Re: [PATCH v2] powerpc/pseries/mobility: ignore ibm, platform-facilities updates
From: Daniel Axtens <hidden>
Date: 2021-10-21 05:48:06
Hi Nathan, Thanks for the detailed explanation. I've not really worked with the partition migration code before I was able to follow your logic.
On VMs with NX encryption, compression, and/or RNG offload, these capabilities are described by nodes in the ibm,platform-facilities device tree hierarchy: $ tree -d /sys/firmware/devicetree/base/ibm,platform-facilities/ /sys/firmware/devicetree/base/ibm,platform-facilities/ ├── ibm,compression-v1 ├── ibm,random-v1 └── ibm,sym-encryption-v1 3 directories The acceleration functions that these nodes describe are not disrupted by live migration, not even temporarily. But the post-migration ibm,update-nodes sequence firmware always sends "delete" messages for this hierarchy, followed by an "add" directive to reconstruct it via ibm,configure-connector (log with debugging statements enabled in mobility.c): mobility: removing node /ibm,platform-facilities/ibm,random-v1:4294967285 mobility: removing node /ibm,platform-facilities/ibm,compression-v1:4294967284 mobility: removing node /ibm,platform-facilities/ibm,sym-encryption-v1:4294967283 mobility: removing node /ibm,platform-facilities:4294967286 ... mobility: added node /ibm,platform-facilities:4294967286 Note we receive a single "add" message for the entire hierarchy, and what we receive from the ibm,configure-connector sequence is the top-level platform-facilities node along with its three children. The debug message simply reports the parent node and not the whole subtree.
If I understand correctly, (and again, this is not my area at all!) we still have to go out to the firmware and call the ibm,configure-connector sequence in order to figure out that the node we're supposed to add is the ibm,platform-facilites node, right? We can't save enough information at delete time to avoid the trip out to firmware?
Also, significantly, the nodes added are almost completely equivalent to
the ones removed; even phandles are unchanged. ibm,shared-interrupt-pool in
the leaf nodes is the only property I've observed to differ, and Linux does
not use that. So in practice, the sum of update messages Linux receives for
this hierarchy is equivalent to minor property updates.
We succeed in removing the original hierarchy from the device tree. But the
vio bus code is ignorant of this, and does not unbind or relinquish its
references. The leaf nodes, still reachable through sysfs, of course still
refer to the now-freed ibm,platform-facilities parent node, which makes
use-after-free possible:
refcount_t: addition on 0; use-after-free.
WARNING: CPU: 3 PID: 1706 at lib/refcount.c:25 refcount_warn_saturate+0x164/0x1f0
refcount_warn_saturate+0x160/0x1f0 (unreliable)
kobject_get+0xf0/0x100
of_node_get+0x30/0x50
of_get_parent+0x50/0xb0
of_fwnode_get_parent+0x54/0x90
fwnode_count_parents+0x50/0x150
fwnode_full_name_string+0x30/0x110
device_node_string+0x49c/0x790
vsnprintf+0x1c0/0x4c0
sprintf+0x44/0x60
devspec_show+0x34/0x50
dev_attr_show+0x40/0xa0
sysfs_kf_seq_show+0xbc/0x200
kernfs_seq_show+0x44/0x60
seq_read_iter+0x2a4/0x740
kernfs_fop_read_iter+0x254/0x2e0
new_sync_read+0x120/0x190
vfs_read+0x1d0/0x240
Moreover, the "new" replacement subtree is not correctly added to the
device tree, resulting in ibm,platform-facilities parent node without the
appropriate leaf nodes, and broken symlinks in the sysfs device hierarchy:
$ tree -d /sys/firmware/devicetree/base/ibm,platform-facilities/
/sys/firmware/devicetree/base/ibm,platform-facilities/
0 directories
$ cd /sys/devices/vio ; find . -xtype l -exec file {} +
./ibm,sym-encryption-v1/of_node: broken symbolic link to
../../../firmware/devicetree/base/ibm,platform-facilities/ibm,sym-encryption-v1
./ibm,random-v1/of_node: broken symbolic link to
../../../firmware/devicetree/base/ibm,platform-facilities/ibm,random-v1
./ibm,compression-v1/of_node: broken symbolic link to
../../../firmware/devicetree/base/ibm,platform-facilities/ibm,compression-v1
This is because add_dt_node() -> dlpar_attach_node() attaches only the
parent node returned from configure-connector, ignoring any children. This
should be corrected for the general case, but fixing that won't help with
the stale OF node references, which is the more urgent problem.
One way to address that would be to make the drivers respond to node
removal notifications, so that node references can be dropped
appropriately. But this would likely force the drivers to disrupt active
clients for no useful purpose: equivalent nodes are immediately re-added.
And recall that the acceleration capabilities described by the nodes remain
available throughout the whole process.
The solution I believe to be robust for this situation is to convert
remove+add of a node with an unchanged phandle to an update of the node's
properties in the Linux device tree structure. That would involve changing
and adding a fair amount of code, and may take several iterations to land.
Until that can be realized we have a confirmed use-after-free and the
possibility of memory corruption. So add a limited workaround that
discriminates on the node type, ignoring adds and removes. This should be
amenable to backporting in the meantime.
Yeah it's an unpleasant situation to find ourselves in. It's a bit icky
but as I think you said in a previous email, at least this isn't worse:
in the common case it should now succeed and and if properties change
significantly it will still fail.
My one question (from more of a security point of view) is:
1) Say you start using the facilities with a particular set of
parameters.
2) Say you then get migrated and the parameters change.
3) If you keep using the platform facilities as if the original
properties are still valid, can this cause any Interesting,
unexpected or otherwise Bad consequences? Are we going to end up
(for example) scribbling over random memory somehow?
Apart from that, the code seems to do what it says, it seems to solve a
real problem, the error and memory handling makes sense, you _put the DT
nodes that you _get, the comments are helpful and descriptive, and it
passes the automated tests on patchwork/snowpatch.
Kind regards,
Daniel
quoted hunk ↗ jump to hunk
Signed-off-by: Nathan Lynch <redacted> Fixes: 410bccf97881 ("powerpc/pseries: Partition migration in the kernel") Cc: stable@vger.kernel.org --- Notes: Changes since v1: * Clarify that the vio bus code maintains references to removed nodes, per Tyrel. arch/powerpc/platforms/pseries/mobility.c | 34 +++++++++++++++++++++++ 1 file changed, 34 insertions(+)diff --git a/arch/powerpc/platforms/pseries/mobility.c b/arch/powerpc/platforms/pseries/mobility.c index e83e0891272d..210a37a065fb 100644 --- a/arch/powerpc/platforms/pseries/mobility.c +++ b/arch/powerpc/platforms/pseries/mobility.c@@ -63,6 +63,27 @@ static int mobility_rtas_call(int token, char *buf, s32 scope) static int delete_dt_node(struct device_node *dn) { + struct device_node *pdn; + bool is_platfac; + + pdn = of_get_parent(dn); + is_platfac = of_node_is_type(dn, "ibm,platform-facilities") || + of_node_is_type(pdn, "ibm,platform-facilities"); + of_node_put(pdn); + + /* + * The drivers that bind to nodes in the platform-facilities + * hierarchy don't support node removal, and the removal directive + * from firmware is always followed by an add of an equivalent + * node. The capability (e.g. RNG, encryption, compression) + * represented by the node is never interrupted by the migration. + * So ignore changes to this part of the tree. + */ + if (is_platfac) { + pr_notice("ignoring remove operation for %pOFfp\n", dn); + return 0; + } + pr_debug("removing node %pOFfp\n", dn); dlpar_detach_node(dn); return 0;@@ -222,6 +243,19 @@ static int add_dt_node(struct device_node *parent_dn, __be32 drc_index) if (!dn) return -ENOENT; + /* + * Since delete_dt_node() ignores this node type, this is the + * necessary counterpart. We also know that a platform-facilities + * node returned from dlpar_configure_connector() has children + * attached, and dlpar_attach_node() only adds the parent, leaking + * the children. So ignore these on the add side for now. + */ + if (of_node_is_type(dn, "ibm,platform-facilities")) { + pr_notice("ignoring add operation for %pOF\n", dn); + dlpar_free_cc_nodes(dn); + return 0; + } + rc = dlpar_attach_node(dn, parent_dn); if (rc) dlpar_free_cc_nodes(dn);-- 2.31.1