ND_INTEL_STATUS_INVALID_PASS: > + rc = -EINVAL; > + goto out; > + case ND_INTEL_STATUS_INVALID_STATE: > + default: > + rc = -ENXIO; > + goto out; > + } > + > + /* > + * TODO: define a cross arch wbinvd when/if NVDIMM_FAMILY_INTEL > + * support arrives on another arch. > + */ > + /* DIMM unlocked, invalidate all CPU caches before we read it */ > + wbinvd_on_all_cpus(); > + > + out: > + return rc; > +} > + > +static int intel_dimm_security_state(struct nvdimm_bus *nvdimm_bus, > + struct nvdimm *nvdimm, enum nvdimm_security_state *state) > +{ > + struct nvdimm_bus_descriptor *nd_desc = to_nd_desc(nvdimm_bus); > + int cmd_rc, rc = 0; > + struct nfit_mem *nfit_mem = nvdimm_provider_data(nvdimm); > + struct { > + struct nd_cmd_pkg pkg; > + struct nd_intel_get_security_state cmd; > + } nd_cmd = { > + .pkg = { > + .nd_command = NVDIMM_INTEL_GET_SECURITY_STATE, > + .nd_family = NVDIMM_FAMILY_INTEL, > + .nd_size_in = 0, > + .nd_size_out = > + sizeof(struct nd_intel_get_security_state), > + .nd_fw_size = > + sizeof(struct nd_intel_get_security_state), > + }, > + .cmd = { > + .status = 0, > + .state = 0, > + }, > + }; > + > + if (!test_bit(NVDIMM_INTEL_GET_SECURITY_STATE, &nfit_mem->dsm_mask)) { > + *state = NVDIMM_SECURITY_UNSUPPORTED; > + return 0; > + } > + > + *state = NVDIMM_SECURITY_DISABLED; > + rc = nd_desc->ndctl(nd_desc, nvdimm, ND_CMD_CALL, &nd_cmd, > + sizeof(nd_cmd), &cmd_rc); > + if (rc < 0) > + goto out; > + if (cmd_rc < 0) { > + rc = cmd_rc; > + goto out; > + } > + > + switch (nd_cmd.cmd.status) { > + case 0: > + break; > + case ND_INTEL_STATUS_RETRY: > + rc = -EAGAIN; > + goto out; > + case ND_INTEL_STATUS_NOT_READY: > + default: > + rc = -ENXIO; > + goto out; > + } > + > + /* check and see if security is enabled and locked */ > + if (nd_cmd.cmd.state & ND_INTEL_SEC_STATE_UNSUPPORTED) > + *state = NVDIMM_SECURITY_UNSUPPORTED; > + else if (nd_cmd.cmd.state & ND_INTEL_SEC_STATE_ENABLED) { > + if (nd_cmd.cmd.state & ND_INTEL_SEC_STATE_LOCKED) > + *state = NVDIMM_SECURITY_LOCKED; > + else > + *state = NVDIMM_SECURITY_UNLOCKED; > + } else > + *state = NVDIMM_SECURITY_DISABLED; > + > + out: > + if (rc < 0) > + *state = NVDIMM_SECURITY_INVALID; > + return rc; > +} > + > +const struct nvdimm_security_ops intel_security_ops = { > + .state = intel_dimm_security_state, > + .unlock = intel_dimm_security_unlock, > +}; > diff --git a/drivers/acpi/nfit/intel.h b/drivers/acpi/nfit/intel.h > index a6f8f4bc8ebb..f9ac718776ca 100644 > --- a/drivers/acpi/nfit/intel.h > +++ b/drivers/acpi/nfit/intel.h > @@ -8,6 +8,8 @@ > > #ifdef CONFIG_X86 > > +extern const struct nvdimm_security_ops intel_security_ops; > + > #define ND_INTEL_STATUS_SIZE 4 > #define ND_INTEL_PASSPHRASE_SIZE 32 > > @@ -64,4 +66,17 @@ struct nd_intel_query_overwrite { > } __packed; > #endif /* CONFIG_X86 */ > > +static inline const struct nvdimm_security_ops * > +acpi_nfit_get_security_ops(int family) > +{ > + switch (family) { > +#ifdef CONFIG_X86 > + case NVDIMM_FAMILY_INTEL: > + return &intel_security_ops; > +#endif > + default: > + return NULL; > + } > +} > + > #endif > diff --git a/drivers/nvdimm/Kconfig b/drivers/nvdimm/Kconfig > index 9d36473dc2a2..830a7a5342e1 100644 > --- a/drivers/nvdimm/Kconfig > +++ b/drivers/nvdimm/Kconfig > @@ -3,6 +3,7 @@ menuconfig LIBNVDIMM > depends on PHYS_ADDR_T_64BIT > depends on HAS_IOMEM > depends on BLK_DEV > + depends on KEYS > help > Generic support for non-volatile memory devices including > ACPI-6-NFIT defined resources. On platforms that define an > diff --git a/drivers/nvdimm/dimm.c b/drivers/nvdimm/dimm.c > index 6c8fb7590838..b6381ddbd6c1 100644 > --- a/drivers/nvdimm/dimm.c > +++ b/drivers/nvdimm/dimm.c > @@ -51,6 +51,13 @@ static int nvdimm_probe(struct device *dev) > get_device(dev); > kref_init(&ndd->kref); > > + nvdimm_security_get_state(dev); > + > + /* unlock DIMM here before touch label */ > + rc = nvdimm_security_unlock_dimm(dev); > + if (rc < 0) > + dev_warn(dev, "failed to unlock dimm %s\n", dev_name(dev)); > + > /* > * EACCES failures reading the namespace label-area-properties > * are interpreted as the DIMM capacity being locked but the > diff --git a/drivers/nvdimm/dimm_devs.c b/drivers/nvdimm/dimm_devs.c > index e753b2095402..fe61f2a2ad5d 100644 > --- a/drivers/nvdimm/dimm_devs.c > +++ b/drivers/nvdimm/dimm_devs.c > @@ -20,6 +20,7 @@ > #include > #include > #include > +#include > #include "nd-core.h" > #include "label.h" > #include "pmem.h" > @@ -28,6 +29,129 @@ > static DEFINE_IDA(dimm_ida); > static struct key *nvdimm_keyring; > > +#define NVDIMM_PREFIX "nvdimm:" > + > +/* > + * Find key in kernel keyring > + */ > +static struct key *nvdimm_get_key(struct device *dev) > +{ > + struct nvdimm *nvdimm = to_nvdimm(dev); > + > + if (!nvdimm->key) > + return NULL; > + > + if (key_validate(nvdimm->key) < 0) > + return NULL; > + > + key_get(nvdimm->key); > + > + dev_dbg(dev, "%s: key found: %d\n", __func__, > + key_serial(nvdimm->key)); > + return nvdimm->key; > +} > + > +/* > + * Retrieve kernel key for DIMM and request from user space if necessary. > + */ > +static struct key *nvdimm_request_key(struct device *dev) > +{ > + struct nvdimm *nvdimm = to_nvdimm(dev); > + struct key *key = NULL; > + char desc[NVDIMM_KEY_DESC_LEN + strlen(NVDIMM_PREFIX)]; > + > + sprintf(desc, "%s%s", NVDIMM_PREFIX, nvdimm->dimm_id); > + key = request_key(&key_type_logon, desc, desc); > + if (IS_ERR(key)) > + key = NULL; > + > + return key; > +} > + > +static int nvdimm_check_key_len(unsigned short len, bool update) > +{ > + if (len == (NVDIMM_PASSPHRASE_LEN * 2) && update) > + return 0; > + > + if (len == NVDIMM_PASSPHRASE_LEN) > + return 0; > + > + return -EINVAL; > +} > + > +int nvdimm_security_get_state(struct device *dev) > +{ > + struct nvdimm *nvdimm = to_nvdimm(dev); > + struct nvdimm_bus *nvdimm_bus = walk_to_nvdimm_bus(dev); > + > + if (!nvdimm->security_ops) > + return 0; > + > + return nvdimm->security_ops->state(nvdimm_bus, nvdimm, > + &nvdimm->state); > +} > + > +int nvdimm_security_unlock_dimm(struct device *dev) > +{ > + struct nvdimm *nvdimm = to_nvdimm(dev); > + struct nvdimm_bus *nvdimm_bus = walk_to_nvdimm_bus(dev); > + struct key *key; > + struct user_key_payload *payload; > + int rc; > + bool cached_key = false; > + > + if (!nvdimm->security_ops) > + return 0; > + > + if (nvdimm->state == NVDIMM_SECURITY_UNLOCKED || Hmm, if the DIMM is unlocked before the driver loads that means we'll never ask for the current key which means we can't count on having it around for change key requests. I think there needs to be a sysfs interface to trigger loading the current key into the keyring cache so it is available for future operations that need it. _______________________________________________ Linux-nvdimm mailing list Linux-nvdimm@lists.01.org https://lists.01.org/mailman/listinfo/linux-nvdimm