* [PATCH v2 0/9] SELinux support for Infiniband RDMA
@ 2016-07-14 22:56 ` Dan Jurgens
0 siblings, 0 replies; 42+ messages in thread
From: Dan Jurgens @ 2016-07-14 22:56 UTC (permalink / raw)
To: chrisw-69jw2NvuJkxg9hUCZPvPmw, paul-r2n+y4ga6xFZroRs9YW3xA,
sds-+05T5uksL2qpZYMLLGbcSA, eparis-FjpueFixGhCM4zKIHC2jIg,
dledford-H+wXaHxf7aLQT0dZR+AlfA,
sean.hefty-ral2JQCrhuEAvxtiuMwx3w,
hal.rosenstock-Re5JQEeQqe8AvxtiuMwx3w
Cc: selinux-+05T5uksL2qpZYMLLGbcSA,
linux-security-module-u79uwXL29TY76Z2rM5mHXA,
linux-rdma-u79uwXL29TY76Z2rM5mHXA,
yevgenyp-VPRAkNaXOzVWk0Htik3J/w, Daniel Jurgens
From: Daniel Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
The selinux next tree is missing some patches for IB/core. This series
applies cleanly to ib-next, and should apply cleanly to selinux-next once
the IB patches are merged.
Currently there is no way to provide granular access control to an
Infiniband fabric. By providing an ability to restrict user access to
specific virtual subfabrics, administrators can limit access to bandwidth
and isolate users on the fabric.
The approach for controlling access for Infiniband is to control access to
partitions. A partition is similar in concept to a VLAN where each data
packet carries the partition key (PKey) in its header and isolation is
enforced by the hardware. The partition key is not a cryptographic key,
it's a 16 bit number identifying the partition. By controlling access to
PKeys, users can be isolated on the fabric.
Every Infiniband fabric must have a subnet manager. The subnet manager
provisions the partitions and configures the end nodes. Each end port has
a PKey table containing the partitions it can access. In order to enforce
access to partitions, the subnet management interface (SMI) must also be
controlled to prevent unauthorized changes to the fabric configuration.
In order to support this there must be a capability to provide security
contexts for two new types of objects - PKeys and IB ports.
A PKey label consists of a subnet prefix and a range of PKey values and is
similar to the labeling mechanism for netports. Each Infiniband port can
reside on a different subnet, labeling the PKey values for specific subnet
prefixes provides the user maximum flexibility. There is a single access
vector for PKeys called "access".
An Infiniband port is labeled by name and port number. There is a single
access vector for IB ports called "manage_subnet".
Because RDMA allows kernel bypass, enforcement must be done during
connection setup. Communication over RDMA requires a send and receive
queue called a queue pair (QP). During the creation of a QP it is
initialized before it can be used to send or receive data. During
initialization the user must provide the PKey and port the QP will use, at
this time access can be enforced.
Because there is a possibility that the enforcement settings or security
policy can change, a means of notifying the ib_core module of such changes
is required. To facilitate this a generic notification callback mechanism
is added to the LSM. One callback is registered for checking the QP PKey
associations when the policy changes. Mad agents also register a callback,
they cache the permission to send and receive SMPs to avoid another per
packet call to the LSM.
Because frequent accesses to the same PKey's SID is expected a cache is
implemented which is very similar to the netport cache.
In order to properly enforce security when changes to the PKey table or
security policy or enforcement occur ib_core must track which QPs are
using which port, pkey index, and alternate path for every IB device.
This makes operations that used to be atomic transactional.
When modifying a QP, ib_core must associate it with the PKey index, port,
and alternate path specified. If the QP was already associated with
different settings, the QP is added to the new list prior to the
modification. If the modify succeeds then the old listing is removed. If
the modify fails the new listing is removed and the old listing remains
unchanged.
When destroying a QP the ib_qp structure is freed by the decive specific
driver (i.e. mlx4_ib) if the 'destroy' is successful. This requires storing
security related information in a separate structure. When a 'destroy'
request is in process the ib_qp structure is in an undefined state so if
there are changes to the security policy or PKey table, the security checks
cannot reset the QP if it doesn't have permission for the new setting. If
the 'destroy' fails, security for that QP must be enforced again and its
status in the list is restored. If the 'destroy' succeeds the security info
can be cleaned up and freed.
There are a number of locks required to protect the QP security structure
and the QP to device/port/pkey index lists. If multiple locks are required,
the safe locking order is: QP security structure mutex first, followed by
any list locks needed, which are sorted first by port followed by pkey
index.
---
v2:
- Use void* blobs in the LSM hooks. Paul Moore
- Make the policy change callback generic. Yuval Shaia, Paul Moore
- Squash LSM changes into the patches where the calls are added. Paul Moore
- Don't add new initial SIDs. Stephen Smalley
- Squash MAD agent PKey and SMI patches. Dan Jurgens
- Changed ib_end_port to ib_port. Paul Moore
- Changed ib_port access vector from smp to manage_subnet. Paul Moore
- Added pkey and ib_port details to the audit log. Paul Moore
- See individual patches for more detail.
Daniel Jurgens (9):
IB/core: IB cache enhancements to support Infiniband security
IB/core: Enforce PKey security on QPs
selinux lsm IB/core: Implement LSM notification system
IB/core: Enforce security on management datagrams
selinux: Create policydb version for Infiniband support
selinux: Allocate and free infiniband security hooks
selinux: Implement Infiniband PKey "Access" access vector
selinux: Add IB Port SMP access vector
selinux: Add a cache for quicker retreival of PKey SIDs
drivers/infiniband/core/Makefile | 3 +-
drivers/infiniband/core/cache.c | 57 +++-
drivers/infiniband/core/core_priv.h | 115 ++++++
drivers/infiniband/core/device.c | 79 ++++
drivers/infiniband/core/mad.c | 52 +++-
drivers/infiniband/core/security.c | 707 ++++++++++++++++++++++++++++++++++
drivers/infiniband/core/uverbs_cmd.c | 20 +-
drivers/infiniband/core/verbs.c | 27 ++-
include/linux/lsm_audit.h | 37 ++-
include/linux/lsm_hooks.h | 34 ++
include/linux/security.h | 37 ++
include/rdma/ib_mad.h | 4 +
include/rdma/ib_verbs.h | 49 +++
security/Kconfig | 9 +
security/lsm_audit.c | 18 +
security/security.c | 97 +++++
security/selinux/Makefile | 2 +-
security/selinux/hooks.c | 79 ++++-
security/selinux/ibpkey.c | 245 ++++++++++++
security/selinux/include/classmap.h | 4 +
security/selinux/include/ibpkey.h | 31 ++
security/selinux/include/objsec.h | 11 +
security/selinux/include/security.h | 7 +-
security/selinux/selinuxfs.c | 2 +
security/selinux/ss/policydb.c | 130 ++++++-
security/selinux/ss/policydb.h | 27 +-
security/selinux/ss/services.c | 83 ++++
27 files changed, 1908 insertions(+), 58 deletions(-)
create mode 100644 drivers/infiniband/core/security.c
create mode 100644 security/selinux/ibpkey.c
create mode 100644 security/selinux/include/ibpkey.h
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 42+ messages in thread
* [PATCH v2 0/9] SELinux support for Infiniband RDMA
@ 2016-07-14 22:56 ` Dan Jurgens
0 siblings, 0 replies; 42+ messages in thread
From: Dan Jurgens @ 2016-07-14 22:56 UTC (permalink / raw)
To: chrisw, paul, sds, eparis, dledford, sean.hefty, hal.rosenstock
Cc: selinux, linux-security-module, linux-rdma, yevgenyp, Daniel Jurgens
From: Daniel Jurgens <danielj@mellanox.com>
The selinux next tree is missing some patches for IB/core. This series
applies cleanly to ib-next, and should apply cleanly to selinux-next once
the IB patches are merged.
Currently there is no way to provide granular access control to an
Infiniband fabric. By providing an ability to restrict user access to
specific virtual subfabrics, administrators can limit access to bandwidth
and isolate users on the fabric.
The approach for controlling access for Infiniband is to control access to
partitions. A partition is similar in concept to a VLAN where each data
packet carries the partition key (PKey) in its header and isolation is
enforced by the hardware. The partition key is not a cryptographic key,
it's a 16 bit number identifying the partition. By controlling access to
PKeys, users can be isolated on the fabric.
Every Infiniband fabric must have a subnet manager. The subnet manager
provisions the partitions and configures the end nodes. Each end port has
a PKey table containing the partitions it can access. In order to enforce
access to partitions, the subnet management interface (SMI) must also be
controlled to prevent unauthorized changes to the fabric configuration.
In order to support this there must be a capability to provide security
contexts for two new types of objects - PKeys and IB ports.
A PKey label consists of a subnet prefix and a range of PKey values and is
similar to the labeling mechanism for netports. Each Infiniband port can
reside on a different subnet, labeling the PKey values for specific subnet
prefixes provides the user maximum flexibility. There is a single access
vector for PKeys called "access".
An Infiniband port is labeled by name and port number. There is a single
access vector for IB ports called "manage_subnet".
Because RDMA allows kernel bypass, enforcement must be done during
connection setup. Communication over RDMA requires a send and receive
queue called a queue pair (QP). During the creation of a QP it is
initialized before it can be used to send or receive data. During
initialization the user must provide the PKey and port the QP will use, at
this time access can be enforced.
Because there is a possibility that the enforcement settings or security
policy can change, a means of notifying the ib_core module of such changes
is required. To facilitate this a generic notification callback mechanism
is added to the LSM. One callback is registered for checking the QP PKey
associations when the policy changes. Mad agents also register a callback,
they cache the permission to send and receive SMPs to avoid another per
packet call to the LSM.
Because frequent accesses to the same PKey's SID is expected a cache is
implemented which is very similar to the netport cache.
In order to properly enforce security when changes to the PKey table or
security policy or enforcement occur ib_core must track which QPs are
using which port, pkey index, and alternate path for every IB device.
This makes operations that used to be atomic transactional.
When modifying a QP, ib_core must associate it with the PKey index, port,
and alternate path specified. If the QP was already associated with
different settings, the QP is added to the new list prior to the
modification. If the modify succeeds then the old listing is removed. If
the modify fails the new listing is removed and the old listing remains
unchanged.
When destroying a QP the ib_qp structure is freed by the decive specific
driver (i.e. mlx4_ib) if the 'destroy' is successful. This requires storing
security related information in a separate structure. When a 'destroy'
request is in process the ib_qp structure is in an undefined state so if
there are changes to the security policy or PKey table, the security checks
cannot reset the QP if it doesn't have permission for the new setting. If
the 'destroy' fails, security for that QP must be enforced again and its
status in the list is restored. If the 'destroy' succeeds the security info
can be cleaned up and freed.
There are a number of locks required to protect the QP security structure
and the QP to device/port/pkey index lists. If multiple locks are required,
the safe locking order is: QP security structure mutex first, followed by
any list locks needed, which are sorted first by port followed by pkey
index.
---
v2:
- Use void* blobs in the LSM hooks. Paul Moore
- Make the policy change callback generic. Yuval Shaia, Paul Moore
- Squash LSM changes into the patches where the calls are added. Paul Moore
- Don't add new initial SIDs. Stephen Smalley
- Squash MAD agent PKey and SMI patches. Dan Jurgens
- Changed ib_end_port to ib_port. Paul Moore
- Changed ib_port access vector from smp to manage_subnet. Paul Moore
- Added pkey and ib_port details to the audit log. Paul Moore
- See individual patches for more detail.
Daniel Jurgens (9):
IB/core: IB cache enhancements to support Infiniband security
IB/core: Enforce PKey security on QPs
selinux lsm IB/core: Implement LSM notification system
IB/core: Enforce security on management datagrams
selinux: Create policydb version for Infiniband support
selinux: Allocate and free infiniband security hooks
selinux: Implement Infiniband PKey "Access" access vector
selinux: Add IB Port SMP access vector
selinux: Add a cache for quicker retreival of PKey SIDs
drivers/infiniband/core/Makefile | 3 +-
drivers/infiniband/core/cache.c | 57 +++-
drivers/infiniband/core/core_priv.h | 115 ++++++
drivers/infiniband/core/device.c | 79 ++++
drivers/infiniband/core/mad.c | 52 +++-
drivers/infiniband/core/security.c | 707 ++++++++++++++++++++++++++++++++++
drivers/infiniband/core/uverbs_cmd.c | 20 +-
drivers/infiniband/core/verbs.c | 27 ++-
include/linux/lsm_audit.h | 37 ++-
include/linux/lsm_hooks.h | 34 ++
include/linux/security.h | 37 ++
include/rdma/ib_mad.h | 4 +
include/rdma/ib_verbs.h | 49 +++
security/Kconfig | 9 +
security/lsm_audit.c | 18 +
security/security.c | 97 +++++
security/selinux/Makefile | 2 +-
security/selinux/hooks.c | 79 ++++-
security/selinux/ibpkey.c | 245 ++++++++++++
security/selinux/include/classmap.h | 4 +
security/selinux/include/ibpkey.h | 31 ++
security/selinux/include/objsec.h | 11 +
security/selinux/include/security.h | 7 +-
security/selinux/selinuxfs.c | 2 +
security/selinux/ss/policydb.c | 130 ++++++-
security/selinux/ss/policydb.h | 27 +-
security/selinux/ss/services.c | 83 ++++
27 files changed, 1908 insertions(+), 58 deletions(-)
create mode 100644 drivers/infiniband/core/security.c
create mode 100644 security/selinux/ibpkey.c
create mode 100644 security/selinux/include/ibpkey.h
^ permalink raw reply [flat|nested] 42+ messages in thread
* [PATCH v2 1/9] IB/core: IB cache enhancements to support Infiniband security
2016-07-14 22:56 ` Dan Jurgens
@ 2016-07-14 22:56 ` Dan Jurgens
-1 siblings, 0 replies; 42+ messages in thread
From: Dan Jurgens @ 2016-07-14 22:56 UTC (permalink / raw)
To: chrisw-69jw2NvuJkxg9hUCZPvPmw, paul-r2n+y4ga6xFZroRs9YW3xA,
sds-+05T5uksL2qpZYMLLGbcSA, eparis-FjpueFixGhCM4zKIHC2jIg,
dledford-H+wXaHxf7aLQT0dZR+AlfA,
sean.hefty-ral2JQCrhuEAvxtiuMwx3w,
hal.rosenstock-Re5JQEeQqe8AvxtiuMwx3w
Cc: selinux-+05T5uksL2qpZYMLLGbcSA,
linux-security-module-u79uwXL29TY76Z2rM5mHXA,
linux-rdma-u79uwXL29TY76Z2rM5mHXA,
yevgenyp-VPRAkNaXOzVWk0Htik3J/w, Daniel Jurgens
From: Daniel Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Cache the subnet prefix and add a function to access it. Enforcing
security requires frequent queries of the subnet prefix and the pkeys in
the pkey table.
Also removed an unneded pr_warn about memory allocation failure.
Signed-off-by: Daniel Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Reviewed-by: Eli Cohen <eli-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Reviewed-by: Leon Romanovsky <leonro-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---
v2:
- In ib_get_cached_subnet_prefix wait to initialize p until after
validation. Yuval Shaia
---
drivers/infiniband/core/cache.c | 36 +++++++++++++++++++++++++++++++++-
drivers/infiniband/core/core_priv.h | 3 ++
include/rdma/ib_verbs.h | 1 +
3 files changed, 38 insertions(+), 2 deletions(-)
diff --git a/drivers/infiniband/core/cache.c b/drivers/infiniband/core/cache.c
index 0409667..8a44894 100644
--- a/drivers/infiniband/core/cache.c
+++ b/drivers/infiniband/core/cache.c
@@ -932,6 +932,26 @@ int ib_get_cached_pkey(struct ib_device *device,
}
EXPORT_SYMBOL(ib_get_cached_pkey);
+int ib_get_cached_subnet_prefix(struct ib_device *device,
+ u8 port_num,
+ u64 *sn_pfx)
+{
+ unsigned long flags;
+ int p;
+
+ if (port_num < rdma_start_port(device) ||
+ port_num > rdma_end_port(device))
+ return -EINVAL;
+
+ p = port_num - rdma_start_port(device);
+ read_lock_irqsave(&device->cache.lock, flags);
+ *sn_pfx = device->cache.subnet_prefix_cache[p];
+ read_unlock_irqrestore(&device->cache.lock, flags);
+
+ return 0;
+}
+EXPORT_SYMBOL(ib_get_cached_subnet_prefix);
+
int ib_find_cached_pkey(struct ib_device *device,
u8 port_num,
u16 pkey,
@@ -1108,6 +1128,8 @@ static void ib_cache_update(struct ib_device *device,
device->cache.lmc_cache[port - rdma_start_port(device)] = tprops->lmc;
+ device->cache.subnet_prefix_cache[port - rdma_start_port(device)] =
+ tprops->subnet_prefix;
write_unlock_irq(&device->cache.lock);
kfree(gid_cache);
@@ -1166,9 +1188,18 @@ int ib_cache_setup_one(struct ib_device *device)
(rdma_end_port(device) -
rdma_start_port(device) + 1),
GFP_KERNEL);
+
+ device->cache.subnet_prefix_cache =
+ kcalloc((rdma_end_port(device) - rdma_start_port(device) + 1),
+ sizeof(*device->cache.subnet_prefix_cache),
+ GFP_KERNEL);
+
if (!device->cache.pkey_cache ||
- !device->cache.lmc_cache) {
- pr_warn("Couldn't allocate cache for %s\n", device->name);
+ !device->cache.lmc_cache ||
+ !device->cache.subnet_prefix_cache) {
+ kfree(device->cache.pkey_cache);
+ kfree(device->cache.lmc_cache);
+ kfree(device->cache.subnet_prefix_cache);
return -ENOMEM;
}
@@ -1211,6 +1242,7 @@ void ib_cache_release_one(struct ib_device *device)
gid_table_release_one(device);
kfree(device->cache.pkey_cache);
kfree(device->cache.lmc_cache);
+ kfree(device->cache.subnet_prefix_cache);
}
void ib_cache_cleanup_one(struct ib_device *device)
diff --git a/drivers/infiniband/core/core_priv.h b/drivers/infiniband/core/core_priv.h
index 19d499d..ce826e4 100644
--- a/drivers/infiniband/core/core_priv.h
+++ b/drivers/infiniband/core/core_priv.h
@@ -153,4 +153,7 @@ int ib_nl_handle_set_timeout(struct sk_buff *skb,
int ib_nl_handle_ip_res_resp(struct sk_buff *skb,
struct netlink_callback *cb);
+int ib_get_cached_subnet_prefix(struct ib_device *device,
+ u8 port_num,
+ u64 *sn_pfx);
#endif /* _CORE_PRIV_H */
diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h
index 156673a..205ea09 100644
--- a/include/rdma/ib_verbs.h
+++ b/include/rdma/ib_verbs.h
@@ -1709,6 +1709,7 @@ struct ib_cache {
struct ib_pkey_cache **pkey_cache;
struct ib_gid_table **gid_cache;
u8 *lmc_cache;
+ u64 *subnet_prefix_cache;
};
struct ib_dma_mapping_ops {
--
1.7.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related [flat|nested] 42+ messages in thread
* [PATCH v2 1/9] IB/core: IB cache enhancements to support Infiniband security
@ 2016-07-14 22:56 ` Dan Jurgens
0 siblings, 0 replies; 42+ messages in thread
From: Dan Jurgens @ 2016-07-14 22:56 UTC (permalink / raw)
To: chrisw, paul, sds, eparis, dledford, sean.hefty, hal.rosenstock
Cc: selinux, linux-security-module, linux-rdma, yevgenyp, Daniel Jurgens
From: Daniel Jurgens <danielj@mellanox.com>
Cache the subnet prefix and add a function to access it. Enforcing
security requires frequent queries of the subnet prefix and the pkeys in
the pkey table.
Also removed an unneded pr_warn about memory allocation failure.
Signed-off-by: Daniel Jurgens <danielj@mellanox.com>
Reviewed-by: Eli Cohen <eli@mellanox.com>
Reviewed-by: Leon Romanovsky <leonro@mellanox.com>
---
v2:
- In ib_get_cached_subnet_prefix wait to initialize p until after
validation. Yuval Shaia
---
drivers/infiniband/core/cache.c | 36 +++++++++++++++++++++++++++++++++-
drivers/infiniband/core/core_priv.h | 3 ++
include/rdma/ib_verbs.h | 1 +
3 files changed, 38 insertions(+), 2 deletions(-)
diff --git a/drivers/infiniband/core/cache.c b/drivers/infiniband/core/cache.c
index 0409667..8a44894 100644
--- a/drivers/infiniband/core/cache.c
+++ b/drivers/infiniband/core/cache.c
@@ -932,6 +932,26 @@ int ib_get_cached_pkey(struct ib_device *device,
}
EXPORT_SYMBOL(ib_get_cached_pkey);
+int ib_get_cached_subnet_prefix(struct ib_device *device,
+ u8 port_num,
+ u64 *sn_pfx)
+{
+ unsigned long flags;
+ int p;
+
+ if (port_num < rdma_start_port(device) ||
+ port_num > rdma_end_port(device))
+ return -EINVAL;
+
+ p = port_num - rdma_start_port(device);
+ read_lock_irqsave(&device->cache.lock, flags);
+ *sn_pfx = device->cache.subnet_prefix_cache[p];
+ read_unlock_irqrestore(&device->cache.lock, flags);
+
+ return 0;
+}
+EXPORT_SYMBOL(ib_get_cached_subnet_prefix);
+
int ib_find_cached_pkey(struct ib_device *device,
u8 port_num,
u16 pkey,
@@ -1108,6 +1128,8 @@ static void ib_cache_update(struct ib_device *device,
device->cache.lmc_cache[port - rdma_start_port(device)] = tprops->lmc;
+ device->cache.subnet_prefix_cache[port - rdma_start_port(device)] =
+ tprops->subnet_prefix;
write_unlock_irq(&device->cache.lock);
kfree(gid_cache);
@@ -1166,9 +1188,18 @@ int ib_cache_setup_one(struct ib_device *device)
(rdma_end_port(device) -
rdma_start_port(device) + 1),
GFP_KERNEL);
+
+ device->cache.subnet_prefix_cache =
+ kcalloc((rdma_end_port(device) - rdma_start_port(device) + 1),
+ sizeof(*device->cache.subnet_prefix_cache),
+ GFP_KERNEL);
+
if (!device->cache.pkey_cache ||
- !device->cache.lmc_cache) {
- pr_warn("Couldn't allocate cache for %s\n", device->name);
+ !device->cache.lmc_cache ||
+ !device->cache.subnet_prefix_cache) {
+ kfree(device->cache.pkey_cache);
+ kfree(device->cache.lmc_cache);
+ kfree(device->cache.subnet_prefix_cache);
return -ENOMEM;
}
@@ -1211,6 +1242,7 @@ void ib_cache_release_one(struct ib_device *device)
gid_table_release_one(device);
kfree(device->cache.pkey_cache);
kfree(device->cache.lmc_cache);
+ kfree(device->cache.subnet_prefix_cache);
}
void ib_cache_cleanup_one(struct ib_device *device)
diff --git a/drivers/infiniband/core/core_priv.h b/drivers/infiniband/core/core_priv.h
index 19d499d..ce826e4 100644
--- a/drivers/infiniband/core/core_priv.h
+++ b/drivers/infiniband/core/core_priv.h
@@ -153,4 +153,7 @@ int ib_nl_handle_set_timeout(struct sk_buff *skb,
int ib_nl_handle_ip_res_resp(struct sk_buff *skb,
struct netlink_callback *cb);
+int ib_get_cached_subnet_prefix(struct ib_device *device,
+ u8 port_num,
+ u64 *sn_pfx);
#endif /* _CORE_PRIV_H */
diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h
index 156673a..205ea09 100644
--- a/include/rdma/ib_verbs.h
+++ b/include/rdma/ib_verbs.h
@@ -1709,6 +1709,7 @@ struct ib_cache {
struct ib_pkey_cache **pkey_cache;
struct ib_gid_table **gid_cache;
u8 *lmc_cache;
+ u64 *subnet_prefix_cache;
};
struct ib_dma_mapping_ops {
--
1.7.1
^ permalink raw reply related [flat|nested] 42+ messages in thread
* [PATCH v2 2/9] IB/core: Enforce PKey security on QPs
2016-07-14 22:56 ` Dan Jurgens
@ 2016-07-14 22:56 ` Dan Jurgens
-1 siblings, 0 replies; 42+ messages in thread
From: Dan Jurgens @ 2016-07-14 22:56 UTC (permalink / raw)
To: chrisw-69jw2NvuJkxg9hUCZPvPmw, paul-r2n+y4ga6xFZroRs9YW3xA,
sds-+05T5uksL2qpZYMLLGbcSA, eparis-FjpueFixGhCM4zKIHC2jIg,
dledford-H+wXaHxf7aLQT0dZR+AlfA,
sean.hefty-ral2JQCrhuEAvxtiuMwx3w,
hal.rosenstock-Re5JQEeQqe8AvxtiuMwx3w
Cc: selinux-+05T5uksL2qpZYMLLGbcSA,
linux-security-module-u79uwXL29TY76Z2rM5mHXA,
linux-rdma-u79uwXL29TY76Z2rM5mHXA,
yevgenyp-VPRAkNaXOzVWk0Htik3J/w, Daniel Jurgens
From: Daniel Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Add new LSM hooks to allocate and free security contexts and check for
permission to access a PKey.
Allocate and free a security context when creating and destroying a QP.
This context is used for controlling access to PKeys.
When a request is made to modify a QP that changes the port, PKey index,
or alternate path, check that the QP has permission for the PKey in the
PKey table index on the subnet prefix of the port. If the QP is shared
make sure all handles to the QP also have access.
Store which port and PKey index a QP is using. After the reset to init
transition the user can modify the port, PKey index and alternate path
independently. So port and PKey settings changes can be a merge of the
previous settings and the new ones.
In order to maintain access control if there are PKey table or subnet
prefix change keep a list of all QPs are using each PKey index on
each port. If a change occurs all QPs using that device and port must
have access enforced for the new cache settings.
These changes add a transaction to the QP modify process. Association
with the old port and PKey index must be maintained if the modify fails,
and must be removed if it succeeds. Association with the new port and
PKey index must be established prior to the modify and removed if the
modify fails.
1. When a QP is modified to a particular Port, PKey index or alternate
path insert that QP into the appropriate lists.
2. Check permission to access the new settings.
3. If step 2 grants access attempt to modify the QP.
4a. If steps 2 and 3 succeed remove any prior associations.
4b. If ether fails remove the new setting associations.
If a PKey table or subnet prefix changes walk the list of QPs and
check that they have permission. If not send the QP to the error state
and raise a fatal error event. If it's a shared QP make sure all the
QPs that share the real_qp have permission as well. If the QP that
owns a security structure is denied access the security structure is
marked as such and the QP is added to an error_list. Once the moving
the QP to error is complete the security structure mark is cleared.
Maintaining the lists correctly turns QP destroy into a transaction.
The hardware driver for the device frees the ib_qp structure, so while
the destroy is in progress the ib_qp pointer in the ib_qp_security
struct is undefined. When the destroy process begins the ib_qp_security
structure is marked as destroying. This prevents any action from being
taken on the QP pointer. After the QP is destroyed successfully it
could still listed on an error_list wait for it to be processed by that
flow before cleaning up the structure.
If the destroy fails the QPs port and PKey settings are reinserted into
the appropriate lists, the destroying flag is cleared, and access control
is enforced, in case there were any cache changes during the destroy
flow.
To keep the security changes isolated a new file is used to hold security
related functionality.
Signed-off-by: Daniel Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Reviewed-by: Leon Romanovsky <leonro-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---
v2:
- Squashed LSM hook additions. Paul Moore
- Changed security blobs to void*. Paul Moore
---
drivers/infiniband/core/Makefile | 3 +-
drivers/infiniband/core/cache.c | 21 +-
drivers/infiniband/core/core_priv.h | 77 +++++
drivers/infiniband/core/device.c | 33 ++
drivers/infiniband/core/security.c | 617 ++++++++++++++++++++++++++++++++++
drivers/infiniband/core/uverbs_cmd.c | 20 +-
drivers/infiniband/core/verbs.c | 27 ++-
include/linux/lsm_hooks.h | 27 ++
include/linux/security.h | 21 ++
include/rdma/ib_verbs.h | 48 +++
security/Kconfig | 9 +
security/security.c | 31 ++
12 files changed, 925 insertions(+), 9 deletions(-)
create mode 100644 drivers/infiniband/core/security.c
diff --git a/drivers/infiniband/core/Makefile b/drivers/infiniband/core/Makefile
index edaae9f..da4e2c1 100644
--- a/drivers/infiniband/core/Makefile
+++ b/drivers/infiniband/core/Makefile
@@ -10,7 +10,8 @@ obj-$(CONFIG_INFINIBAND_USER_ACCESS) += ib_uverbs.o ib_ucm.o \
ib_core-y := packer.o ud_header.o verbs.o cq.o rw.o sysfs.o \
device.o fmr_pool.o cache.o netlink.o \
roce_gid_mgmt.o mr_pool.o addr.o sa_query.o \
- multicast.o mad.o smi.o agent.o mad_rmpp.o
+ multicast.o mad.o smi.o agent.o mad_rmpp.o \
+ security.o
ib_core-$(CONFIG_INFINIBAND_USER_MEM) += umem.o
ib_core-$(CONFIG_INFINIBAND_ON_DEMAND_PAGING) += umem_odp.o umem_rbtree.o
diff --git a/drivers/infiniband/core/cache.c b/drivers/infiniband/core/cache.c
index 8a44894..ebe0337 100644
--- a/drivers/infiniband/core/cache.c
+++ b/drivers/infiniband/core/cache.c
@@ -53,6 +53,7 @@ struct ib_update_work {
struct work_struct work;
struct ib_device *device;
u8 port_num;
+ bool enforce_security;
};
union ib_gid zgid;
@@ -1044,7 +1045,8 @@ int ib_get_cached_lmc(struct ib_device *device,
EXPORT_SYMBOL(ib_get_cached_lmc);
static void ib_cache_update(struct ib_device *device,
- u8 port)
+ u8 port,
+ bool enforce_security)
{
struct ib_port_attr *tprops = NULL;
struct ib_pkey_cache *pkey_cache = NULL, *old_pkey_cache;
@@ -1132,6 +1134,11 @@ static void ib_cache_update(struct ib_device *device,
tprops->subnet_prefix;
write_unlock_irq(&device->cache.lock);
+ if (enforce_security)
+ ib_security_cache_change(device,
+ port,
+ tprops->subnet_prefix);
+
kfree(gid_cache);
kfree(old_pkey_cache);
kfree(tprops);
@@ -1148,7 +1155,9 @@ static void ib_cache_task(struct work_struct *_work)
struct ib_update_work *work =
container_of(_work, struct ib_update_work, work);
- ib_cache_update(work->device, work->port_num);
+ ib_cache_update(work->device,
+ work->port_num,
+ work->enforce_security);
kfree(work);
}
@@ -1169,6 +1178,12 @@ static void ib_cache_event(struct ib_event_handler *handler,
INIT_WORK(&work->work, ib_cache_task);
work->device = event->device;
work->port_num = event->element.port_num;
+ if (event->event == IB_EVENT_PKEY_CHANGE ||
+ event->event == IB_EVENT_GID_CHANGE)
+ work->enforce_security = true;
+ else
+ work->enforce_security = false;
+
queue_work(ib_wq, &work->work);
}
}
@@ -1209,7 +1224,7 @@ int ib_cache_setup_one(struct ib_device *device)
return err;
for (p = 0; p <= rdma_end_port(device) - rdma_start_port(device); ++p)
- ib_cache_update(device, p + rdma_start_port(device));
+ ib_cache_update(device, p + rdma_start_port(device), true);
INIT_IB_EVENT_HANDLER(&device->cache.event_handler,
device, ib_cache_event);
diff --git a/drivers/infiniband/core/core_priv.h b/drivers/infiniband/core/core_priv.h
index ce826e4..68e3de0 100644
--- a/drivers/infiniband/core/core_priv.h
+++ b/drivers/infiniband/core/core_priv.h
@@ -38,6 +38,14 @@
#include <rdma/ib_verbs.h>
+struct pkey_index_qp_list {
+ struct list_head pkey_index_list;
+ u16 pkey_index;
+ /* Lock to hold while iterating the qp_list. */
+ spinlock_t qp_list_lock;
+ struct list_head qp_list;
+};
+
#if IS_ENABLED(CONFIG_INFINIBAND_ADDR_TRANS_CONFIGFS)
int cma_configfs_init(void);
void cma_configfs_exit(void);
@@ -156,4 +164,73 @@ int ib_nl_handle_ip_res_resp(struct sk_buff *skb,
int ib_get_cached_subnet_prefix(struct ib_device *device,
u8 port_num,
u64 *sn_pfx);
+
+#ifdef CONFIG_SECURITY_INFINIBAND
+void ib_security_destroy_port_pkey_list(struct ib_device *device);
+
+void ib_security_cache_change(struct ib_device *device,
+ u8 port_num,
+ u64 subnet_prefix);
+
+int ib_security_modify_qp(struct ib_qp *qp,
+ struct ib_qp_attr *qp_attr,
+ int qp_attr_mask,
+ struct ib_udata *udata);
+
+int ib_create_qp_security(struct ib_qp *qp, struct ib_device *dev);
+void ib_destroy_qp_security_begin(struct ib_qp_security *sec);
+void ib_destroy_qp_security_abort(struct ib_qp_security *sec);
+void ib_destroy_qp_security_end(struct ib_qp_security *sec);
+int ib_open_shared_qp_security(struct ib_qp *qp, struct ib_device *dev);
+void ib_close_shared_qp_security(struct ib_qp_security *sec);
+#else
+static inline void ib_security_destroy_port_pkey_list(struct ib_device *device)
+{
+}
+
+static inline void ib_security_cache_change(struct ib_device *device,
+ u8 port_num,
+ u64 subnet_prefix)
+{
+}
+
+static inline int ib_security_modify_qp(struct ib_qp *qp,
+ struct ib_qp_attr *qp_attr,
+ int qp_attr_mask,
+ struct ib_udata *udata)
+{
+ return qp->device->modify_qp(qp->real_qp,
+ qp_attr,
+ qp_attr_mask,
+ udata);
+}
+
+static inline int ib_create_qp_security(struct ib_qp *qp,
+ struct ib_device *dev)
+{
+ return 0;
+}
+
+static inline void ib_destroy_qp_security_begin(struct ib_qp_security *sec)
+{
+}
+
+static inline void ib_destroy_qp_security_abort(struct ib_qp_security *sec)
+{
+}
+
+static inline void ib_destroy_qp_security_end(struct ib_qp_security *sec)
+{
+}
+
+static inline int ib_open_shared_qp_security(struct ib_qp *qp,
+ struct ib_device *dev)
+{
+ return 0;
+}
+
+static inline void ib_close_shared_qp_security(struct ib_qp_security *sec)
+{
+}
+#endif
#endif /* _CORE_PRIV_H */
diff --git a/drivers/infiniband/core/device.c b/drivers/infiniband/core/device.c
index 760ef60..5b42e83 100644
--- a/drivers/infiniband/core/device.c
+++ b/drivers/infiniband/core/device.c
@@ -320,6 +320,30 @@ void ib_get_device_fw_str(struct ib_device *dev, char *str, size_t str_len)
}
EXPORT_SYMBOL(ib_get_device_fw_str);
+static int setup_port_pkey_list(struct ib_device *device)
+{
+ int i;
+
+ /**
+ * device->port_pkey_list is indexed directly by the port number,
+ * Therefore it is declared as a 1 based array with potential empty
+ * slots at the beginning.
+ */
+ device->port_pkey_list = kzalloc(sizeof(*device->port_pkey_list)
+ * (rdma_end_port(device) + 1),
+ GFP_KERNEL);
+
+ if (!device->port_pkey_list)
+ return -ENOMEM;
+
+ for (i = 0; i < (rdma_end_port(device) + 1); i++) {
+ spin_lock_init(&device->port_pkey_list[i].list_lock);
+ INIT_LIST_HEAD(&device->port_pkey_list[i].pkey_list);
+ }
+
+ return 0;
+}
+
/**
* ib_register_device - Register an IB device with IB core
* @device:Device to register
@@ -357,6 +381,12 @@ int ib_register_device(struct ib_device *device,
goto out;
}
+ ret = setup_port_pkey_list(device);
+ if (ret) {
+ dev_warn(device->dma_device, "Couldn't create per port_pkey_list\n");
+ goto out;
+ }
+
ret = ib_cache_setup_one(device);
if (ret) {
pr_warn("Couldn't set up InfiniBand P_Key/GID cache\n");
@@ -427,6 +457,9 @@ void ib_unregister_device(struct ib_device *device)
ib_device_unregister_sysfs(device);
ib_cache_cleanup_one(device);
+ ib_security_destroy_port_pkey_list(device);
+ kfree(device->port_pkey_list);
+
down_write(&lists_rwsem);
spin_lock_irqsave(&device->client_data_lock, flags);
list_for_each_entry_safe(context, tmp, &device->client_data_list, list)
diff --git a/drivers/infiniband/core/security.c b/drivers/infiniband/core/security.c
new file mode 100644
index 0000000..44ff452
--- /dev/null
+++ b/drivers/infiniband/core/security.c
@@ -0,0 +1,617 @@
+/*
+ * Copyright (c) 2016 Mellanox Technologies Ltd. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses. You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ * Redistribution and use in source and binary forms, with or
+ * without modification, are permitted provided that the following
+ * conditions are met:
+ *
+ * - Redistributions of source code must retain the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer.
+ *
+ * - Redistributions in binary form must reproduce the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer in the documentation and/or other materials
+ * provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifdef CONFIG_SECURITY_INFINIBAND
+
+#include <linux/security.h>
+#include <linux/completion.h>
+#include <linux/list.h>
+
+#include <rdma/ib_verbs.h>
+#include <rdma/ib_cache.h>
+#include "core_priv.h"
+
+static struct pkey_index_qp_list *get_pkey_idx_qp_list(struct ib_port_pkey *pp)
+{
+ struct pkey_index_qp_list *pkey = NULL;
+ struct pkey_index_qp_list *tmp_pkey;
+ struct ib_device *dev = pp->sec->dev;
+
+ spin_lock(&dev->port_pkey_list[pp->port_num].list_lock);
+ list_for_each_entry(tmp_pkey,
+ &dev->port_pkey_list[pp->port_num].pkey_list,
+ pkey_index_list) {
+ if (tmp_pkey->pkey_index == pp->pkey_index) {
+ pkey = tmp_pkey;
+ break;
+ }
+ }
+ spin_unlock(&dev->port_pkey_list[pp->port_num].list_lock);
+ return pkey;
+}
+
+static int get_pkey_and_subnet_prefix(struct ib_port_pkey *pp,
+ u16 *pkey,
+ u64 *subnet_prefix)
+{
+ struct ib_device *dev = pp->sec->dev;
+ int ret;
+
+ ret = ib_get_cached_pkey(dev, pp->port_num, pp->pkey_index, pkey);
+ if (ret)
+ return ret;
+
+ ret = ib_get_cached_subnet_prefix(dev, pp->port_num, subnet_prefix);
+
+ return ret;
+}
+
+static int enforce_qp_pkey_security(u16 pkey,
+ u64 subnet_prefix,
+ struct ib_qp_security *qp_sec)
+{
+ struct ib_qp_security *shared_qp_sec;
+ int ret;
+
+ ret = security_ib_pkey_access(subnet_prefix, pkey, qp_sec->security);
+ if (ret)
+ return ret;
+
+ if (qp_sec->qp == qp_sec->qp->real_qp) {
+ list_for_each_entry(shared_qp_sec,
+ &qp_sec->shared_qp_list,
+ shared_qp_list) {
+ ret = security_ib_pkey_access(subnet_prefix,
+ pkey,
+ shared_qp_sec->security);
+ if (ret)
+ return ret;
+ }
+ }
+ return 0;
+}
+
+/* The caller of this function must hold the QP security
+ * mutex of the QP of the security structure in *pps.
+ *
+ * It takes separate ports_pkeys and security structure
+ * because in some cases the pps will be for a new settings
+ * or the pps will be for the real QP and security structure
+ * will be for a shared QP.
+ */
+static int check_qp_port_pkey_settings(struct ib_ports_pkeys *pps,
+ struct ib_qp_security *sec)
+{
+ u64 subnet_prefix;
+ u16 pkey;
+ int ret = 0;
+
+ if (!pps)
+ return 0;
+
+ if (pps->main.state != IB_PORT_PKEY_NOT_VALID) {
+ get_pkey_and_subnet_prefix(&pps->main,
+ &pkey,
+ &subnet_prefix);
+
+ ret = enforce_qp_pkey_security(pkey,
+ subnet_prefix,
+ sec);
+ }
+ if (ret)
+ goto out;
+
+ if (pps->alt.state != IB_PORT_PKEY_NOT_VALID) {
+ get_pkey_and_subnet_prefix(&pps->alt,
+ &pkey,
+ &subnet_prefix);
+
+ ret = enforce_qp_pkey_security(pkey,
+ subnet_prefix,
+ sec);
+ }
+
+ if (ret)
+ goto out;
+
+out:
+ return ret;
+}
+
+/* The caller of this function must hold the QP security
+ * mutex.
+ */
+static void qp_to_error(struct ib_qp_security *sec)
+{
+ struct ib_qp_security *shared_qp_sec;
+ struct ib_qp_attr attr = {
+ .qp_state = IB_QPS_ERR
+ };
+ struct ib_event event = {
+ .event = IB_EVENT_QP_FATAL
+ };
+
+ /* If the QP is in the process of being destroyed
+ * the qp pointer in the security structure is
+ * undefined. It cannot be modified now.
+ */
+ if (sec->destroying)
+ return;
+
+ ib_modify_qp(sec->qp,
+ &attr,
+ IB_QP_STATE);
+
+ if (sec->qp->event_handler && sec->qp->qp_context) {
+ event.element.qp = sec->qp;
+ sec->qp->event_handler(&event,
+ sec->qp->qp_context);
+ }
+
+ list_for_each_entry(shared_qp_sec,
+ &sec->shared_qp_list,
+ shared_qp_list) {
+ struct ib_qp *qp = shared_qp_sec->qp;
+
+ if (qp->event_handler && qp->qp_context) {
+ event.element.qp = qp;
+ event.device = qp->device;
+ qp->event_handler(&event,
+ qp->qp_context);
+ }
+ }
+}
+
+static inline void check_pkey_qps(struct pkey_index_qp_list *pkey,
+ struct ib_device *device,
+ u8 port_num,
+ u64 subnet_prefix)
+{
+ struct ib_port_pkey *pp, *tmp_pp;
+ bool comp;
+ LIST_HEAD(to_error_list);
+ u16 pkey_val;
+
+ if (!ib_get_cached_pkey(device,
+ port_num,
+ pkey->pkey_index,
+ &pkey_val)) {
+ spin_lock(&pkey->qp_list_lock);
+ list_for_each_entry(pp, &pkey->qp_list, qp_list) {
+ if (atomic_read(&pp->sec->error_list_count))
+ continue;
+
+ if (enforce_qp_pkey_security(pkey_val,
+ subnet_prefix,
+ pp->sec)) {
+ atomic_inc(&pp->sec->error_list_count);
+ list_add(&pp->to_error_list,
+ &to_error_list);
+ }
+ }
+ spin_unlock(&pkey->qp_list_lock);
+ }
+
+ list_for_each_entry_safe(pp,
+ tmp_pp,
+ &to_error_list,
+ to_error_list) {
+ mutex_lock(&pp->sec->mutex);
+ qp_to_error(pp->sec);
+ list_del(&pp->to_error_list);
+ atomic_dec(&pp->sec->error_list_count);
+ comp = pp->sec->destroying;
+ mutex_unlock(&pp->sec->mutex);
+
+ if (comp)
+ complete(&pp->sec->error_complete);
+ }
+}
+
+/* The caller of this function must hold the QP security
+ * mutex.
+ */
+static int port_pkey_list_insert(struct ib_port_pkey *pp)
+{
+ struct pkey_index_qp_list *tmp_pkey;
+ struct pkey_index_qp_list *pkey;
+ struct ib_device *dev;
+ u8 port_num = pp->port_num;
+ int ret = 0;
+
+ if (pp->state != IB_PORT_PKEY_VALID)
+ return 0;
+
+ dev = pp->sec->dev;
+
+ pkey = get_pkey_idx_qp_list(pp);
+
+ if (!pkey) {
+ bool found = false;
+
+ pkey = kzalloc(sizeof(*pkey), GFP_KERNEL);
+ if (!pkey)
+ return -ENOMEM;
+
+ spin_lock(&dev->port_pkey_list[port_num].list_lock);
+ /* Check for the PKey again. A racing process may
+ * have created it.
+ */
+ list_for_each_entry(tmp_pkey,
+ &dev->port_pkey_list[port_num].pkey_list,
+ pkey_index_list) {
+ if (tmp_pkey->pkey_index == pp->pkey_index) {
+ kfree(pkey);
+ pkey = tmp_pkey;
+ found = true;
+ break;
+ }
+ }
+
+ if (!found) {
+ pkey->pkey_index = pp->pkey_index;
+ spin_lock_init(&pkey->qp_list_lock);
+ INIT_LIST_HEAD(&pkey->qp_list);
+ list_add(&pkey->pkey_index_list,
+ &dev->port_pkey_list[port_num].pkey_list);
+ }
+ spin_unlock(&dev->port_pkey_list[port_num].list_lock);
+ }
+
+ spin_lock(&pkey->qp_list_lock);
+ list_add(&pp->qp_list, &pkey->qp_list);
+ spin_unlock(&pkey->qp_list_lock);
+
+ pp->state = IB_PORT_PKEY_LISTED;
+
+ return ret;
+}
+
+/* The caller of this function must hold the QP security
+ * mutex.
+ */
+static void port_pkey_list_remove(struct ib_port_pkey *pp)
+{
+ struct pkey_index_qp_list *pkey;
+
+ if (pp->state != IB_PORT_PKEY_LISTED)
+ return;
+
+ pkey = get_pkey_idx_qp_list(pp);
+
+ spin_lock(&pkey->qp_list_lock);
+ list_del(&pp->qp_list);
+ spin_unlock(&pkey->qp_list_lock);
+
+ /* The setting may still be valid, i.e. after
+ * a destroy has failed for example.
+ */
+ pp->state = IB_PORT_PKEY_VALID;
+}
+
+static void destroy_qp_security(struct ib_qp_security *sec)
+{
+ security_ib_free_security(sec->security);
+ kfree(sec->ports_pkeys);
+ kfree(sec);
+}
+
+/* The caller of this function must hold the QP security
+ * mutex.
+ */
+static struct ib_ports_pkeys *get_new_pps(const struct ib_qp *qp,
+ const struct ib_qp_attr *qp_attr,
+ int qp_attr_mask)
+{
+ struct ib_ports_pkeys *new_pps;
+ struct ib_ports_pkeys *qp_pps = qp->qp_sec->ports_pkeys;
+
+ new_pps = kzalloc(sizeof(*new_pps), GFP_KERNEL);
+ if (!new_pps)
+ return NULL;
+
+ if (qp_attr_mask & (IB_QP_PKEY_INDEX | IB_QP_PORT)) {
+ if (!qp_pps) {
+ new_pps->main.port_num = qp_attr->port_num;
+ new_pps->main.pkey_index = qp_attr->pkey_index;
+ } else {
+ new_pps->main.port_num = (qp_attr_mask & IB_QP_PORT) ?
+ qp_attr->port_num :
+ qp_pps->main.port_num;
+
+ new_pps->main.pkey_index =
+ (qp_attr_mask & IB_QP_PKEY_INDEX) ?
+ qp_attr->pkey_index :
+ qp_pps->main.pkey_index;
+ }
+ new_pps->main.state = IB_PORT_PKEY_VALID;
+ } else if (qp_pps) {
+ new_pps->main.port_num = qp_pps->main.port_num;
+ new_pps->main.pkey_index = qp_pps->main.pkey_index;
+ if (qp_pps->main.state != IB_PORT_PKEY_NOT_VALID)
+ new_pps->main.state = IB_PORT_PKEY_VALID;
+ }
+
+ if (qp_attr_mask & IB_QP_ALT_PATH) {
+ new_pps->alt.port_num = qp_attr->alt_port_num;
+ new_pps->alt.pkey_index = qp_attr->alt_pkey_index;
+ new_pps->alt.state = IB_PORT_PKEY_VALID;
+ } else if (qp_pps) {
+ new_pps->alt.port_num = qp_pps->alt.port_num;
+ new_pps->alt.pkey_index = qp_pps->alt.pkey_index;
+ if (qp_pps->alt.state != IB_PORT_PKEY_NOT_VALID)
+ new_pps->alt.state = IB_PORT_PKEY_VALID;
+ }
+
+ new_pps->main.sec = qp->qp_sec;
+ new_pps->alt.sec = qp->qp_sec;
+ return new_pps;
+}
+
+int ib_open_shared_qp_security(struct ib_qp *qp, struct ib_device *dev)
+{
+ struct ib_qp *real_qp = qp->real_qp;
+ int ret;
+
+ ret = ib_create_qp_security(qp, dev);
+
+ if (ret)
+ goto out;
+
+ mutex_lock(&real_qp->qp_sec->mutex);
+ ret = check_qp_port_pkey_settings(real_qp->qp_sec->ports_pkeys,
+ qp->qp_sec);
+
+ if (ret)
+ goto ret;
+
+ if (qp != real_qp)
+ list_add(&qp->qp_sec->shared_qp_list,
+ &real_qp->qp_sec->shared_qp_list);
+ret:
+ mutex_unlock(&real_qp->qp_sec->mutex);
+ if (ret)
+ destroy_qp_security(qp->qp_sec);
+
+out:
+ return ret;
+}
+
+void ib_close_shared_qp_security(struct ib_qp_security *sec)
+{
+ struct ib_qp *real_qp = sec->qp->real_qp;
+
+ mutex_lock(&real_qp->qp_sec->mutex);
+ list_del(&sec->shared_qp_list);
+ mutex_unlock(&real_qp->qp_sec->mutex);
+
+ destroy_qp_security(sec);
+}
+
+int ib_create_qp_security(struct ib_qp *qp, struct ib_device *dev)
+{
+ int ret;
+
+ qp->qp_sec = kzalloc(sizeof(*qp->qp_sec), GFP_KERNEL);
+ if (!qp->qp_sec)
+ return -ENOMEM;
+
+ qp->qp_sec->qp = qp;
+ qp->qp_sec->dev = dev;
+ mutex_init(&qp->qp_sec->mutex);
+ INIT_LIST_HEAD(&qp->qp_sec->shared_qp_list);
+ atomic_set(&qp->qp_sec->error_list_count, 0);
+ init_completion(&qp->qp_sec->error_complete);
+ ret = security_ib_alloc_security(&qp->qp_sec->security);
+ if (ret)
+ kfree(qp->qp_sec);
+
+ return ret;
+}
+EXPORT_SYMBOL(ib_create_qp_security);
+
+void ib_destroy_qp_security_begin(struct ib_qp_security *sec)
+{
+ mutex_lock(&sec->mutex);
+
+ /* Remove the QP from the lists so it won't get added to
+ * a to_error_list during the destroy process.
+ */
+ if (sec->ports_pkeys) {
+ port_pkey_list_remove(&sec->ports_pkeys->main);
+ port_pkey_list_remove(&sec->ports_pkeys->alt);
+ }
+
+ /* If the QP is already in one or more of those lists
+ * the destroying flag will ensure the to error flow
+ * doesn't operate on an undefined QP.
+ */
+ sec->destroying = true;
+
+ /* Record the error list count to know how many completions
+ * to wait for.
+ */
+ sec->error_comps_pending = atomic_read(&sec->error_list_count);
+
+ mutex_unlock(&sec->mutex);
+}
+
+void ib_destroy_qp_security_abort(struct ib_qp_security *sec)
+{
+ int ret;
+ int i;
+
+ /* If a concurrent cache update is in progress this
+ * QP security could be marked for an error state
+ * transition. Wait for this to complete.
+ */
+ for (i = 0; i < sec->error_comps_pending; i++)
+ wait_for_completion(&sec->error_complete);
+
+ mutex_lock(&sec->mutex);
+ sec->destroying = false;
+
+ /* Restore the position in the lists and verify
+ * access is still allowed in case a cache update
+ * occurred while attempting to destroy.
+ *
+ * Because these setting were listed already
+ * and removed during ib_destroy_qp_security_begin
+ * we know the pkey_index_qp_list for the PKey
+ * already exists so port_pkey_list_insert won't fail.
+ */
+ if (sec->ports_pkeys) {
+ port_pkey_list_insert(&sec->ports_pkeys->main);
+ port_pkey_list_insert(&sec->ports_pkeys->alt);
+ }
+
+ ret = check_qp_port_pkey_settings(sec->ports_pkeys, sec);
+ if (ret)
+ qp_to_error(sec);
+
+ mutex_unlock(&sec->mutex);
+}
+
+void ib_destroy_qp_security_end(struct ib_qp_security *sec)
+{
+ int i;
+
+ /* If a concurrent cache update is occurring we must
+ * wait until this QP security structure is processed
+ * in the QP to error flow before destroying it because
+ * the to_error_list is in use.
+ */
+ for (i = 0; i < sec->error_comps_pending; i++)
+ wait_for_completion(&sec->error_complete);
+
+ destroy_qp_security(sec);
+}
+
+void ib_security_cache_change(struct ib_device *device,
+ u8 port_num,
+ u64 subnet_prefix)
+{
+ struct pkey_index_qp_list *pkey;
+
+ list_for_each_entry(pkey,
+ &device->port_pkey_list[port_num].pkey_list,
+ pkey_index_list) {
+ check_pkey_qps(pkey,
+ device,
+ port_num,
+ subnet_prefix);
+ }
+}
+
+void ib_security_destroy_port_pkey_list(struct ib_device *device)
+{
+ struct pkey_index_qp_list *pkey, *tmp_pkey;
+ int i;
+
+ for (i = rdma_start_port(device); i <= rdma_end_port(device); i++) {
+ spin_lock(&device->port_pkey_list[i].list_lock);
+ list_for_each_entry_safe(pkey,
+ tmp_pkey,
+ &device->port_pkey_list[i].pkey_list,
+ pkey_index_list) {
+ list_del(&pkey->pkey_index_list);
+ kfree(pkey);
+ }
+ spin_unlock(&device->port_pkey_list[i].list_lock);
+ }
+}
+
+int ib_security_modify_qp(struct ib_qp *qp,
+ struct ib_qp_attr *qp_attr,
+ int qp_attr_mask,
+ struct ib_udata *udata)
+{
+ int ret = 0;
+ struct ib_ports_pkeys *tmp_pps;
+ struct ib_ports_pkeys *new_pps;
+ bool special_qp = (qp->qp_type == IB_QPT_SMI ||
+ qp->qp_type == IB_QPT_GSI);
+ bool pps_change = ((qp_attr_mask & (IB_QP_PKEY_INDEX | IB_QP_PORT)) ||
+ (qp_attr_mask & IB_QP_ALT_PATH));
+
+ if (pps_change && !special_qp) {
+ mutex_lock(&qp->qp_sec->mutex);
+ new_pps = get_new_pps(qp,
+ qp_attr,
+ qp_attr_mask);
+
+ /* Add this QP to the lists for the new port
+ * and pkey settings before checking for permission
+ * in case there is a concurrent cache update
+ * occurring. Walking the list for a cache change
+ * doesn't acquire the security mutex unless it's
+ * sending the QP to error.
+ */
+ ret = port_pkey_list_insert(&new_pps->main);
+
+ if (!ret)
+ ret = port_pkey_list_insert(&new_pps->alt);
+
+ if (!ret)
+ ret = check_qp_port_pkey_settings(new_pps,
+ qp->qp_sec);
+ }
+
+ if (!ret)
+ ret = qp->device->modify_qp(qp->real_qp,
+ qp_attr,
+ qp_attr_mask,
+ udata);
+
+ if (pps_change && !special_qp) {
+ /* Clean up the lists and free the appropriate
+ * ports_pkeys structure.
+ */
+ if (ret) {
+ tmp_pps = new_pps;
+ } else {
+ tmp_pps = qp->qp_sec->ports_pkeys;
+ qp->qp_sec->ports_pkeys = new_pps;
+ }
+
+ if (tmp_pps) {
+ port_pkey_list_remove(&tmp_pps->main);
+ port_pkey_list_remove(&tmp_pps->alt);
+ }
+ kfree(tmp_pps);
+ mutex_unlock(&qp->qp_sec->mutex);
+ }
+ return ret;
+}
+EXPORT_SYMBOL(ib_security_modify_qp);
+
+#endif /* CONFIG_SECURITY_INFINIBAND */
diff --git a/drivers/infiniband/core/uverbs_cmd.c b/drivers/infiniband/core/uverbs_cmd.c
index f664731..abeb391 100644
--- a/drivers/infiniband/core/uverbs_cmd.c
+++ b/drivers/infiniband/core/uverbs_cmd.c
@@ -37,7 +37,7 @@
#include <linux/fs.h>
#include <linux/slab.h>
#include <linux/sched.h>
-
+#include <linux/security.h>
#include <asm/uaccess.h>
#include "uverbs.h"
@@ -1915,6 +1915,10 @@ static int create_qp(struct ib_uverbs_file *file,
}
if (cmd->qp_type != IB_QPT_XRC_TGT) {
+ ret = ib_create_qp_security(qp, device);
+ if (ret)
+ goto err_destroy;
+
qp->real_qp = qp;
qp->device = device;
qp->pd = pd;
@@ -2405,10 +2409,18 @@ ssize_t ib_uverbs_modify_qp(struct ib_uverbs_file *file,
ret = ib_resolve_eth_dmac(qp, attr, &cmd.attr_mask);
if (ret)
goto release_qp;
- ret = qp->device->modify_qp(qp, attr,
- modify_qp_mask(qp->qp_type, cmd.attr_mask), &udata);
+
+ ret = ib_security_modify_qp(qp,
+ attr,
+ modify_qp_mask(qp->qp_type,
+ cmd.attr_mask),
+ &udata);
} else {
- ret = ib_modify_qp(qp, attr, modify_qp_mask(qp->qp_type, cmd.attr_mask));
+ ret = ib_security_modify_qp(qp,
+ attr,
+ modify_qp_mask(qp->qp_type,
+ cmd.attr_mask),
+ NULL);
}
if (ret)
diff --git a/drivers/infiniband/core/verbs.c b/drivers/infiniband/core/verbs.c
index 6916d5c..e783f54 100644
--- a/drivers/infiniband/core/verbs.c
+++ b/drivers/infiniband/core/verbs.c
@@ -44,6 +44,7 @@
#include <linux/in.h>
#include <linux/in6.h>
#include <net/addrconf.h>
+#include <linux/security.h>
#include <rdma/ib_verbs.h>
#include <rdma/ib_cache.h>
@@ -682,12 +683,20 @@ static struct ib_qp *__ib_open_qp(struct ib_qp *real_qp,
{
struct ib_qp *qp;
unsigned long flags;
+ int err;
qp = kzalloc(sizeof *qp, GFP_KERNEL);
if (!qp)
return ERR_PTR(-ENOMEM);
qp->real_qp = real_qp;
+ err = ib_open_shared_qp_security(qp, real_qp->device);
+ if (err) {
+ kfree(qp);
+ return ERR_PTR(err);
+ }
+
+ qp->real_qp = real_qp;
atomic_inc(&real_qp->usecnt);
qp->device = real_qp->device;
qp->event_handler = event_handler;
@@ -773,6 +782,12 @@ struct ib_qp *ib_create_qp(struct ib_pd *pd,
if (IS_ERR(qp))
return qp;
+ ret = ib_create_qp_security(qp, device);
+ if (ret) {
+ ib_destroy_qp(qp);
+ return ERR_PTR(ret);
+ }
+
qp->device = device;
qp->real_qp = qp;
qp->uobject = NULL;
@@ -1222,7 +1237,7 @@ int ib_modify_qp(struct ib_qp *qp,
if (ret)
return ret;
- return qp->device->modify_qp(qp->real_qp, qp_attr, qp_attr_mask, NULL);
+ return ib_security_modify_qp(qp->real_qp, qp_attr, qp_attr_mask, NULL);
}
EXPORT_SYMBOL(ib_modify_qp);
@@ -1251,6 +1266,7 @@ int ib_close_qp(struct ib_qp *qp)
spin_unlock_irqrestore(&real_qp->device->event_handler_lock, flags);
atomic_dec(&real_qp->usecnt);
+ ib_close_shared_qp_security(qp->qp_sec);
kfree(qp);
return 0;
@@ -1291,6 +1307,7 @@ int ib_destroy_qp(struct ib_qp *qp)
struct ib_cq *scq, *rcq;
struct ib_srq *srq;
struct ib_rwq_ind_table *ind_tbl;
+ struct ib_qp_security *sec;
int ret;
WARN_ON_ONCE(qp->mrs_used > 0);
@@ -1306,6 +1323,9 @@ int ib_destroy_qp(struct ib_qp *qp)
rcq = qp->recv_cq;
srq = qp->srq;
ind_tbl = qp->rwq_ind_tbl;
+ sec = qp->qp_sec;
+ if (sec)
+ ib_destroy_qp_security_begin(sec);
if (!qp->uobject)
rdma_rw_cleanup_mrs(qp);
@@ -1322,6 +1342,11 @@ int ib_destroy_qp(struct ib_qp *qp)
atomic_dec(&srq->usecnt);
if (ind_tbl)
atomic_dec(&ind_tbl->usecnt);
+ if (sec)
+ ib_destroy_qp_security_end(sec);
+ } else {
+ if (sec)
+ ib_destroy_qp_security_abort(sec);
}
return ret;
diff --git a/include/linux/lsm_hooks.h b/include/linux/lsm_hooks.h
index 7ae3976..f85a8f3 100644
--- a/include/linux/lsm_hooks.h
+++ b/include/linux/lsm_hooks.h
@@ -8,6 +8,7 @@
* Copyright (C) 2001 Silicon Graphics, Inc. (Trust Technology Group)
* Copyright (C) 2015 Intel Corporation.
* Copyright (C) 2015 Casey Schaufler <casey-iSGtlc1asvQWG2LlvL+J4A@public.gmane.org>
+ * Copyright (C) 2016 Mellanox Techonologies
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
@@ -876,6 +877,21 @@
* associated with the TUN device's security structure.
* @security pointer to the TUN devices's security structure.
*
+ * Security hooks for Infiniband
+ *
+ * @ib_pkey_access:
+ * Check permission to access a pkey when modifing a QP.
+ * @subnet_prefix the subnet prefix of the port being used.
+ * @pkey the pkey to be accessed.
+ * @sec pointer to a security structure.
+ * @ib_alloc_security:
+ * Allocate a security structure for Infiniband objects.
+ * @sec pointer to a security structure pointer.
+ * Returns 0 on success, non-zero on failure
+ * @ib_free_security:
+ * Deallocate an Infiniband security structure.
+ * @sec contains the security structure to be freed.
+ *
* Security hooks for XFRM operations.
*
* @xfrm_policy_alloc_security:
@@ -1579,6 +1595,12 @@ union security_list_options {
int (*tun_dev_open)(void *security);
#endif /* CONFIG_SECURITY_NETWORK */
+#ifdef CONFIG_SECURITY_INFINIBAND
+ int (*ib_pkey_access)(u64 subnet_prefix, u16 pkey, void *sec);
+ int (*ib_alloc_security)(void **sec);
+ void (*ib_free_security)(void *sec);
+#endif /* CONFIG_SECURITY_INFINIBAND */
+
#ifdef CONFIG_SECURITY_NETWORK_XFRM
int (*xfrm_policy_alloc_security)(struct xfrm_sec_ctx **ctxp,
struct xfrm_user_sec_ctx *sec_ctx,
@@ -1806,6 +1828,11 @@ struct security_hook_heads {
struct list_head tun_dev_attach;
struct list_head tun_dev_open;
#endif /* CONFIG_SECURITY_NETWORK */
+#ifdef CONFIG_SECURITY_INFINIBAND
+ struct list_head ib_pkey_access;
+ struct list_head ib_alloc_security;
+ struct list_head ib_free_security;
+#endif /* CONFIG_SECURITY_INFINIBAND */
#ifdef CONFIG_SECURITY_NETWORK_XFRM
struct list_head xfrm_policy_alloc_security;
struct list_head xfrm_policy_clone_security;
diff --git a/include/linux/security.h b/include/linux/security.h
index 14df373..33e23c4 100644
--- a/include/linux/security.h
+++ b/include/linux/security.h
@@ -6,6 +6,7 @@
* Copyright (C) 2001 Networks Associates Technology, Inc <ssmalley-M06CiZnz2FM@public.gmane.org>
* Copyright (C) 2001 James Morris <jmorris-G2x6lROWQUcJY7gZg3T8ig@public.gmane.org>
* Copyright (C) 2001 Silicon Graphics, Inc. (Trust Technology Group)
+ * Copyright (C) 2016 Mellanox Techonologies
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
@@ -1370,6 +1371,26 @@ static inline int security_tun_dev_open(void *security)
}
#endif /* CONFIG_SECURITY_NETWORK */
+#ifdef CONFIG_SECURITY_INFINIBAND
+int security_ib_pkey_access(u64 subnet_prefix, u16 pkey, void *sec);
+int security_ib_alloc_security(void **sec);
+void security_ib_free_security(void *sec);
+#else /* CONFIG_SECURITY_INFINIBAND */
+static inline int security_ib_pkey_access(u64 subnet_prefix, u16 pkey, void *sec)
+{
+ return 0;
+}
+
+static inline int security_ib_alloc_security(void **sec)
+{
+ return 0;
+}
+
+static inline void security_ib_free_security(void *sec)
+{
+}
+#endif /* CONFIG_SECURITY_INFINIBAND */
+
#ifdef CONFIG_SECURITY_NETWORK_XFRM
int security_xfrm_policy_alloc(struct xfrm_sec_ctx **ctxp,
diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h
index 205ea09..9042ba3 100644
--- a/include/rdma/ib_verbs.h
+++ b/include/rdma/ib_verbs.h
@@ -1490,6 +1490,45 @@ struct ib_rwq_ind_table_init_attr {
struct ib_wq **ind_tbl;
};
+enum port_pkey_state {
+ IB_PORT_PKEY_NOT_VALID = 0,
+ IB_PORT_PKEY_VALID = 1,
+ IB_PORT_PKEY_LISTED = 2,
+};
+
+struct ib_qp_security;
+
+struct ib_port_pkey {
+ enum port_pkey_state state;
+ u16 pkey_index;
+ u8 port_num;
+ struct list_head qp_list;
+ struct list_head to_error_list;
+ struct ib_qp_security *sec;
+};
+
+struct ib_ports_pkeys {
+ struct ib_port_pkey main;
+ struct ib_port_pkey alt;
+};
+
+struct ib_qp_security {
+ struct ib_qp *qp;
+ struct ib_device *dev;
+ /* Hold this mutex when changing port and pkey settings. */
+ struct mutex mutex;
+ struct ib_ports_pkeys *ports_pkeys;
+ /* A list of all open shared QP handles. Required to enforce security
+ * properly for all users of a shared QP.
+ */
+ struct list_head shared_qp_list;
+ void *security;
+ bool destroying;
+ atomic_t error_list_count;
+ struct completion error_complete;
+ int error_comps_pending;
+};
+
struct ib_qp {
struct ib_device *device;
struct ib_pd *pd;
@@ -1513,6 +1552,7 @@ struct ib_qp {
u32 qp_num;
enum ib_qp_type qp_type;
struct ib_rwq_ind_table *rwq_ind_tbl;
+ struct ib_qp_security *qp_sec;
};
struct ib_mr {
@@ -1760,6 +1800,12 @@ struct ib_port_immutable {
u32 max_mad_size;
};
+struct ib_port_pkey_list {
+ /* Lock to hold while modifying the list. */
+ spinlock_t list_lock;
+ struct list_head pkey_list;
+};
+
struct ib_device {
struct device *dma_device;
@@ -1782,6 +1828,8 @@ struct ib_device {
int num_comp_vectors;
+ struct ib_port_pkey_list *port_pkey_list;
+
struct iw_cm_verbs *iwcm;
/**
diff --git a/security/Kconfig b/security/Kconfig
index 176758c..ce965c6 100644
--- a/security/Kconfig
+++ b/security/Kconfig
@@ -49,6 +49,15 @@ config SECURITY_NETWORK
implement socket and networking access controls.
If you are unsure how to answer this question, answer N.
+config SECURITY_INFINIBAND
+ bool "Infiniband Security Hooks"
+ depends on SECURITY && INFINIBAND
+ help
+ This enables the Infiniband security hooks.
+ If enabled, a security module can use these hooks to
+ implement Infiniband access controls.
+ If you are unsure how to answer this question, answer N.
+
config SECURITY_NETWORK_XFRM
bool "XFRM (IPSec) Networking Security Hooks"
depends on XFRM && SECURITY_NETWORK
diff --git a/security/security.c b/security/security.c
index 7095693..234982d 100644
--- a/security/security.c
+++ b/security/security.c
@@ -4,6 +4,7 @@
* Copyright (C) 2001 WireX Communications, Inc <chris-ZMHXrckZAt0AvxtiuMwx3w@public.gmane.org>
* Copyright (C) 2001-2002 Greg Kroah-Hartman <greg-U8xfFu+wG4EAvxtiuMwx3w@public.gmane.org>
* Copyright (C) 2001 Networks Associates Technology, Inc <ssmalley-M06CiZnz2FM@public.gmane.org>
+ * Copyright (C) 2016 Mellanox Technologies
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
@@ -1399,6 +1400,27 @@ EXPORT_SYMBOL(security_tun_dev_open);
#endif /* CONFIG_SECURITY_NETWORK */
+#ifdef CONFIG_SECURITY_INFINIBAND
+
+int security_ib_pkey_access(u64 subnet_prefix, u16 pkey, void *sec)
+{
+ return call_int_hook(ib_pkey_access, 0, subnet_prefix, pkey, sec);
+}
+EXPORT_SYMBOL(security_ib_pkey_access);
+
+int security_ib_alloc_security(void **sec)
+{
+ return call_int_hook(ib_alloc_security, 0, sec);
+}
+EXPORT_SYMBOL(security_ib_alloc_security);
+
+void security_ib_free_security(void *sec)
+{
+ call_void_hook(ib_free_security, sec);
+}
+EXPORT_SYMBOL(security_ib_free_security);
+#endif /* CONFIG_SECURITY_INFINIBAND */
+
#ifdef CONFIG_SECURITY_NETWORK_XFRM
int security_xfrm_policy_alloc(struct xfrm_sec_ctx **ctxp,
@@ -1850,6 +1872,15 @@ struct security_hook_heads security_hook_heads = {
LIST_HEAD_INIT(security_hook_heads.tun_dev_attach),
.tun_dev_open = LIST_HEAD_INIT(security_hook_heads.tun_dev_open),
#endif /* CONFIG_SECURITY_NETWORK */
+
+#ifdef CONFIG_SECURITY_INFINIBAND
+ .ib_pkey_access = LIST_HEAD_INIT(security_hook_heads.ib_pkey_access),
+ .ib_alloc_security =
+ LIST_HEAD_INIT(security_hook_heads.ib_alloc_security),
+ .ib_free_security =
+ LIST_HEAD_INIT(security_hook_heads.ib_free_security),
+#endif /* CONFIG_SECURITY_INFINIBAND */
+
#ifdef CONFIG_SECURITY_NETWORK_XFRM
.xfrm_policy_alloc_security =
LIST_HEAD_INIT(security_hook_heads.xfrm_policy_alloc_security),
--
1.7.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related [flat|nested] 42+ messages in thread
* [PATCH v2 2/9] IB/core: Enforce PKey security on QPs
@ 2016-07-14 22:56 ` Dan Jurgens
0 siblings, 0 replies; 42+ messages in thread
From: Dan Jurgens @ 2016-07-14 22:56 UTC (permalink / raw)
To: chrisw, paul, sds, eparis, dledford, sean.hefty, hal.rosenstock
Cc: selinux, linux-security-module, linux-rdma, yevgenyp, Daniel Jurgens
From: Daniel Jurgens <danielj@mellanox.com>
Add new LSM hooks to allocate and free security contexts and check for
permission to access a PKey.
Allocate and free a security context when creating and destroying a QP.
This context is used for controlling access to PKeys.
When a request is made to modify a QP that changes the port, PKey index,
or alternate path, check that the QP has permission for the PKey in the
PKey table index on the subnet prefix of the port. If the QP is shared
make sure all handles to the QP also have access.
Store which port and PKey index a QP is using. After the reset to init
transition the user can modify the port, PKey index and alternate path
independently. So port and PKey settings changes can be a merge of the
previous settings and the new ones.
In order to maintain access control if there are PKey table or subnet
prefix change keep a list of all QPs are using each PKey index on
each port. If a change occurs all QPs using that device and port must
have access enforced for the new cache settings.
These changes add a transaction to the QP modify process. Association
with the old port and PKey index must be maintained if the modify fails,
and must be removed if it succeeds. Association with the new port and
PKey index must be established prior to the modify and removed if the
modify fails.
1. When a QP is modified to a particular Port, PKey index or alternate
path insert that QP into the appropriate lists.
2. Check permission to access the new settings.
3. If step 2 grants access attempt to modify the QP.
4a. If steps 2 and 3 succeed remove any prior associations.
4b. If ether fails remove the new setting associations.
If a PKey table or subnet prefix changes walk the list of QPs and
check that they have permission. If not send the QP to the error state
and raise a fatal error event. If it's a shared QP make sure all the
QPs that share the real_qp have permission as well. If the QP that
owns a security structure is denied access the security structure is
marked as such and the QP is added to an error_list. Once the moving
the QP to error is complete the security structure mark is cleared.
Maintaining the lists correctly turns QP destroy into a transaction.
The hardware driver for the device frees the ib_qp structure, so while
the destroy is in progress the ib_qp pointer in the ib_qp_security
struct is undefined. When the destroy process begins the ib_qp_security
structure is marked as destroying. This prevents any action from being
taken on the QP pointer. After the QP is destroyed successfully it
could still listed on an error_list wait for it to be processed by that
flow before cleaning up the structure.
If the destroy fails the QPs port and PKey settings are reinserted into
the appropriate lists, the destroying flag is cleared, and access control
is enforced, in case there were any cache changes during the destroy
flow.
To keep the security changes isolated a new file is used to hold security
related functionality.
Signed-off-by: Daniel Jurgens <danielj@mellanox.com>
Reviewed-by: Leon Romanovsky <leonro@mellanox.com>
---
v2:
- Squashed LSM hook additions. Paul Moore
- Changed security blobs to void*. Paul Moore
---
drivers/infiniband/core/Makefile | 3 +-
drivers/infiniband/core/cache.c | 21 +-
drivers/infiniband/core/core_priv.h | 77 +++++
drivers/infiniband/core/device.c | 33 ++
drivers/infiniband/core/security.c | 617 ++++++++++++++++++++++++++++++++++
drivers/infiniband/core/uverbs_cmd.c | 20 +-
drivers/infiniband/core/verbs.c | 27 ++-
include/linux/lsm_hooks.h | 27 ++
include/linux/security.h | 21 ++
include/rdma/ib_verbs.h | 48 +++
security/Kconfig | 9 +
security/security.c | 31 ++
12 files changed, 925 insertions(+), 9 deletions(-)
create mode 100644 drivers/infiniband/core/security.c
diff --git a/drivers/infiniband/core/Makefile b/drivers/infiniband/core/Makefile
index edaae9f..da4e2c1 100644
--- a/drivers/infiniband/core/Makefile
+++ b/drivers/infiniband/core/Makefile
@@ -10,7 +10,8 @@ obj-$(CONFIG_INFINIBAND_USER_ACCESS) += ib_uverbs.o ib_ucm.o \
ib_core-y := packer.o ud_header.o verbs.o cq.o rw.o sysfs.o \
device.o fmr_pool.o cache.o netlink.o \
roce_gid_mgmt.o mr_pool.o addr.o sa_query.o \
- multicast.o mad.o smi.o agent.o mad_rmpp.o
+ multicast.o mad.o smi.o agent.o mad_rmpp.o \
+ security.o
ib_core-$(CONFIG_INFINIBAND_USER_MEM) += umem.o
ib_core-$(CONFIG_INFINIBAND_ON_DEMAND_PAGING) += umem_odp.o umem_rbtree.o
diff --git a/drivers/infiniband/core/cache.c b/drivers/infiniband/core/cache.c
index 8a44894..ebe0337 100644
--- a/drivers/infiniband/core/cache.c
+++ b/drivers/infiniband/core/cache.c
@@ -53,6 +53,7 @@ struct ib_update_work {
struct work_struct work;
struct ib_device *device;
u8 port_num;
+ bool enforce_security;
};
union ib_gid zgid;
@@ -1044,7 +1045,8 @@ int ib_get_cached_lmc(struct ib_device *device,
EXPORT_SYMBOL(ib_get_cached_lmc);
static void ib_cache_update(struct ib_device *device,
- u8 port)
+ u8 port,
+ bool enforce_security)
{
struct ib_port_attr *tprops = NULL;
struct ib_pkey_cache *pkey_cache = NULL, *old_pkey_cache;
@@ -1132,6 +1134,11 @@ static void ib_cache_update(struct ib_device *device,
tprops->subnet_prefix;
write_unlock_irq(&device->cache.lock);
+ if (enforce_security)
+ ib_security_cache_change(device,
+ port,
+ tprops->subnet_prefix);
+
kfree(gid_cache);
kfree(old_pkey_cache);
kfree(tprops);
@@ -1148,7 +1155,9 @@ static void ib_cache_task(struct work_struct *_work)
struct ib_update_work *work =
container_of(_work, struct ib_update_work, work);
- ib_cache_update(work->device, work->port_num);
+ ib_cache_update(work->device,
+ work->port_num,
+ work->enforce_security);
kfree(work);
}
@@ -1169,6 +1178,12 @@ static void ib_cache_event(struct ib_event_handler *handler,
INIT_WORK(&work->work, ib_cache_task);
work->device = event->device;
work->port_num = event->element.port_num;
+ if (event->event == IB_EVENT_PKEY_CHANGE ||
+ event->event == IB_EVENT_GID_CHANGE)
+ work->enforce_security = true;
+ else
+ work->enforce_security = false;
+
queue_work(ib_wq, &work->work);
}
}
@@ -1209,7 +1224,7 @@ int ib_cache_setup_one(struct ib_device *device)
return err;
for (p = 0; p <= rdma_end_port(device) - rdma_start_port(device); ++p)
- ib_cache_update(device, p + rdma_start_port(device));
+ ib_cache_update(device, p + rdma_start_port(device), true);
INIT_IB_EVENT_HANDLER(&device->cache.event_handler,
device, ib_cache_event);
diff --git a/drivers/infiniband/core/core_priv.h b/drivers/infiniband/core/core_priv.h
index ce826e4..68e3de0 100644
--- a/drivers/infiniband/core/core_priv.h
+++ b/drivers/infiniband/core/core_priv.h
@@ -38,6 +38,14 @@
#include <rdma/ib_verbs.h>
+struct pkey_index_qp_list {
+ struct list_head pkey_index_list;
+ u16 pkey_index;
+ /* Lock to hold while iterating the qp_list. */
+ spinlock_t qp_list_lock;
+ struct list_head qp_list;
+};
+
#if IS_ENABLED(CONFIG_INFINIBAND_ADDR_TRANS_CONFIGFS)
int cma_configfs_init(void);
void cma_configfs_exit(void);
@@ -156,4 +164,73 @@ int ib_nl_handle_ip_res_resp(struct sk_buff *skb,
int ib_get_cached_subnet_prefix(struct ib_device *device,
u8 port_num,
u64 *sn_pfx);
+
+#ifdef CONFIG_SECURITY_INFINIBAND
+void ib_security_destroy_port_pkey_list(struct ib_device *device);
+
+void ib_security_cache_change(struct ib_device *device,
+ u8 port_num,
+ u64 subnet_prefix);
+
+int ib_security_modify_qp(struct ib_qp *qp,
+ struct ib_qp_attr *qp_attr,
+ int qp_attr_mask,
+ struct ib_udata *udata);
+
+int ib_create_qp_security(struct ib_qp *qp, struct ib_device *dev);
+void ib_destroy_qp_security_begin(struct ib_qp_security *sec);
+void ib_destroy_qp_security_abort(struct ib_qp_security *sec);
+void ib_destroy_qp_security_end(struct ib_qp_security *sec);
+int ib_open_shared_qp_security(struct ib_qp *qp, struct ib_device *dev);
+void ib_close_shared_qp_security(struct ib_qp_security *sec);
+#else
+static inline void ib_security_destroy_port_pkey_list(struct ib_device *device)
+{
+}
+
+static inline void ib_security_cache_change(struct ib_device *device,
+ u8 port_num,
+ u64 subnet_prefix)
+{
+}
+
+static inline int ib_security_modify_qp(struct ib_qp *qp,
+ struct ib_qp_attr *qp_attr,
+ int qp_attr_mask,
+ struct ib_udata *udata)
+{
+ return qp->device->modify_qp(qp->real_qp,
+ qp_attr,
+ qp_attr_mask,
+ udata);
+}
+
+static inline int ib_create_qp_security(struct ib_qp *qp,
+ struct ib_device *dev)
+{
+ return 0;
+}
+
+static inline void ib_destroy_qp_security_begin(struct ib_qp_security *sec)
+{
+}
+
+static inline void ib_destroy_qp_security_abort(struct ib_qp_security *sec)
+{
+}
+
+static inline void ib_destroy_qp_security_end(struct ib_qp_security *sec)
+{
+}
+
+static inline int ib_open_shared_qp_security(struct ib_qp *qp,
+ struct ib_device *dev)
+{
+ return 0;
+}
+
+static inline void ib_close_shared_qp_security(struct ib_qp_security *sec)
+{
+}
+#endif
#endif /* _CORE_PRIV_H */
diff --git a/drivers/infiniband/core/device.c b/drivers/infiniband/core/device.c
index 760ef60..5b42e83 100644
--- a/drivers/infiniband/core/device.c
+++ b/drivers/infiniband/core/device.c
@@ -320,6 +320,30 @@ void ib_get_device_fw_str(struct ib_device *dev, char *str, size_t str_len)
}
EXPORT_SYMBOL(ib_get_device_fw_str);
+static int setup_port_pkey_list(struct ib_device *device)
+{
+ int i;
+
+ /**
+ * device->port_pkey_list is indexed directly by the port number,
+ * Therefore it is declared as a 1 based array with potential empty
+ * slots at the beginning.
+ */
+ device->port_pkey_list = kzalloc(sizeof(*device->port_pkey_list)
+ * (rdma_end_port(device) + 1),
+ GFP_KERNEL);
+
+ if (!device->port_pkey_list)
+ return -ENOMEM;
+
+ for (i = 0; i < (rdma_end_port(device) + 1); i++) {
+ spin_lock_init(&device->port_pkey_list[i].list_lock);
+ INIT_LIST_HEAD(&device->port_pkey_list[i].pkey_list);
+ }
+
+ return 0;
+}
+
/**
* ib_register_device - Register an IB device with IB core
* @device:Device to register
@@ -357,6 +381,12 @@ int ib_register_device(struct ib_device *device,
goto out;
}
+ ret = setup_port_pkey_list(device);
+ if (ret) {
+ dev_warn(device->dma_device, "Couldn't create per port_pkey_list\n");
+ goto out;
+ }
+
ret = ib_cache_setup_one(device);
if (ret) {
pr_warn("Couldn't set up InfiniBand P_Key/GID cache\n");
@@ -427,6 +457,9 @@ void ib_unregister_device(struct ib_device *device)
ib_device_unregister_sysfs(device);
ib_cache_cleanup_one(device);
+ ib_security_destroy_port_pkey_list(device);
+ kfree(device->port_pkey_list);
+
down_write(&lists_rwsem);
spin_lock_irqsave(&device->client_data_lock, flags);
list_for_each_entry_safe(context, tmp, &device->client_data_list, list)
diff --git a/drivers/infiniband/core/security.c b/drivers/infiniband/core/security.c
new file mode 100644
index 0000000..44ff452
--- /dev/null
+++ b/drivers/infiniband/core/security.c
@@ -0,0 +1,617 @@
+/*
+ * Copyright (c) 2016 Mellanox Technologies Ltd. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses. You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ * Redistribution and use in source and binary forms, with or
+ * without modification, are permitted provided that the following
+ * conditions are met:
+ *
+ * - Redistributions of source code must retain the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer.
+ *
+ * - Redistributions in binary form must reproduce the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer in the documentation and/or other materials
+ * provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifdef CONFIG_SECURITY_INFINIBAND
+
+#include <linux/security.h>
+#include <linux/completion.h>
+#include <linux/list.h>
+
+#include <rdma/ib_verbs.h>
+#include <rdma/ib_cache.h>
+#include "core_priv.h"
+
+static struct pkey_index_qp_list *get_pkey_idx_qp_list(struct ib_port_pkey *pp)
+{
+ struct pkey_index_qp_list *pkey = NULL;
+ struct pkey_index_qp_list *tmp_pkey;
+ struct ib_device *dev = pp->sec->dev;
+
+ spin_lock(&dev->port_pkey_list[pp->port_num].list_lock);
+ list_for_each_entry(tmp_pkey,
+ &dev->port_pkey_list[pp->port_num].pkey_list,
+ pkey_index_list) {
+ if (tmp_pkey->pkey_index == pp->pkey_index) {
+ pkey = tmp_pkey;
+ break;
+ }
+ }
+ spin_unlock(&dev->port_pkey_list[pp->port_num].list_lock);
+ return pkey;
+}
+
+static int get_pkey_and_subnet_prefix(struct ib_port_pkey *pp,
+ u16 *pkey,
+ u64 *subnet_prefix)
+{
+ struct ib_device *dev = pp->sec->dev;
+ int ret;
+
+ ret = ib_get_cached_pkey(dev, pp->port_num, pp->pkey_index, pkey);
+ if (ret)
+ return ret;
+
+ ret = ib_get_cached_subnet_prefix(dev, pp->port_num, subnet_prefix);
+
+ return ret;
+}
+
+static int enforce_qp_pkey_security(u16 pkey,
+ u64 subnet_prefix,
+ struct ib_qp_security *qp_sec)
+{
+ struct ib_qp_security *shared_qp_sec;
+ int ret;
+
+ ret = security_ib_pkey_access(subnet_prefix, pkey, qp_sec->security);
+ if (ret)
+ return ret;
+
+ if (qp_sec->qp == qp_sec->qp->real_qp) {
+ list_for_each_entry(shared_qp_sec,
+ &qp_sec->shared_qp_list,
+ shared_qp_list) {
+ ret = security_ib_pkey_access(subnet_prefix,
+ pkey,
+ shared_qp_sec->security);
+ if (ret)
+ return ret;
+ }
+ }
+ return 0;
+}
+
+/* The caller of this function must hold the QP security
+ * mutex of the QP of the security structure in *pps.
+ *
+ * It takes separate ports_pkeys and security structure
+ * because in some cases the pps will be for a new settings
+ * or the pps will be for the real QP and security structure
+ * will be for a shared QP.
+ */
+static int check_qp_port_pkey_settings(struct ib_ports_pkeys *pps,
+ struct ib_qp_security *sec)
+{
+ u64 subnet_prefix;
+ u16 pkey;
+ int ret = 0;
+
+ if (!pps)
+ return 0;
+
+ if (pps->main.state != IB_PORT_PKEY_NOT_VALID) {
+ get_pkey_and_subnet_prefix(&pps->main,
+ &pkey,
+ &subnet_prefix);
+
+ ret = enforce_qp_pkey_security(pkey,
+ subnet_prefix,
+ sec);
+ }
+ if (ret)
+ goto out;
+
+ if (pps->alt.state != IB_PORT_PKEY_NOT_VALID) {
+ get_pkey_and_subnet_prefix(&pps->alt,
+ &pkey,
+ &subnet_prefix);
+
+ ret = enforce_qp_pkey_security(pkey,
+ subnet_prefix,
+ sec);
+ }
+
+ if (ret)
+ goto out;
+
+out:
+ return ret;
+}
+
+/* The caller of this function must hold the QP security
+ * mutex.
+ */
+static void qp_to_error(struct ib_qp_security *sec)
+{
+ struct ib_qp_security *shared_qp_sec;
+ struct ib_qp_attr attr = {
+ .qp_state = IB_QPS_ERR
+ };
+ struct ib_event event = {
+ .event = IB_EVENT_QP_FATAL
+ };
+
+ /* If the QP is in the process of being destroyed
+ * the qp pointer in the security structure is
+ * undefined. It cannot be modified now.
+ */
+ if (sec->destroying)
+ return;
+
+ ib_modify_qp(sec->qp,
+ &attr,
+ IB_QP_STATE);
+
+ if (sec->qp->event_handler && sec->qp->qp_context) {
+ event.element.qp = sec->qp;
+ sec->qp->event_handler(&event,
+ sec->qp->qp_context);
+ }
+
+ list_for_each_entry(shared_qp_sec,
+ &sec->shared_qp_list,
+ shared_qp_list) {
+ struct ib_qp *qp = shared_qp_sec->qp;
+
+ if (qp->event_handler && qp->qp_context) {
+ event.element.qp = qp;
+ event.device = qp->device;
+ qp->event_handler(&event,
+ qp->qp_context);
+ }
+ }
+}
+
+static inline void check_pkey_qps(struct pkey_index_qp_list *pkey,
+ struct ib_device *device,
+ u8 port_num,
+ u64 subnet_prefix)
+{
+ struct ib_port_pkey *pp, *tmp_pp;
+ bool comp;
+ LIST_HEAD(to_error_list);
+ u16 pkey_val;
+
+ if (!ib_get_cached_pkey(device,
+ port_num,
+ pkey->pkey_index,
+ &pkey_val)) {
+ spin_lock(&pkey->qp_list_lock);
+ list_for_each_entry(pp, &pkey->qp_list, qp_list) {
+ if (atomic_read(&pp->sec->error_list_count))
+ continue;
+
+ if (enforce_qp_pkey_security(pkey_val,
+ subnet_prefix,
+ pp->sec)) {
+ atomic_inc(&pp->sec->error_list_count);
+ list_add(&pp->to_error_list,
+ &to_error_list);
+ }
+ }
+ spin_unlock(&pkey->qp_list_lock);
+ }
+
+ list_for_each_entry_safe(pp,
+ tmp_pp,
+ &to_error_list,
+ to_error_list) {
+ mutex_lock(&pp->sec->mutex);
+ qp_to_error(pp->sec);
+ list_del(&pp->to_error_list);
+ atomic_dec(&pp->sec->error_list_count);
+ comp = pp->sec->destroying;
+ mutex_unlock(&pp->sec->mutex);
+
+ if (comp)
+ complete(&pp->sec->error_complete);
+ }
+}
+
+/* The caller of this function must hold the QP security
+ * mutex.
+ */
+static int port_pkey_list_insert(struct ib_port_pkey *pp)
+{
+ struct pkey_index_qp_list *tmp_pkey;
+ struct pkey_index_qp_list *pkey;
+ struct ib_device *dev;
+ u8 port_num = pp->port_num;
+ int ret = 0;
+
+ if (pp->state != IB_PORT_PKEY_VALID)
+ return 0;
+
+ dev = pp->sec->dev;
+
+ pkey = get_pkey_idx_qp_list(pp);
+
+ if (!pkey) {
+ bool found = false;
+
+ pkey = kzalloc(sizeof(*pkey), GFP_KERNEL);
+ if (!pkey)
+ return -ENOMEM;
+
+ spin_lock(&dev->port_pkey_list[port_num].list_lock);
+ /* Check for the PKey again. A racing process may
+ * have created it.
+ */
+ list_for_each_entry(tmp_pkey,
+ &dev->port_pkey_list[port_num].pkey_list,
+ pkey_index_list) {
+ if (tmp_pkey->pkey_index == pp->pkey_index) {
+ kfree(pkey);
+ pkey = tmp_pkey;
+ found = true;
+ break;
+ }
+ }
+
+ if (!found) {
+ pkey->pkey_index = pp->pkey_index;
+ spin_lock_init(&pkey->qp_list_lock);
+ INIT_LIST_HEAD(&pkey->qp_list);
+ list_add(&pkey->pkey_index_list,
+ &dev->port_pkey_list[port_num].pkey_list);
+ }
+ spin_unlock(&dev->port_pkey_list[port_num].list_lock);
+ }
+
+ spin_lock(&pkey->qp_list_lock);
+ list_add(&pp->qp_list, &pkey->qp_list);
+ spin_unlock(&pkey->qp_list_lock);
+
+ pp->state = IB_PORT_PKEY_LISTED;
+
+ return ret;
+}
+
+/* The caller of this function must hold the QP security
+ * mutex.
+ */
+static void port_pkey_list_remove(struct ib_port_pkey *pp)
+{
+ struct pkey_index_qp_list *pkey;
+
+ if (pp->state != IB_PORT_PKEY_LISTED)
+ return;
+
+ pkey = get_pkey_idx_qp_list(pp);
+
+ spin_lock(&pkey->qp_list_lock);
+ list_del(&pp->qp_list);
+ spin_unlock(&pkey->qp_list_lock);
+
+ /* The setting may still be valid, i.e. after
+ * a destroy has failed for example.
+ */
+ pp->state = IB_PORT_PKEY_VALID;
+}
+
+static void destroy_qp_security(struct ib_qp_security *sec)
+{
+ security_ib_free_security(sec->security);
+ kfree(sec->ports_pkeys);
+ kfree(sec);
+}
+
+/* The caller of this function must hold the QP security
+ * mutex.
+ */
+static struct ib_ports_pkeys *get_new_pps(const struct ib_qp *qp,
+ const struct ib_qp_attr *qp_attr,
+ int qp_attr_mask)
+{
+ struct ib_ports_pkeys *new_pps;
+ struct ib_ports_pkeys *qp_pps = qp->qp_sec->ports_pkeys;
+
+ new_pps = kzalloc(sizeof(*new_pps), GFP_KERNEL);
+ if (!new_pps)
+ return NULL;
+
+ if (qp_attr_mask & (IB_QP_PKEY_INDEX | IB_QP_PORT)) {
+ if (!qp_pps) {
+ new_pps->main.port_num = qp_attr->port_num;
+ new_pps->main.pkey_index = qp_attr->pkey_index;
+ } else {
+ new_pps->main.port_num = (qp_attr_mask & IB_QP_PORT) ?
+ qp_attr->port_num :
+ qp_pps->main.port_num;
+
+ new_pps->main.pkey_index =
+ (qp_attr_mask & IB_QP_PKEY_INDEX) ?
+ qp_attr->pkey_index :
+ qp_pps->main.pkey_index;
+ }
+ new_pps->main.state = IB_PORT_PKEY_VALID;
+ } else if (qp_pps) {
+ new_pps->main.port_num = qp_pps->main.port_num;
+ new_pps->main.pkey_index = qp_pps->main.pkey_index;
+ if (qp_pps->main.state != IB_PORT_PKEY_NOT_VALID)
+ new_pps->main.state = IB_PORT_PKEY_VALID;
+ }
+
+ if (qp_attr_mask & IB_QP_ALT_PATH) {
+ new_pps->alt.port_num = qp_attr->alt_port_num;
+ new_pps->alt.pkey_index = qp_attr->alt_pkey_index;
+ new_pps->alt.state = IB_PORT_PKEY_VALID;
+ } else if (qp_pps) {
+ new_pps->alt.port_num = qp_pps->alt.port_num;
+ new_pps->alt.pkey_index = qp_pps->alt.pkey_index;
+ if (qp_pps->alt.state != IB_PORT_PKEY_NOT_VALID)
+ new_pps->alt.state = IB_PORT_PKEY_VALID;
+ }
+
+ new_pps->main.sec = qp->qp_sec;
+ new_pps->alt.sec = qp->qp_sec;
+ return new_pps;
+}
+
+int ib_open_shared_qp_security(struct ib_qp *qp, struct ib_device *dev)
+{
+ struct ib_qp *real_qp = qp->real_qp;
+ int ret;
+
+ ret = ib_create_qp_security(qp, dev);
+
+ if (ret)
+ goto out;
+
+ mutex_lock(&real_qp->qp_sec->mutex);
+ ret = check_qp_port_pkey_settings(real_qp->qp_sec->ports_pkeys,
+ qp->qp_sec);
+
+ if (ret)
+ goto ret;
+
+ if (qp != real_qp)
+ list_add(&qp->qp_sec->shared_qp_list,
+ &real_qp->qp_sec->shared_qp_list);
+ret:
+ mutex_unlock(&real_qp->qp_sec->mutex);
+ if (ret)
+ destroy_qp_security(qp->qp_sec);
+
+out:
+ return ret;
+}
+
+void ib_close_shared_qp_security(struct ib_qp_security *sec)
+{
+ struct ib_qp *real_qp = sec->qp->real_qp;
+
+ mutex_lock(&real_qp->qp_sec->mutex);
+ list_del(&sec->shared_qp_list);
+ mutex_unlock(&real_qp->qp_sec->mutex);
+
+ destroy_qp_security(sec);
+}
+
+int ib_create_qp_security(struct ib_qp *qp, struct ib_device *dev)
+{
+ int ret;
+
+ qp->qp_sec = kzalloc(sizeof(*qp->qp_sec), GFP_KERNEL);
+ if (!qp->qp_sec)
+ return -ENOMEM;
+
+ qp->qp_sec->qp = qp;
+ qp->qp_sec->dev = dev;
+ mutex_init(&qp->qp_sec->mutex);
+ INIT_LIST_HEAD(&qp->qp_sec->shared_qp_list);
+ atomic_set(&qp->qp_sec->error_list_count, 0);
+ init_completion(&qp->qp_sec->error_complete);
+ ret = security_ib_alloc_security(&qp->qp_sec->security);
+ if (ret)
+ kfree(qp->qp_sec);
+
+ return ret;
+}
+EXPORT_SYMBOL(ib_create_qp_security);
+
+void ib_destroy_qp_security_begin(struct ib_qp_security *sec)
+{
+ mutex_lock(&sec->mutex);
+
+ /* Remove the QP from the lists so it won't get added to
+ * a to_error_list during the destroy process.
+ */
+ if (sec->ports_pkeys) {
+ port_pkey_list_remove(&sec->ports_pkeys->main);
+ port_pkey_list_remove(&sec->ports_pkeys->alt);
+ }
+
+ /* If the QP is already in one or more of those lists
+ * the destroying flag will ensure the to error flow
+ * doesn't operate on an undefined QP.
+ */
+ sec->destroying = true;
+
+ /* Record the error list count to know how many completions
+ * to wait for.
+ */
+ sec->error_comps_pending = atomic_read(&sec->error_list_count);
+
+ mutex_unlock(&sec->mutex);
+}
+
+void ib_destroy_qp_security_abort(struct ib_qp_security *sec)
+{
+ int ret;
+ int i;
+
+ /* If a concurrent cache update is in progress this
+ * QP security could be marked for an error state
+ * transition. Wait for this to complete.
+ */
+ for (i = 0; i < sec->error_comps_pending; i++)
+ wait_for_completion(&sec->error_complete);
+
+ mutex_lock(&sec->mutex);
+ sec->destroying = false;
+
+ /* Restore the position in the lists and verify
+ * access is still allowed in case a cache update
+ * occurred while attempting to destroy.
+ *
+ * Because these setting were listed already
+ * and removed during ib_destroy_qp_security_begin
+ * we know the pkey_index_qp_list for the PKey
+ * already exists so port_pkey_list_insert won't fail.
+ */
+ if (sec->ports_pkeys) {
+ port_pkey_list_insert(&sec->ports_pkeys->main);
+ port_pkey_list_insert(&sec->ports_pkeys->alt);
+ }
+
+ ret = check_qp_port_pkey_settings(sec->ports_pkeys, sec);
+ if (ret)
+ qp_to_error(sec);
+
+ mutex_unlock(&sec->mutex);
+}
+
+void ib_destroy_qp_security_end(struct ib_qp_security *sec)
+{
+ int i;
+
+ /* If a concurrent cache update is occurring we must
+ * wait until this QP security structure is processed
+ * in the QP to error flow before destroying it because
+ * the to_error_list is in use.
+ */
+ for (i = 0; i < sec->error_comps_pending; i++)
+ wait_for_completion(&sec->error_complete);
+
+ destroy_qp_security(sec);
+}
+
+void ib_security_cache_change(struct ib_device *device,
+ u8 port_num,
+ u64 subnet_prefix)
+{
+ struct pkey_index_qp_list *pkey;
+
+ list_for_each_entry(pkey,
+ &device->port_pkey_list[port_num].pkey_list,
+ pkey_index_list) {
+ check_pkey_qps(pkey,
+ device,
+ port_num,
+ subnet_prefix);
+ }
+}
+
+void ib_security_destroy_port_pkey_list(struct ib_device *device)
+{
+ struct pkey_index_qp_list *pkey, *tmp_pkey;
+ int i;
+
+ for (i = rdma_start_port(device); i <= rdma_end_port(device); i++) {
+ spin_lock(&device->port_pkey_list[i].list_lock);
+ list_for_each_entry_safe(pkey,
+ tmp_pkey,
+ &device->port_pkey_list[i].pkey_list,
+ pkey_index_list) {
+ list_del(&pkey->pkey_index_list);
+ kfree(pkey);
+ }
+ spin_unlock(&device->port_pkey_list[i].list_lock);
+ }
+}
+
+int ib_security_modify_qp(struct ib_qp *qp,
+ struct ib_qp_attr *qp_attr,
+ int qp_attr_mask,
+ struct ib_udata *udata)
+{
+ int ret = 0;
+ struct ib_ports_pkeys *tmp_pps;
+ struct ib_ports_pkeys *new_pps;
+ bool special_qp = (qp->qp_type == IB_QPT_SMI ||
+ qp->qp_type == IB_QPT_GSI);
+ bool pps_change = ((qp_attr_mask & (IB_QP_PKEY_INDEX | IB_QP_PORT)) ||
+ (qp_attr_mask & IB_QP_ALT_PATH));
+
+ if (pps_change && !special_qp) {
+ mutex_lock(&qp->qp_sec->mutex);
+ new_pps = get_new_pps(qp,
+ qp_attr,
+ qp_attr_mask);
+
+ /* Add this QP to the lists for the new port
+ * and pkey settings before checking for permission
+ * in case there is a concurrent cache update
+ * occurring. Walking the list for a cache change
+ * doesn't acquire the security mutex unless it's
+ * sending the QP to error.
+ */
+ ret = port_pkey_list_insert(&new_pps->main);
+
+ if (!ret)
+ ret = port_pkey_list_insert(&new_pps->alt);
+
+ if (!ret)
+ ret = check_qp_port_pkey_settings(new_pps,
+ qp->qp_sec);
+ }
+
+ if (!ret)
+ ret = qp->device->modify_qp(qp->real_qp,
+ qp_attr,
+ qp_attr_mask,
+ udata);
+
+ if (pps_change && !special_qp) {
+ /* Clean up the lists and free the appropriate
+ * ports_pkeys structure.
+ */
+ if (ret) {
+ tmp_pps = new_pps;
+ } else {
+ tmp_pps = qp->qp_sec->ports_pkeys;
+ qp->qp_sec->ports_pkeys = new_pps;
+ }
+
+ if (tmp_pps) {
+ port_pkey_list_remove(&tmp_pps->main);
+ port_pkey_list_remove(&tmp_pps->alt);
+ }
+ kfree(tmp_pps);
+ mutex_unlock(&qp->qp_sec->mutex);
+ }
+ return ret;
+}
+EXPORT_SYMBOL(ib_security_modify_qp);
+
+#endif /* CONFIG_SECURITY_INFINIBAND */
diff --git a/drivers/infiniband/core/uverbs_cmd.c b/drivers/infiniband/core/uverbs_cmd.c
index f664731..abeb391 100644
--- a/drivers/infiniband/core/uverbs_cmd.c
+++ b/drivers/infiniband/core/uverbs_cmd.c
@@ -37,7 +37,7 @@
#include <linux/fs.h>
#include <linux/slab.h>
#include <linux/sched.h>
-
+#include <linux/security.h>
#include <asm/uaccess.h>
#include "uverbs.h"
@@ -1915,6 +1915,10 @@ static int create_qp(struct ib_uverbs_file *file,
}
if (cmd->qp_type != IB_QPT_XRC_TGT) {
+ ret = ib_create_qp_security(qp, device);
+ if (ret)
+ goto err_destroy;
+
qp->real_qp = qp;
qp->device = device;
qp->pd = pd;
@@ -2405,10 +2409,18 @@ ssize_t ib_uverbs_modify_qp(struct ib_uverbs_file *file,
ret = ib_resolve_eth_dmac(qp, attr, &cmd.attr_mask);
if (ret)
goto release_qp;
- ret = qp->device->modify_qp(qp, attr,
- modify_qp_mask(qp->qp_type, cmd.attr_mask), &udata);
+
+ ret = ib_security_modify_qp(qp,
+ attr,
+ modify_qp_mask(qp->qp_type,
+ cmd.attr_mask),
+ &udata);
} else {
- ret = ib_modify_qp(qp, attr, modify_qp_mask(qp->qp_type, cmd.attr_mask));
+ ret = ib_security_modify_qp(qp,
+ attr,
+ modify_qp_mask(qp->qp_type,
+ cmd.attr_mask),
+ NULL);
}
if (ret)
diff --git a/drivers/infiniband/core/verbs.c b/drivers/infiniband/core/verbs.c
index 6916d5c..e783f54 100644
--- a/drivers/infiniband/core/verbs.c
+++ b/drivers/infiniband/core/verbs.c
@@ -44,6 +44,7 @@
#include <linux/in.h>
#include <linux/in6.h>
#include <net/addrconf.h>
+#include <linux/security.h>
#include <rdma/ib_verbs.h>
#include <rdma/ib_cache.h>
@@ -682,12 +683,20 @@ static struct ib_qp *__ib_open_qp(struct ib_qp *real_qp,
{
struct ib_qp *qp;
unsigned long flags;
+ int err;
qp = kzalloc(sizeof *qp, GFP_KERNEL);
if (!qp)
return ERR_PTR(-ENOMEM);
qp->real_qp = real_qp;
+ err = ib_open_shared_qp_security(qp, real_qp->device);
+ if (err) {
+ kfree(qp);
+ return ERR_PTR(err);
+ }
+
+ qp->real_qp = real_qp;
atomic_inc(&real_qp->usecnt);
qp->device = real_qp->device;
qp->event_handler = event_handler;
@@ -773,6 +782,12 @@ struct ib_qp *ib_create_qp(struct ib_pd *pd,
if (IS_ERR(qp))
return qp;
+ ret = ib_create_qp_security(qp, device);
+ if (ret) {
+ ib_destroy_qp(qp);
+ return ERR_PTR(ret);
+ }
+
qp->device = device;
qp->real_qp = qp;
qp->uobject = NULL;
@@ -1222,7 +1237,7 @@ int ib_modify_qp(struct ib_qp *qp,
if (ret)
return ret;
- return qp->device->modify_qp(qp->real_qp, qp_attr, qp_attr_mask, NULL);
+ return ib_security_modify_qp(qp->real_qp, qp_attr, qp_attr_mask, NULL);
}
EXPORT_SYMBOL(ib_modify_qp);
@@ -1251,6 +1266,7 @@ int ib_close_qp(struct ib_qp *qp)
spin_unlock_irqrestore(&real_qp->device->event_handler_lock, flags);
atomic_dec(&real_qp->usecnt);
+ ib_close_shared_qp_security(qp->qp_sec);
kfree(qp);
return 0;
@@ -1291,6 +1307,7 @@ int ib_destroy_qp(struct ib_qp *qp)
struct ib_cq *scq, *rcq;
struct ib_srq *srq;
struct ib_rwq_ind_table *ind_tbl;
+ struct ib_qp_security *sec;
int ret;
WARN_ON_ONCE(qp->mrs_used > 0);
@@ -1306,6 +1323,9 @@ int ib_destroy_qp(struct ib_qp *qp)
rcq = qp->recv_cq;
srq = qp->srq;
ind_tbl = qp->rwq_ind_tbl;
+ sec = qp->qp_sec;
+ if (sec)
+ ib_destroy_qp_security_begin(sec);
if (!qp->uobject)
rdma_rw_cleanup_mrs(qp);
@@ -1322,6 +1342,11 @@ int ib_destroy_qp(struct ib_qp *qp)
atomic_dec(&srq->usecnt);
if (ind_tbl)
atomic_dec(&ind_tbl->usecnt);
+ if (sec)
+ ib_destroy_qp_security_end(sec);
+ } else {
+ if (sec)
+ ib_destroy_qp_security_abort(sec);
}
return ret;
diff --git a/include/linux/lsm_hooks.h b/include/linux/lsm_hooks.h
index 7ae3976..f85a8f3 100644
--- a/include/linux/lsm_hooks.h
+++ b/include/linux/lsm_hooks.h
@@ -8,6 +8,7 @@
* Copyright (C) 2001 Silicon Graphics, Inc. (Trust Technology Group)
* Copyright (C) 2015 Intel Corporation.
* Copyright (C) 2015 Casey Schaufler <casey@schaufler-ca.com>
+ * Copyright (C) 2016 Mellanox Techonologies
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
@@ -876,6 +877,21 @@
* associated with the TUN device's security structure.
* @security pointer to the TUN devices's security structure.
*
+ * Security hooks for Infiniband
+ *
+ * @ib_pkey_access:
+ * Check permission to access a pkey when modifing a QP.
+ * @subnet_prefix the subnet prefix of the port being used.
+ * @pkey the pkey to be accessed.
+ * @sec pointer to a security structure.
+ * @ib_alloc_security:
+ * Allocate a security structure for Infiniband objects.
+ * @sec pointer to a security structure pointer.
+ * Returns 0 on success, non-zero on failure
+ * @ib_free_security:
+ * Deallocate an Infiniband security structure.
+ * @sec contains the security structure to be freed.
+ *
* Security hooks for XFRM operations.
*
* @xfrm_policy_alloc_security:
@@ -1579,6 +1595,12 @@ union security_list_options {
int (*tun_dev_open)(void *security);
#endif /* CONFIG_SECURITY_NETWORK */
+#ifdef CONFIG_SECURITY_INFINIBAND
+ int (*ib_pkey_access)(u64 subnet_prefix, u16 pkey, void *sec);
+ int (*ib_alloc_security)(void **sec);
+ void (*ib_free_security)(void *sec);
+#endif /* CONFIG_SECURITY_INFINIBAND */
+
#ifdef CONFIG_SECURITY_NETWORK_XFRM
int (*xfrm_policy_alloc_security)(struct xfrm_sec_ctx **ctxp,
struct xfrm_user_sec_ctx *sec_ctx,
@@ -1806,6 +1828,11 @@ struct security_hook_heads {
struct list_head tun_dev_attach;
struct list_head tun_dev_open;
#endif /* CONFIG_SECURITY_NETWORK */
+#ifdef CONFIG_SECURITY_INFINIBAND
+ struct list_head ib_pkey_access;
+ struct list_head ib_alloc_security;
+ struct list_head ib_free_security;
+#endif /* CONFIG_SECURITY_INFINIBAND */
#ifdef CONFIG_SECURITY_NETWORK_XFRM
struct list_head xfrm_policy_alloc_security;
struct list_head xfrm_policy_clone_security;
diff --git a/include/linux/security.h b/include/linux/security.h
index 14df373..33e23c4 100644
--- a/include/linux/security.h
+++ b/include/linux/security.h
@@ -6,6 +6,7 @@
* Copyright (C) 2001 Networks Associates Technology, Inc <ssmalley@nai.com>
* Copyright (C) 2001 James Morris <jmorris@intercode.com.au>
* Copyright (C) 2001 Silicon Graphics, Inc. (Trust Technology Group)
+ * Copyright (C) 2016 Mellanox Techonologies
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
@@ -1370,6 +1371,26 @@ static inline int security_tun_dev_open(void *security)
}
#endif /* CONFIG_SECURITY_NETWORK */
+#ifdef CONFIG_SECURITY_INFINIBAND
+int security_ib_pkey_access(u64 subnet_prefix, u16 pkey, void *sec);
+int security_ib_alloc_security(void **sec);
+void security_ib_free_security(void *sec);
+#else /* CONFIG_SECURITY_INFINIBAND */
+static inline int security_ib_pkey_access(u64 subnet_prefix, u16 pkey, void *sec)
+{
+ return 0;
+}
+
+static inline int security_ib_alloc_security(void **sec)
+{
+ return 0;
+}
+
+static inline void security_ib_free_security(void *sec)
+{
+}
+#endif /* CONFIG_SECURITY_INFINIBAND */
+
#ifdef CONFIG_SECURITY_NETWORK_XFRM
int security_xfrm_policy_alloc(struct xfrm_sec_ctx **ctxp,
diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h
index 205ea09..9042ba3 100644
--- a/include/rdma/ib_verbs.h
+++ b/include/rdma/ib_verbs.h
@@ -1490,6 +1490,45 @@ struct ib_rwq_ind_table_init_attr {
struct ib_wq **ind_tbl;
};
+enum port_pkey_state {
+ IB_PORT_PKEY_NOT_VALID = 0,
+ IB_PORT_PKEY_VALID = 1,
+ IB_PORT_PKEY_LISTED = 2,
+};
+
+struct ib_qp_security;
+
+struct ib_port_pkey {
+ enum port_pkey_state state;
+ u16 pkey_index;
+ u8 port_num;
+ struct list_head qp_list;
+ struct list_head to_error_list;
+ struct ib_qp_security *sec;
+};
+
+struct ib_ports_pkeys {
+ struct ib_port_pkey main;
+ struct ib_port_pkey alt;
+};
+
+struct ib_qp_security {
+ struct ib_qp *qp;
+ struct ib_device *dev;
+ /* Hold this mutex when changing port and pkey settings. */
+ struct mutex mutex;
+ struct ib_ports_pkeys *ports_pkeys;
+ /* A list of all open shared QP handles. Required to enforce security
+ * properly for all users of a shared QP.
+ */
+ struct list_head shared_qp_list;
+ void *security;
+ bool destroying;
+ atomic_t error_list_count;
+ struct completion error_complete;
+ int error_comps_pending;
+};
+
struct ib_qp {
struct ib_device *device;
struct ib_pd *pd;
@@ -1513,6 +1552,7 @@ struct ib_qp {
u32 qp_num;
enum ib_qp_type qp_type;
struct ib_rwq_ind_table *rwq_ind_tbl;
+ struct ib_qp_security *qp_sec;
};
struct ib_mr {
@@ -1760,6 +1800,12 @@ struct ib_port_immutable {
u32 max_mad_size;
};
+struct ib_port_pkey_list {
+ /* Lock to hold while modifying the list. */
+ spinlock_t list_lock;
+ struct list_head pkey_list;
+};
+
struct ib_device {
struct device *dma_device;
@@ -1782,6 +1828,8 @@ struct ib_device {
int num_comp_vectors;
+ struct ib_port_pkey_list *port_pkey_list;
+
struct iw_cm_verbs *iwcm;
/**
diff --git a/security/Kconfig b/security/Kconfig
index 176758c..ce965c6 100644
--- a/security/Kconfig
+++ b/security/Kconfig
@@ -49,6 +49,15 @@ config SECURITY_NETWORK
implement socket and networking access controls.
If you are unsure how to answer this question, answer N.
+config SECURITY_INFINIBAND
+ bool "Infiniband Security Hooks"
+ depends on SECURITY && INFINIBAND
+ help
+ This enables the Infiniband security hooks.
+ If enabled, a security module can use these hooks to
+ implement Infiniband access controls.
+ If you are unsure how to answer this question, answer N.
+
config SECURITY_NETWORK_XFRM
bool "XFRM (IPSec) Networking Security Hooks"
depends on XFRM && SECURITY_NETWORK
diff --git a/security/security.c b/security/security.c
index 7095693..234982d 100644
--- a/security/security.c
+++ b/security/security.c
@@ -4,6 +4,7 @@
* Copyright (C) 2001 WireX Communications, Inc <chris@wirex.com>
* Copyright (C) 2001-2002 Greg Kroah-Hartman <greg@kroah.com>
* Copyright (C) 2001 Networks Associates Technology, Inc <ssmalley@nai.com>
+ * Copyright (C) 2016 Mellanox Technologies
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
@@ -1399,6 +1400,27 @@ EXPORT_SYMBOL(security_tun_dev_open);
#endif /* CONFIG_SECURITY_NETWORK */
+#ifdef CONFIG_SECURITY_INFINIBAND
+
+int security_ib_pkey_access(u64 subnet_prefix, u16 pkey, void *sec)
+{
+ return call_int_hook(ib_pkey_access, 0, subnet_prefix, pkey, sec);
+}
+EXPORT_SYMBOL(security_ib_pkey_access);
+
+int security_ib_alloc_security(void **sec)
+{
+ return call_int_hook(ib_alloc_security, 0, sec);
+}
+EXPORT_SYMBOL(security_ib_alloc_security);
+
+void security_ib_free_security(void *sec)
+{
+ call_void_hook(ib_free_security, sec);
+}
+EXPORT_SYMBOL(security_ib_free_security);
+#endif /* CONFIG_SECURITY_INFINIBAND */
+
#ifdef CONFIG_SECURITY_NETWORK_XFRM
int security_xfrm_policy_alloc(struct xfrm_sec_ctx **ctxp,
@@ -1850,6 +1872,15 @@ struct security_hook_heads security_hook_heads = {
LIST_HEAD_INIT(security_hook_heads.tun_dev_attach),
.tun_dev_open = LIST_HEAD_INIT(security_hook_heads.tun_dev_open),
#endif /* CONFIG_SECURITY_NETWORK */
+
+#ifdef CONFIG_SECURITY_INFINIBAND
+ .ib_pkey_access = LIST_HEAD_INIT(security_hook_heads.ib_pkey_access),
+ .ib_alloc_security =
+ LIST_HEAD_INIT(security_hook_heads.ib_alloc_security),
+ .ib_free_security =
+ LIST_HEAD_INIT(security_hook_heads.ib_free_security),
+#endif /* CONFIG_SECURITY_INFINIBAND */
+
#ifdef CONFIG_SECURITY_NETWORK_XFRM
.xfrm_policy_alloc_security =
LIST_HEAD_INIT(security_hook_heads.xfrm_policy_alloc_security),
--
1.7.1
^ permalink raw reply related [flat|nested] 42+ messages in thread
* [PATCH v2 3/9] selinux lsm IB/core: Implement LSM notification system
2016-07-14 22:56 ` Dan Jurgens
@ 2016-07-14 22:56 ` Dan Jurgens
-1 siblings, 0 replies; 42+ messages in thread
From: Dan Jurgens @ 2016-07-14 22:56 UTC (permalink / raw)
To: chrisw-69jw2NvuJkxg9hUCZPvPmw, paul-r2n+y4ga6xFZroRs9YW3xA,
sds-+05T5uksL2qpZYMLLGbcSA, eparis-FjpueFixGhCM4zKIHC2jIg,
dledford-H+wXaHxf7aLQT0dZR+AlfA,
sean.hefty-ral2JQCrhuEAvxtiuMwx3w,
hal.rosenstock-Re5JQEeQqe8AvxtiuMwx3w
Cc: selinux-+05T5uksL2qpZYMLLGbcSA,
linux-security-module-u79uwXL29TY76Z2rM5mHXA,
linux-rdma-u79uwXL29TY76Z2rM5mHXA,
yevgenyp-VPRAkNaXOzVWk0Htik3J/w, Daniel Jurgens
From: Daniel Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Add a generic notificaiton mechanism in the LSM. Interested consumers
can register a callback with the LSM and security modules can produce
events.
Because access to Infiniband QPs are enforced in the setup phase of a
connection security should be enforced again if the policy changes.
Register infiniband devices for policy change notification and check all
QPs on that device when the notification is received.
Add a call to the notification mechanism is from SELinux when the AVC
cache changes.
Signed-off-by: Daniel Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---
v2:
- new patch that has the generic notification, replaces selinux and
IB/core patches related to the ib_flush callback. Yuval Shaia and Paul
Moore
---
drivers/infiniband/core/device.c | 46 ++++++++++++++++++++++++++++++
include/linux/security.h | 10 ++++++
security/security.c | 58 ++++++++++++++++++++++++++++++++++++++
security/selinux/hooks.c | 5 ++-
security/selinux/selinuxfs.c | 2 +
5 files changed, 119 insertions(+), 2 deletions(-)
diff --git a/drivers/infiniband/core/device.c b/drivers/infiniband/core/device.c
index 5b42e83..3219a7a 100644
--- a/drivers/infiniband/core/device.c
+++ b/drivers/infiniband/core/device.c
@@ -39,6 +39,7 @@
#include <linux/init.h>
#include <linux/mutex.h>
#include <linux/netdevice.h>
+#include <linux/security.h>
#include <rdma/rdma_netlink.h>
#include <rdma/ib_addr.h>
#include <rdma/ib_cache.h>
@@ -82,6 +83,10 @@ static LIST_HEAD(client_list);
static DEFINE_MUTEX(device_mutex);
static DECLARE_RWSEM(lists_rwsem);
+static void ib_policy_change_task(struct work_struct *work);
+static DECLARE_WORK(ib_policy_change_work, ib_policy_change_task);
+
+static u32 lsm_callback_id;
static int ib_device_check_mandatory(struct ib_device *device)
{
@@ -344,6 +349,37 @@ static int setup_port_pkey_list(struct ib_device *device)
return 0;
}
+static void ib_policy_change_task(struct work_struct *work)
+{
+ struct ib_device *dev;
+
+ down_read(&lists_rwsem);
+ list_for_each_entry(dev, &device_list, core_list) {
+ int i;
+
+ for (i = rdma_start_port(dev); i <= rdma_end_port(dev); i++) {
+ u64 sp;
+ int ret = ib_get_cached_subnet_prefix(dev,
+ i,
+ &sp);
+
+ WARN_ONCE(ret,
+ "ib_get_cached_subnet_prefix err: %d, this should never happen here\n",
+ ret);
+ ib_security_cache_change(dev, i, sp);
+ }
+ }
+ up_read(&lists_rwsem);
+}
+
+static void ib_security_change(enum lsm_event event, void *ctx, void *lsm_data)
+{
+ if (event != LSM_POLICY_CHANGE)
+ return;
+
+ schedule_work(&ib_policy_change_work);
+}
+
/**
* ib_register_device - Register an IB device with IB core
* @device:Device to register
@@ -1075,10 +1111,19 @@ static int __init ib_core_init(void)
goto err_sa;
}
+ ret = security_register_lsm_notifier(&ib_security_change, NULL,
+ &lsm_callback_id);
+ if (ret) {
+ pr_warn("Couldn't register LSM notifier. ret %d\n", ret);
+ goto err_ibnl_clients;
+ }
+
ib_cache_setup();
return 0;
+err_ibnl_clients:
+ ib_remove_ibnl_clients();
err_sa:
ib_sa_cleanup();
err_mad:
@@ -1098,6 +1143,7 @@ err:
static void __exit ib_core_cleanup(void)
{
+ security_unregister_lsm_notifier(lsm_callback_id);
ib_cache_cleanup();
ib_remove_ibnl_clients();
ib_sa_cleanup();
diff --git a/include/linux/security.h b/include/linux/security.h
index 33e23c4..bf53911 100644
--- a/include/linux/security.h
+++ b/include/linux/security.h
@@ -69,6 +69,16 @@ struct audit_krule;
struct user_namespace;
struct timezone;
+enum lsm_event {
+ LSM_POLICY_CHANGE,
+};
+
+typedef void (*lsm_notifier)(enum lsm_event event, void *ctx, void *data);
+
+void security_lsm_notify(enum lsm_event event, void *data);
+int security_register_lsm_notifier(lsm_notifier func, void *ctx, u32 *id);
+void security_unregister_lsm_notifier(u32 id);
+
/* These functions are in security/commoncap.c */
extern int cap_capable(const struct cred *cred, struct user_namespace *ns,
int cap, int audit);
diff --git a/security/security.c b/security/security.c
index 234982d..1263c1d 100644
--- a/security/security.c
+++ b/security/security.c
@@ -33,6 +33,18 @@
/* Maximum number of letters for an LSM name string */
#define SECURITY_NAME_MAX 10
+struct lsm_notifier_entry {
+ u32 callback_id;
+ lsm_notifier func;
+ void *ctx;
+ struct list_head list;
+ struct rcu_head rcu;
+};
+
+static LIST_HEAD(lsm_notifier_list);
+static DEFINE_SPINLOCK(lsm_notifier_lock);
+static u32 next_callback_id;
+
/* Boot-time LSM user choice */
static __initdata char chosen_lsm[SECURITY_NAME_MAX + 1] =
CONFIG_DEFAULT_SECURITY;
@@ -98,6 +110,52 @@ int __init security_module_enable(const char *module)
return !strcmp(module, chosen_lsm);
}
+void security_lsm_notify(enum lsm_event event, void *data)
+{
+ struct lsm_notifier_entry *entry;
+
+ rcu_read_lock();
+ list_for_each_entry_rcu(entry, &lsm_notifier_list, list)
+ entry->func(event, entry->ctx, data);
+ rcu_read_unlock();
+}
+EXPORT_SYMBOL(security_lsm_notify);
+
+int security_register_lsm_notifier(lsm_notifier func, void *ctx, u32 *id)
+{
+ struct lsm_notifier_entry *entry;
+
+ entry = kzalloc(sizeof(*entry), GFP_KERNEL);
+ if (!entry)
+ return -ENOMEM;
+
+ entry->func = func;
+ entry->ctx = ctx;
+
+ spin_lock(&lsm_notifier_lock);
+ entry->callback_id = next_callback_id++;
+ *id = entry->callback_id;
+ list_add_rcu(&entry->list, &lsm_notifier_list);
+ spin_unlock(&lsm_notifier_lock);
+
+ return 0;
+}
+EXPORT_SYMBOL(security_register_lsm_notifier);
+
+void security_unregister_lsm_notifier(u32 id)
+{
+ struct lsm_notifier_entry *entry;
+
+ list_for_each_entry_rcu(entry, &lsm_notifier_list, list) {
+ if (entry->callback_id == id) {
+ list_del_rcu(&entry->list);
+ kfree_rcu(entry, rcu);
+ break;
+ }
+ }
+}
+EXPORT_SYMBOL(security_unregister_lsm_notifier);
+
/*
* Hook list operation macros.
*
diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c
index a86d537..a363202 100644
--- a/security/selinux/hooks.c
+++ b/security/selinux/hooks.c
@@ -159,13 +159,14 @@ static int selinux_peerlbl_enabled(void)
return (selinux_policycap_alwaysnetwork || netlbl_enabled() || selinux_xfrm_enabled());
}
-static int selinux_netcache_avc_callback(u32 event)
+static int selinux_cache_avc_callback(u32 event)
{
if (event == AVC_CALLBACK_RESET) {
sel_netif_flush();
sel_netnode_flush();
sel_netport_flush();
synchronize_net();
+ security_lsm_notify(LSM_POLICY_CHANGE, NULL);
}
return 0;
}
@@ -6235,7 +6236,7 @@ static __init int selinux_init(void)
security_add_hooks(selinux_hooks, ARRAY_SIZE(selinux_hooks));
- if (avc_add_callback(selinux_netcache_avc_callback, AVC_CALLBACK_RESET))
+ if (avc_add_callback(selinux_cache_avc_callback, AVC_CALLBACK_RESET))
panic("SELinux: Unable to register AVC netcache callback\n");
if (selinux_enforcing)
diff --git a/security/selinux/selinuxfs.c b/security/selinux/selinuxfs.c
index 1b1fd27..adc09e7 100644
--- a/security/selinux/selinuxfs.c
+++ b/security/selinux/selinuxfs.c
@@ -177,6 +177,8 @@ static ssize_t sel_write_enforce(struct file *file, const char __user *buf,
avc_ss_reset(0);
selnl_notify_setenforce(selinux_enforcing);
selinux_status_update_setenforce(selinux_enforcing);
+ if (!selinux_enforcing)
+ security_lsm_notify(LSM_POLICY_CHANGE, NULL);
}
length = count;
out:
--
1.7.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related [flat|nested] 42+ messages in thread
* [PATCH v2 3/9] selinux lsm IB/core: Implement LSM notification system
@ 2016-07-14 22:56 ` Dan Jurgens
0 siblings, 0 replies; 42+ messages in thread
From: Dan Jurgens @ 2016-07-14 22:56 UTC (permalink / raw)
To: chrisw, paul, sds, eparis, dledford, sean.hefty, hal.rosenstock
Cc: selinux, linux-security-module, linux-rdma, yevgenyp, Daniel Jurgens
From: Daniel Jurgens <danielj@mellanox.com>
Add a generic notificaiton mechanism in the LSM. Interested consumers
can register a callback with the LSM and security modules can produce
events.
Because access to Infiniband QPs are enforced in the setup phase of a
connection security should be enforced again if the policy changes.
Register infiniband devices for policy change notification and check all
QPs on that device when the notification is received.
Add a call to the notification mechanism is from SELinux when the AVC
cache changes.
Signed-off-by: Daniel Jurgens <danielj@mellanox.com>
---
v2:
- new patch that has the generic notification, replaces selinux and
IB/core patches related to the ib_flush callback. Yuval Shaia and Paul
Moore
---
drivers/infiniband/core/device.c | 46 ++++++++++++++++++++++++++++++
include/linux/security.h | 10 ++++++
security/security.c | 58 ++++++++++++++++++++++++++++++++++++++
security/selinux/hooks.c | 5 ++-
security/selinux/selinuxfs.c | 2 +
5 files changed, 119 insertions(+), 2 deletions(-)
diff --git a/drivers/infiniband/core/device.c b/drivers/infiniband/core/device.c
index 5b42e83..3219a7a 100644
--- a/drivers/infiniband/core/device.c
+++ b/drivers/infiniband/core/device.c
@@ -39,6 +39,7 @@
#include <linux/init.h>
#include <linux/mutex.h>
#include <linux/netdevice.h>
+#include <linux/security.h>
#include <rdma/rdma_netlink.h>
#include <rdma/ib_addr.h>
#include <rdma/ib_cache.h>
@@ -82,6 +83,10 @@ static LIST_HEAD(client_list);
static DEFINE_MUTEX(device_mutex);
static DECLARE_RWSEM(lists_rwsem);
+static void ib_policy_change_task(struct work_struct *work);
+static DECLARE_WORK(ib_policy_change_work, ib_policy_change_task);
+
+static u32 lsm_callback_id;
static int ib_device_check_mandatory(struct ib_device *device)
{
@@ -344,6 +349,37 @@ static int setup_port_pkey_list(struct ib_device *device)
return 0;
}
+static void ib_policy_change_task(struct work_struct *work)
+{
+ struct ib_device *dev;
+
+ down_read(&lists_rwsem);
+ list_for_each_entry(dev, &device_list, core_list) {
+ int i;
+
+ for (i = rdma_start_port(dev); i <= rdma_end_port(dev); i++) {
+ u64 sp;
+ int ret = ib_get_cached_subnet_prefix(dev,
+ i,
+ &sp);
+
+ WARN_ONCE(ret,
+ "ib_get_cached_subnet_prefix err: %d, this should never happen here\n",
+ ret);
+ ib_security_cache_change(dev, i, sp);
+ }
+ }
+ up_read(&lists_rwsem);
+}
+
+static void ib_security_change(enum lsm_event event, void *ctx, void *lsm_data)
+{
+ if (event != LSM_POLICY_CHANGE)
+ return;
+
+ schedule_work(&ib_policy_change_work);
+}
+
/**
* ib_register_device - Register an IB device with IB core
* @device:Device to register
@@ -1075,10 +1111,19 @@ static int __init ib_core_init(void)
goto err_sa;
}
+ ret = security_register_lsm_notifier(&ib_security_change, NULL,
+ &lsm_callback_id);
+ if (ret) {
+ pr_warn("Couldn't register LSM notifier. ret %d\n", ret);
+ goto err_ibnl_clients;
+ }
+
ib_cache_setup();
return 0;
+err_ibnl_clients:
+ ib_remove_ibnl_clients();
err_sa:
ib_sa_cleanup();
err_mad:
@@ -1098,6 +1143,7 @@ err:
static void __exit ib_core_cleanup(void)
{
+ security_unregister_lsm_notifier(lsm_callback_id);
ib_cache_cleanup();
ib_remove_ibnl_clients();
ib_sa_cleanup();
diff --git a/include/linux/security.h b/include/linux/security.h
index 33e23c4..bf53911 100644
--- a/include/linux/security.h
+++ b/include/linux/security.h
@@ -69,6 +69,16 @@ struct audit_krule;
struct user_namespace;
struct timezone;
+enum lsm_event {
+ LSM_POLICY_CHANGE,
+};
+
+typedef void (*lsm_notifier)(enum lsm_event event, void *ctx, void *data);
+
+void security_lsm_notify(enum lsm_event event, void *data);
+int security_register_lsm_notifier(lsm_notifier func, void *ctx, u32 *id);
+void security_unregister_lsm_notifier(u32 id);
+
/* These functions are in security/commoncap.c */
extern int cap_capable(const struct cred *cred, struct user_namespace *ns,
int cap, int audit);
diff --git a/security/security.c b/security/security.c
index 234982d..1263c1d 100644
--- a/security/security.c
+++ b/security/security.c
@@ -33,6 +33,18 @@
/* Maximum number of letters for an LSM name string */
#define SECURITY_NAME_MAX 10
+struct lsm_notifier_entry {
+ u32 callback_id;
+ lsm_notifier func;
+ void *ctx;
+ struct list_head list;
+ struct rcu_head rcu;
+};
+
+static LIST_HEAD(lsm_notifier_list);
+static DEFINE_SPINLOCK(lsm_notifier_lock);
+static u32 next_callback_id;
+
/* Boot-time LSM user choice */
static __initdata char chosen_lsm[SECURITY_NAME_MAX + 1] =
CONFIG_DEFAULT_SECURITY;
@@ -98,6 +110,52 @@ int __init security_module_enable(const char *module)
return !strcmp(module, chosen_lsm);
}
+void security_lsm_notify(enum lsm_event event, void *data)
+{
+ struct lsm_notifier_entry *entry;
+
+ rcu_read_lock();
+ list_for_each_entry_rcu(entry, &lsm_notifier_list, list)
+ entry->func(event, entry->ctx, data);
+ rcu_read_unlock();
+}
+EXPORT_SYMBOL(security_lsm_notify);
+
+int security_register_lsm_notifier(lsm_notifier func, void *ctx, u32 *id)
+{
+ struct lsm_notifier_entry *entry;
+
+ entry = kzalloc(sizeof(*entry), GFP_KERNEL);
+ if (!entry)
+ return -ENOMEM;
+
+ entry->func = func;
+ entry->ctx = ctx;
+
+ spin_lock(&lsm_notifier_lock);
+ entry->callback_id = next_callback_id++;
+ *id = entry->callback_id;
+ list_add_rcu(&entry->list, &lsm_notifier_list);
+ spin_unlock(&lsm_notifier_lock);
+
+ return 0;
+}
+EXPORT_SYMBOL(security_register_lsm_notifier);
+
+void security_unregister_lsm_notifier(u32 id)
+{
+ struct lsm_notifier_entry *entry;
+
+ list_for_each_entry_rcu(entry, &lsm_notifier_list, list) {
+ if (entry->callback_id == id) {
+ list_del_rcu(&entry->list);
+ kfree_rcu(entry, rcu);
+ break;
+ }
+ }
+}
+EXPORT_SYMBOL(security_unregister_lsm_notifier);
+
/*
* Hook list operation macros.
*
diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c
index a86d537..a363202 100644
--- a/security/selinux/hooks.c
+++ b/security/selinux/hooks.c
@@ -159,13 +159,14 @@ static int selinux_peerlbl_enabled(void)
return (selinux_policycap_alwaysnetwork || netlbl_enabled() || selinux_xfrm_enabled());
}
-static int selinux_netcache_avc_callback(u32 event)
+static int selinux_cache_avc_callback(u32 event)
{
if (event == AVC_CALLBACK_RESET) {
sel_netif_flush();
sel_netnode_flush();
sel_netport_flush();
synchronize_net();
+ security_lsm_notify(LSM_POLICY_CHANGE, NULL);
}
return 0;
}
@@ -6235,7 +6236,7 @@ static __init int selinux_init(void)
security_add_hooks(selinux_hooks, ARRAY_SIZE(selinux_hooks));
- if (avc_add_callback(selinux_netcache_avc_callback, AVC_CALLBACK_RESET))
+ if (avc_add_callback(selinux_cache_avc_callback, AVC_CALLBACK_RESET))
panic("SELinux: Unable to register AVC netcache callback\n");
if (selinux_enforcing)
diff --git a/security/selinux/selinuxfs.c b/security/selinux/selinuxfs.c
index 1b1fd27..adc09e7 100644
--- a/security/selinux/selinuxfs.c
+++ b/security/selinux/selinuxfs.c
@@ -177,6 +177,8 @@ static ssize_t sel_write_enforce(struct file *file, const char __user *buf,
avc_ss_reset(0);
selnl_notify_setenforce(selinux_enforcing);
selinux_status_update_setenforce(selinux_enforcing);
+ if (!selinux_enforcing)
+ security_lsm_notify(LSM_POLICY_CHANGE, NULL);
}
length = count;
out:
--
1.7.1
^ permalink raw reply related [flat|nested] 42+ messages in thread
* [PATCH v2 4/9] IB/core: Enforce security on management datagrams
2016-07-14 22:56 ` Dan Jurgens
(?)
@ 2016-07-14 22:56 ` Dan Jurgens
-1 siblings, 0 replies; 42+ messages in thread
From: Dan Jurgens @ 2016-07-14 22:56 UTC (permalink / raw)
To: chrisw, paul, sds, eparis, dledford, sean.hefty, hal.rosenstock
Cc: selinux, linux-security-module, linux-rdma, yevgenyp, Daniel Jurgens
From: Daniel Jurgens <danielj@mellanox.com>
Allocate and free a security context when creating and destroying a MAD
agent. This context is used for controlling access to PKeys and sending
and receiving SMPs.
When sending or receiving a MAD check that the agent has permission to
access the PKey for the Subnet Prefix of the port.
During MAD and snoop agent registration for SMI QPs check that the
calling process has permission to access the manage the subnet and
register a callback with the LSM to be notified of policy changes. When
notificaiton of a policy change occurs recheck permission and set a flag
indicating sending and receiving SMPs is allowed.
When sending and receiving MADs check that the agent has access to the
SMI if it's on an SMI QP. Because security policy can change it's
possible permission was allowed when creating the agent, but no longer
is.
Signed-off-by: Daniel Jurgens <danielj@mellanox.com>
---
v2:
- Squashed LSM hook additions. Paul Moore
- Changed security blobs to void*. Paul Moore
- Shorten end_port to port. Paul Moore
- Change "smp" to "manage_subnet". Paul Moore
- Use the LSM policy change notification and a flag to track permission
instead of calling the LSM hook for every SMP. Dan Jurgens
- Squashed PKey and SMP enforcement into the same patch and moved the
logic into security.c. Dan Jurgens
---
drivers/infiniband/core/core_priv.h | 35 +++++++++++++
drivers/infiniband/core/mad.c | 52 +++++++++++++++++---
drivers/infiniband/core/security.c | 90 +++++++++++++++++++++++++++++++++++
include/linux/lsm_hooks.h | 7 +++
include/linux/security.h | 6 ++
include/rdma/ib_mad.h | 4 ++
security/security.c | 8 +++
7 files changed, 194 insertions(+), 8 deletions(-)
diff --git a/drivers/infiniband/core/core_priv.h b/drivers/infiniband/core/core_priv.h
index 68e3de0..2c35162 100644
--- a/drivers/infiniband/core/core_priv.h
+++ b/drivers/infiniband/core/core_priv.h
@@ -37,6 +37,8 @@
#include <linux/spinlock.h>
#include <rdma/ib_verbs.h>
+#include <rdma/ib_mad.h>
+#include "mad_priv.h"
struct pkey_index_qp_list {
struct list_head pkey_index_list;
@@ -166,6 +168,11 @@ int ib_get_cached_subnet_prefix(struct ib_device *device,
u64 *sn_pfx);
#ifdef CONFIG_SECURITY_INFINIBAND
+int ib_security_pkey_access(struct ib_device *dev,
+ u8 port_num,
+ u16 pkey_index,
+ void *sec);
+
void ib_security_destroy_port_pkey_list(struct ib_device *device);
void ib_security_cache_change(struct ib_device *device,
@@ -183,7 +190,19 @@ void ib_destroy_qp_security_abort(struct ib_qp_security *sec);
void ib_destroy_qp_security_end(struct ib_qp_security *sec);
int ib_open_shared_qp_security(struct ib_qp *qp, struct ib_device *dev);
void ib_close_shared_qp_security(struct ib_qp_security *sec);
+int ib_mad_agent_security_setup(struct ib_mad_agent *agent,
+ enum ib_qp_type qp_type);
+void ib_mad_agent_security_cleanup(struct ib_mad_agent *agent);
+int ib_mad_enforce_security(struct ib_mad_agent_private *map, u16 pkey_index);
#else
+static inline int ib_security_pkey_access(struct ib_device *dev,
+ u8 port_num,
+ u16 pkey_index,
+ void *sec)
+{
+ return 0;
+}
+
static inline void ib_security_destroy_port_pkey_list(struct ib_device *device)
{
}
@@ -232,5 +251,21 @@ static inline int ib_open_shared_qp_security(struct ib_qp *qp,
static inline void ib_close_shared_qp_security(struct ib_qp_security *sec)
{
}
+
+static inline int ib_mad_agent_security_setup(struct ib_mad_agent *agent,
+ enum ib_qp_type qp_type)
+{
+ return 0;
+}
+
+static inline void ib_mad_agent_security_cleanup(struct ib_mad_agent *agent)
+{
+}
+
+static inline int ib_mad_enforce_security(struct ib_mad_agent_private *map,
+ u16 pkey_index)
+{
+ return 0;
+}
#endif
#endif /* _CORE_PRIV_H */
diff --git a/drivers/infiniband/core/mad.c b/drivers/infiniband/core/mad.c
index 2d49228..94ebc3e 100644
--- a/drivers/infiniband/core/mad.c
+++ b/drivers/infiniband/core/mad.c
@@ -40,9 +40,11 @@
#include <linux/dma-mapping.h>
#include <linux/slab.h>
#include <linux/module.h>
+#include <linux/security.h>
#include <rdma/ib_cache.h>
#include "mad_priv.h"
+#include "core_priv.h"
#include "mad_rmpp.h"
#include "smi.h"
#include "opa_smi.h"
@@ -367,6 +369,12 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device,
atomic_set(&mad_agent_priv->refcount, 1);
init_completion(&mad_agent_priv->comp);
+ ret2 = ib_mad_agent_security_setup(&mad_agent_priv->agent, qp_type);
+ if (ret2) {
+ ret = ERR_PTR(ret2);
+ goto error4;
+ }
+
spin_lock_irqsave(&port_priv->reg_lock, flags);
mad_agent_priv->agent.hi_tid = ++ib_mad_client_id;
@@ -384,7 +392,7 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device,
if (method) {
if (method_in_use(&method,
mad_reg_req))
- goto error4;
+ goto error5;
}
}
ret2 = add_nonoui_reg_req(mad_reg_req, mad_agent_priv,
@@ -400,14 +408,14 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device,
if (is_vendor_method_in_use(
vendor_class,
mad_reg_req))
- goto error4;
+ goto error5;
}
}
ret2 = add_oui_reg_req(mad_reg_req, mad_agent_priv);
}
if (ret2) {
ret = ERR_PTR(ret2);
- goto error4;
+ goto error5;
}
}
@@ -416,9 +424,10 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device,
spin_unlock_irqrestore(&port_priv->reg_lock, flags);
return &mad_agent_priv->agent;
-
-error4:
+error5:
spin_unlock_irqrestore(&port_priv->reg_lock, flags);
+ ib_mad_agent_security_cleanup(&mad_agent_priv->agent);
+error4:
kfree(reg_req);
error3:
kfree(mad_agent_priv);
@@ -489,6 +498,7 @@ struct ib_mad_agent *ib_register_mad_snoop(struct ib_device *device,
struct ib_mad_agent *ret;
struct ib_mad_snoop_private *mad_snoop_priv;
int qpn;
+ int err;
/* Validate parameters */
if ((is_snooping_sends(mad_snoop_flags) && !snoop_handler) ||
@@ -523,17 +533,25 @@ struct ib_mad_agent *ib_register_mad_snoop(struct ib_device *device,
mad_snoop_priv->agent.port_num = port_num;
mad_snoop_priv->mad_snoop_flags = mad_snoop_flags;
init_completion(&mad_snoop_priv->comp);
+
+ err = ib_mad_agent_security_setup(&mad_snoop_priv->agent, qp_type);
+ if (err) {
+ ret = ERR_PTR(err);
+ goto error2;
+ }
+
mad_snoop_priv->snoop_index = register_snoop_agent(
&port_priv->qp_info[qpn],
mad_snoop_priv);
if (mad_snoop_priv->snoop_index < 0) {
ret = ERR_PTR(mad_snoop_priv->snoop_index);
- goto error2;
+ goto error3;
}
atomic_set(&mad_snoop_priv->refcount, 1);
return &mad_snoop_priv->agent;
-
+error3:
+ ib_mad_agent_security_cleanup(&mad_snoop_priv->agent);
error2:
kfree(mad_snoop_priv);
error1:
@@ -579,6 +597,8 @@ static void unregister_mad_agent(struct ib_mad_agent_private *mad_agent_priv)
deref_mad_agent(mad_agent_priv);
wait_for_completion(&mad_agent_priv->comp);
+ ib_mad_agent_security_cleanup(&mad_agent_priv->agent);
+
kfree(mad_agent_priv->reg_req);
kfree(mad_agent_priv);
}
@@ -597,6 +617,8 @@ static void unregister_mad_snoop(struct ib_mad_snoop_private *mad_snoop_priv)
deref_snoop_agent(mad_snoop_priv);
wait_for_completion(&mad_snoop_priv->comp);
+ ib_mad_agent_security_cleanup(&mad_snoop_priv->agent);
+
kfree(mad_snoop_priv);
}
@@ -1219,12 +1241,16 @@ int ib_post_send_mad(struct ib_mad_send_buf *send_buf,
/* Walk list of send WRs and post each on send list */
for (; send_buf; send_buf = next_send_buf) {
-
mad_send_wr = container_of(send_buf,
struct ib_mad_send_wr_private,
send_buf);
mad_agent_priv = mad_send_wr->mad_agent_priv;
+ ret = ib_mad_enforce_security(mad_agent_priv,
+ mad_send_wr->send_wr.pkey_index);
+ if (ret)
+ goto error;
+
if (!send_buf->mad_agent->send_handler ||
(send_buf->timeout_ms &&
!send_buf->mad_agent->recv_handler)) {
@@ -1958,6 +1984,14 @@ static void ib_mad_complete_recv(struct ib_mad_agent_private *mad_agent_priv,
struct ib_mad_send_wr_private *mad_send_wr;
struct ib_mad_send_wc mad_send_wc;
unsigned long flags;
+ int ret;
+
+ ret = ib_mad_enforce_security(mad_agent_priv,
+ mad_recv_wc->wc->pkey_index);
+ if (ret) {
+ ib_free_recv_mad(mad_recv_wc);
+ deref_mad_agent(mad_agent_priv);
+ }
INIT_LIST_HEAD(&mad_recv_wc->rmpp_list);
list_add(&mad_recv_wc->recv_buf.list, &mad_recv_wc->rmpp_list);
@@ -2015,6 +2049,8 @@ static void ib_mad_complete_recv(struct ib_mad_agent_private *mad_agent_priv,
mad_recv_wc);
deref_mad_agent(mad_agent_priv);
}
+
+ return;
}
static enum smi_action handle_ib_smi(const struct ib_mad_port_private *port_priv,
diff --git a/drivers/infiniband/core/security.c b/drivers/infiniband/core/security.c
index 44ff452..915e3ac 100644
--- a/drivers/infiniband/core/security.c
+++ b/drivers/infiniband/core/security.c
@@ -39,6 +39,7 @@
#include <rdma/ib_verbs.h>
#include <rdma/ib_cache.h>
#include "core_priv.h"
+#include "mad_priv.h"
static struct pkey_index_qp_list *get_pkey_idx_qp_list(struct ib_port_pkey *pp)
{
@@ -614,4 +615,93 @@ int ib_security_modify_qp(struct ib_qp *qp,
}
EXPORT_SYMBOL(ib_security_modify_qp);
+int ib_security_pkey_access(struct ib_device *dev,
+ u8 port_num,
+ u16 pkey_index,
+ void *sec)
+{
+ u64 subnet_prefix;
+ u16 pkey;
+ int ret;
+
+ ret = ib_get_cached_pkey(dev, port_num, pkey_index, &pkey);
+ if (ret)
+ return ret;
+
+ ret = ib_get_cached_subnet_prefix(dev, port_num, &subnet_prefix);
+
+ if (ret)
+ return ret;
+
+ return security_ib_pkey_access(subnet_prefix, pkey, sec);
+}
+EXPORT_SYMBOL(ib_security_pkey_access);
+
+static void ib_mad_agent_security_change(enum lsm_event event, void *ctx,
+ void *data)
+{
+ struct ib_mad_agent *ag = (struct ib_mad_agent *)ctx;
+
+ if (event != LSM_POLICY_CHANGE)
+ return;
+
+ ag->smp_allowed = !security_ib_port_manage_subnet(ag->device->name,
+ ag->port_num,
+ ag->security);
+}
+
+int ib_mad_agent_security_setup(struct ib_mad_agent *agent,
+ enum ib_qp_type qp_type)
+{
+ int ret;
+
+ ret = security_ib_alloc_security(&agent->security);
+ if (ret)
+ return ret;
+
+ if (qp_type != IB_QPT_SMI)
+ return 0;
+
+ ret = security_ib_port_manage_subnet(agent->device->name,
+ agent->port_num,
+ agent->security);
+ if (ret)
+ return ret;
+
+ ret = security_register_lsm_notifier(&ib_mad_agent_security_change,
+ agent,
+ &agent->lsm_id);
+ if (ret)
+ return ret;
+
+ agent->smp_allowed = true;
+ agent->lsm_reg = true;
+ return 0;
+}
+
+void ib_mad_agent_security_cleanup(struct ib_mad_agent *agent)
+{
+ security_ib_free_security(agent->security);
+ if (agent->lsm_reg)
+ security_unregister_lsm_notifier(agent->lsm_id);
+}
+
+int ib_mad_enforce_security(struct ib_mad_agent_private *map, u16 pkey_index)
+{
+ int ret;
+
+ if (map->agent.qp->qp_type == IB_QPT_SMI && !map->agent.smp_allowed)
+ return -EACCES;
+
+ ret = ib_security_pkey_access(map->agent.device,
+ map->agent.port_num,
+ pkey_index,
+ map->agent.security);
+
+ if (ret)
+ return ret;
+
+ return 0;
+}
+
#endif /* CONFIG_SECURITY_INFINIBAND */
diff --git a/include/linux/lsm_hooks.h b/include/linux/lsm_hooks.h
index f85a8f3..97c81ad 100644
--- a/include/linux/lsm_hooks.h
+++ b/include/linux/lsm_hooks.h
@@ -884,6 +884,11 @@
* @subnet_prefix the subnet prefix of the port being used.
* @pkey the pkey to be accessed.
* @sec pointer to a security structure.
+ * @ib_port_manage_subnet:
+ * Check permissions to send and receive SMPs on a end port.
+ * @dev_name the IB device name (i.e. mlx4_0).
+ * @port_num the port number.
+ * @sec pointer to a security structure.
* @ib_alloc_security:
* Allocate a security structure for Infiniband objects.
* @sec pointer to a security structure pointer.
@@ -1597,6 +1602,7 @@ union security_list_options {
#ifdef CONFIG_SECURITY_INFINIBAND
int (*ib_pkey_access)(u64 subnet_prefix, u16 pkey, void *sec);
+ int (*ib_port_manage_subnet)(const char *dev_name, u8 port, void *sec);
int (*ib_alloc_security)(void **sec);
void (*ib_free_security)(void *sec);
#endif /* CONFIG_SECURITY_INFINIBAND */
@@ -1830,6 +1836,7 @@ struct security_hook_heads {
#endif /* CONFIG_SECURITY_NETWORK */
#ifdef CONFIG_SECURITY_INFINIBAND
struct list_head ib_pkey_access;
+ struct list_head ib_port_manage_subnet;
struct list_head ib_alloc_security;
struct list_head ib_free_security;
#endif /* CONFIG_SECURITY_INFINIBAND */
diff --git a/include/linux/security.h b/include/linux/security.h
index bf53911..2600098 100644
--- a/include/linux/security.h
+++ b/include/linux/security.h
@@ -1383,6 +1383,7 @@ static inline int security_tun_dev_open(void *security)
#ifdef CONFIG_SECURITY_INFINIBAND
int security_ib_pkey_access(u64 subnet_prefix, u16 pkey, void *sec);
+int security_ib_port_manage_subnet(const char *name, u8 port, void *sec);
int security_ib_alloc_security(void **sec);
void security_ib_free_security(void *sec);
#else /* CONFIG_SECURITY_INFINIBAND */
@@ -1391,6 +1392,11 @@ static inline int security_ib_pkey_access(u64 subnet_prefix, u16 pkey, void *sec
return 0;
}
+static inline int security_ib_port_manage_subnet(const char *dev_name, u8 port, void *sec)
+{
+ return 0;
+}
+
static inline int security_ib_alloc_security(void **sec)
{
return 0;
diff --git a/include/rdma/ib_mad.h b/include/rdma/ib_mad.h
index c8a773f..f846a61 100644
--- a/include/rdma/ib_mad.h
+++ b/include/rdma/ib_mad.h
@@ -537,6 +537,10 @@ struct ib_mad_agent {
u32 flags;
u8 port_num;
u8 rmpp_version;
+ void *security;
+ bool smp_allowed;
+ u32 lsm_id;
+ bool lsm_reg;
};
/**
diff --git a/security/security.c b/security/security.c
index 1263c1d..fb46e90 100644
--- a/security/security.c
+++ b/security/security.c
@@ -1466,6 +1466,12 @@ int security_ib_pkey_access(u64 subnet_prefix, u16 pkey, void *sec)
}
EXPORT_SYMBOL(security_ib_pkey_access);
+int security_ib_port_manage_subnet(const char *dev_name, u8 port, void *sec)
+{
+ return call_int_hook(ib_port_manage_subnet, 0, dev_name, port, sec);
+}
+EXPORT_SYMBOL(security_ib_port_manage_subnet);
+
int security_ib_alloc_security(void **sec)
{
return call_int_hook(ib_alloc_security, 0, sec);
@@ -1933,6 +1939,8 @@ struct security_hook_heads security_hook_heads = {
#ifdef CONFIG_SECURITY_INFINIBAND
.ib_pkey_access = LIST_HEAD_INIT(security_hook_heads.ib_pkey_access),
+ .ib_port_manage_subnet =
+ LIST_HEAD_INIT(security_hook_heads.ib_port_manage_subnet),
.ib_alloc_security =
LIST_HEAD_INIT(security_hook_heads.ib_alloc_security),
.ib_free_security =
--
1.7.1
^ permalink raw reply related [flat|nested] 42+ messages in thread
* [PATCH v2 5/9] selinux: Create policydb version for Infiniband support
2016-07-14 22:56 ` Dan Jurgens
@ 2016-07-14 22:56 ` Dan Jurgens
-1 siblings, 0 replies; 42+ messages in thread
From: Dan Jurgens @ 2016-07-14 22:56 UTC (permalink / raw)
To: chrisw-69jw2NvuJkxg9hUCZPvPmw, paul-r2n+y4ga6xFZroRs9YW3xA,
sds-+05T5uksL2qpZYMLLGbcSA, eparis-FjpueFixGhCM4zKIHC2jIg,
dledford-H+wXaHxf7aLQT0dZR+AlfA,
sean.hefty-ral2JQCrhuEAvxtiuMwx3w,
hal.rosenstock-Re5JQEeQqe8AvxtiuMwx3w
Cc: selinux-+05T5uksL2qpZYMLLGbcSA,
linux-security-module-u79uwXL29TY76Z2rM5mHXA,
linux-rdma-u79uwXL29TY76Z2rM5mHXA,
yevgenyp-VPRAkNaXOzVWk0Htik3J/w, Daniel Jurgens
From: Daniel Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Support for Infiniband requires the addition of two new object contexts,
one for infiniband PKeys and another IB Ports. Added handlers to read
and write the new ocontext types when reading or writing a binary policy
representation.
Signed-off-by: Daniel Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Reviewed-by: Eli Cohen <eli-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---
v2:
- Shorten ib_end_port to ib_port. Paul Moore
- Added bounds checking to port number. Paul Moore
- Eliminated {} in OCON_PKEY case statement. Yuval Shaia
---
security/selinux/include/security.h | 3 +-
security/selinux/ss/policydb.c | 130 +++++++++++++++++++++++++++++++----
security/selinux/ss/policydb.h | 27 +++++--
3 files changed, 136 insertions(+), 24 deletions(-)
diff --git a/security/selinux/include/security.h b/security/selinux/include/security.h
index 38feb55..a7e6ed2 100644
--- a/security/selinux/include/security.h
+++ b/security/selinux/include/security.h
@@ -36,13 +36,14 @@
#define POLICYDB_VERSION_DEFAULT_TYPE 28
#define POLICYDB_VERSION_CONSTRAINT_NAMES 29
#define POLICYDB_VERSION_XPERMS_IOCTL 30
+#define POLICYDB_VERSION_INFINIBAND 31
/* Range of policy versions we understand*/
#define POLICYDB_VERSION_MIN POLICYDB_VERSION_BASE
#ifdef CONFIG_SECURITY_SELINUX_POLICYDB_VERSION_MAX
#define POLICYDB_VERSION_MAX CONFIG_SECURITY_SELINUX_POLICYDB_VERSION_MAX_VALUE
#else
-#define POLICYDB_VERSION_MAX POLICYDB_VERSION_XPERMS_IOCTL
+#define POLICYDB_VERSION_MAX POLICYDB_VERSION_INFINIBAND
#endif
/* Mask for just the mount related flags */
diff --git a/security/selinux/ss/policydb.c b/security/selinux/ss/policydb.c
index 992a315..3aea761 100644
--- a/security/selinux/ss/policydb.c
+++ b/security/selinux/ss/policydb.c
@@ -17,6 +17,11 @@
*
* Added support for the policy capability bitmap
*
+ * Update: Mellanox Techonologies
+ *
+ * Added Infiniband support
+ *
+ * Copyright (C) 2016 Mellanox Techonologies
* Copyright (C) 2007 Hewlett-Packard Development Company, L.P.
* Copyright (C) 2004-2005 Trusted Computer Solutions, Inc.
* Copyright (C) 2003 - 2004 Tresys Technology, LLC
@@ -76,81 +81,86 @@ static struct policydb_compat_info policydb_compat[] = {
{
.version = POLICYDB_VERSION_BASE,
.sym_num = SYM_NUM - 3,
- .ocon_num = OCON_NUM - 1,
+ .ocon_num = OCON_NUM - 3,
},
{
.version = POLICYDB_VERSION_BOOL,
.sym_num = SYM_NUM - 2,
- .ocon_num = OCON_NUM - 1,
+ .ocon_num = OCON_NUM - 3,
},
{
.version = POLICYDB_VERSION_IPV6,
.sym_num = SYM_NUM - 2,
- .ocon_num = OCON_NUM,
+ .ocon_num = OCON_NUM - 2,
},
{
.version = POLICYDB_VERSION_NLCLASS,
.sym_num = SYM_NUM - 2,
- .ocon_num = OCON_NUM,
+ .ocon_num = OCON_NUM - 2,
},
{
.version = POLICYDB_VERSION_MLS,
.sym_num = SYM_NUM,
- .ocon_num = OCON_NUM,
+ .ocon_num = OCON_NUM - 2,
},
{
.version = POLICYDB_VERSION_AVTAB,
.sym_num = SYM_NUM,
- .ocon_num = OCON_NUM,
+ .ocon_num = OCON_NUM - 2,
},
{
.version = POLICYDB_VERSION_RANGETRANS,
.sym_num = SYM_NUM,
- .ocon_num = OCON_NUM,
+ .ocon_num = OCON_NUM - 2,
},
{
.version = POLICYDB_VERSION_POLCAP,
.sym_num = SYM_NUM,
- .ocon_num = OCON_NUM,
+ .ocon_num = OCON_NUM - 2,
},
{
.version = POLICYDB_VERSION_PERMISSIVE,
.sym_num = SYM_NUM,
- .ocon_num = OCON_NUM,
+ .ocon_num = OCON_NUM - 2,
},
{
.version = POLICYDB_VERSION_BOUNDARY,
.sym_num = SYM_NUM,
- .ocon_num = OCON_NUM,
+ .ocon_num = OCON_NUM - 2,
},
{
.version = POLICYDB_VERSION_FILENAME_TRANS,
.sym_num = SYM_NUM,
- .ocon_num = OCON_NUM,
+ .ocon_num = OCON_NUM - 2,
},
{
.version = POLICYDB_VERSION_ROLETRANS,
.sym_num = SYM_NUM,
- .ocon_num = OCON_NUM,
+ .ocon_num = OCON_NUM - 2,
},
{
.version = POLICYDB_VERSION_NEW_OBJECT_DEFAULTS,
.sym_num = SYM_NUM,
- .ocon_num = OCON_NUM,
+ .ocon_num = OCON_NUM - 2,
},
{
.version = POLICYDB_VERSION_DEFAULT_TYPE,
.sym_num = SYM_NUM,
- .ocon_num = OCON_NUM,
+ .ocon_num = OCON_NUM - 2,
},
{
.version = POLICYDB_VERSION_CONSTRAINT_NAMES,
.sym_num = SYM_NUM,
- .ocon_num = OCON_NUM,
+ .ocon_num = OCON_NUM - 2,
},
{
.version = POLICYDB_VERSION_XPERMS_IOCTL,
.sym_num = SYM_NUM,
+ .ocon_num = OCON_NUM - 2,
+ },
+ {
+ .version = POLICYDB_VERSION_INFINIBAND,
+ .sym_num = SYM_NUM,
.ocon_num = OCON_NUM,
},
};
@@ -2219,6 +2229,61 @@ static int ocontext_read(struct policydb *p, struct policydb_compat_info *info,
goto out;
break;
}
+ case OCON_PKEY: {
+ rc = next_entry(nodebuf, fp, sizeof(u32) * 6);
+ if (rc)
+ goto out;
+
+ c->u.pkey.subnet_prefix = be64_to_cpu(*((__be64 *)nodebuf));
+ /* The subnet prefix is stored as an IPv6
+ * address in the policy.
+ *
+ * Check that the lower 2 DWORDS are 0.
+ */
+ if (nodebuf[2] || nodebuf[3]) {
+ rc = -EINVAL;
+ goto out;
+ }
+
+ if (nodebuf[4] > 0xffff ||
+ nodebuf[5] > 0xffff) {
+ rc = -EINVAL;
+ goto out;
+ }
+
+ c->u.pkey.low_pkey = le32_to_cpu(nodebuf[4]);
+ c->u.pkey.high_pkey = le32_to_cpu(nodebuf[5]);
+
+ rc = context_read_and_validate(&c->context[0],
+ p,
+ fp);
+ if (rc)
+ goto out;
+ break;
+ }
+ case OCON_IB_PORT:
+ rc = next_entry(buf, fp, sizeof(u32) * 2);
+ if (rc)
+ goto out;
+ len = le32_to_cpu(buf[0]);
+
+ rc = str_read(&c->u.ib_port.dev_name, GFP_KERNEL, fp, len);
+ if (rc)
+ goto out;
+
+ if (buf[1] > 0xff || buf[1] == 0) {
+ rc = -EINVAL;
+ goto out;
+ }
+
+ c->u.ib_port.port = le32_to_cpu(buf[1]);
+
+ rc = context_read_and_validate(&c->context[0],
+ p,
+ fp);
+ if (rc)
+ goto out;
+ break;
}
}
}
@@ -3147,6 +3212,41 @@ static int ocontext_write(struct policydb *p, struct policydb_compat_info *info,
if (rc)
return rc;
break;
+ case OCON_PKEY:
+ *((__be64 *)nodebuf) = cpu_to_be64(c->u.pkey.subnet_prefix);
+
+ /*
+ * The low order 2 bits were confirmed to be 0
+ * when the policy was loaded. Write them out
+ * as zero
+ */
+ nodebuf[2] = 0;
+ nodebuf[3] = 0;
+
+ nodebuf[4] = cpu_to_le32(c->u.pkey.low_pkey);
+ nodebuf[5] = cpu_to_le32(c->u.pkey.high_pkey);
+
+ rc = put_entry(nodebuf, sizeof(u32), 6, fp);
+ if (rc)
+ return rc;
+ rc = context_write(p, &c->context[0], fp);
+ if (rc)
+ return rc;
+ break;
+ case OCON_IB_PORT:
+ len = strlen(c->u.ib_port.dev_name);
+ buf[0] = cpu_to_le32(len);
+ buf[1] = cpu_to_le32(c->u.ib_port.port);
+ rc = put_entry(buf, sizeof(u32), 2, fp);
+ if (rc)
+ return rc;
+ rc = put_entry(c->u.ib_port.dev_name, 1, len, fp);
+ if (rc)
+ return rc;
+ rc = context_write(p, &c->context[0], fp);
+ if (rc)
+ return rc;
+ break;
}
}
}
diff --git a/security/selinux/ss/policydb.h b/security/selinux/ss/policydb.h
index 725d594..9f20774 100644
--- a/security/selinux/ss/policydb.h
+++ b/security/selinux/ss/policydb.h
@@ -187,6 +187,15 @@ struct ocontext {
u32 addr[4];
u32 mask[4];
} node6; /* IPv6 node information */
+ struct {
+ u64 subnet_prefix;
+ u16 low_pkey;
+ u16 high_pkey;
+ } pkey;
+ struct {
+ char *dev_name;
+ u8 port;
+ } ib_port;
} u;
union {
u32 sclass; /* security class for genfs */
@@ -215,14 +224,16 @@ struct genfs {
#define SYM_NUM 8
/* object context array indices */
-#define OCON_ISID 0 /* initial SIDs */
-#define OCON_FS 1 /* unlabeled file systems */
-#define OCON_PORT 2 /* TCP and UDP port numbers */
-#define OCON_NETIF 3 /* network interfaces */
-#define OCON_NODE 4 /* nodes */
-#define OCON_FSUSE 5 /* fs_use */
-#define OCON_NODE6 6 /* IPv6 nodes */
-#define OCON_NUM 7
+#define OCON_ISID 0 /* initial SIDs */
+#define OCON_FS 1 /* unlabeled file systems */
+#define OCON_PORT 2 /* TCP and UDP port numbers */
+#define OCON_NETIF 3 /* network interfaces */
+#define OCON_NODE 4 /* nodes */
+#define OCON_FSUSE 5 /* fs_use */
+#define OCON_NODE6 6 /* IPv6 nodes */
+#define OCON_PKEY 7 /* Infiniband PKeys */
+#define OCON_IB_PORT 8 /* Infiniband ports */
+#define OCON_NUM 9
/* The policy database */
struct policydb {
--
1.7.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related [flat|nested] 42+ messages in thread
* [PATCH v2 5/9] selinux: Create policydb version for Infiniband support
@ 2016-07-14 22:56 ` Dan Jurgens
0 siblings, 0 replies; 42+ messages in thread
From: Dan Jurgens @ 2016-07-14 22:56 UTC (permalink / raw)
To: chrisw, paul, sds, eparis, dledford, sean.hefty, hal.rosenstock
Cc: selinux, linux-security-module, linux-rdma, yevgenyp, Daniel Jurgens
From: Daniel Jurgens <danielj@mellanox.com>
Support for Infiniband requires the addition of two new object contexts,
one for infiniband PKeys and another IB Ports. Added handlers to read
and write the new ocontext types when reading or writing a binary policy
representation.
Signed-off-by: Daniel Jurgens <danielj@mellanox.com>
Reviewed-by: Eli Cohen <eli@mellanox.com>
---
v2:
- Shorten ib_end_port to ib_port. Paul Moore
- Added bounds checking to port number. Paul Moore
- Eliminated {} in OCON_PKEY case statement. Yuval Shaia
---
security/selinux/include/security.h | 3 +-
security/selinux/ss/policydb.c | 130 +++++++++++++++++++++++++++++++----
security/selinux/ss/policydb.h | 27 +++++--
3 files changed, 136 insertions(+), 24 deletions(-)
diff --git a/security/selinux/include/security.h b/security/selinux/include/security.h
index 38feb55..a7e6ed2 100644
--- a/security/selinux/include/security.h
+++ b/security/selinux/include/security.h
@@ -36,13 +36,14 @@
#define POLICYDB_VERSION_DEFAULT_TYPE 28
#define POLICYDB_VERSION_CONSTRAINT_NAMES 29
#define POLICYDB_VERSION_XPERMS_IOCTL 30
+#define POLICYDB_VERSION_INFINIBAND 31
/* Range of policy versions we understand*/
#define POLICYDB_VERSION_MIN POLICYDB_VERSION_BASE
#ifdef CONFIG_SECURITY_SELINUX_POLICYDB_VERSION_MAX
#define POLICYDB_VERSION_MAX CONFIG_SECURITY_SELINUX_POLICYDB_VERSION_MAX_VALUE
#else
-#define POLICYDB_VERSION_MAX POLICYDB_VERSION_XPERMS_IOCTL
+#define POLICYDB_VERSION_MAX POLICYDB_VERSION_INFINIBAND
#endif
/* Mask for just the mount related flags */
diff --git a/security/selinux/ss/policydb.c b/security/selinux/ss/policydb.c
index 992a315..3aea761 100644
--- a/security/selinux/ss/policydb.c
+++ b/security/selinux/ss/policydb.c
@@ -17,6 +17,11 @@
*
* Added support for the policy capability bitmap
*
+ * Update: Mellanox Techonologies
+ *
+ * Added Infiniband support
+ *
+ * Copyright (C) 2016 Mellanox Techonologies
* Copyright (C) 2007 Hewlett-Packard Development Company, L.P.
* Copyright (C) 2004-2005 Trusted Computer Solutions, Inc.
* Copyright (C) 2003 - 2004 Tresys Technology, LLC
@@ -76,81 +81,86 @@ static struct policydb_compat_info policydb_compat[] = {
{
.version = POLICYDB_VERSION_BASE,
.sym_num = SYM_NUM - 3,
- .ocon_num = OCON_NUM - 1,
+ .ocon_num = OCON_NUM - 3,
},
{
.version = POLICYDB_VERSION_BOOL,
.sym_num = SYM_NUM - 2,
- .ocon_num = OCON_NUM - 1,
+ .ocon_num = OCON_NUM - 3,
},
{
.version = POLICYDB_VERSION_IPV6,
.sym_num = SYM_NUM - 2,
- .ocon_num = OCON_NUM,
+ .ocon_num = OCON_NUM - 2,
},
{
.version = POLICYDB_VERSION_NLCLASS,
.sym_num = SYM_NUM - 2,
- .ocon_num = OCON_NUM,
+ .ocon_num = OCON_NUM - 2,
},
{
.version = POLICYDB_VERSION_MLS,
.sym_num = SYM_NUM,
- .ocon_num = OCON_NUM,
+ .ocon_num = OCON_NUM - 2,
},
{
.version = POLICYDB_VERSION_AVTAB,
.sym_num = SYM_NUM,
- .ocon_num = OCON_NUM,
+ .ocon_num = OCON_NUM - 2,
},
{
.version = POLICYDB_VERSION_RANGETRANS,
.sym_num = SYM_NUM,
- .ocon_num = OCON_NUM,
+ .ocon_num = OCON_NUM - 2,
},
{
.version = POLICYDB_VERSION_POLCAP,
.sym_num = SYM_NUM,
- .ocon_num = OCON_NUM,
+ .ocon_num = OCON_NUM - 2,
},
{
.version = POLICYDB_VERSION_PERMISSIVE,
.sym_num = SYM_NUM,
- .ocon_num = OCON_NUM,
+ .ocon_num = OCON_NUM - 2,
},
{
.version = POLICYDB_VERSION_BOUNDARY,
.sym_num = SYM_NUM,
- .ocon_num = OCON_NUM,
+ .ocon_num = OCON_NUM - 2,
},
{
.version = POLICYDB_VERSION_FILENAME_TRANS,
.sym_num = SYM_NUM,
- .ocon_num = OCON_NUM,
+ .ocon_num = OCON_NUM - 2,
},
{
.version = POLICYDB_VERSION_ROLETRANS,
.sym_num = SYM_NUM,
- .ocon_num = OCON_NUM,
+ .ocon_num = OCON_NUM - 2,
},
{
.version = POLICYDB_VERSION_NEW_OBJECT_DEFAULTS,
.sym_num = SYM_NUM,
- .ocon_num = OCON_NUM,
+ .ocon_num = OCON_NUM - 2,
},
{
.version = POLICYDB_VERSION_DEFAULT_TYPE,
.sym_num = SYM_NUM,
- .ocon_num = OCON_NUM,
+ .ocon_num = OCON_NUM - 2,
},
{
.version = POLICYDB_VERSION_CONSTRAINT_NAMES,
.sym_num = SYM_NUM,
- .ocon_num = OCON_NUM,
+ .ocon_num = OCON_NUM - 2,
},
{
.version = POLICYDB_VERSION_XPERMS_IOCTL,
.sym_num = SYM_NUM,
+ .ocon_num = OCON_NUM - 2,
+ },
+ {
+ .version = POLICYDB_VERSION_INFINIBAND,
+ .sym_num = SYM_NUM,
.ocon_num = OCON_NUM,
},
};
@@ -2219,6 +2229,61 @@ static int ocontext_read(struct policydb *p, struct policydb_compat_info *info,
goto out;
break;
}
+ case OCON_PKEY: {
+ rc = next_entry(nodebuf, fp, sizeof(u32) * 6);
+ if (rc)
+ goto out;
+
+ c->u.pkey.subnet_prefix = be64_to_cpu(*((__be64 *)nodebuf));
+ /* The subnet prefix is stored as an IPv6
+ * address in the policy.
+ *
+ * Check that the lower 2 DWORDS are 0.
+ */
+ if (nodebuf[2] || nodebuf[3]) {
+ rc = -EINVAL;
+ goto out;
+ }
+
+ if (nodebuf[4] > 0xffff ||
+ nodebuf[5] > 0xffff) {
+ rc = -EINVAL;
+ goto out;
+ }
+
+ c->u.pkey.low_pkey = le32_to_cpu(nodebuf[4]);
+ c->u.pkey.high_pkey = le32_to_cpu(nodebuf[5]);
+
+ rc = context_read_and_validate(&c->context[0],
+ p,
+ fp);
+ if (rc)
+ goto out;
+ break;
+ }
+ case OCON_IB_PORT:
+ rc = next_entry(buf, fp, sizeof(u32) * 2);
+ if (rc)
+ goto out;
+ len = le32_to_cpu(buf[0]);
+
+ rc = str_read(&c->u.ib_port.dev_name, GFP_KERNEL, fp, len);
+ if (rc)
+ goto out;
+
+ if (buf[1] > 0xff || buf[1] == 0) {
+ rc = -EINVAL;
+ goto out;
+ }
+
+ c->u.ib_port.port = le32_to_cpu(buf[1]);
+
+ rc = context_read_and_validate(&c->context[0],
+ p,
+ fp);
+ if (rc)
+ goto out;
+ break;
}
}
}
@@ -3147,6 +3212,41 @@ static int ocontext_write(struct policydb *p, struct policydb_compat_info *info,
if (rc)
return rc;
break;
+ case OCON_PKEY:
+ *((__be64 *)nodebuf) = cpu_to_be64(c->u.pkey.subnet_prefix);
+
+ /*
+ * The low order 2 bits were confirmed to be 0
+ * when the policy was loaded. Write them out
+ * as zero
+ */
+ nodebuf[2] = 0;
+ nodebuf[3] = 0;
+
+ nodebuf[4] = cpu_to_le32(c->u.pkey.low_pkey);
+ nodebuf[5] = cpu_to_le32(c->u.pkey.high_pkey);
+
+ rc = put_entry(nodebuf, sizeof(u32), 6, fp);
+ if (rc)
+ return rc;
+ rc = context_write(p, &c->context[0], fp);
+ if (rc)
+ return rc;
+ break;
+ case OCON_IB_PORT:
+ len = strlen(c->u.ib_port.dev_name);
+ buf[0] = cpu_to_le32(len);
+ buf[1] = cpu_to_le32(c->u.ib_port.port);
+ rc = put_entry(buf, sizeof(u32), 2, fp);
+ if (rc)
+ return rc;
+ rc = put_entry(c->u.ib_port.dev_name, 1, len, fp);
+ if (rc)
+ return rc;
+ rc = context_write(p, &c->context[0], fp);
+ if (rc)
+ return rc;
+ break;
}
}
}
diff --git a/security/selinux/ss/policydb.h b/security/selinux/ss/policydb.h
index 725d594..9f20774 100644
--- a/security/selinux/ss/policydb.h
+++ b/security/selinux/ss/policydb.h
@@ -187,6 +187,15 @@ struct ocontext {
u32 addr[4];
u32 mask[4];
} node6; /* IPv6 node information */
+ struct {
+ u64 subnet_prefix;
+ u16 low_pkey;
+ u16 high_pkey;
+ } pkey;
+ struct {
+ char *dev_name;
+ u8 port;
+ } ib_port;
} u;
union {
u32 sclass; /* security class for genfs */
@@ -215,14 +224,16 @@ struct genfs {
#define SYM_NUM 8
/* object context array indices */
-#define OCON_ISID 0 /* initial SIDs */
-#define OCON_FS 1 /* unlabeled file systems */
-#define OCON_PORT 2 /* TCP and UDP port numbers */
-#define OCON_NETIF 3 /* network interfaces */
-#define OCON_NODE 4 /* nodes */
-#define OCON_FSUSE 5 /* fs_use */
-#define OCON_NODE6 6 /* IPv6 nodes */
-#define OCON_NUM 7
+#define OCON_ISID 0 /* initial SIDs */
+#define OCON_FS 1 /* unlabeled file systems */
+#define OCON_PORT 2 /* TCP and UDP port numbers */
+#define OCON_NETIF 3 /* network interfaces */
+#define OCON_NODE 4 /* nodes */
+#define OCON_FSUSE 5 /* fs_use */
+#define OCON_NODE6 6 /* IPv6 nodes */
+#define OCON_PKEY 7 /* Infiniband PKeys */
+#define OCON_IB_PORT 8 /* Infiniband ports */
+#define OCON_NUM 9
/* The policy database */
struct policydb {
--
1.7.1
^ permalink raw reply related [flat|nested] 42+ messages in thread
* [PATCH v2 6/9] selinux: Allocate and free infiniband security hooks
2016-07-14 22:56 ` Dan Jurgens
@ 2016-07-14 22:56 ` Dan Jurgens
-1 siblings, 0 replies; 42+ messages in thread
From: Dan Jurgens @ 2016-07-14 22:56 UTC (permalink / raw)
To: chrisw-69jw2NvuJkxg9hUCZPvPmw, paul-r2n+y4ga6xFZroRs9YW3xA,
sds-+05T5uksL2qpZYMLLGbcSA, eparis-FjpueFixGhCM4zKIHC2jIg,
dledford-H+wXaHxf7aLQT0dZR+AlfA,
sean.hefty-ral2JQCrhuEAvxtiuMwx3w,
hal.rosenstock-Re5JQEeQqe8AvxtiuMwx3w
Cc: selinux-+05T5uksL2qpZYMLLGbcSA,
linux-security-module-u79uwXL29TY76Z2rM5mHXA,
linux-rdma-u79uwXL29TY76Z2rM5mHXA,
yevgenyp-VPRAkNaXOzVWk0Htik3J/w, Daniel Jurgens
From: Daniel Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Implement and attach hooks to allocate and free Infiniband object
security structures.
Signed-off-by: Daniel Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---
v2:
- Use void * blobs for security structs. Paul Moore
- Shorten ib_end_port to ib_port. Paul Moore
- Allocate memory for security struct with GFP_KERNEL. Yuval Shaia
---
security/selinux/hooks.c | 25 ++++++++++++++++++++++++-
security/selinux/include/objsec.h | 5 +++++
2 files changed, 29 insertions(+), 1 deletions(-)
diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c
index a363202..fe3d20d 100644
--- a/security/selinux/hooks.c
+++ b/security/selinux/hooks.c
@@ -17,6 +17,7 @@
* Paul Moore <paul-r2n+y4ga6xFZroRs9YW3xA@public.gmane.org>
* Copyright (C) 2007 Hitachi Software Engineering Co., Ltd.
* Yuichi Nakamura <ynakam-FkO1umbPgv4fag7Bw7Dlfw@public.gmane.org>
+ * Copyright (C) 2016 Mellanox Technologies
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License version 2,
@@ -5991,7 +5992,26 @@ static int selinux_key_getsecurity(struct key *key, char **_buffer)
*_buffer = context;
return rc;
}
+#endif
+
+#ifdef CONFIG_SECURITY_INFINIBAND
+static int selinux_ib_alloc_security(void **ib_sec)
+{
+ struct ib_security_struct *sec;
+
+ sec = kzalloc(sizeof(*sec), GFP_KERNEL);
+ if (!sec)
+ return -ENOMEM;
+ sec->sid = current_sid();
+
+ *ib_sec = sec;
+ return 0;
+}
+static void selinux_ib_free_security(void *ib_sec)
+{
+ kfree(ib_sec);
+}
#endif
static struct security_hook_list selinux_hooks[] = {
@@ -6175,7 +6195,10 @@ static struct security_hook_list selinux_hooks[] = {
LSM_HOOK_INIT(tun_dev_attach_queue, selinux_tun_dev_attach_queue),
LSM_HOOK_INIT(tun_dev_attach, selinux_tun_dev_attach),
LSM_HOOK_INIT(tun_dev_open, selinux_tun_dev_open),
-
+#ifdef CONFIG_SECURITY_INFINIBAND
+ LSM_HOOK_INIT(ib_alloc_security, selinux_ib_alloc_security),
+ LSM_HOOK_INIT(ib_free_security, selinux_ib_free_security),
+#endif
#ifdef CONFIG_SECURITY_NETWORK_XFRM
LSM_HOOK_INIT(xfrm_policy_alloc_security, selinux_xfrm_policy_alloc),
LSM_HOOK_INIT(xfrm_policy_clone_security, selinux_xfrm_policy_clone),
diff --git a/security/selinux/include/objsec.h b/security/selinux/include/objsec.h
index c21e135..8e7db43 100644
--- a/security/selinux/include/objsec.h
+++ b/security/selinux/include/objsec.h
@@ -10,6 +10,7 @@
*
* Copyright (C) 2001,2002 Networks Associates Technology, Inc.
* Copyright (C) 2003 Red Hat, Inc., James Morris <jmorris-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
+ * Copyright (C) 2016 Mellanox Technologies
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License version 2,
@@ -128,6 +129,10 @@ struct key_security_struct {
u32 sid; /* SID of key */
};
+struct ib_security_struct {
+ u32 sid; /* SID of the queue pair or MAD agent */
+};
+
extern unsigned int selinux_checkreqprot;
#endif /* _SELINUX_OBJSEC_H_ */
--
1.7.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related [flat|nested] 42+ messages in thread
* [PATCH v2 6/9] selinux: Allocate and free infiniband security hooks
@ 2016-07-14 22:56 ` Dan Jurgens
0 siblings, 0 replies; 42+ messages in thread
From: Dan Jurgens @ 2016-07-14 22:56 UTC (permalink / raw)
To: chrisw, paul, sds, eparis, dledford, sean.hefty, hal.rosenstock
Cc: selinux, linux-security-module, linux-rdma, yevgenyp, Daniel Jurgens
From: Daniel Jurgens <danielj@mellanox.com>
Implement and attach hooks to allocate and free Infiniband object
security structures.
Signed-off-by: Daniel Jurgens <danielj@mellanox.com>
---
v2:
- Use void * blobs for security structs. Paul Moore
- Shorten ib_end_port to ib_port. Paul Moore
- Allocate memory for security struct with GFP_KERNEL. Yuval Shaia
---
security/selinux/hooks.c | 25 ++++++++++++++++++++++++-
security/selinux/include/objsec.h | 5 +++++
2 files changed, 29 insertions(+), 1 deletions(-)
diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c
index a363202..fe3d20d 100644
--- a/security/selinux/hooks.c
+++ b/security/selinux/hooks.c
@@ -17,6 +17,7 @@
* Paul Moore <paul@paul-moore.com>
* Copyright (C) 2007 Hitachi Software Engineering Co., Ltd.
* Yuichi Nakamura <ynakam@hitachisoft.jp>
+ * Copyright (C) 2016 Mellanox Technologies
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License version 2,
@@ -5991,7 +5992,26 @@ static int selinux_key_getsecurity(struct key *key, char **_buffer)
*_buffer = context;
return rc;
}
+#endif
+
+#ifdef CONFIG_SECURITY_INFINIBAND
+static int selinux_ib_alloc_security(void **ib_sec)
+{
+ struct ib_security_struct *sec;
+
+ sec = kzalloc(sizeof(*sec), GFP_KERNEL);
+ if (!sec)
+ return -ENOMEM;
+ sec->sid = current_sid();
+
+ *ib_sec = sec;
+ return 0;
+}
+static void selinux_ib_free_security(void *ib_sec)
+{
+ kfree(ib_sec);
+}
#endif
static struct security_hook_list selinux_hooks[] = {
@@ -6175,7 +6195,10 @@ static struct security_hook_list selinux_hooks[] = {
LSM_HOOK_INIT(tun_dev_attach_queue, selinux_tun_dev_attach_queue),
LSM_HOOK_INIT(tun_dev_attach, selinux_tun_dev_attach),
LSM_HOOK_INIT(tun_dev_open, selinux_tun_dev_open),
-
+#ifdef CONFIG_SECURITY_INFINIBAND
+ LSM_HOOK_INIT(ib_alloc_security, selinux_ib_alloc_security),
+ LSM_HOOK_INIT(ib_free_security, selinux_ib_free_security),
+#endif
#ifdef CONFIG_SECURITY_NETWORK_XFRM
LSM_HOOK_INIT(xfrm_policy_alloc_security, selinux_xfrm_policy_alloc),
LSM_HOOK_INIT(xfrm_policy_clone_security, selinux_xfrm_policy_clone),
diff --git a/security/selinux/include/objsec.h b/security/selinux/include/objsec.h
index c21e135..8e7db43 100644
--- a/security/selinux/include/objsec.h
+++ b/security/selinux/include/objsec.h
@@ -10,6 +10,7 @@
*
* Copyright (C) 2001,2002 Networks Associates Technology, Inc.
* Copyright (C) 2003 Red Hat, Inc., James Morris <jmorris@redhat.com>
+ * Copyright (C) 2016 Mellanox Technologies
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License version 2,
@@ -128,6 +129,10 @@ struct key_security_struct {
u32 sid; /* SID of key */
};
+struct ib_security_struct {
+ u32 sid; /* SID of the queue pair or MAD agent */
+};
+
extern unsigned int selinux_checkreqprot;
#endif /* _SELINUX_OBJSEC_H_ */
--
1.7.1
^ permalink raw reply related [flat|nested] 42+ messages in thread
* [PATCH v2 7/9] selinux: Implement Infiniband PKey "Access" access vector
2016-07-14 22:56 ` Dan Jurgens
@ 2016-07-14 22:56 ` Dan Jurgens
-1 siblings, 0 replies; 42+ messages in thread
From: Dan Jurgens @ 2016-07-14 22:56 UTC (permalink / raw)
To: chrisw-69jw2NvuJkxg9hUCZPvPmw, paul-r2n+y4ga6xFZroRs9YW3xA,
sds-+05T5uksL2qpZYMLLGbcSA, eparis-FjpueFixGhCM4zKIHC2jIg,
dledford-H+wXaHxf7aLQT0dZR+AlfA,
sean.hefty-ral2JQCrhuEAvxtiuMwx3w,
hal.rosenstock-Re5JQEeQqe8AvxtiuMwx3w
Cc: selinux-+05T5uksL2qpZYMLLGbcSA,
linux-security-module-u79uwXL29TY76Z2rM5mHXA,
linux-rdma-u79uwXL29TY76Z2rM5mHXA,
yevgenyp-VPRAkNaXOzVWk0Htik3J/w, Daniel Jurgens
From: Daniel Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Add a type and access vector for PKeys. Implement the ib_pkey_access
hook to check that the caller has permission to access the PKey on the
given subnet prefix. Add an interface to get the PKey SID. Walk the PKey
ocontexts to find an entry for the given subnet prefix and pkey.
Signed-off-by: Daniel Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---
v2:
- Use void* blobs for security structs. Paul Moore
- Add pkey specific data to the audit log. Paul Moore
- Don't introduce a new initial sid, use unlabeled. Stephen Smalley
---
include/linux/lsm_audit.h | 7 ++++++
security/lsm_audit.c | 13 +++++++++++
security/selinux/hooks.c | 23 +++++++++++++++++++
security/selinux/include/classmap.h | 2 +
security/selinux/include/security.h | 2 +
security/selinux/ss/services.c | 41 +++++++++++++++++++++++++++++++++++
6 files changed, 88 insertions(+), 0 deletions(-)
diff --git a/include/linux/lsm_audit.h b/include/linux/lsm_audit.h
index ffb9c9d..8ff7eae 100644
--- a/include/linux/lsm_audit.h
+++ b/include/linux/lsm_audit.h
@@ -45,6 +45,11 @@ struct lsm_ioctlop_audit {
u16 cmd;
};
+struct lsm_pkey_audit {
+ u64 subnet_prefix;
+ u16 pkey;
+};
+
/* Auxiliary data to use in generating the audit record. */
struct common_audit_data {
char type;
@@ -59,6 +64,7 @@ struct common_audit_data {
#define LSM_AUDIT_DATA_INODE 9
#define LSM_AUDIT_DATA_DENTRY 10
#define LSM_AUDIT_DATA_IOCTL_OP 11
+#define LSM_AUDIT_DATA_PKEY 12
union {
struct path path;
struct dentry *dentry;
@@ -75,6 +81,7 @@ struct common_audit_data {
#endif
char *kmod_name;
struct lsm_ioctlop_audit *op;
+ struct lsm_pkey_audit *pkey;
} u;
/* this union contains LSM specific data */
union {
diff --git a/security/lsm_audit.c b/security/lsm_audit.c
index cccbf30..2546d82 100644
--- a/security/lsm_audit.c
+++ b/security/lsm_audit.c
@@ -397,6 +397,19 @@ static void dump_common_audit_data(struct audit_buffer *ab,
audit_log_format(ab, " kmod=");
audit_log_untrustedstring(ab, a->u.kmod_name);
break;
+ case LSM_AUDIT_DATA_PKEY: {
+ struct in6_addr sbn_pfx;
+
+ memset(&sbn_pfx.s6_addr, 0,
+ sizeof(sbn_pfx.s6_addr));
+
+ memcpy(&sbn_pfx.s6_addr, &a->u.pkey->subnet_prefix,
+ sizeof(a->u.pkey->subnet_prefix));
+
+ audit_log_format(ab, " pkey=0x%x subnet_prefix=%pI6c",
+ a->u.pkey->pkey, &sbn_pfx);
+ break;
+ }
} /* switch (a->type) */
}
diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c
index fe3d20d..d9c4bfa 100644
--- a/security/selinux/hooks.c
+++ b/security/selinux/hooks.c
@@ -5995,6 +5995,28 @@ static int selinux_key_getsecurity(struct key *key, char **_buffer)
#endif
#ifdef CONFIG_SECURITY_INFINIBAND
+static int selinux_ib_pkey_access(u64 subnet_prefix, u16 pkey_val, void *ib_sec)
+{
+ struct common_audit_data ad;
+ int err;
+ u32 sid = 0;
+ struct ib_security_struct *sec = ib_sec;
+ struct lsm_pkey_audit pkey;
+
+ err = security_pkey_sid(subnet_prefix, pkey_val, &sid);
+
+ if (err)
+ return err;
+
+ ad.type = LSM_AUDIT_DATA_PKEY;
+ pkey.subnet_prefix = subnet_prefix;
+ pkey.pkey = pkey_val;
+ ad.u.pkey = &pkey;
+ return avc_has_perm(sec->sid, sid,
+ SECCLASS_INFINIBAND_PKEY,
+ INFINIBAND_PKEY__ACCESS, &ad);
+}
+
static int selinux_ib_alloc_security(void **ib_sec)
{
struct ib_security_struct *sec;
@@ -6196,6 +6218,7 @@ static struct security_hook_list selinux_hooks[] = {
LSM_HOOK_INIT(tun_dev_attach, selinux_tun_dev_attach),
LSM_HOOK_INIT(tun_dev_open, selinux_tun_dev_open),
#ifdef CONFIG_SECURITY_INFINIBAND
+ LSM_HOOK_INIT(ib_pkey_access, selinux_ib_pkey_access),
LSM_HOOK_INIT(ib_alloc_security, selinux_ib_alloc_security),
LSM_HOOK_INIT(ib_free_security, selinux_ib_free_security),
#endif
diff --git a/security/selinux/include/classmap.h b/security/selinux/include/classmap.h
index 1f1f4b2..d42dd4d 100644
--- a/security/selinux/include/classmap.h
+++ b/security/selinux/include/classmap.h
@@ -165,5 +165,7 @@ struct security_class_mapping secclass_map[] = {
{ COMMON_CAP_PERMS, NULL } },
{ "cap2_userns",
{ COMMON_CAP2_PERMS, NULL } },
+ { "infiniband_pkey",
+ { "access", NULL } },
{ NULL }
};
diff --git a/security/selinux/include/security.h b/security/selinux/include/security.h
index a7e6ed2..8f1a66e 100644
--- a/security/selinux/include/security.h
+++ b/security/selinux/include/security.h
@@ -180,6 +180,8 @@ int security_get_user_sids(u32 callsid, char *username,
int security_port_sid(u8 protocol, u16 port, u32 *out_sid);
+int security_pkey_sid(u64 subnet_prefix, u16 pkey_num, u32 *out_sid);
+
int security_netif_sid(char *name, u32 *if_sid);
int security_node_sid(u16 domain, void *addr, u32 addrlen,
diff --git a/security/selinux/ss/services.c b/security/selinux/ss/services.c
index 89df646..ba346da 100644
--- a/security/selinux/ss/services.c
+++ b/security/selinux/ss/services.c
@@ -2229,6 +2229,47 @@ out:
}
/**
+ * security_pkey_sid - Obtain the SID for a pkey.
+ * @subnet_prefix: Subnet Prefix
+ * @pkey_num: pkey number
+ * @out_sid: security identifier
+ */
+int security_pkey_sid(u64 subnet_prefix, u16 pkey_num, u32 *out_sid)
+{
+ struct ocontext *c;
+ int rc = 0;
+
+ read_lock(&policy_rwlock);
+
+ c = policydb.ocontexts[OCON_PKEY];
+ while (c) {
+ if (c->u.pkey.low_pkey <= pkey_num &&
+ c->u.pkey.high_pkey >= pkey_num &&
+ c->u.pkey.subnet_prefix == subnet_prefix)
+ break;
+
+ c = c->next;
+ }
+
+ if (c) {
+ if (!c->sid[0]) {
+ rc = sidtab_context_to_sid(&sidtab,
+ &c->context[0],
+ &c->sid[0]);
+ if (rc)
+ goto out;
+ }
+ *out_sid = c->sid[0];
+ } else {
+ *out_sid = SECINITSID_UNLABELED;
+ }
+
+out:
+ read_unlock(&policy_rwlock);
+ return rc;
+}
+
+/**
* security_netif_sid - Obtain the SID for a network interface.
* @name: interface name
* @if_sid: interface SID
--
1.7.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related [flat|nested] 42+ messages in thread
* [PATCH v2 7/9] selinux: Implement Infiniband PKey "Access" access vector
@ 2016-07-14 22:56 ` Dan Jurgens
0 siblings, 0 replies; 42+ messages in thread
From: Dan Jurgens @ 2016-07-14 22:56 UTC (permalink / raw)
To: chrisw, paul, sds, eparis, dledford, sean.hefty, hal.rosenstock
Cc: selinux, linux-security-module, linux-rdma, yevgenyp, Daniel Jurgens
From: Daniel Jurgens <danielj@mellanox.com>
Add a type and access vector for PKeys. Implement the ib_pkey_access
hook to check that the caller has permission to access the PKey on the
given subnet prefix. Add an interface to get the PKey SID. Walk the PKey
ocontexts to find an entry for the given subnet prefix and pkey.
Signed-off-by: Daniel Jurgens <danielj@mellanox.com>
---
v2:
- Use void* blobs for security structs. Paul Moore
- Add pkey specific data to the audit log. Paul Moore
- Don't introduce a new initial sid, use unlabeled. Stephen Smalley
---
include/linux/lsm_audit.h | 7 ++++++
security/lsm_audit.c | 13 +++++++++++
security/selinux/hooks.c | 23 +++++++++++++++++++
security/selinux/include/classmap.h | 2 +
security/selinux/include/security.h | 2 +
security/selinux/ss/services.c | 41 +++++++++++++++++++++++++++++++++++
6 files changed, 88 insertions(+), 0 deletions(-)
diff --git a/include/linux/lsm_audit.h b/include/linux/lsm_audit.h
index ffb9c9d..8ff7eae 100644
--- a/include/linux/lsm_audit.h
+++ b/include/linux/lsm_audit.h
@@ -45,6 +45,11 @@ struct lsm_ioctlop_audit {
u16 cmd;
};
+struct lsm_pkey_audit {
+ u64 subnet_prefix;
+ u16 pkey;
+};
+
/* Auxiliary data to use in generating the audit record. */
struct common_audit_data {
char type;
@@ -59,6 +64,7 @@ struct common_audit_data {
#define LSM_AUDIT_DATA_INODE 9
#define LSM_AUDIT_DATA_DENTRY 10
#define LSM_AUDIT_DATA_IOCTL_OP 11
+#define LSM_AUDIT_DATA_PKEY 12
union {
struct path path;
struct dentry *dentry;
@@ -75,6 +81,7 @@ struct common_audit_data {
#endif
char *kmod_name;
struct lsm_ioctlop_audit *op;
+ struct lsm_pkey_audit *pkey;
} u;
/* this union contains LSM specific data */
union {
diff --git a/security/lsm_audit.c b/security/lsm_audit.c
index cccbf30..2546d82 100644
--- a/security/lsm_audit.c
+++ b/security/lsm_audit.c
@@ -397,6 +397,19 @@ static void dump_common_audit_data(struct audit_buffer *ab,
audit_log_format(ab, " kmod=");
audit_log_untrustedstring(ab, a->u.kmod_name);
break;
+ case LSM_AUDIT_DATA_PKEY: {
+ struct in6_addr sbn_pfx;
+
+ memset(&sbn_pfx.s6_addr, 0,
+ sizeof(sbn_pfx.s6_addr));
+
+ memcpy(&sbn_pfx.s6_addr, &a->u.pkey->subnet_prefix,
+ sizeof(a->u.pkey->subnet_prefix));
+
+ audit_log_format(ab, " pkey=0x%x subnet_prefix=%pI6c",
+ a->u.pkey->pkey, &sbn_pfx);
+ break;
+ }
} /* switch (a->type) */
}
diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c
index fe3d20d..d9c4bfa 100644
--- a/security/selinux/hooks.c
+++ b/security/selinux/hooks.c
@@ -5995,6 +5995,28 @@ static int selinux_key_getsecurity(struct key *key, char **_buffer)
#endif
#ifdef CONFIG_SECURITY_INFINIBAND
+static int selinux_ib_pkey_access(u64 subnet_prefix, u16 pkey_val, void *ib_sec)
+{
+ struct common_audit_data ad;
+ int err;
+ u32 sid = 0;
+ struct ib_security_struct *sec = ib_sec;
+ struct lsm_pkey_audit pkey;
+
+ err = security_pkey_sid(subnet_prefix, pkey_val, &sid);
+
+ if (err)
+ return err;
+
+ ad.type = LSM_AUDIT_DATA_PKEY;
+ pkey.subnet_prefix = subnet_prefix;
+ pkey.pkey = pkey_val;
+ ad.u.pkey = &pkey;
+ return avc_has_perm(sec->sid, sid,
+ SECCLASS_INFINIBAND_PKEY,
+ INFINIBAND_PKEY__ACCESS, &ad);
+}
+
static int selinux_ib_alloc_security(void **ib_sec)
{
struct ib_security_struct *sec;
@@ -6196,6 +6218,7 @@ static struct security_hook_list selinux_hooks[] = {
LSM_HOOK_INIT(tun_dev_attach, selinux_tun_dev_attach),
LSM_HOOK_INIT(tun_dev_open, selinux_tun_dev_open),
#ifdef CONFIG_SECURITY_INFINIBAND
+ LSM_HOOK_INIT(ib_pkey_access, selinux_ib_pkey_access),
LSM_HOOK_INIT(ib_alloc_security, selinux_ib_alloc_security),
LSM_HOOK_INIT(ib_free_security, selinux_ib_free_security),
#endif
diff --git a/security/selinux/include/classmap.h b/security/selinux/include/classmap.h
index 1f1f4b2..d42dd4d 100644
--- a/security/selinux/include/classmap.h
+++ b/security/selinux/include/classmap.h
@@ -165,5 +165,7 @@ struct security_class_mapping secclass_map[] = {
{ COMMON_CAP_PERMS, NULL } },
{ "cap2_userns",
{ COMMON_CAP2_PERMS, NULL } },
+ { "infiniband_pkey",
+ { "access", NULL } },
{ NULL }
};
diff --git a/security/selinux/include/security.h b/security/selinux/include/security.h
index a7e6ed2..8f1a66e 100644
--- a/security/selinux/include/security.h
+++ b/security/selinux/include/security.h
@@ -180,6 +180,8 @@ int security_get_user_sids(u32 callsid, char *username,
int security_port_sid(u8 protocol, u16 port, u32 *out_sid);
+int security_pkey_sid(u64 subnet_prefix, u16 pkey_num, u32 *out_sid);
+
int security_netif_sid(char *name, u32 *if_sid);
int security_node_sid(u16 domain, void *addr, u32 addrlen,
diff --git a/security/selinux/ss/services.c b/security/selinux/ss/services.c
index 89df646..ba346da 100644
--- a/security/selinux/ss/services.c
+++ b/security/selinux/ss/services.c
@@ -2229,6 +2229,47 @@ out:
}
/**
+ * security_pkey_sid - Obtain the SID for a pkey.
+ * @subnet_prefix: Subnet Prefix
+ * @pkey_num: pkey number
+ * @out_sid: security identifier
+ */
+int security_pkey_sid(u64 subnet_prefix, u16 pkey_num, u32 *out_sid)
+{
+ struct ocontext *c;
+ int rc = 0;
+
+ read_lock(&policy_rwlock);
+
+ c = policydb.ocontexts[OCON_PKEY];
+ while (c) {
+ if (c->u.pkey.low_pkey <= pkey_num &&
+ c->u.pkey.high_pkey >= pkey_num &&
+ c->u.pkey.subnet_prefix == subnet_prefix)
+ break;
+
+ c = c->next;
+ }
+
+ if (c) {
+ if (!c->sid[0]) {
+ rc = sidtab_context_to_sid(&sidtab,
+ &c->context[0],
+ &c->sid[0]);
+ if (rc)
+ goto out;
+ }
+ *out_sid = c->sid[0];
+ } else {
+ *out_sid = SECINITSID_UNLABELED;
+ }
+
+out:
+ read_unlock(&policy_rwlock);
+ return rc;
+}
+
+/**
* security_netif_sid - Obtain the SID for a network interface.
* @name: interface name
* @if_sid: interface SID
--
1.7.1
^ permalink raw reply related [flat|nested] 42+ messages in thread
* [PATCH v2 8/9] selinux: Add IB Port SMP access vector
2016-07-14 22:56 ` Dan Jurgens
` (2 preceding siblings ...)
(?)
@ 2016-07-14 22:56 ` Dan Jurgens
[not found] ` <1468537011-20407-9-git-send-email-danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
-1 siblings, 1 reply; 42+ messages in thread
From: Dan Jurgens @ 2016-07-14 22:56 UTC (permalink / raw)
To: chrisw, paul, sds, eparis, dledford, sean.hefty, hal.rosenstock
Cc: selinux, linux-security-module, linux-rdma, yevgenyp, Daniel Jurgens
From: Daniel Jurgens <danielj@mellanox.com>
Add a type for Infiniband ports and an access vector for subnet
management packets. Implement the ib_port_smp hook to check that the
caller has permission to send and receive SMPs on the end port specified
by the device name and port. Add interface to query the SID for a IB
port, which walks the IB_PORT ocontexts to find an entry for the
given name and port.
Signed-off-by: Daniel Jurgens <danielj@mellanox.com>
---
v2:
- Shorted ib_end_port. Paul Moore
- Pass void blobs to security hooks. Paul Moore
- Log specific IB port info in audit log. Paul Moore
- Don't create a new intial sid, use unlabeled. Stephen Smalley
- Changed "smp" to "manage_subnet". Paul Moore
---
include/linux/lsm_audit.h | 32 ++++++++++++++++----------
security/lsm_audit.c | 5 ++++
security/selinux/hooks.c | 23 +++++++++++++++++++
security/selinux/include/classmap.h | 2 +
security/selinux/include/security.h | 2 +
security/selinux/ss/services.c | 42 +++++++++++++++++++++++++++++++++++
6 files changed, 94 insertions(+), 12 deletions(-)
diff --git a/include/linux/lsm_audit.h b/include/linux/lsm_audit.h
index 8ff7eae..d2f6bda 100644
--- a/include/linux/lsm_audit.h
+++ b/include/linux/lsm_audit.h
@@ -21,6 +21,7 @@
#include <linux/path.h>
#include <linux/key.h>
#include <linux/skbuff.h>
+#include <rdma/ib_verbs.h>
struct lsm_network_audit {
int netif;
@@ -50,21 +51,27 @@ struct lsm_pkey_audit {
u16 pkey;
};
+struct lsm_ib_port_audit {
+ char dev_name[IB_DEVICE_NAME_MAX];
+ u8 port;
+};
+
/* Auxiliary data to use in generating the audit record. */
struct common_audit_data {
char type;
-#define LSM_AUDIT_DATA_PATH 1
-#define LSM_AUDIT_DATA_NET 2
-#define LSM_AUDIT_DATA_CAP 3
-#define LSM_AUDIT_DATA_IPC 4
-#define LSM_AUDIT_DATA_TASK 5
-#define LSM_AUDIT_DATA_KEY 6
-#define LSM_AUDIT_DATA_NONE 7
-#define LSM_AUDIT_DATA_KMOD 8
-#define LSM_AUDIT_DATA_INODE 9
-#define LSM_AUDIT_DATA_DENTRY 10
-#define LSM_AUDIT_DATA_IOCTL_OP 11
-#define LSM_AUDIT_DATA_PKEY 12
+#define LSM_AUDIT_DATA_PATH 1
+#define LSM_AUDIT_DATA_NET 2
+#define LSM_AUDIT_DATA_CAP 3
+#define LSM_AUDIT_DATA_IPC 4
+#define LSM_AUDIT_DATA_TASK 5
+#define LSM_AUDIT_DATA_KEY 6
+#define LSM_AUDIT_DATA_NONE 7
+#define LSM_AUDIT_DATA_KMOD 8
+#define LSM_AUDIT_DATA_INODE 9
+#define LSM_AUDIT_DATA_DENTRY 10
+#define LSM_AUDIT_DATA_IOCTL_OP 11
+#define LSM_AUDIT_DATA_PKEY 12
+#define LSM_AUDIT_DATA_IB_PORT 13
union {
struct path path;
struct dentry *dentry;
@@ -82,6 +89,7 @@ struct common_audit_data {
char *kmod_name;
struct lsm_ioctlop_audit *op;
struct lsm_pkey_audit *pkey;
+ struct lsm_ib_port_audit *ib_port;
} u;
/* this union contains LSM specific data */
union {
diff --git a/security/lsm_audit.c b/security/lsm_audit.c
index 2546d82..95632a9 100644
--- a/security/lsm_audit.c
+++ b/security/lsm_audit.c
@@ -410,6 +410,11 @@ static void dump_common_audit_data(struct audit_buffer *ab,
a->u.pkey->pkey, &sbn_pfx);
break;
}
+ case LSM_AUDIT_DATA_IB_PORT:
+ audit_log_format(ab, " device=");
+ audit_log_untrustedstring(ab, a->u.ib_port->dev_name);
+ audit_log_format(ab, " port=%u", a->u.ib_port->port);
+ break;
} /* switch (a->type) */
}
diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c
index d9c4bfa..40ad0e8 100644
--- a/security/selinux/hooks.c
+++ b/security/selinux/hooks.c
@@ -6017,6 +6017,28 @@ static int selinux_ib_pkey_access(u64 subnet_prefix, u16 pkey_val, void *ib_sec)
INFINIBAND_PKEY__ACCESS, &ad);
}
+static int selinux_ib_port_manage_subnet(const char *dev_name, u8 port, void *ib_sec)
+{
+ struct common_audit_data ad;
+ int err;
+ u32 sid = 0;
+ struct ib_security_struct *sec = ib_sec;
+ struct lsm_ib_port_audit ib_port;
+
+ err = security_ib_port_sid(dev_name, port, &sid);
+
+ if (err)
+ return err;
+
+ ad.type = LSM_AUDIT_DATA_IB_PORT;
+ strncpy(ib_port.dev_name, dev_name, sizeof(ib_port.dev_name));
+ ib_port.port = port;
+ ad.u.ib_port = &ib_port;
+ return avc_has_perm(sec->sid, sid,
+ SECCLASS_INFINIBAND_PORT,
+ INFINIBAND_PORT__MANAGE_SUBNET, &ad);
+}
+
static int selinux_ib_alloc_security(void **ib_sec)
{
struct ib_security_struct *sec;
@@ -6219,6 +6241,7 @@ static struct security_hook_list selinux_hooks[] = {
LSM_HOOK_INIT(tun_dev_open, selinux_tun_dev_open),
#ifdef CONFIG_SECURITY_INFINIBAND
LSM_HOOK_INIT(ib_pkey_access, selinux_ib_pkey_access),
+ LSM_HOOK_INIT(ib_port_manage_subnet, selinux_ib_port_manage_subnet),
LSM_HOOK_INIT(ib_alloc_security, selinux_ib_alloc_security),
LSM_HOOK_INIT(ib_free_security, selinux_ib_free_security),
#endif
diff --git a/security/selinux/include/classmap.h b/security/selinux/include/classmap.h
index d42dd4d..9ba1238 100644
--- a/security/selinux/include/classmap.h
+++ b/security/selinux/include/classmap.h
@@ -167,5 +167,7 @@ struct security_class_mapping secclass_map[] = {
{ COMMON_CAP2_PERMS, NULL } },
{ "infiniband_pkey",
{ "access", NULL } },
+ { "infiniband_port",
+ { "manage_subnet", NULL } },
{ NULL }
};
diff --git a/security/selinux/include/security.h b/security/selinux/include/security.h
index 8f1a66e..1d58583 100644
--- a/security/selinux/include/security.h
+++ b/security/selinux/include/security.h
@@ -182,6 +182,8 @@ int security_port_sid(u8 protocol, u16 port, u32 *out_sid);
int security_pkey_sid(u64 subnet_prefix, u16 pkey_num, u32 *out_sid);
+int security_ib_port_sid(const char *dev_name, u8 port, u32 *out_sid);
+
int security_netif_sid(char *name, u32 *if_sid);
int security_node_sid(u16 domain, void *addr, u32 addrlen,
diff --git a/security/selinux/ss/services.c b/security/selinux/ss/services.c
index ba346da..4c97952 100644
--- a/security/selinux/ss/services.c
+++ b/security/selinux/ss/services.c
@@ -2270,6 +2270,48 @@ out:
}
/**
+ * security_ib_port_sid - Obtain the SID for a subnet management interface.
+ * @dev_name: device name
+ * @port: port number
+ * @out_sid: security identifier
+ */
+int security_ib_port_sid(const char *dev_name, u8 port, u32 *out_sid)
+{
+ struct ocontext *c;
+ int rc = 0;
+
+ read_lock(&policy_rwlock);
+
+ c = policydb.ocontexts[OCON_IB_PORT];
+ while (c) {
+ if (c->u.ib_port.port == port &&
+ !strncmp(c->u.ib_port.dev_name,
+ dev_name,
+ IB_DEVICE_NAME_MAX))
+ break;
+
+ c = c->next;
+ }
+
+ if (c) {
+ if (!c->sid[0]) {
+ rc = sidtab_context_to_sid(&sidtab,
+ &c->context[0],
+ &c->sid[0]);
+ if (rc)
+ goto out;
+ }
+ *out_sid = c->sid[0];
+ } else {
+ *out_sid = SECINITSID_UNLABELED;
+ }
+
+out:
+ read_unlock(&policy_rwlock);
+ return rc;
+}
+
+/**
* security_netif_sid - Obtain the SID for a network interface.
* @name: interface name
* @if_sid: interface SID
--
1.7.1
^ permalink raw reply related [flat|nested] 42+ messages in thread
* [PATCH v2 9/9] selinux: Add a cache for quicker retreival of PKey SIDs
2016-07-14 22:56 ` Dan Jurgens
@ 2016-07-14 22:56 ` Dan Jurgens
-1 siblings, 0 replies; 42+ messages in thread
From: Dan Jurgens @ 2016-07-14 22:56 UTC (permalink / raw)
To: chrisw-69jw2NvuJkxg9hUCZPvPmw, paul-r2n+y4ga6xFZroRs9YW3xA,
sds-+05T5uksL2qpZYMLLGbcSA, eparis-FjpueFixGhCM4zKIHC2jIg,
dledford-H+wXaHxf7aLQT0dZR+AlfA,
sean.hefty-ral2JQCrhuEAvxtiuMwx3w,
hal.rosenstock-Re5JQEeQqe8AvxtiuMwx3w
Cc: selinux-+05T5uksL2qpZYMLLGbcSA,
linux-security-module-u79uwXL29TY76Z2rM5mHXA,
linux-rdma-u79uwXL29TY76Z2rM5mHXA,
yevgenyp-VPRAkNaXOzVWk0Htik3J/w, Daniel Jurgens
From: Daniel Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
It is likely that the SID for the same PKey will be requested many
times. To reduce the time to modify QPs and process MADs use a cache to
store PKey SIDs.
This code is heavily based on the "netif" and "netport" concept
originally developed by James Morris <jmorris-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> and Paul Moore
<paul-r2n+y4ga6xFZroRs9YW3xA@public.gmane.org> (see security/selinux/netif.c and
security/selinux/netport.c for more information)
Signed-off-by: Daniel Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---
v2:
- Renamed the files to ibpkey. Paul Moore
- Fixed a braket indentation mismatch in sel_pkey_find. Yuval Shaia
- Change spin_lock_bh to spin_lock_irqsave to resolve HARDIRQ lockdep
warning. Dan Jurgens
---
security/selinux/Makefile | 2 +-
security/selinux/hooks.c | 5 +-
security/selinux/ibpkey.c | 245 +++++++++++++++++++++++++++++++++++++
security/selinux/include/ibpkey.h | 31 +++++
security/selinux/include/objsec.h | 6 +
5 files changed, 287 insertions(+), 2 deletions(-)
create mode 100644 security/selinux/ibpkey.c
create mode 100644 security/selinux/include/ibpkey.h
diff --git a/security/selinux/Makefile b/security/selinux/Makefile
index 3411c33..ff5895e 100644
--- a/security/selinux/Makefile
+++ b/security/selinux/Makefile
@@ -5,7 +5,7 @@
obj-$(CONFIG_SECURITY_SELINUX) := selinux.o
selinux-y := avc.o hooks.o selinuxfs.o netlink.o nlmsgtab.o netif.o \
- netnode.o netport.o exports.o \
+ netnode.o netport.o ibpkey.o exports.o \
ss/ebitmap.o ss/hashtab.o ss/symtab.o ss/sidtab.o ss/avtab.o \
ss/policydb.o ss/services.o ss/conditional.o ss/mls.o ss/status.o
diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c
index 40ad0e8..3423c0c 100644
--- a/security/selinux/hooks.c
+++ b/security/selinux/hooks.c
@@ -90,6 +90,7 @@
#include "netif.h"
#include "netnode.h"
#include "netport.h"
+#include "ibpkey.h"
#include "xfrm.h"
#include "netlabel.h"
#include "audit.h"
@@ -167,6 +168,8 @@ static int selinux_cache_avc_callback(u32 event)
sel_netnode_flush();
sel_netport_flush();
synchronize_net();
+
+ sel_pkey_flush();
security_lsm_notify(LSM_POLICY_CHANGE, NULL);
}
return 0;
@@ -6003,7 +6006,7 @@ static int selinux_ib_pkey_access(u64 subnet_prefix, u16 pkey_val, void *ib_sec)
struct ib_security_struct *sec = ib_sec;
struct lsm_pkey_audit pkey;
- err = security_pkey_sid(subnet_prefix, pkey_val, &sid);
+ err = sel_pkey_sid(subnet_prefix, pkey_val, &sid);
if (err)
return err;
diff --git a/security/selinux/ibpkey.c b/security/selinux/ibpkey.c
new file mode 100644
index 0000000..6e52c54
--- /dev/null
+++ b/security/selinux/ibpkey.c
@@ -0,0 +1,245 @@
+/*
+ * Pkey table
+ *
+ * SELinux must keep a mapping of Infinband PKEYs to labels/SIDs. This
+ * mapping is maintained as part of the normal policy but a fast cache is
+ * needed to reduce the lookup overhead.
+ *
+ * This code is heavily based on the "netif" and "netport" concept originally
+ * developed by
+ * James Morris <jmorris-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> and
+ * Paul Moore <paul-r2n+y4ga6xFZroRs9YW3xA@public.gmane.org>
+ * (see security/selinux/netif.c and security/selinux/netport.c for more
+ * information)
+ *
+ */
+
+/*
+ * (c) Mellanox Technologies, 2016
+ *
+ * This program is free software: you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ */
+
+#include <linux/types.h>
+#include <linux/rcupdate.h>
+#include <linux/list.h>
+#include <linux/spinlock.h>
+
+#include "ibpkey.h"
+#include "objsec.h"
+
+#define SEL_PKEY_HASH_SIZE 256
+#define SEL_PKEY_HASH_BKT_LIMIT 16
+
+struct sel_pkey_bkt {
+ int size;
+ struct list_head list;
+};
+
+struct sel_pkey {
+ struct pkey_security_struct psec;
+ struct list_head list;
+ struct rcu_head rcu;
+};
+
+static LIST_HEAD(sel_pkey_list);
+static DEFINE_SPINLOCK(sel_pkey_lock);
+static struct sel_pkey_bkt sel_pkey_hash[SEL_PKEY_HASH_SIZE];
+
+/**
+ * sel_pkey_hashfn - Hashing function for the pkey table
+ * @pkey: pkey number
+ *
+ * Description:
+ * This is the hashing function for the pkey table, it returns the bucket
+ * number for the given pkey.
+ *
+ */
+static unsigned int sel_pkey_hashfn(u16 pkey)
+{
+ return (pkey & (SEL_PKEY_HASH_SIZE - 1));
+}
+
+/**
+ * sel_pkey_find - Search for a pkey record
+ * @subnet_prefix: subnet_prefix
+ * @pkey_num: pkey_num
+ *
+ * Description:
+ * Search the pkey table and return the matching record. If an entry
+ * can not be found in the table return NULL.
+ *
+ */
+static struct sel_pkey *sel_pkey_find(u64 subnet_prefix, u16 pkey_num)
+{
+ unsigned int idx;
+ struct sel_pkey *pkey;
+
+ idx = sel_pkey_hashfn(pkey_num);
+ list_for_each_entry_rcu(pkey, &sel_pkey_hash[idx].list, list) {
+ if (pkey->psec.pkey == pkey_num &&
+ pkey->psec.subnet_prefix == subnet_prefix)
+ return pkey;
+ }
+
+ return NULL;
+}
+
+/**
+ * sel_pkey_insert - Insert a new pkey into the table
+ * @pkey: the new pkey record
+ *
+ * Description:
+ * Add a new pkey record to the hash table.
+ *
+ */
+static void sel_pkey_insert(struct sel_pkey *pkey)
+{
+ unsigned int idx;
+
+ /* we need to impose a limit on the growth of the hash table so check
+ * this bucket to make sure it is within the specified bounds
+ */
+ idx = sel_pkey_hashfn(pkey->psec.pkey);
+ list_add_rcu(&pkey->list, &sel_pkey_hash[idx].list);
+ if (sel_pkey_hash[idx].size == SEL_PKEY_HASH_BKT_LIMIT) {
+ struct sel_pkey *tail;
+
+ tail = list_entry(
+ rcu_dereference_protected(
+ sel_pkey_hash[idx].list.prev,
+ lockdep_is_held(&sel_pkey_lock)),
+ struct sel_pkey, list);
+ list_del_rcu(&tail->list);
+ kfree_rcu(tail, rcu);
+ } else {
+ sel_pkey_hash[idx].size++;
+ }
+}
+
+/**
+ * sel_pkey_sid_slow - Lookup the SID of a pkey using the policy
+ * @subnet_prefix: subnet prefix
+ * @pkey_num: pkey number
+ * @sid: pkey SID
+ *
+ * Description:
+ * This function determines the SID of a pkey by querying the security
+ * policy. The result is added to the pkey table to speedup future
+ * queries. Returns zero on success, negative values on failure.
+ *
+ */
+static int sel_pkey_sid_slow(u64 subnet_prefix, u16 pkey_num, u32 *sid)
+{
+ int ret = -ENOMEM;
+ struct sel_pkey *pkey;
+ struct sel_pkey *new = NULL;
+ unsigned long flags;
+
+ spin_lock_irqsave(&sel_pkey_lock, flags);
+ pkey = sel_pkey_find(subnet_prefix, pkey_num);
+ if (pkey) {
+ *sid = pkey->psec.sid;
+ spin_unlock_irqrestore(&sel_pkey_lock, flags);
+ return 0;
+ }
+
+ ret = security_pkey_sid(subnet_prefix, pkey_num, sid);
+ if (ret != 0)
+ goto out;
+
+ new = kzalloc(sizeof(*new), GFP_ATOMIC);
+ if (!new)
+ goto out;
+
+ new->psec.subnet_prefix = subnet_prefix;
+ new->psec.pkey = pkey_num;
+ new->psec.sid = *sid;
+ sel_pkey_insert(new);
+
+out:
+ spin_unlock_irqrestore(&sel_pkey_lock, flags);
+ if (unlikely(ret))
+ kfree(new);
+
+ return ret;
+}
+
+/**
+ * sel_pkey_sid - Lookup the SID of a PKEY
+ * @subnet_prefix: subnet_prefix
+ * @pkey_num: pkey number
+ * @sid: pkey SID
+ *
+ * Description:
+ * This function determines the SID of a PKEY using the fastest method
+ * possible. First the pkey table is queried, but if an entry can't be found
+ * then the policy is queried and the result is added to the table to speedup
+ * future queries. Returns zero on success, negative values on failure.
+ *
+ */
+int sel_pkey_sid(u64 subnet_prefix, u16 pkey_num, u32 *sid)
+{
+ struct sel_pkey *pkey;
+
+ rcu_read_lock();
+ pkey = sel_pkey_find(subnet_prefix, pkey_num);
+ if (pkey) {
+ *sid = pkey->psec.sid;
+ rcu_read_unlock();
+ return 0;
+ }
+ rcu_read_unlock();
+
+ return sel_pkey_sid_slow(subnet_prefix, pkey_num, sid);
+}
+
+/**
+ * sel_pkey_flush - Flush the entire pkey table
+ *
+ * Description:
+ * Remove all entries from the pkey table
+ *
+ */
+void sel_pkey_flush(void)
+{
+ unsigned int idx;
+ struct sel_pkey *pkey, *pkey_tmp;
+ unsigned long flags;
+
+ spin_lock_irqsave(&sel_pkey_lock, flags);
+ for (idx = 0; idx < SEL_PKEY_HASH_SIZE; idx++) {
+ list_for_each_entry_safe(pkey, pkey_tmp,
+ &sel_pkey_hash[idx].list, list) {
+ list_del_rcu(&pkey->list);
+ kfree_rcu(pkey, rcu);
+ }
+ sel_pkey_hash[idx].size = 0;
+ }
+ spin_unlock_irqrestore(&sel_pkey_lock, flags);
+}
+
+static __init int sel_pkey_init(void)
+{
+ int iter;
+
+ if (!selinux_enabled)
+ return 0;
+
+ for (iter = 0; iter < SEL_PKEY_HASH_SIZE; iter++) {
+ INIT_LIST_HEAD(&sel_pkey_hash[iter].list);
+ sel_pkey_hash[iter].size = 0;
+ }
+
+ return 0;
+}
+
+subsys_initcall(sel_pkey_init);
diff --git a/security/selinux/include/ibpkey.h b/security/selinux/include/ibpkey.h
new file mode 100644
index 0000000..387885a
--- /dev/null
+++ b/security/selinux/include/ibpkey.h
@@ -0,0 +1,31 @@
+/*
+ * pkey table
+ *
+ * SELinux must keep a mapping of pkeys to labels/SIDs. This
+ * mapping is maintained as part of the normal policy but a fast cache is
+ * needed to reduce the lookup overhead.
+ *
+ */
+
+/*
+ * (c) Mellanox Technologies, 2016
+ *
+ * This program is free software: you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ */
+
+#ifndef _SELINUX_IB_PKEY_H
+#define _SELINUX_IB_PKEY_H
+
+void sel_pkey_flush(void);
+
+int sel_pkey_sid(u64 subnet_prefix, u16 pkey, u32 *sid);
+
+#endif
diff --git a/security/selinux/include/objsec.h b/security/selinux/include/objsec.h
index 8e7db43..4139f28 100644
--- a/security/selinux/include/objsec.h
+++ b/security/selinux/include/objsec.h
@@ -133,6 +133,12 @@ struct ib_security_struct {
u32 sid; /* SID of the queue pair or MAD agent */
};
+struct pkey_security_struct {
+ u64 subnet_prefix; /* Port subnet prefix */
+ u16 pkey; /* PKey number */
+ u32 sid; /* SID of pkey */
+};
+
extern unsigned int selinux_checkreqprot;
#endif /* _SELINUX_OBJSEC_H_ */
--
1.7.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related [flat|nested] 42+ messages in thread
* [PATCH v2 9/9] selinux: Add a cache for quicker retreival of PKey SIDs
@ 2016-07-14 22:56 ` Dan Jurgens
0 siblings, 0 replies; 42+ messages in thread
From: Dan Jurgens @ 2016-07-14 22:56 UTC (permalink / raw)
To: chrisw, paul, sds, eparis, dledford, sean.hefty, hal.rosenstock
Cc: selinux, linux-security-module, linux-rdma, yevgenyp, Daniel Jurgens
From: Daniel Jurgens <danielj@mellanox.com>
It is likely that the SID for the same PKey will be requested many
times. To reduce the time to modify QPs and process MADs use a cache to
store PKey SIDs.
This code is heavily based on the "netif" and "netport" concept
originally developed by James Morris <jmorris@redhat.com> and Paul Moore
<paul@paul-moore.com> (see security/selinux/netif.c and
security/selinux/netport.c for more information)
Signed-off-by: Daniel Jurgens <danielj@mellanox.com>
---
v2:
- Renamed the files to ibpkey. Paul Moore
- Fixed a braket indentation mismatch in sel_pkey_find. Yuval Shaia
- Change spin_lock_bh to spin_lock_irqsave to resolve HARDIRQ lockdep
warning. Dan Jurgens
---
security/selinux/Makefile | 2 +-
security/selinux/hooks.c | 5 +-
security/selinux/ibpkey.c | 245 +++++++++++++++++++++++++++++++++++++
security/selinux/include/ibpkey.h | 31 +++++
security/selinux/include/objsec.h | 6 +
5 files changed, 287 insertions(+), 2 deletions(-)
create mode 100644 security/selinux/ibpkey.c
create mode 100644 security/selinux/include/ibpkey.h
diff --git a/security/selinux/Makefile b/security/selinux/Makefile
index 3411c33..ff5895e 100644
--- a/security/selinux/Makefile
+++ b/security/selinux/Makefile
@@ -5,7 +5,7 @@
obj-$(CONFIG_SECURITY_SELINUX) := selinux.o
selinux-y := avc.o hooks.o selinuxfs.o netlink.o nlmsgtab.o netif.o \
- netnode.o netport.o exports.o \
+ netnode.o netport.o ibpkey.o exports.o \
ss/ebitmap.o ss/hashtab.o ss/symtab.o ss/sidtab.o ss/avtab.o \
ss/policydb.o ss/services.o ss/conditional.o ss/mls.o ss/status.o
diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c
index 40ad0e8..3423c0c 100644
--- a/security/selinux/hooks.c
+++ b/security/selinux/hooks.c
@@ -90,6 +90,7 @@
#include "netif.h"
#include "netnode.h"
#include "netport.h"
+#include "ibpkey.h"
#include "xfrm.h"
#include "netlabel.h"
#include "audit.h"
@@ -167,6 +168,8 @@ static int selinux_cache_avc_callback(u32 event)
sel_netnode_flush();
sel_netport_flush();
synchronize_net();
+
+ sel_pkey_flush();
security_lsm_notify(LSM_POLICY_CHANGE, NULL);
}
return 0;
@@ -6003,7 +6006,7 @@ static int selinux_ib_pkey_access(u64 subnet_prefix, u16 pkey_val, void *ib_sec)
struct ib_security_struct *sec = ib_sec;
struct lsm_pkey_audit pkey;
- err = security_pkey_sid(subnet_prefix, pkey_val, &sid);
+ err = sel_pkey_sid(subnet_prefix, pkey_val, &sid);
if (err)
return err;
diff --git a/security/selinux/ibpkey.c b/security/selinux/ibpkey.c
new file mode 100644
index 0000000..6e52c54
--- /dev/null
+++ b/security/selinux/ibpkey.c
@@ -0,0 +1,245 @@
+/*
+ * Pkey table
+ *
+ * SELinux must keep a mapping of Infinband PKEYs to labels/SIDs. This
+ * mapping is maintained as part of the normal policy but a fast cache is
+ * needed to reduce the lookup overhead.
+ *
+ * This code is heavily based on the "netif" and "netport" concept originally
+ * developed by
+ * James Morris <jmorris@redhat.com> and
+ * Paul Moore <paul@paul-moore.com>
+ * (see security/selinux/netif.c and security/selinux/netport.c for more
+ * information)
+ *
+ */
+
+/*
+ * (c) Mellanox Technologies, 2016
+ *
+ * This program is free software: you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ */
+
+#include <linux/types.h>
+#include <linux/rcupdate.h>
+#include <linux/list.h>
+#include <linux/spinlock.h>
+
+#include "ibpkey.h"
+#include "objsec.h"
+
+#define SEL_PKEY_HASH_SIZE 256
+#define SEL_PKEY_HASH_BKT_LIMIT 16
+
+struct sel_pkey_bkt {
+ int size;
+ struct list_head list;
+};
+
+struct sel_pkey {
+ struct pkey_security_struct psec;
+ struct list_head list;
+ struct rcu_head rcu;
+};
+
+static LIST_HEAD(sel_pkey_list);
+static DEFINE_SPINLOCK(sel_pkey_lock);
+static struct sel_pkey_bkt sel_pkey_hash[SEL_PKEY_HASH_SIZE];
+
+/**
+ * sel_pkey_hashfn - Hashing function for the pkey table
+ * @pkey: pkey number
+ *
+ * Description:
+ * This is the hashing function for the pkey table, it returns the bucket
+ * number for the given pkey.
+ *
+ */
+static unsigned int sel_pkey_hashfn(u16 pkey)
+{
+ return (pkey & (SEL_PKEY_HASH_SIZE - 1));
+}
+
+/**
+ * sel_pkey_find - Search for a pkey record
+ * @subnet_prefix: subnet_prefix
+ * @pkey_num: pkey_num
+ *
+ * Description:
+ * Search the pkey table and return the matching record. If an entry
+ * can not be found in the table return NULL.
+ *
+ */
+static struct sel_pkey *sel_pkey_find(u64 subnet_prefix, u16 pkey_num)
+{
+ unsigned int idx;
+ struct sel_pkey *pkey;
+
+ idx = sel_pkey_hashfn(pkey_num);
+ list_for_each_entry_rcu(pkey, &sel_pkey_hash[idx].list, list) {
+ if (pkey->psec.pkey == pkey_num &&
+ pkey->psec.subnet_prefix == subnet_prefix)
+ return pkey;
+ }
+
+ return NULL;
+}
+
+/**
+ * sel_pkey_insert - Insert a new pkey into the table
+ * @pkey: the new pkey record
+ *
+ * Description:
+ * Add a new pkey record to the hash table.
+ *
+ */
+static void sel_pkey_insert(struct sel_pkey *pkey)
+{
+ unsigned int idx;
+
+ /* we need to impose a limit on the growth of the hash table so check
+ * this bucket to make sure it is within the specified bounds
+ */
+ idx = sel_pkey_hashfn(pkey->psec.pkey);
+ list_add_rcu(&pkey->list, &sel_pkey_hash[idx].list);
+ if (sel_pkey_hash[idx].size == SEL_PKEY_HASH_BKT_LIMIT) {
+ struct sel_pkey *tail;
+
+ tail = list_entry(
+ rcu_dereference_protected(
+ sel_pkey_hash[idx].list.prev,
+ lockdep_is_held(&sel_pkey_lock)),
+ struct sel_pkey, list);
+ list_del_rcu(&tail->list);
+ kfree_rcu(tail, rcu);
+ } else {
+ sel_pkey_hash[idx].size++;
+ }
+}
+
+/**
+ * sel_pkey_sid_slow - Lookup the SID of a pkey using the policy
+ * @subnet_prefix: subnet prefix
+ * @pkey_num: pkey number
+ * @sid: pkey SID
+ *
+ * Description:
+ * This function determines the SID of a pkey by querying the security
+ * policy. The result is added to the pkey table to speedup future
+ * queries. Returns zero on success, negative values on failure.
+ *
+ */
+static int sel_pkey_sid_slow(u64 subnet_prefix, u16 pkey_num, u32 *sid)
+{
+ int ret = -ENOMEM;
+ struct sel_pkey *pkey;
+ struct sel_pkey *new = NULL;
+ unsigned long flags;
+
+ spin_lock_irqsave(&sel_pkey_lock, flags);
+ pkey = sel_pkey_find(subnet_prefix, pkey_num);
+ if (pkey) {
+ *sid = pkey->psec.sid;
+ spin_unlock_irqrestore(&sel_pkey_lock, flags);
+ return 0;
+ }
+
+ ret = security_pkey_sid(subnet_prefix, pkey_num, sid);
+ if (ret != 0)
+ goto out;
+
+ new = kzalloc(sizeof(*new), GFP_ATOMIC);
+ if (!new)
+ goto out;
+
+ new->psec.subnet_prefix = subnet_prefix;
+ new->psec.pkey = pkey_num;
+ new->psec.sid = *sid;
+ sel_pkey_insert(new);
+
+out:
+ spin_unlock_irqrestore(&sel_pkey_lock, flags);
+ if (unlikely(ret))
+ kfree(new);
+
+ return ret;
+}
+
+/**
+ * sel_pkey_sid - Lookup the SID of a PKEY
+ * @subnet_prefix: subnet_prefix
+ * @pkey_num: pkey number
+ * @sid: pkey SID
+ *
+ * Description:
+ * This function determines the SID of a PKEY using the fastest method
+ * possible. First the pkey table is queried, but if an entry can't be found
+ * then the policy is queried and the result is added to the table to speedup
+ * future queries. Returns zero on success, negative values on failure.
+ *
+ */
+int sel_pkey_sid(u64 subnet_prefix, u16 pkey_num, u32 *sid)
+{
+ struct sel_pkey *pkey;
+
+ rcu_read_lock();
+ pkey = sel_pkey_find(subnet_prefix, pkey_num);
+ if (pkey) {
+ *sid = pkey->psec.sid;
+ rcu_read_unlock();
+ return 0;
+ }
+ rcu_read_unlock();
+
+ return sel_pkey_sid_slow(subnet_prefix, pkey_num, sid);
+}
+
+/**
+ * sel_pkey_flush - Flush the entire pkey table
+ *
+ * Description:
+ * Remove all entries from the pkey table
+ *
+ */
+void sel_pkey_flush(void)
+{
+ unsigned int idx;
+ struct sel_pkey *pkey, *pkey_tmp;
+ unsigned long flags;
+
+ spin_lock_irqsave(&sel_pkey_lock, flags);
+ for (idx = 0; idx < SEL_PKEY_HASH_SIZE; idx++) {
+ list_for_each_entry_safe(pkey, pkey_tmp,
+ &sel_pkey_hash[idx].list, list) {
+ list_del_rcu(&pkey->list);
+ kfree_rcu(pkey, rcu);
+ }
+ sel_pkey_hash[idx].size = 0;
+ }
+ spin_unlock_irqrestore(&sel_pkey_lock, flags);
+}
+
+static __init int sel_pkey_init(void)
+{
+ int iter;
+
+ if (!selinux_enabled)
+ return 0;
+
+ for (iter = 0; iter < SEL_PKEY_HASH_SIZE; iter++) {
+ INIT_LIST_HEAD(&sel_pkey_hash[iter].list);
+ sel_pkey_hash[iter].size = 0;
+ }
+
+ return 0;
+}
+
+subsys_initcall(sel_pkey_init);
diff --git a/security/selinux/include/ibpkey.h b/security/selinux/include/ibpkey.h
new file mode 100644
index 0000000..387885a
--- /dev/null
+++ b/security/selinux/include/ibpkey.h
@@ -0,0 +1,31 @@
+/*
+ * pkey table
+ *
+ * SELinux must keep a mapping of pkeys to labels/SIDs. This
+ * mapping is maintained as part of the normal policy but a fast cache is
+ * needed to reduce the lookup overhead.
+ *
+ */
+
+/*
+ * (c) Mellanox Technologies, 2016
+ *
+ * This program is free software: you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ */
+
+#ifndef _SELINUX_IB_PKEY_H
+#define _SELINUX_IB_PKEY_H
+
+void sel_pkey_flush(void);
+
+int sel_pkey_sid(u64 subnet_prefix, u16 pkey, u32 *sid);
+
+#endif
diff --git a/security/selinux/include/objsec.h b/security/selinux/include/objsec.h
index 8e7db43..4139f28 100644
--- a/security/selinux/include/objsec.h
+++ b/security/selinux/include/objsec.h
@@ -133,6 +133,12 @@ struct ib_security_struct {
u32 sid; /* SID of the queue pair or MAD agent */
};
+struct pkey_security_struct {
+ u64 subnet_prefix; /* Port subnet prefix */
+ u16 pkey; /* PKey number */
+ u32 sid; /* SID of pkey */
+};
+
extern unsigned int selinux_checkreqprot;
#endif /* _SELINUX_OBJSEC_H_ */
--
1.7.1
^ permalink raw reply related [flat|nested] 42+ messages in thread
* Re: [PATCH v2 9/9] selinux: Add a cache for quicker retreival of PKey SIDs
2016-07-14 22:56 ` Dan Jurgens
@ 2016-07-15 9:06 ` kbuild test robot
-1 siblings, 0 replies; 42+ messages in thread
From: kbuild test robot @ 2016-07-15 9:06 UTC (permalink / raw)
Cc: kbuild-all-JC7UmRfGjtg, chrisw-69jw2NvuJkxg9hUCZPvPmw,
paul-r2n+y4ga6xFZroRs9YW3xA, sds-+05T5uksL2qpZYMLLGbcSA,
eparis-FjpueFixGhCM4zKIHC2jIg, dledford-H+wXaHxf7aLQT0dZR+AlfA,
sean.hefty-ral2JQCrhuEAvxtiuMwx3w,
hal.rosenstock-Re5JQEeQqe8AvxtiuMwx3w,
selinux-+05T5uksL2qpZYMLLGbcSA,
linux-security-module-u79uwXL29TY76Z2rM5mHXA,
linux-rdma-u79uwXL29TY76Z2rM5mHXA,
yevgenyp-VPRAkNaXOzVWk0Htik3J/w, Daniel Jurgens
Hi,
[auto build test WARNING on next-20160713]
[cannot apply to pcmoore-selinux/next rdma/master v4.7-rc7 v4.7-rc6 v4.7-rc5 v4.7-rc7]
[if your patch is applied to the wrong git tree, please drop us a note to help improve the system]
url: https://github.com/0day-ci/linux/commits/Dan-Jurgens/SELinux-support-for-Infiniband-RDMA/20160715-122805
reproduce:
# apt-get install sparse
make ARCH=x86_64 allmodconfig
make C=1 CF=-D__CHECK_ENDIAN__
sparse warnings: (new ones prefixed by >>)
include/linux/compiler.h:232:8: sparse: attribute 'no_sanitize_address': unknown attribute
>> security/selinux/ibpkey.c:116:24: sparse: incompatible types in comparison expression (different address spaces)
vim +116 security/selinux/ibpkey.c
100 * Description:
101 * Add a new pkey record to the hash table.
102 *
103 */
104 static void sel_pkey_insert(struct sel_pkey *pkey)
105 {
106 unsigned int idx;
107
108 /* we need to impose a limit on the growth of the hash table so check
109 * this bucket to make sure it is within the specified bounds
110 */
111 idx = sel_pkey_hashfn(pkey->psec.pkey);
112 list_add_rcu(&pkey->list, &sel_pkey_hash[idx].list);
113 if (sel_pkey_hash[idx].size == SEL_PKEY_HASH_BKT_LIMIT) {
114 struct sel_pkey *tail;
115
> 116 tail = list_entry(
117 rcu_dereference_protected(
118 sel_pkey_hash[idx].list.prev,
119 lockdep_is_held(&sel_pkey_lock)),
120 struct sel_pkey, list);
121 list_del_rcu(&tail->list);
122 kfree_rcu(tail, rcu);
123 } else {
124 sel_pkey_hash[idx].size++;
---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 42+ messages in thread
* Re: [PATCH v2 9/9] selinux: Add a cache for quicker retreival of PKey SIDs
@ 2016-07-15 9:06 ` kbuild test robot
0 siblings, 0 replies; 42+ messages in thread
From: kbuild test robot @ 2016-07-15 9:06 UTC (permalink / raw)
To: Dan Jurgens
Cc: kbuild-all, chrisw, paul, sds, eparis, dledford, sean.hefty,
hal.rosenstock, selinux, linux-security-module, linux-rdma,
yevgenyp, Daniel Jurgens
Hi,
[auto build test WARNING on next-20160713]
[cannot apply to pcmoore-selinux/next rdma/master v4.7-rc7 v4.7-rc6 v4.7-rc5 v4.7-rc7]
[if your patch is applied to the wrong git tree, please drop us a note to help improve the system]
url: https://github.com/0day-ci/linux/commits/Dan-Jurgens/SELinux-support-for-Infiniband-RDMA/20160715-122805
reproduce:
# apt-get install sparse
make ARCH=x86_64 allmodconfig
make C=1 CF=-D__CHECK_ENDIAN__
sparse warnings: (new ones prefixed by >>)
include/linux/compiler.h:232:8: sparse: attribute 'no_sanitize_address': unknown attribute
>> security/selinux/ibpkey.c:116:24: sparse: incompatible types in comparison expression (different address spaces)
vim +116 security/selinux/ibpkey.c
100 * Description:
101 * Add a new pkey record to the hash table.
102 *
103 */
104 static void sel_pkey_insert(struct sel_pkey *pkey)
105 {
106 unsigned int idx;
107
108 /* we need to impose a limit on the growth of the hash table so check
109 * this bucket to make sure it is within the specified bounds
110 */
111 idx = sel_pkey_hashfn(pkey->psec.pkey);
112 list_add_rcu(&pkey->list, &sel_pkey_hash[idx].list);
113 if (sel_pkey_hash[idx].size == SEL_PKEY_HASH_BKT_LIMIT) {
114 struct sel_pkey *tail;
115
> 116 tail = list_entry(
117 rcu_dereference_protected(
118 sel_pkey_hash[idx].list.prev,
119 lockdep_is_held(&sel_pkey_lock)),
120 struct sel_pkey, list);
121 list_del_rcu(&tail->list);
122 kfree_rcu(tail, rcu);
123 } else {
124 sel_pkey_hash[idx].size++;
---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation
^ permalink raw reply [flat|nested] 42+ messages in thread
* Re: [PATCH v2 3/9] selinux lsm IB/core: Implement LSM notification system
2016-07-14 22:56 ` Dan Jurgens
@ 2016-07-15 10:54 ` kbuild test robot
-1 siblings, 0 replies; 42+ messages in thread
From: kbuild test robot @ 2016-07-15 10:54 UTC (permalink / raw)
Cc: kbuild-all-JC7UmRfGjtg, chrisw-69jw2NvuJkxg9hUCZPvPmw,
paul-r2n+y4ga6xFZroRs9YW3xA, sds-+05T5uksL2qpZYMLLGbcSA,
eparis-FjpueFixGhCM4zKIHC2jIg, dledford-H+wXaHxf7aLQT0dZR+AlfA,
sean.hefty-ral2JQCrhuEAvxtiuMwx3w,
hal.rosenstock-Re5JQEeQqe8AvxtiuMwx3w,
selinux-+05T5uksL2qpZYMLLGbcSA,
linux-security-module-u79uwXL29TY76Z2rM5mHXA,
linux-rdma-u79uwXL29TY76Z2rM5mHXA,
yevgenyp-VPRAkNaXOzVWk0Htik3J/w, Daniel Jurgens
[-- Attachment #1: Type: text/plain, Size: 1062 bytes --]
Hi,
[auto build test ERROR on next-20160713]
[cannot apply to pcmoore-selinux/next rdma/master v4.7-rc7 v4.7-rc6 v4.7-rc5 v4.7-rc7]
[if your patch is applied to the wrong git tree, please drop us a note to help improve the system]
url: https://github.com/0day-ci/linux/commits/Dan-Jurgens/SELinux-support-for-Infiniband-RDMA/20160715-122805
config: ia64-defconfig (attached as .config)
compiler: ia64-linux-gcc (GCC) 4.9.0
reproduce:
wget https://git.kernel.org/cgit/linux/kernel/git/wfg/lkp-tests.git/plain/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# save the attached .config to linux build tree
make.cross ARCH=ia64
All errors (new ones prefixed by >>):
>> ERROR: "security_register_lsm_notifier" [drivers/infiniband/core/ib_core.ko] undefined!
>> ERROR: "security_unregister_lsm_notifier" [drivers/infiniband/core/ib_core.ko] undefined!
---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation
[-- Attachment #2: .config.gz --]
[-- Type: application/octet-stream, Size: 18326 bytes --]
^ permalink raw reply [flat|nested] 42+ messages in thread
* Re: [PATCH v2 3/9] selinux lsm IB/core: Implement LSM notification system
@ 2016-07-15 10:54 ` kbuild test robot
0 siblings, 0 replies; 42+ messages in thread
From: kbuild test robot @ 2016-07-15 10:54 UTC (permalink / raw)
To: Dan Jurgens
Cc: kbuild-all, chrisw, paul, sds, eparis, dledford, sean.hefty,
hal.rosenstock, selinux, linux-security-module, linux-rdma,
yevgenyp, Daniel Jurgens
[-- Attachment #1: Type: text/plain, Size: 1062 bytes --]
Hi,
[auto build test ERROR on next-20160713]
[cannot apply to pcmoore-selinux/next rdma/master v4.7-rc7 v4.7-rc6 v4.7-rc5 v4.7-rc7]
[if your patch is applied to the wrong git tree, please drop us a note to help improve the system]
url: https://github.com/0day-ci/linux/commits/Dan-Jurgens/SELinux-support-for-Infiniband-RDMA/20160715-122805
config: ia64-defconfig (attached as .config)
compiler: ia64-linux-gcc (GCC) 4.9.0
reproduce:
wget https://git.kernel.org/cgit/linux/kernel/git/wfg/lkp-tests.git/plain/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# save the attached .config to linux build tree
make.cross ARCH=ia64
All errors (new ones prefixed by >>):
>> ERROR: "security_register_lsm_notifier" [drivers/infiniband/core/ib_core.ko] undefined!
>> ERROR: "security_unregister_lsm_notifier" [drivers/infiniband/core/ib_core.ko] undefined!
---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation
[-- Attachment #2: .config.gz --]
[-- Type: application/octet-stream, Size: 18326 bytes --]
^ permalink raw reply [flat|nested] 42+ messages in thread
* Re: [PATCH v2 0/9] SELinux support for Infiniband RDMA
2016-07-14 22:56 ` Dan Jurgens
@ 2016-07-22 15:46 ` Paul Moore
-1 siblings, 0 replies; 42+ messages in thread
From: Paul Moore @ 2016-07-22 15:46 UTC (permalink / raw)
To: Dan Jurgens
Cc: chrisw-69jw2NvuJkxg9hUCZPvPmw, Stephen Smalley, Eric Paris,
dledford-H+wXaHxf7aLQT0dZR+AlfA,
sean.hefty-ral2JQCrhuEAvxtiuMwx3w,
hal.rosenstock-Re5JQEeQqe8AvxtiuMwx3w,
selinux-+05T5uksL2qpZYMLLGbcSA,
linux-security-module-u79uwXL29TY76Z2rM5mHXA,
linux-rdma-u79uwXL29TY76Z2rM5mHXA,
yevgenyp-VPRAkNaXOzVWk0Htik3J/w
On Thu, Jul 14, 2016 at 6:56 PM, Dan Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org> wrote:
> From: Daniel Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
>
> The selinux next tree is missing some patches for IB/core. This series
> applies cleanly to ib-next, and should apply cleanly to selinux-next once
> the IB patches are merged.
Hi Dan,
Thanks for the updated patchset, I'm taking a look at it today. I do
have one question, related but independent to your patches: is there a
way to test the LSM controlled portions of the Infiniband stack
without any IB hardware (e.g. is there an IB "loopback" device)? I'm
asking because we are really trying to make sure we have tests for any
new code/functionality we add to SELinux, and requiring IB hardware to
test the IB access controls would make this difficult.
* https://github.com/SELinuxProject/selinux-testsuite
-Paul
--
paul moore
www.paul-moore.com
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 42+ messages in thread
* Re: [PATCH v2 0/9] SELinux support for Infiniband RDMA
@ 2016-07-22 15:46 ` Paul Moore
0 siblings, 0 replies; 42+ messages in thread
From: Paul Moore @ 2016-07-22 15:46 UTC (permalink / raw)
To: Dan Jurgens
Cc: chrisw, Stephen Smalley, Eric Paris, dledford, sean.hefty,
hal.rosenstock, selinux, linux-security-module, linux-rdma,
yevgenyp
On Thu, Jul 14, 2016 at 6:56 PM, Dan Jurgens <danielj@mellanox.com> wrote:
> From: Daniel Jurgens <danielj@mellanox.com>
>
> The selinux next tree is missing some patches for IB/core. This series
> applies cleanly to ib-next, and should apply cleanly to selinux-next once
> the IB patches are merged.
Hi Dan,
Thanks for the updated patchset, I'm taking a look at it today. I do
have one question, related but independent to your patches: is there a
way to test the LSM controlled portions of the Infiniband stack
without any IB hardware (e.g. is there an IB "loopback" device)? I'm
asking because we are really trying to make sure we have tests for any
new code/functionality we add to SELinux, and requiring IB hardware to
test the IB access controls would make this difficult.
* https://github.com/SELinuxProject/selinux-testsuite
-Paul
--
paul moore
www.paul-moore.com
^ permalink raw reply [flat|nested] 42+ messages in thread
* Re: [PATCH v2 0/9] SELinux support for Infiniband RDMA
2016-07-22 15:46 ` Paul Moore
@ 2016-07-22 16:02 ` Daniel Jurgens
-1 siblings, 0 replies; 42+ messages in thread
From: Daniel Jurgens @ 2016-07-22 16:02 UTC (permalink / raw)
To: Paul Moore
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA,
linux-security-module-u79uwXL29TY76Z2rM5mHXA,
chrisw-69jw2NvuJkxg9hUCZPvPmw, dledford-H+wXaHxf7aLQT0dZR+AlfA,
selinux-+05T5uksL2qpZYMLLGbcSA,
sean.hefty-ral2JQCrhuEAvxtiuMwx3w, Stephen Smalley,
hal.rosenstock-Re5JQEeQqe8AvxtiuMwx3w
On 7/22/2016 10:46 AM, Paul Moore wrote:
> On Thu, Jul 14, 2016 at 6:56 PM, Dan Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org> wrote:
>> From: Daniel Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
>>
>> The selinux next tree is missing some patches for IB/core. This series
>> applies cleanly to ib-next, and should apply cleanly to selinux-next once
>> the IB patches are merged.
> Hi Dan,
>
> Thanks for the updated patchset, I'm taking a look at it today. I do
> have one question, related but independent to your patches: is there a
> way to test the LSM controlled portions of the Infiniband stack
> without any IB hardware (e.g. is there an IB "loopback" device)? I'm
> asking because we are really trying to make sure we have tests for any
> new code/functionality we add to SELinux, and requiring IB hardware to
> test the IB access controls would make this difficult.
>
> * https://github.com/SELinuxProject/selinux-testsuite
>
> -Paul
>
I found this: https://github.com/nminoru/pib but haven't used it. I'll try playing with it and see if it can serve this purpose.
_______________________________________________
Selinux mailing list
Selinux-+05T5uksL2qpZYMLLGbcSA@public.gmane.org
To unsubscribe, send email to Selinux-leave-+05T5uksL2pAGbPMOrvdOA@public.gmane.org
To get help, send an email containing "help" to Selinux-request-+05T5uksL2pAGbPMOrvdOA@public.gmane.org
^ permalink raw reply [flat|nested] 42+ messages in thread
* Re: [PATCH v2 0/9] SELinux support for Infiniband RDMA
@ 2016-07-22 16:02 ` Daniel Jurgens
0 siblings, 0 replies; 42+ messages in thread
From: Daniel Jurgens @ 2016-07-22 16:02 UTC (permalink / raw)
To: Paul Moore
Cc: chrisw, Stephen Smalley, Eric Paris, dledford, sean.hefty,
hal.rosenstock, selinux, linux-security-module, linux-rdma,
Yevgeny Petrilin
On 7/22/2016 10:46 AM, Paul Moore wrote:
> On Thu, Jul 14, 2016 at 6:56 PM, Dan Jurgens <danielj@mellanox.com> wrote:
>> From: Daniel Jurgens <danielj@mellanox.com>
>>
>> The selinux next tree is missing some patches for IB/core. This series
>> applies cleanly to ib-next, and should apply cleanly to selinux-next once
>> the IB patches are merged.
> Hi Dan,
>
> Thanks for the updated patchset, I'm taking a look at it today. I do
> have one question, related but independent to your patches: is there a
> way to test the LSM controlled portions of the Infiniband stack
> without any IB hardware (e.g. is there an IB "loopback" device)? I'm
> asking because we are really trying to make sure we have tests for any
> new code/functionality we add to SELinux, and requiring IB hardware to
> test the IB access controls would make this difficult.
>
> * https://github.com/SELinuxProject/selinux-testsuite
>
> -Paul
>
I found this: https://github.com/nminoru/pib but haven't used it. I'll try playing with it and see if it can serve this purpose.
^ permalink raw reply [flat|nested] 42+ messages in thread
* Re: [PATCH v2 3/9] selinux lsm IB/core: Implement LSM notification system
2016-07-14 22:56 ` Dan Jurgens
@ 2016-07-22 16:21 ` Paul Moore
-1 siblings, 0 replies; 42+ messages in thread
From: Paul Moore @ 2016-07-22 16:21 UTC (permalink / raw)
To: Dan Jurgens
Cc: chrisw-69jw2NvuJkxg9hUCZPvPmw, Stephen Smalley, Eric Paris,
dledford-H+wXaHxf7aLQT0dZR+AlfA,
sean.hefty-ral2JQCrhuEAvxtiuMwx3w,
hal.rosenstock-Re5JQEeQqe8AvxtiuMwx3w,
selinux-+05T5uksL2qpZYMLLGbcSA,
linux-security-module-u79uwXL29TY76Z2rM5mHXA,
linux-rdma-u79uwXL29TY76Z2rM5mHXA,
yevgenyp-VPRAkNaXOzVWk0Htik3J/w
On Thu, Jul 14, 2016 at 6:56 PM, Dan Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org> wrote:
> From: Daniel Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
>
> Add a generic notificaiton mechanism in the LSM. Interested consumers
> can register a callback with the LSM and security modules can produce
> events.
>
> Because access to Infiniband QPs are enforced in the setup phase of a
> connection security should be enforced again if the policy changes.
> Register infiniband devices for policy change notification and check all
> QPs on that device when the notification is received.
>
> Add a call to the notification mechanism is from SELinux when the AVC
> cache changes.
>
> Signed-off-by: Daniel Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
>
> ---
> v2:
> - new patch that has the generic notification, replaces selinux and
> IB/core patches related to the ib_flush callback. Yuval Shaia and Paul
> Moore
> ---
> drivers/infiniband/core/device.c | 46 ++++++++++++++++++++++++++++++
> include/linux/security.h | 10 ++++++
> security/security.c | 58 ++++++++++++++++++++++++++++++++++++++
> security/selinux/hooks.c | 5 ++-
> security/selinux/selinuxfs.c | 2 +
> 5 files changed, 119 insertions(+), 2 deletions(-)
>
Thanks for making this more generic. I've got some comments below,
but in the course of reviewing this I realized that the kernel has an
existing, general purpose notifier mechanism (see
include/linux/notifier.h); while I've seen the notifier code in the
network stack, I never realized it was created as a general purpose
mechanism. My apologies, I should have noticed this earlier and
mentioned it.
As far as how that impacts this patch; it seems like we should be good
citizens and use the general mechanism, I don't see any obvious
reasons why it wouldn't work. That said, I do feel bad for not
realizing this earlier. If it isn't too annoying, would you mind
updating this patch *again*?
> diff --git a/include/linux/security.h b/include/linux/security.h
> index 33e23c4..bf53911 100644
> --- a/include/linux/security.h
> +++ b/include/linux/security.h
> @@ -69,6 +69,16 @@ struct audit_krule;
> struct user_namespace;
> struct timezone;
>
> +enum lsm_event {
> + LSM_POLICY_CHANGE,
> +};
> +
> +typedef void (*lsm_notifier)(enum lsm_event event, void *ctx, void *data);
> +
> +void security_lsm_notify(enum lsm_event event, void *data);
> +int security_register_lsm_notifier(lsm_notifier func, void *ctx, u32 *id);
> +void security_unregister_lsm_notifier(u32 id);
> +
> /* These functions are in security/commoncap.c */
> extern int cap_capable(const struct cred *cred, struct user_namespace *ns,
> int cap, int audit);
> diff --git a/security/security.c b/security/security.c
> index 234982d..1263c1d 100644
> --- a/security/security.c
> +++ b/security/security.c
> @@ -33,6 +33,18 @@
> /* Maximum number of letters for an LSM name string */
> #define SECURITY_NAME_MAX 10
>
> +struct lsm_notifier_entry {
> + u32 callback_id;
> + lsm_notifier func;
> + void *ctx;
> + struct list_head list;
> + struct rcu_head rcu;
> +};
> +
> +static LIST_HEAD(lsm_notifier_list);
> +static DEFINE_SPINLOCK(lsm_notifier_lock);
> +static u32 next_callback_id;
I'd rather see the callback ID named something that makes it obvious
it is tied to the lsm_notifier_* bits, perhaps lsm_notifier_nextid?
> /* Boot-time LSM user choice */
> static __initdata char chosen_lsm[SECURITY_NAME_MAX + 1] =
> CONFIG_DEFAULT_SECURITY;
> @@ -98,6 +110,52 @@ int __init security_module_enable(const char *module)
> return !strcmp(module, chosen_lsm);
> }
>
> +void security_lsm_notify(enum lsm_event event, void *data)
> +{
> + struct lsm_notifier_entry *entry;
> +
> + rcu_read_lock();
> + list_for_each_entry_rcu(entry, &lsm_notifier_list, list)
> + entry->func(event, entry->ctx, data);
> + rcu_read_unlock();
> +}
> +EXPORT_SYMBOL(security_lsm_notify);
> +
> +int security_register_lsm_notifier(lsm_notifier func, void *ctx, u32 *id)
> +{
> + struct lsm_notifier_entry *entry;
> +
> + entry = kzalloc(sizeof(*entry), GFP_KERNEL);
> + if (!entry)
> + return -ENOMEM;
> +
> + entry->func = func;
> + entry->ctx = ctx;
Do we need ctx? We do need to ability to pass a blob to the callback
which will vary based on the event, but I'm not sure the ctx here has
much value.
> + spin_lock(&lsm_notifier_lock);
> + entry->callback_id = next_callback_id++;
> + *id = entry->callback_id;
> + list_add_rcu(&entry->list, &lsm_notifier_list);
> + spin_unlock(&lsm_notifier_lock);
> +
> + return 0;
> +}
> +EXPORT_SYMBOL(security_register_lsm_notifier);
--
paul moore
www.paul-moore.com
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 42+ messages in thread
* Re: [PATCH v2 3/9] selinux lsm IB/core: Implement LSM notification system
@ 2016-07-22 16:21 ` Paul Moore
0 siblings, 0 replies; 42+ messages in thread
From: Paul Moore @ 2016-07-22 16:21 UTC (permalink / raw)
To: Dan Jurgens
Cc: chrisw, Stephen Smalley, Eric Paris, dledford, sean.hefty,
hal.rosenstock, selinux, linux-security-module, linux-rdma,
yevgenyp
On Thu, Jul 14, 2016 at 6:56 PM, Dan Jurgens <danielj@mellanox.com> wrote:
> From: Daniel Jurgens <danielj@mellanox.com>
>
> Add a generic notificaiton mechanism in the LSM. Interested consumers
> can register a callback with the LSM and security modules can produce
> events.
>
> Because access to Infiniband QPs are enforced in the setup phase of a
> connection security should be enforced again if the policy changes.
> Register infiniband devices for policy change notification and check all
> QPs on that device when the notification is received.
>
> Add a call to the notification mechanism is from SELinux when the AVC
> cache changes.
>
> Signed-off-by: Daniel Jurgens <danielj@mellanox.com>
>
> ---
> v2:
> - new patch that has the generic notification, replaces selinux and
> IB/core patches related to the ib_flush callback. Yuval Shaia and Paul
> Moore
> ---
> drivers/infiniband/core/device.c | 46 ++++++++++++++++++++++++++++++
> include/linux/security.h | 10 ++++++
> security/security.c | 58 ++++++++++++++++++++++++++++++++++++++
> security/selinux/hooks.c | 5 ++-
> security/selinux/selinuxfs.c | 2 +
> 5 files changed, 119 insertions(+), 2 deletions(-)
>
Thanks for making this more generic. I've got some comments below,
but in the course of reviewing this I realized that the kernel has an
existing, general purpose notifier mechanism (see
include/linux/notifier.h); while I've seen the notifier code in the
network stack, I never realized it was created as a general purpose
mechanism. My apologies, I should have noticed this earlier and
mentioned it.
As far as how that impacts this patch; it seems like we should be good
citizens and use the general mechanism, I don't see any obvious
reasons why it wouldn't work. That said, I do feel bad for not
realizing this earlier. If it isn't too annoying, would you mind
updating this patch *again*?
> diff --git a/include/linux/security.h b/include/linux/security.h
> index 33e23c4..bf53911 100644
> --- a/include/linux/security.h
> +++ b/include/linux/security.h
> @@ -69,6 +69,16 @@ struct audit_krule;
> struct user_namespace;
> struct timezone;
>
> +enum lsm_event {
> + LSM_POLICY_CHANGE,
> +};
> +
> +typedef void (*lsm_notifier)(enum lsm_event event, void *ctx, void *data);
> +
> +void security_lsm_notify(enum lsm_event event, void *data);
> +int security_register_lsm_notifier(lsm_notifier func, void *ctx, u32 *id);
> +void security_unregister_lsm_notifier(u32 id);
> +
> /* These functions are in security/commoncap.c */
> extern int cap_capable(const struct cred *cred, struct user_namespace *ns,
> int cap, int audit);
> diff --git a/security/security.c b/security/security.c
> index 234982d..1263c1d 100644
> --- a/security/security.c
> +++ b/security/security.c
> @@ -33,6 +33,18 @@
> /* Maximum number of letters for an LSM name string */
> #define SECURITY_NAME_MAX 10
>
> +struct lsm_notifier_entry {
> + u32 callback_id;
> + lsm_notifier func;
> + void *ctx;
> + struct list_head list;
> + struct rcu_head rcu;
> +};
> +
> +static LIST_HEAD(lsm_notifier_list);
> +static DEFINE_SPINLOCK(lsm_notifier_lock);
> +static u32 next_callback_id;
I'd rather see the callback ID named something that makes it obvious
it is tied to the lsm_notifier_* bits, perhaps lsm_notifier_nextid?
> /* Boot-time LSM user choice */
> static __initdata char chosen_lsm[SECURITY_NAME_MAX + 1] =
> CONFIG_DEFAULT_SECURITY;
> @@ -98,6 +110,52 @@ int __init security_module_enable(const char *module)
> return !strcmp(module, chosen_lsm);
> }
>
> +void security_lsm_notify(enum lsm_event event, void *data)
> +{
> + struct lsm_notifier_entry *entry;
> +
> + rcu_read_lock();
> + list_for_each_entry_rcu(entry, &lsm_notifier_list, list)
> + entry->func(event, entry->ctx, data);
> + rcu_read_unlock();
> +}
> +EXPORT_SYMBOL(security_lsm_notify);
> +
> +int security_register_lsm_notifier(lsm_notifier func, void *ctx, u32 *id)
> +{
> + struct lsm_notifier_entry *entry;
> +
> + entry = kzalloc(sizeof(*entry), GFP_KERNEL);
> + if (!entry)
> + return -ENOMEM;
> +
> + entry->func = func;
> + entry->ctx = ctx;
Do we need ctx? We do need to ability to pass a blob to the callback
which will vary based on the event, but I'm not sure the ctx here has
much value.
> + spin_lock(&lsm_notifier_lock);
> + entry->callback_id = next_callback_id++;
> + *id = entry->callback_id;
> + list_add_rcu(&entry->list, &lsm_notifier_list);
> + spin_unlock(&lsm_notifier_lock);
> +
> + return 0;
> +}
> +EXPORT_SYMBOL(security_register_lsm_notifier);
--
paul moore
www.paul-moore.com
^ permalink raw reply [flat|nested] 42+ messages in thread
* Re: [PATCH v2 5/9] selinux: Create policydb version for Infiniband support
2016-07-14 22:56 ` Dan Jurgens
@ 2016-07-22 16:29 ` Paul Moore
-1 siblings, 0 replies; 42+ messages in thread
From: Paul Moore @ 2016-07-22 16:29 UTC (permalink / raw)
To: Dan Jurgens
Cc: chrisw-69jw2NvuJkxg9hUCZPvPmw, Stephen Smalley, Eric Paris,
dledford-H+wXaHxf7aLQT0dZR+AlfA,
sean.hefty-ral2JQCrhuEAvxtiuMwx3w,
hal.rosenstock-Re5JQEeQqe8AvxtiuMwx3w,
selinux-+05T5uksL2qpZYMLLGbcSA,
linux-security-module-u79uwXL29TY76Z2rM5mHXA,
linux-rdma-u79uwXL29TY76Z2rM5mHXA,
yevgenyp-VPRAkNaXOzVWk0Htik3J/w
On Thu, Jul 14, 2016 at 6:56 PM, Dan Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org> wrote:
> From: Daniel Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
>
> Support for Infiniband requires the addition of two new object contexts,
> one for infiniband PKeys and another IB Ports. Added handlers to read
> and write the new ocontext types when reading or writing a binary policy
> representation.
>
> Signed-off-by: Daniel Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
> Reviewed-by: Eli Cohen <eli-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
>
> ---
> v2:
> - Shorten ib_end_port to ib_port. Paul Moore
We had a discussion about this in the last patchset and I think things
may have gotten confused. From what I remember, according to the IB
developers the proper term is "end port"; if that is true, we probably
should use "end port", but I would prefer if we drop the underscore in
most of the places, e.g. "ib_endport" instead of "ib_end_port".
Sorry for the continuing nit pickiness.
> - Added bounds checking to port number. Paul Moore
> - Eliminated {} in OCON_PKEY case statement. Yuval Shaia
> ---
> security/selinux/include/security.h | 3 +-
> security/selinux/ss/policydb.c | 130 +++++++++++++++++++++++++++++++----
> security/selinux/ss/policydb.h | 27 +++++--
> 3 files changed, 136 insertions(+), 24 deletions(-)
--
paul moore
www.paul-moore.com
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 42+ messages in thread
* Re: [PATCH v2 5/9] selinux: Create policydb version for Infiniband support
@ 2016-07-22 16:29 ` Paul Moore
0 siblings, 0 replies; 42+ messages in thread
From: Paul Moore @ 2016-07-22 16:29 UTC (permalink / raw)
To: Dan Jurgens
Cc: chrisw, Stephen Smalley, Eric Paris, dledford, sean.hefty,
hal.rosenstock, selinux, linux-security-module, linux-rdma,
yevgenyp
On Thu, Jul 14, 2016 at 6:56 PM, Dan Jurgens <danielj@mellanox.com> wrote:
> From: Daniel Jurgens <danielj@mellanox.com>
>
> Support for Infiniband requires the addition of two new object contexts,
> one for infiniband PKeys and another IB Ports. Added handlers to read
> and write the new ocontext types when reading or writing a binary policy
> representation.
>
> Signed-off-by: Daniel Jurgens <danielj@mellanox.com>
> Reviewed-by: Eli Cohen <eli@mellanox.com>
>
> ---
> v2:
> - Shorten ib_end_port to ib_port. Paul Moore
We had a discussion about this in the last patchset and I think things
may have gotten confused. From what I remember, according to the IB
developers the proper term is "end port"; if that is true, we probably
should use "end port", but I would prefer if we drop the underscore in
most of the places, e.g. "ib_endport" instead of "ib_end_port".
Sorry for the continuing nit pickiness.
> - Added bounds checking to port number. Paul Moore
> - Eliminated {} in OCON_PKEY case statement. Yuval Shaia
> ---
> security/selinux/include/security.h | 3 +-
> security/selinux/ss/policydb.c | 130 +++++++++++++++++++++++++++++++----
> security/selinux/ss/policydb.h | 27 +++++--
> 3 files changed, 136 insertions(+), 24 deletions(-)
--
paul moore
www.paul-moore.com
^ permalink raw reply [flat|nested] 42+ messages in thread
* Re: [PATCH v2 5/9] selinux: Create policydb version for Infiniband support
2016-07-22 16:29 ` Paul Moore
@ 2016-07-22 16:46 ` Jason Gunthorpe
-1 siblings, 0 replies; 42+ messages in thread
From: Jason Gunthorpe @ 2016-07-22 16:46 UTC (permalink / raw)
To: Paul Moore
Cc: Dan Jurgens, chrisw-69jw2NvuJkxg9hUCZPvPmw, Stephen Smalley,
Eric Paris, dledford-H+wXaHxf7aLQT0dZR+AlfA,
sean.hefty-ral2JQCrhuEAvxtiuMwx3w,
hal.rosenstock-Re5JQEeQqe8AvxtiuMwx3w,
selinux-+05T5uksL2qpZYMLLGbcSA,
linux-security-module-u79uwXL29TY76Z2rM5mHXA,
linux-rdma-u79uwXL29TY76Z2rM5mHXA,
yevgenyp-VPRAkNaXOzVWk0Htik3J/w
On Fri, Jul 22, 2016 at 12:29:25PM -0400, Paul Moore wrote:
> We had a discussion about this in the last patchset and I think things
> may have gotten confused. From what I remember, according to the IB
> developers the proper term is "end port"; if that is true, we probably
> should use "end port", but I would prefer if we drop the underscore in
> most of the places, e.g. "ib_endport" instead of "ib_end_port".
Yes, the proper term is end port, (contrasted with switch port).
Jason
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 42+ messages in thread
* Re: [PATCH v2 5/9] selinux: Create policydb version for Infiniband support
@ 2016-07-22 16:46 ` Jason Gunthorpe
0 siblings, 0 replies; 42+ messages in thread
From: Jason Gunthorpe @ 2016-07-22 16:46 UTC (permalink / raw)
To: Paul Moore
Cc: Dan Jurgens, chrisw, Stephen Smalley, Eric Paris, dledford,
sean.hefty, hal.rosenstock, selinux, linux-security-module,
linux-rdma, yevgenyp
On Fri, Jul 22, 2016 at 12:29:25PM -0400, Paul Moore wrote:
> We had a discussion about this in the last patchset and I think things
> may have gotten confused. From what I remember, according to the IB
> developers the proper term is "end port"; if that is true, we probably
> should use "end port", but I would prefer if we drop the underscore in
> most of the places, e.g. "ib_endport" instead of "ib_end_port".
Yes, the proper term is end port, (contrasted with switch port).
Jason
^ permalink raw reply [flat|nested] 42+ messages in thread
* Re: [PATCH v2 3/9] selinux lsm IB/core: Implement LSM notification system
2016-07-22 16:21 ` Paul Moore
@ 2016-07-22 16:50 ` Daniel Jurgens
-1 siblings, 0 replies; 42+ messages in thread
From: Daniel Jurgens @ 2016-07-22 16:50 UTC (permalink / raw)
To: Paul Moore
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA,
linux-security-module-u79uwXL29TY76Z2rM5mHXA,
chrisw-69jw2NvuJkxg9hUCZPvPmw, dledford-H+wXaHxf7aLQT0dZR+AlfA,
selinux-+05T5uksL2qpZYMLLGbcSA,
sean.hefty-ral2JQCrhuEAvxtiuMwx3w, Stephen Smalley,
hal.rosenstock-Re5JQEeQqe8AvxtiuMwx3w
On 7/22/2016 11:21 AM, Paul Moore wrote:
> On Thu, Jul 14, 2016 at 6:56 PM, Dan Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org> wrote:
>> v2:
>> - new patch that has the generic notification, replaces selinux and
>> IB/core patches related to the ib_flush callback. Yuval Shaia and Paul
>> Moore
>> ---
>> drivers/infiniband/core/device.c | 46 ++++++++++++++++++++++++++++++
>> include/linux/security.h | 10 ++++++
>> security/security.c | 58 ++++++++++++++++++++++++++++++++++++++
>> security/selinux/hooks.c | 5 ++-
>> security/selinux/selinuxfs.c | 2 +
>> 5 files changed, 119 insertions(+), 2 deletions(-)
>>
> Thanks for making this more generic. I've got some comments below,
> but in the course of reviewing this I realized that the kernel has an
> existing, general purpose notifier mechanism (see
> include/linux/notifier.h); while I've seen the notifier code in the
> network stack, I never realized it was created as a general purpose
> mechanism. My apologies, I should have noticed this earlier and
> mentioned it.
>
> As far as how that impacts this patch; it seems like we should be good
> citizens and use the general mechanism, I don't see any obvious
> reasons why it wouldn't work. That said, I do feel bad for not
> realizing this earlier. If it isn't too annoying, would you mind
> updating this patch *again*?
Sure, I'll can do that for v3.
>> +
>> +static LIST_HEAD(lsm_notifier_list);
>> +static DEFINE_SPINLOCK(lsm_notifier_lock);
>> +static u32 next_callback_id;
> I'd rather see the callback ID named something that makes it obvious
> it is tied to the lsm_notifier_* bits, perhaps lsm_notifier_nextid?
I think this will go away with the switch to the generic notification, but if it stays I'll make the change.
>> /* Boot-time LSM user choice */
>> static __initdata char chosen_lsm[SECURITY_NAME_MAX + 1] =
>> CONFIG_DEFAULT_SECURITY;
>> @@ -98,6 +110,52 @@ int __init security_module_enable(const char *module)
>> return !strcmp(module, chosen_lsm);
>> }
>>
>> +void security_lsm_notify(enum lsm_event event, void *data)
>> +{
>> + struct lsm_notifier_entry *entry;
>> +
>> + rcu_read_lock();
>> + list_for_each_entry_rcu(entry, &lsm_notifier_list, list)
>> + entry->func(event, entry->ctx, data);
>> + rcu_read_unlock();
>> +}
>> +EXPORT_SYMBOL(security_lsm_notify);
>> +
>> +int security_register_lsm_notifier(lsm_notifier func, void *ctx, u32 *id)
>> +{
>> + struct lsm_notifier_entry *entry;
>> +
>> + entry = kzalloc(sizeof(*entry), GFP_KERNEL);
>> + if (!entry)
>> + return -ENOMEM;
>> +
>> + entry->func = func;
>> + entry->ctx = ctx;
> Do we need ctx? We do need to ability to pass a blob to the callback
> which will vary based on the event, but I'm not sure the ctx here has
> much value.
You'll see how I used ctx in "[PATCH v2 4/9] IB/core: Enforce security on management datagrams". Many mad agents can be created and each will register for a callback. I used the context to store a pointer to the specific mad agent.
>> + spin_lock(&lsm_notifier_lock);
>> + entry->callback_id = next_callback_id++;
>> + *id = entry->callback_id;
>> + list_add_rcu(&entry->list, &lsm_notifier_list);
>> + spin_unlock(&lsm_notifier_lock);
>> +
>> + return 0;
>> +}
>> +EXPORT_SYMBOL(security_register_lsm_notifier);
_______________________________________________
Selinux mailing list
Selinux-+05T5uksL2qpZYMLLGbcSA@public.gmane.org
To unsubscribe, send email to Selinux-leave-+05T5uksL2pAGbPMOrvdOA@public.gmane.org
To get help, send an email containing "help" to Selinux-request-+05T5uksL2pAGbPMOrvdOA@public.gmane.org
^ permalink raw reply [flat|nested] 42+ messages in thread
* Re: [PATCH v2 3/9] selinux lsm IB/core: Implement LSM notification system
@ 2016-07-22 16:50 ` Daniel Jurgens
0 siblings, 0 replies; 42+ messages in thread
From: Daniel Jurgens @ 2016-07-22 16:50 UTC (permalink / raw)
To: Paul Moore
Cc: chrisw, Stephen Smalley, Eric Paris, dledford, sean.hefty,
hal.rosenstock, selinux, linux-security-module, linux-rdma,
Yevgeny Petrilin
On 7/22/2016 11:21 AM, Paul Moore wrote:
> On Thu, Jul 14, 2016 at 6:56 PM, Dan Jurgens <danielj@mellanox.com> wrote:
>> v2:
>> - new patch that has the generic notification, replaces selinux and
>> IB/core patches related to the ib_flush callback. Yuval Shaia and Paul
>> Moore
>> ---
>> drivers/infiniband/core/device.c | 46 ++++++++++++++++++++++++++++++
>> include/linux/security.h | 10 ++++++
>> security/security.c | 58 ++++++++++++++++++++++++++++++++++++++
>> security/selinux/hooks.c | 5 ++-
>> security/selinux/selinuxfs.c | 2 +
>> 5 files changed, 119 insertions(+), 2 deletions(-)
>>
> Thanks for making this more generic. I've got some comments below,
> but in the course of reviewing this I realized that the kernel has an
> existing, general purpose notifier mechanism (see
> include/linux/notifier.h); while I've seen the notifier code in the
> network stack, I never realized it was created as a general purpose
> mechanism. My apologies, I should have noticed this earlier and
> mentioned it.
>
> As far as how that impacts this patch; it seems like we should be good
> citizens and use the general mechanism, I don't see any obvious
> reasons why it wouldn't work. That said, I do feel bad for not
> realizing this earlier. If it isn't too annoying, would you mind
> updating this patch *again*?
Sure, I'll can do that for v3.
>> +
>> +static LIST_HEAD(lsm_notifier_list);
>> +static DEFINE_SPINLOCK(lsm_notifier_lock);
>> +static u32 next_callback_id;
> I'd rather see the callback ID named something that makes it obvious
> it is tied to the lsm_notifier_* bits, perhaps lsm_notifier_nextid?
I think this will go away with the switch to the generic notification, but if it stays I'll make the change.
>> /* Boot-time LSM user choice */
>> static __initdata char chosen_lsm[SECURITY_NAME_MAX + 1] =
>> CONFIG_DEFAULT_SECURITY;
>> @@ -98,6 +110,52 @@ int __init security_module_enable(const char *module)
>> return !strcmp(module, chosen_lsm);
>> }
>>
>> +void security_lsm_notify(enum lsm_event event, void *data)
>> +{
>> + struct lsm_notifier_entry *entry;
>> +
>> + rcu_read_lock();
>> + list_for_each_entry_rcu(entry, &lsm_notifier_list, list)
>> + entry->func(event, entry->ctx, data);
>> + rcu_read_unlock();
>> +}
>> +EXPORT_SYMBOL(security_lsm_notify);
>> +
>> +int security_register_lsm_notifier(lsm_notifier func, void *ctx, u32 *id)
>> +{
>> + struct lsm_notifier_entry *entry;
>> +
>> + entry = kzalloc(sizeof(*entry), GFP_KERNEL);
>> + if (!entry)
>> + return -ENOMEM;
>> +
>> + entry->func = func;
>> + entry->ctx = ctx;
> Do we need ctx? We do need to ability to pass a blob to the callback
> which will vary based on the event, but I'm not sure the ctx here has
> much value.
You'll see how I used ctx in "[PATCH v2 4/9] IB/core: Enforce security on management datagrams". Many mad agents can be created and each will register for a callback. I used the context to store a pointer to the specific mad agent.
>> + spin_lock(&lsm_notifier_lock);
>> + entry->callback_id = next_callback_id++;
>> + *id = entry->callback_id;
>> + list_add_rcu(&entry->list, &lsm_notifier_list);
>> + spin_unlock(&lsm_notifier_lock);
>> +
>> + return 0;
>> +}
>> +EXPORT_SYMBOL(security_register_lsm_notifier);
^ permalink raw reply [flat|nested] 42+ messages in thread
* Re: [PATCH v2 5/9] selinux: Create policydb version for Infiniband support
2016-07-22 16:46 ` Jason Gunthorpe
@ 2016-07-22 16:52 ` Daniel Jurgens
-1 siblings, 0 replies; 42+ messages in thread
From: Daniel Jurgens @ 2016-07-22 16:52 UTC (permalink / raw)
To: Jason Gunthorpe, Paul Moore
Cc: chrisw-69jw2NvuJkxg9hUCZPvPmw, Stephen Smalley, Eric Paris,
dledford-H+wXaHxf7aLQT0dZR+AlfA,
sean.hefty-ral2JQCrhuEAvxtiuMwx3w,
hal.rosenstock-Re5JQEeQqe8AvxtiuMwx3w,
selinux-+05T5uksL2qpZYMLLGbcSA,
linux-security-module-u79uwXL29TY76Z2rM5mHXA,
linux-rdma-u79uwXL29TY76Z2rM5mHXA, Yevgeny Petrilin
On 7/22/2016 11:47 AM, Jason Gunthorpe wrote:
> On Fri, Jul 22, 2016 at 12:29:25PM -0400, Paul Moore wrote:
>> We had a discussion about this in the last patchset and I think things
>> may have gotten confused. From what I remember, according to the IB
>> developers the proper term is "end port"; if that is true, we probably
>> should use "end port", but I would prefer if we drop the underscore in
>> most of the places, e.g. "ib_endport" instead of "ib_end_port".
> Yes, the proper term is end port, (contrasted with switch port).
>
> Jason
>
I can add the 'end'.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 42+ messages in thread
* Re: [PATCH v2 5/9] selinux: Create policydb version for Infiniband support
@ 2016-07-22 16:52 ` Daniel Jurgens
0 siblings, 0 replies; 42+ messages in thread
From: Daniel Jurgens @ 2016-07-22 16:52 UTC (permalink / raw)
To: Jason Gunthorpe, Paul Moore
Cc: chrisw, Stephen Smalley, Eric Paris, dledford, sean.hefty,
hal.rosenstock, selinux, linux-security-module, linux-rdma,
Yevgeny Petrilin
On 7/22/2016 11:47 AM, Jason Gunthorpe wrote:
> On Fri, Jul 22, 2016 at 12:29:25PM -0400, Paul Moore wrote:
>> We had a discussion about this in the last patchset and I think things
>> may have gotten confused. From what I remember, according to the IB
>> developers the proper term is "end port"; if that is true, we probably
>> should use "end port", but I would prefer if we drop the underscore in
>> most of the places, e.g. "ib_endport" instead of "ib_end_port".
> Yes, the proper term is end port, (contrasted with switch port).
>
> Jason
>
I can add the 'end'.
^ permalink raw reply [flat|nested] 42+ messages in thread
* Re: [PATCH v2 7/9] selinux: Implement Infiniband PKey "Access" access vector
2016-07-14 22:56 ` Dan Jurgens
@ 2016-07-22 18:53 ` Paul Moore
-1 siblings, 0 replies; 42+ messages in thread
From: Paul Moore @ 2016-07-22 18:53 UTC (permalink / raw)
To: Dan Jurgens
Cc: chrisw-69jw2NvuJkxg9hUCZPvPmw, Stephen Smalley, Eric Paris,
dledford-H+wXaHxf7aLQT0dZR+AlfA,
sean.hefty-ral2JQCrhuEAvxtiuMwx3w,
hal.rosenstock-Re5JQEeQqe8AvxtiuMwx3w,
selinux-+05T5uksL2qpZYMLLGbcSA,
linux-security-module-u79uwXL29TY76Z2rM5mHXA,
linux-rdma-u79uwXL29TY76Z2rM5mHXA,
yevgenyp-VPRAkNaXOzVWk0Htik3J/w
On Thu, Jul 14, 2016 at 6:56 PM, Dan Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org> wrote:
> From: Daniel Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
>
> Add a type and access vector for PKeys. Implement the ib_pkey_access
> hook to check that the caller has permission to access the PKey on the
> given subnet prefix. Add an interface to get the PKey SID. Walk the PKey
> ocontexts to find an entry for the given subnet prefix and pkey.
>
> Signed-off-by: Daniel Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
>
> ---
> v2:
> - Use void* blobs for security structs. Paul Moore
> - Add pkey specific data to the audit log. Paul Moore
> - Don't introduce a new initial sid, use unlabeled. Stephen Smalley
> ---
> include/linux/lsm_audit.h | 7 ++++++
> security/lsm_audit.c | 13 +++++++++++
> security/selinux/hooks.c | 23 +++++++++++++++++++
> security/selinux/include/classmap.h | 2 +
> security/selinux/include/security.h | 2 +
> security/selinux/ss/services.c | 41 +++++++++++++++++++++++++++++++++++
> 6 files changed, 88 insertions(+), 0 deletions(-)
...
> diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c
> index fe3d20d..d9c4bfa 100644
> --- a/security/selinux/hooks.c
> +++ b/security/selinux/hooks.c
> @@ -5995,6 +5995,28 @@ static int selinux_key_getsecurity(struct key *key, char **_buffer)
> #endif
>
> #ifdef CONFIG_SECURITY_INFINIBAND
> +static int selinux_ib_pkey_access(u64 subnet_prefix, u16 pkey_val, void *ib_sec)
> +{
I'm only mentioning this because it looks like there is going to be
another spin of the patches; perhaps move the ib_sec argument before
the subnet_prefix and pkey_val arguments. This way it matches the
subject-object ordering we have in the avc_has_perm() call.
> + struct common_audit_data ad;
> + int err;
> + u32 sid = 0;
> + struct ib_security_struct *sec = ib_sec;
> + struct lsm_pkey_audit pkey;
> +
> + err = security_pkey_sid(subnet_prefix, pkey_val, &sid);
> +
> + if (err)
> + return err;
> +
> + ad.type = LSM_AUDIT_DATA_PKEY;
> + pkey.subnet_prefix = subnet_prefix;
> + pkey.pkey = pkey_val;
> + ad.u.pkey = &pkey;
> + return avc_has_perm(sec->sid, sid,
> + SECCLASS_INFINIBAND_PKEY,
> + INFINIBAND_PKEY__ACCESS, &ad);
> +}
--
paul moore
www.paul-moore.com
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 42+ messages in thread
* Re: [PATCH v2 7/9] selinux: Implement Infiniband PKey "Access" access vector
@ 2016-07-22 18:53 ` Paul Moore
0 siblings, 0 replies; 42+ messages in thread
From: Paul Moore @ 2016-07-22 18:53 UTC (permalink / raw)
To: Dan Jurgens
Cc: chrisw, Stephen Smalley, Eric Paris, dledford, sean.hefty,
hal.rosenstock, selinux, linux-security-module, linux-rdma,
yevgenyp
On Thu, Jul 14, 2016 at 6:56 PM, Dan Jurgens <danielj@mellanox.com> wrote:
> From: Daniel Jurgens <danielj@mellanox.com>
>
> Add a type and access vector for PKeys. Implement the ib_pkey_access
> hook to check that the caller has permission to access the PKey on the
> given subnet prefix. Add an interface to get the PKey SID. Walk the PKey
> ocontexts to find an entry for the given subnet prefix and pkey.
>
> Signed-off-by: Daniel Jurgens <danielj@mellanox.com>
>
> ---
> v2:
> - Use void* blobs for security structs. Paul Moore
> - Add pkey specific data to the audit log. Paul Moore
> - Don't introduce a new initial sid, use unlabeled. Stephen Smalley
> ---
> include/linux/lsm_audit.h | 7 ++++++
> security/lsm_audit.c | 13 +++++++++++
> security/selinux/hooks.c | 23 +++++++++++++++++++
> security/selinux/include/classmap.h | 2 +
> security/selinux/include/security.h | 2 +
> security/selinux/ss/services.c | 41 +++++++++++++++++++++++++++++++++++
> 6 files changed, 88 insertions(+), 0 deletions(-)
...
> diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c
> index fe3d20d..d9c4bfa 100644
> --- a/security/selinux/hooks.c
> +++ b/security/selinux/hooks.c
> @@ -5995,6 +5995,28 @@ static int selinux_key_getsecurity(struct key *key, char **_buffer)
> #endif
>
> #ifdef CONFIG_SECURITY_INFINIBAND
> +static int selinux_ib_pkey_access(u64 subnet_prefix, u16 pkey_val, void *ib_sec)
> +{
I'm only mentioning this because it looks like there is going to be
another spin of the patches; perhaps move the ib_sec argument before
the subnet_prefix and pkey_val arguments. This way it matches the
subject-object ordering we have in the avc_has_perm() call.
> + struct common_audit_data ad;
> + int err;
> + u32 sid = 0;
> + struct ib_security_struct *sec = ib_sec;
> + struct lsm_pkey_audit pkey;
> +
> + err = security_pkey_sid(subnet_prefix, pkey_val, &sid);
> +
> + if (err)
> + return err;
> +
> + ad.type = LSM_AUDIT_DATA_PKEY;
> + pkey.subnet_prefix = subnet_prefix;
> + pkey.pkey = pkey_val;
> + ad.u.pkey = &pkey;
> + return avc_has_perm(sec->sid, sid,
> + SECCLASS_INFINIBAND_PKEY,
> + INFINIBAND_PKEY__ACCESS, &ad);
> +}
--
paul moore
www.paul-moore.com
^ permalink raw reply [flat|nested] 42+ messages in thread
* Re: [PATCH v2 8/9] selinux: Add IB Port SMP access vector
2016-07-14 22:56 ` [PATCH v2 8/9] selinux: Add IB Port SMP access vector Dan Jurgens
@ 2016-07-22 19:26 ` Paul Moore
0 siblings, 0 replies; 42+ messages in thread
From: Paul Moore @ 2016-07-22 19:26 UTC (permalink / raw)
To: Dan Jurgens
Cc: chrisw-69jw2NvuJkxg9hUCZPvPmw, Stephen Smalley, Eric Paris,
dledford-H+wXaHxf7aLQT0dZR+AlfA,
sean.hefty-ral2JQCrhuEAvxtiuMwx3w,
hal.rosenstock-Re5JQEeQqe8AvxtiuMwx3w,
selinux-+05T5uksL2qpZYMLLGbcSA,
linux-security-module-u79uwXL29TY76Z2rM5mHXA,
linux-rdma-u79uwXL29TY76Z2rM5mHXA,
yevgenyp-VPRAkNaXOzVWk0Htik3J/w
On Thu, Jul 14, 2016 at 6:56 PM, Dan Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org> wrote:
> From: Daniel Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
>
> Add a type for Infiniband ports and an access vector for subnet
> management packets. Implement the ib_port_smp hook to check that the
> caller has permission to send and receive SMPs on the end port specified
> by the device name and port. Add interface to query the SID for a IB
> port, which walks the IB_PORT ocontexts to find an entry for the
> given name and port.
>
> Signed-off-by: Daniel Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
...
> diff --git a/security/lsm_audit.c b/security/lsm_audit.c
> index 2546d82..95632a9 100644
> --- a/security/lsm_audit.c
> +++ b/security/lsm_audit.c
> @@ -410,6 +410,11 @@ static void dump_common_audit_data(struct audit_buffer *ab,
> a->u.pkey->pkey, &sbn_pfx);
> break;
> }
> + case LSM_AUDIT_DATA_IB_PORT:
> + audit_log_format(ab, " device=");
> + audit_log_untrustedstring(ab, a->u.ib_port->dev_name);
The device name comes from the driver, right? We only need to use the
"untrustedstring" logging variant when we are dealing with strings
from userspace that have no guarantees on what they might contain.
Check the audit_string_contains_control() function to see what we
consider "bad" characters.
> + audit_log_format(ab, " port=%u", a->u.ib_port->port);
Based on our other conversations, I'm guessing that should be " endport=%u"?
> + break;
> } /* switch (a->type) */
> }
>
> diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c
> index d9c4bfa..40ad0e8 100644
> --- a/security/selinux/hooks.c
> +++ b/security/selinux/hooks.c
> @@ -6017,6 +6017,28 @@ static int selinux_ib_pkey_access(u64 subnet_prefix, u16 pkey_val, void *ib_sec)
> INFINIBAND_PKEY__ACCESS, &ad);
> }
>
> +static int selinux_ib_port_manage_subnet(const char *dev_name, u8 port, void *ib_sec)
> +{
See my previous comments about argument ordering, in other words, put
ib_sec first to help preserve my sanity in the future.
> + struct common_audit_data ad;
> + int err;
> + u32 sid = 0;
> + struct ib_security_struct *sec = ib_sec;
> + struct lsm_ib_port_audit ib_port;
> +
> + err = security_ib_port_sid(dev_name, port, &sid);
> +
> + if (err)
> + return err;
> +
> + ad.type = LSM_AUDIT_DATA_IB_PORT;
> + strncpy(ib_port.dev_name, dev_name, sizeof(ib_port.dev_name));
> + ib_port.port = port;
> + ad.u.ib_port = &ib_port;
> + return avc_has_perm(sec->sid, sid,
> + SECCLASS_INFINIBAND_PORT,
> + INFINIBAND_PORT__MANAGE_SUBNET, &ad);
> +}
> +
--
paul moore
www.paul-moore.com
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 42+ messages in thread
* Re: [PATCH v2 8/9] selinux: Add IB Port SMP access vector
@ 2016-07-22 19:26 ` Paul Moore
0 siblings, 0 replies; 42+ messages in thread
From: Paul Moore @ 2016-07-22 19:26 UTC (permalink / raw)
To: Dan Jurgens
Cc: chrisw, Stephen Smalley, Eric Paris, dledford, sean.hefty,
hal.rosenstock, selinux, linux-security-module, linux-rdma,
yevgenyp
On Thu, Jul 14, 2016 at 6:56 PM, Dan Jurgens <danielj@mellanox.com> wrote:
> From: Daniel Jurgens <danielj@mellanox.com>
>
> Add a type for Infiniband ports and an access vector for subnet
> management packets. Implement the ib_port_smp hook to check that the
> caller has permission to send and receive SMPs on the end port specified
> by the device name and port. Add interface to query the SID for a IB
> port, which walks the IB_PORT ocontexts to find an entry for the
> given name and port.
>
> Signed-off-by: Daniel Jurgens <danielj@mellanox.com>
...
> diff --git a/security/lsm_audit.c b/security/lsm_audit.c
> index 2546d82..95632a9 100644
> --- a/security/lsm_audit.c
> +++ b/security/lsm_audit.c
> @@ -410,6 +410,11 @@ static void dump_common_audit_data(struct audit_buffer *ab,
> a->u.pkey->pkey, &sbn_pfx);
> break;
> }
> + case LSM_AUDIT_DATA_IB_PORT:
> + audit_log_format(ab, " device=");
> + audit_log_untrustedstring(ab, a->u.ib_port->dev_name);
The device name comes from the driver, right? We only need to use the
"untrustedstring" logging variant when we are dealing with strings
from userspace that have no guarantees on what they might contain.
Check the audit_string_contains_control() function to see what we
consider "bad" characters.
> + audit_log_format(ab, " port=%u", a->u.ib_port->port);
Based on our other conversations, I'm guessing that should be " endport=%u"?
> + break;
> } /* switch (a->type) */
> }
>
> diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c
> index d9c4bfa..40ad0e8 100644
> --- a/security/selinux/hooks.c
> +++ b/security/selinux/hooks.c
> @@ -6017,6 +6017,28 @@ static int selinux_ib_pkey_access(u64 subnet_prefix, u16 pkey_val, void *ib_sec)
> INFINIBAND_PKEY__ACCESS, &ad);
> }
>
> +static int selinux_ib_port_manage_subnet(const char *dev_name, u8 port, void *ib_sec)
> +{
See my previous comments about argument ordering, in other words, put
ib_sec first to help preserve my sanity in the future.
> + struct common_audit_data ad;
> + int err;
> + u32 sid = 0;
> + struct ib_security_struct *sec = ib_sec;
> + struct lsm_ib_port_audit ib_port;
> +
> + err = security_ib_port_sid(dev_name, port, &sid);
> +
> + if (err)
> + return err;
> +
> + ad.type = LSM_AUDIT_DATA_IB_PORT;
> + strncpy(ib_port.dev_name, dev_name, sizeof(ib_port.dev_name));
> + ib_port.port = port;
> + ad.u.ib_port = &ib_port;
> + return avc_has_perm(sec->sid, sid,
> + SECCLASS_INFINIBAND_PORT,
> + INFINIBAND_PORT__MANAGE_SUBNET, &ad);
> +}
> +
--
paul moore
www.paul-moore.com
^ permalink raw reply [flat|nested] 42+ messages in thread
* Re: [PATCH v2 8/9] selinux: Add IB Port SMP access vector
2016-07-22 19:26 ` Paul Moore
@ 2016-07-27 16:27 ` Daniel Jurgens
-1 siblings, 0 replies; 42+ messages in thread
From: Daniel Jurgens @ 2016-07-27 16:27 UTC (permalink / raw)
To: Paul Moore
Cc: chrisw-69jw2NvuJkxg9hUCZPvPmw, Stephen Smalley, Eric Paris,
dledford-H+wXaHxf7aLQT0dZR+AlfA,
sean.hefty-ral2JQCrhuEAvxtiuMwx3w,
hal.rosenstock-Re5JQEeQqe8AvxtiuMwx3w,
selinux-+05T5uksL2qpZYMLLGbcSA,
linux-security-module-u79uwXL29TY76Z2rM5mHXA,
linux-rdma-u79uwXL29TY76Z2rM5mHXA, Yevgeny Petrilin
On 7/22/2016 2:26 PM, Paul Moore wrote:
> On Thu, Jul 14, 2016 at 6:56 PM, Dan Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org> wrote:
>
>> + audit_log_format(ab, " port=%u", a->u.ib_port->port);
> Based on our other conversations, I'm guessing that should be " endport=%u"?
I think port is fine there, device name and port number. Together they are an "endport".
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 42+ messages in thread
* Re: [PATCH v2 8/9] selinux: Add IB Port SMP access vector
@ 2016-07-27 16:27 ` Daniel Jurgens
0 siblings, 0 replies; 42+ messages in thread
From: Daniel Jurgens @ 2016-07-27 16:27 UTC (permalink / raw)
To: Paul Moore
Cc: chrisw, Stephen Smalley, Eric Paris, dledford, sean.hefty,
hal.rosenstock, selinux, linux-security-module, linux-rdma,
Yevgeny Petrilin
On 7/22/2016 2:26 PM, Paul Moore wrote:
> On Thu, Jul 14, 2016 at 6:56 PM, Dan Jurgens <danielj@mellanox.com> wrote:
>
>> + audit_log_format(ab, " port=%u", a->u.ib_port->port);
> Based on our other conversations, I'm guessing that should be " endport=%u"?
I think port is fine there, device name and port number. Together they are an "endport".
^ permalink raw reply [flat|nested] 42+ messages in thread
end of thread, other threads:[~2016-07-27 16:27 UTC | newest]
Thread overview: 42+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2016-07-14 22:56 [PATCH v2 0/9] SELinux support for Infiniband RDMA Dan Jurgens
2016-07-14 22:56 ` Dan Jurgens
2016-07-14 22:56 ` [PATCH v2 4/9] IB/core: Enforce security on management datagrams Dan Jurgens
[not found] ` <1468537011-20407-1-git-send-email-danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
2016-07-14 22:56 ` [PATCH v2 1/9] IB/core: IB cache enhancements to support Infiniband security Dan Jurgens
2016-07-14 22:56 ` Dan Jurgens
2016-07-14 22:56 ` [PATCH v2 2/9] IB/core: Enforce PKey security on QPs Dan Jurgens
2016-07-14 22:56 ` Dan Jurgens
2016-07-14 22:56 ` [PATCH v2 3/9] selinux lsm IB/core: Implement LSM notification system Dan Jurgens
2016-07-14 22:56 ` Dan Jurgens
[not found] ` <1468537011-20407-4-git-send-email-danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
2016-07-15 10:54 ` kbuild test robot
2016-07-15 10:54 ` kbuild test robot
2016-07-22 16:21 ` Paul Moore
2016-07-22 16:21 ` Paul Moore
2016-07-22 16:50 ` Daniel Jurgens
2016-07-22 16:50 ` Daniel Jurgens
2016-07-14 22:56 ` [PATCH v2 5/9] selinux: Create policydb version for Infiniband support Dan Jurgens
2016-07-14 22:56 ` Dan Jurgens
[not found] ` <1468537011-20407-6-git-send-email-danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
2016-07-22 16:29 ` Paul Moore
2016-07-22 16:29 ` Paul Moore
[not found] ` <CAHC9VhTRQ_jvdOPn-eszzHkb1+HL7ySaM5qK2pgX=9GFdjOcpg-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2016-07-22 16:46 ` Jason Gunthorpe
2016-07-22 16:46 ` Jason Gunthorpe
2016-07-22 16:52 ` Daniel Jurgens
2016-07-22 16:52 ` Daniel Jurgens
2016-07-14 22:56 ` [PATCH v2 6/9] selinux: Allocate and free infiniband security hooks Dan Jurgens
2016-07-14 22:56 ` Dan Jurgens
2016-07-14 22:56 ` [PATCH v2 7/9] selinux: Implement Infiniband PKey "Access" access vector Dan Jurgens
2016-07-14 22:56 ` Dan Jurgens
[not found] ` <1468537011-20407-8-git-send-email-danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
2016-07-22 18:53 ` Paul Moore
2016-07-22 18:53 ` Paul Moore
2016-07-14 22:56 ` [PATCH v2 9/9] selinux: Add a cache for quicker retreival of PKey SIDs Dan Jurgens
2016-07-14 22:56 ` Dan Jurgens
[not found] ` <1468537011-20407-10-git-send-email-danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
2016-07-15 9:06 ` kbuild test robot
2016-07-15 9:06 ` kbuild test robot
2016-07-22 15:46 ` [PATCH v2 0/9] SELinux support for Infiniband RDMA Paul Moore
2016-07-22 15:46 ` Paul Moore
2016-07-22 16:02 ` Daniel Jurgens
2016-07-22 16:02 ` Daniel Jurgens
2016-07-14 22:56 ` [PATCH v2 8/9] selinux: Add IB Port SMP access vector Dan Jurgens
[not found] ` <1468537011-20407-9-git-send-email-danielj-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
2016-07-22 19:26 ` Paul Moore
2016-07-22 19:26 ` Paul Moore
2016-07-27 16:27 ` Daniel Jurgens
2016-07-27 16:27 ` Daniel Jurgens
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.