svcrdma: Fast Memory Registration Support

This patchset implements support for Fast Memory Registration in the NFS server. Fast Memory Regstration is the ability to quickly map a kernel memory page list as a logically contiguous memory region from the perspective of the adapter. This mapping is created and invalidated using work requests posted on the SQ. This allows for large amounts of data to transferred between the client and server with a single work request as well as the ability to invalidate a previously mapped memory region. For iWARP, this allows for "one-shot" memory regions to be mapped for a single NFS-RDMA data transfer. This improves security since a byzantine app listening on the net will have a very short window during which the RKEY is valid.

This capability is only enabled if the underlying device advertises that it is supported.

This patches are also available here:
git://git.linux-nfs.org/projects/tomtucker/xprt-switch-2.6.git

Signed-off-by Tom Tucker <[email protected]>

include/linux/sunrpc/svc_rdma.h | 27 ++++++++++++++++++++++++++-

1 files changed, 26 insertions(+), 1 deletions(-)

[PATCH 02/09] svcrdma: Add FRMR get/put services

include/linux/sunrpc/svc_rdma.h | 3 +
net/sunrpc/xprtrdma/svc_rdma_transport.c | 125 ++++++++++++++++++++++++++++-

2 files changed, 123 insertions(+), 5 deletions(-)

[PATCH 03/09] svcrdma: Query device for Fast Reg support during connection setup

net/sunrpc/xprtrdma/svc_rdma_transport.c | 86 +++++++++++++++++++++++++++--

1 files changed, 80 insertions(+), 6 deletions(-)

[PATCH 04/09] svcrdma: Add a service to register a Fast Reg MR with the device

include/linux/sunrpc/svc_rdma.h | 1 +
net/sunrpc/xprtrdma/svc_rdma_transport.c | 53 ++++++++++++++++++++++++++---

2 files changed, 48 insertions(+), 6 deletions(-)

[PATCH 05/09] svcrdma: Modify post recv path to use local dma key

net/sunrpc/xprtrdma/svc_rdma_transport.c | 10 +++++++---

1 files changed, 7 insertions(+), 3 deletions(-)

[PATCH 06/09] svcrdma: Add support to svc_rdma_send to handle chained WR

net/sunrpc/xprtrdma/svc_rdma_transport.c | 29 +++++++++++++++++++++--------

1 files changed, 21 insertions(+), 8 deletions(-)

[PATCH 07/09] svcrdma: Modify the RPC recv path to use FRMR when available

include/linux/sunrpc/svc_rdma.h          |    1 +
net/sunrpc/xprtrdma/svc_rdma_recvfrom.c  |  187 ++++++++++++++++++++++++++----
net/sunrpc/xprtrdma/svc_rdma_transport.c |    5 +-
3 files changed, 171 insertions(+), 22 deletions(-)

[PATCH 08/09] svcrdma: Modify the RPC reply path to use FRMR when available

net/sunrpc/xprtrdma/svc_rdma_sendto.c | 263 +++++++++++++++++++++++++-----
net/sunrpc/xprtrdma/svc_rdma_transport.c | 2 +

2 files changed, 225 insertions(+), 40 deletions(-)

[PATCH 09/09] svcrdma: Update svc_rdma_send_error to use DMA LKEY

net/sunrpc/xprtrdma/svc_rdma_transport.c | 11 +++++++++--

1 files changed, 9 insertions(+), 2 deletions(-)

Quoted 12 lines · Tom Tucker

This patchset implements support for Fast Memory Registration in the NFS server. Fast Memory Regstration is the ability to quickly map a kernel memory page list as a logically contiguous memory region from the perspective of the adapter. This mapping is created and invalidated using work requests posted on the SQ. This allows for large amounts of data to transferred between the client and server with a single work request as well as the ability to invalidate a previously mapped memory region. For iWARP, this allows for "one-shot" memory regions to be mapped for a single NFS-RDMA data transfer. This improves security since a byzantine app listening on the net will have a very short window during which the RKEY is valid.

Quoted 2 lines

This capability is only enabled if the underlying device advertises that it is supported.

Thanks for your continuing work on this.

I think we really need to document the security assumptions, though.

(Currently is your entire memory at the mercy of anyone on the same local network as your rdma adapter?

make this stuff useful in more situations, but language like "a very short window" doesn't sound promising. Also, we've got to make sure users understand where it's safe to use this stuff....)

--b.

Quoted 53 lines

This patches are also available here:
git://git.linux-nfs.org/projects/tomtucker/xprt-switch-2.6.git

Signed-off-by Tom Tucker <[email protected]>

include/linux/sunrpc/svc_rdma.h | 27 ++++++++++++++++++++++++++-

1 files changed, 26 insertions(+), 1 deletions(-)

[PATCH 02/09] svcrdma: Add FRMR get/put services

include/linux/sunrpc/svc_rdma.h | 3 +
net/sunrpc/xprtrdma/svc_rdma_transport.c | 125 ++++++++++++++++++++++++++++-

2 files changed, 123 insertions(+), 5 deletions(-)

[PATCH 03/09] svcrdma: Query device for Fast Reg support during connection setup

net/sunrpc/xprtrdma/svc_rdma_transport.c | 86 +++++++++++++++++++++++++++--

1 files changed, 80 insertions(+), 6 deletions(-)

[PATCH 04/09] svcrdma: Add a service to register a Fast Reg MR with the device

include/linux/sunrpc/svc_rdma.h | 1 +
net/sunrpc/xprtrdma/svc_rdma_transport.c | 53 ++++++++++++++++++++++++++---

2 files changed, 48 insertions(+), 6 deletions(-)

[PATCH 05/09] svcrdma: Modify post recv path to use local dma key

net/sunrpc/xprtrdma/svc_rdma_transport.c | 10 +++++++---

1 files changed, 7 insertions(+), 3 deletions(-)

[PATCH 06/09] svcrdma: Add support to svc_rdma_send to handle chained WR

net/sunrpc/xprtrdma/svc_rdma_transport.c | 29 +++++++++++++++++++++--------

1 files changed, 21 insertions(+), 8 deletions(-)

[PATCH 07/09] svcrdma: Modify the RPC recv path to use FRMR when available

include/linux/sunrpc/svc_rdma.h          |    1 +
net/sunrpc/xprtrdma/svc_rdma_recvfrom.c  |  187 ++++++++++++++++++++++++++----
net/sunrpc/xprtrdma/svc_rdma_transport.c |    5 +-
3 files changed, 171 insertions(+), 22 deletions(-)

[PATCH 08/09] svcrdma: Modify the RPC reply path to use FRMR when available

net/sunrpc/xprtrdma/svc_rdma_sendto.c | 263 +++++++++++++++++++++++++-----
net/sunrpc/xprtrdma/svc_rdma_transport.c | 2 +

2 files changed, 225 insertions(+), 40 deletions(-)

[PATCH 09/09] svcrdma: Update svc_rdma_send_error to use DMA LKEY

net/sunrpc/xprtrdma/svc_rdma_transport.c | 11 +++++++++--

1 files changed, 9 insertions(+), 2 deletions(-)

Quoted 16 lines · J. Bruce Fields

Quoted 14 lines · Tom Tucker

This patchset implements support for Fast Memory Registration in the NFS server. Fast Memory Regstration is the ability to quickly map a kernel memory page list as a logically contiguous memory region from the perspective of the adapter. This mapping is created and invalidated using work requests posted on the SQ. This allows for large amounts of data to transferred between the client and server with a single work request as well as the ability to invalidate a previously mapped memory region. For iWARP, this allows for "one-shot" memory regions to be mapped for a single NFS-RDMA data transfer. This improves security since a byzantine app listening on the net will have a very short window during which the RKEY is valid.

This capability is only enabled if the underlying device advertises that it is supported.

Quoted 4 lines

Thanks for your continuing work on this.

I think we really need to document the security assumptions, though.

Yes, that's a good idea. Maybe a file in Documentation/svcrdma?

Quoted 3 lines

(Currently is your entire memory at the mercy of anyone on the same local network as your rdma adapter?

------------------------------------------------------------------------

A principal exploit is that a node listening on a mirror port of a switch could snoop RDMA packets containing RKEY and then forge a packet with this RKEY to write or read the memory of the peer to which the RKEY referred.

The NFSRDMA protocol is defined such that a) only the server initiates RDMA, and b) only the client's memory is exposed via RKEY. This is why the server reads to fetch RPC data from the client even though it would be more efficient for the client to write the data to the server's memory.

The above design goal is not entirely realized with iWARP, however, because the RKEY (called an STag on iWARP) for the data sink of an RDMA_READ is actually placed on the wire!

Not only that, iWARP (RDDP) requires that this
RKEY have Remote Write! This means that the server's memory is exposed by virtue of having placed the RKEY for it's local memory on the wire in order to receive the result of the RDMA_READ.

By contrast, IB uses an opaque
transaction ID# to associate the READ_RPL with the READ_REQ _and_ the data

sink of an RDMA_READ does not require remote access. That said, the evil node in question, for example, could potentially forge a packet with this transaction ID and corrupt the target memory, however, the duration of the exploit is this single READ_REQ.

The newer RDMA adapters (both iWARP and IB) support "Fast Memory Registration". This capability allows memory to be quickly registered and de-registered by submitting WR on the SQ. So the idea is to create an RKEY that ONLY maps the single RPC. So the WR sequence is post_map, post_rdma_read, post_invalidate. This has two benefits, a) it restricts the domain of the exploit to the memory of a single RPC,

and b) it limits the
duration of the exploit to the time it takes to satisfy the RDMA_READ.

If so, fixing that would certainly

Quoted 4 lines

make this stuff useful in more situations, but language like "a very short window" doesn't sound promising. Also, we've got to make sure users understand where it's safe to use this stuff....)

There are those who argue that a one-shot STag/RKEY is no less secure than TCP.

Consider that the exact same evil application could more easily corrupt RPC payload by simply forging a packet with the correct TCP sequence number -- in fact it's easier than the RDMA exploit because the RDMA exploit requires that you correctly forge both the TCP packet _and_ the RDMA payload. In addition the duration of the TCP exploit is the lifetime of the connection, not the lifetime of a single WR.

So if you buy the argument above, RDMA on IB or iWARP using Fast Reg is no less secure than TCP. That is the goal of this patch series.

Quoted 14 lines · Tom Tucker

Quoted 13 lines · J. Bruce Fields

Quoted 12 lines · Tom Tucker

This patchset implements support for Fast Memory Registration in the NFS server. Fast Memory Regstration is the ability to quickly map a kernel memory page list as a logically contiguous memory region from the perspective of the adapter. This mapping is created and invalidated using work requests posted on the SQ. This allows for large amounts of data to transferred between the client and server with a single work request as well as the ability to invalidate a previously mapped memory region. For iWARP, this allows for "one-shot" memory regions to be mapped for a single NFS-RDMA data transfer. This improves security since a byzantine app listening on the net will have a very short window during which the RKEY is valid.

Quoted 7 lines

Quoted 6 lines

Quoted 2 lines

This capability is only enabled if the underlying device advertises that it is supported.

Thanks for your continuing work on this.

I think we really need to document the security assumptions, though.

Yes, that's a good idea. Maybe a file in Documentation/svcrdma?

That would be great! (Actually Documentation/filesystems/nfs-something might be good. Adding it to nfs-rdma.txt would be ideal.)

Quoted 4 lines

Quoted 2 lines

(Currently is your entire memory at the mercy of anyone on the same local network as your rdma adapter?

------------------------------------------------------------------------

Quoted 3 lines

A principal exploit is that a node listening on a mirror port of a switch could snoop RDMA packets containing RKEY and then forge a packet with this RKEY to write or read the memory of the peer to which the RKEY referred.

RKEY? Apologies, I've been avoiding up till now asking about every bit of jargon in hopes I could still get the gist, but I think I have to give up and learn it. Remind me what I need to read?

Quoted 6 lines

The NFSRDMA protocol is defined such that a) only the server initiates RDMA, and b) only the client's memory is exposed via RKEY. This is why the server reads to fetch RPC data from the client even though it would be more efficient for the client to write the data to the server's memory.

Quoted 3 lines

The above design goal is not entirely realized with iWARP, however, because the RKEY (called an STag on iWARP) for the data sink of an RDMA_READ is actually placed on the wire!

So "data sink" here means the server memory that the data's being copied to? OK.

Quoted 5 lines

Not only that, iWARP (RDDP) requires that this
RKEY have Remote Write! This means that the server's memory is exposed by virtue of having placed the RKEY for it's local memory on the wire in order to receive the result of the RDMA_READ.

Got it.

Quoted 2 lines

By contrast, IB uses an opaque
transaction ID# to associate the READ_RPL with the READ_REQ _and_ the data

Google gives only two hits for READ_RPL. Help!

Quoted 5 lines

sink of an RDMA_READ does not require remote access. That said, the evil node in question, for example, could potentially forge a packet with this transaction ID and corrupt the target memory, however, the duration of the exploit is this single READ_REQ.

OK. So what would be the effect of the forged packet?

Quoted 6 lines

The newer RDMA adapters (both iWARP and IB) support "Fast Memory Registration". This capability allows memory to be quickly registered and de-registered by submitting WR on the SQ. So the idea is to create an RKEY that ONLY maps the single RPC. So the WR sequence is post_map, post_rdma_read, post_invalidate. This has two benefits, a) it restricts the domain of the exploit to the memory of a single RPC,

So the only server-side memory that can be remotely written to is the pages that will hold the write data?

Quoted 2 lines

and b) it limits the
duration of the exploit to the time it takes to satisfy the RDMA_READ.

OK.

Quoted 6 lines

If so, fixing that would certainly

Quoted 4 lines

make this stuff useful in more situations, but language like "a very short window" doesn't sound promising. Also, we've got to make sure users understand where it's safe to use this stuff....)

There are those who argue that a one-shot STag/RKEY is no less secure than TCP.

If the attacker is limited to corrupting some write data, then, yes, that doesn't sound so different from nfs/rcp with auth_unix.

Quoted 7 lines

Consider that the exact same evil application could more easily corrupt RPC payload by simply forging a packet with the correct TCP sequence number -- in fact it's easier than the RDMA exploit because the RDMA exploit requires that you correctly forge both the TCP packet _and_ the RDMA payload. In addition the duration of the TCP exploit is the lifetime of the connection, not the lifetime of a single WR.

Quoted 2 lines

So if you buy the argument above, RDMA on IB or iWARP using Fast Reg is no less secure than TCP. That is the goal of this patch series.

OK, thanks.

--b.

Here are the relevant specs:

Unfortunately, the IB specs are private to the IBTA.
http://www.infinibandta.org/home

The OFA verbs most closely resemble IB verbs.

Here are the iWARP protocol specs:

http://www.ietf.org/rfc/rfc5040.txt
http://www.ietf.org/rfc/rfc5041.txt
http://www.ietf.org/rfc/rfc5044.txt

Here are the iWARP verbs specs. This is what the adapters where written to.

http://www.rdmaconsortium.org/home/RNIC_Verbs_Overview2.pdf
http://www.rdmaconsortium.org/home/draft-hilland-iwarp-verbs-v1.0-RDMAC.pdf

Quoted 9 lines · Tom Tucker

Here are the relevant specs:

Unfortunately, the IB specs are private to the IBTA.
http://www.infinibandta.org/home

The OFA verbs most closely resemble IB verbs.

Here are the iWARP protocol specs:

Quoted 8 lines

http://www.ietf.org/rfc/rfc5040.txt
http://www.ietf.org/rfc/rfc5041.txt
http://www.ietf.org/rfc/rfc5044.txt

Here are the iWARP verbs specs. This is what the adapters where written to.

http://www.rdmaconsortium.org/home/RNIC_Verbs_Overview2.pdf
http://www.rdmaconsortium.org/home/draft-hilland-iwarp-verbs-v1.0-RDMAC.pdf

OK, thanks very much to you and James for all the explanation. And actually the IB specs do appear to be available if you give them your name and address?

So could one of you add some of this (the explanation of the security issues, at least) to Documentation/filesystems/nfs-rdma.txt?

I'm worried about the fact that the amount of trust required in the network will vary so much depending on exactly which hardware and kernel version the user has.

As long as this is relatively specialized hardware, perhaps it's not such a big deal.... But at the very least I think we should make sure this is carefully documented.

--b.
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

Quoted 17 lines · Tom Tucker

Quoted 16 lines · J. Bruce Fields

Quoted 14 lines · Tom Tucker

This patchset implements support for Fast Memory Registration in the NFS server. Fast Memory Regstration is the ability to quickly map a kernel memory page list as a logically contiguous memory region from the perspective of the adapter. This mapping is created and invalidated using work requests posted on the SQ. This allows for large amounts of data to transferred between the client and server with a single work request as well as the ability to invalidate a previously mapped memory region. For iWARP, this allows for "one-shot" memory regions to be mapped for a single NFS-RDMA data transfer. This improves security since a byzantine app listening on the net will have a very short window during which the RKEY is valid.

This capability is only enabled if the underlying device advertises that it is supported.

Quoted 6 lines

Quoted 4 lines

Thanks for your continuing work on this.

I think we really need to document the security assumptions, though.

Yes, that's a good idea. Maybe a file in Documentation/svcrdma?

That would be great! (Actually Documentation/filesystems/nfs-something might be good. Adding it to nfs-rdma.txt would be ideal.)

Quoted 6 lines

Quoted 3 lines

(Currently is your entire memory at the mercy of anyone on the same local network as your rdma adapter?

------------------------------------------------------------------------

Quoted 3 lines

A principal exploit is that a node listening on a mirror port of a switch could snoop RDMA packets containing RKEY and then forge a packet with this RKEY to write or read the memory of the peer to which the RKEY referred.

RKEY? Apologies, I've been avoiding up till now asking about every bit of jargon in hopes I could still get the gist, but I think I have to give up and learn it. Remind me what I need to read?

Quoted 8 lines

The NFSRDMA protocol is defined such that a) only the server initiates RDMA, and b) only the client's memory is exposed via RKEY. This is why the server reads to fetch RPC data from the client even though it would be more efficient for the client to write the data to the server's memory.

The above design goal is not entirely realized with iWARP, however, because the RKEY (called an STag on iWARP) for the data sink of an RDMA_READ is actually placed on the wire!

So "data sink" here means the server memory that the data's being copied to? OK.

Quoted 4 lines

Not only that, iWARP (RDDP) requires that this
RKEY have Remote Write! This means that the server's memory is exposed by virtue of having placed the RKEY for it's local memory on the wire in order to receive the result of the RDMA_READ.

Got it.

Quoted 2 lines

By contrast, IB uses an opaque
transaction ID# to associate the READ_RPL with the READ_REQ _and_ the data

Google gives only two hits for READ_RPL. Help!

Quoted 5 lines

sink of an RDMA_READ does not require remote access. That said, the evil node in question, for example, could potentially forge a packet with this transaction ID and corrupt the target memory, however, the duration of the exploit is this single READ_REQ.

OK. So what would be the effect of the forged packet?

Quoted 7 lines

The newer RDMA adapters (both iWARP and IB) support "Fast Memory Registration". This capability allows memory to be quickly registered and de-registered by submitting WR on the SQ. So the idea is to create an RKEY that ONLY maps the single RPC. So the WR sequence is post_map, post_rdma_read, post_invalidate. This has two benefits, a) it restricts the domain of the exploit to the memory of a single RPC,

So the only server-side memory that can be remotely written to is the pages that will hold the write data?

Quoted 3 lines

and b) it limits the
duration of the exploit to the time it takes to satisfy the RDMA_READ.

OK.

Quoted 7 lines

If so, fixing that would certainly

Quoted 4 lines

make this stuff useful in more situations, but language like "a very short window" doesn't sound promising. Also, we've got to make sure users understand where it's safe to use this stuff....)

There are those who argue that a one-shot STag/RKEY is no less secure than TCP.

If the attacker is limited to corrupting some write data, then, yes, that doesn't sound so different from nfs/rcp with auth_unix.

Quoted 9 lines

Consider that the exact same evil application could more easily corrupt RPC payload by simply forging a packet with the correct TCP sequence number -- in fact it's easier than the RDMA exploit because the RDMA exploit requires that you correctly forge both the TCP packet _and_ the RDMA payload. In addition the duration of the TCP exploit is the lifetime of the connection, not the lifetime of a single WR.

So if you buy the argument above, RDMA on IB or iWARP using Fast Reg is no less secure than TCP. That is the goal of this patch series.

OK, thanks.

--b.
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html