| Message ID | 20230131182855.4027499-12-dhowells@redhat.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | smb3: Use iov_iters down to the network transport and fix DIO page pinning | expand |
On Tue, Jan 31, 2023 at 06:28:54PM +0000, David Howells wrote: > When the cifs client is talking to the ksmbd server by RDMA and the ksmbd > server has "smb3 encryption = yes" in its config file, the normal PDU > stream is encrypted, but the directly-delivered data isn't in the stream > (and isn't encrypted), but is rather delivered by DDP/RDMA packets (at > least with IWarp). This really needs to be split into a separate backportable fix series. And it seems like Stefan has just send such a series.
Am 01.02.23 um 14:36 schrieb Christoph Hellwig: > On Tue, Jan 31, 2023 at 06:28:54PM +0000, David Howells wrote: >> When the cifs client is talking to the ksmbd server by RDMA and the ksmbd >> server has "smb3 encryption = yes" in its config file, the normal PDU >> stream is encrypted, but the directly-delivered data isn't in the stream >> (and isn't encrypted), but is rather delivered by DDP/RDMA packets (at >> least with IWarp). > > This really needs to be split into a separate backportable fix series. > And it seems like Stefan has just send such a series. My changes are triggered by this commit... Now I'm looking more closely as I didn't understand it...
Am 31.01.23 um 19:28 schrieb David Howells: > When the cifs client is talking to the ksmbd server by RDMA and the ksmbd > server has "smb3 encryption = yes" in its config file, the normal PDU > stream is encrypted, but the directly-delivered data isn't in the stream > (and isn't encrypted), but is rather delivered by DDP/RDMA packets (at > least with IWarp). > > Currently, the direct delivery fails with: > > buf can not contain only a part of read data > WARNING: CPU: 0 PID: 4619 at fs/cifs/smb2ops.c:4731 handle_read_data+0x393/0x405 > ... > RIP: 0010:handle_read_data+0x393/0x405 > ... > smb3_handle_read_data+0x30/0x37 > receive_encrypted_standard+0x141/0x224 > cifs_demultiplex_thread+0x21a/0x63b > kthread+0xe7/0xef > ret_from_fork+0x22/0x30 > > The problem apparently stemming from the fact that it's trying to manage > the decryption, but the data isn't in the smallbuf, the bigbuf or the page > array). > > This can be fixed simply by inserting an extra case into handle_read_data() > that checks to see if use_rdma_mr is true, and if it is, just setting > rdata->got_bytes to the length of data delivered and allowing normal > continuation. > > This can be seen in an IWarp packet trace. With the upstream code, it does > a DDP/RDMA packet, which produces the warning above and then retries, > retrieving the data inline, spread across several SMBDirect messages that > get glued together into a single PDU. With the patch applied, only the > DDP/RDMA packet is seen. > > Note that this doesn't happen if the server isn't told to encrypt stuff and > it does also happen with softRoCE. > > Signed-off-by: David Howells <dhowells@redhat.com> > cc: Steve French <smfrench@gmail.com> > cc: Tom Talpey <tom@talpey.com> > cc: Long Li <longli@microsoft.com> > cc: Namjae Jeon <linkinjeon@kernel.org> > cc: Stefan Metzmacher <metze@samba.org> > cc: linux-cifs@vger.kernel.org > > Link: https://lore.kernel.org/r/166855224228.1998592.2212551359609792175.stgit@warthog.procyon.org.uk/ # v1 > --- > fs/cifs/smb2ops.c | 3 +++ > 1 file changed, 3 insertions(+) > > diff --git a/fs/cifs/smb2ops.c b/fs/cifs/smb2ops.c > index cea578a45ed8..73b66ac86abf 100644 > --- a/fs/cifs/smb2ops.c > +++ b/fs/cifs/smb2ops.c > @@ -4733,6 +4733,9 @@ handle_read_data(struct TCP_Server_Info *server, struct mid_q_entry *mid, > if (length < 0) > return length; > rdata->got_bytes = data_len; > + } else if (use_rdma_mr) { > + /* The data was delivered directly by RDMA. */ > + rdata->got_bytes = data_len; I actually don't understand why this would only be a problem with encryption. I guess there's much more needed and data_offset should most likely be ignored completely in the rdma offload case. So I'd guess its just luck that we don't trigger the below warning/error more often. > } else { > /* read response payload cannot be in both buf and pages */ > WARN_ONCE(1, "buf can not contain only a part of read data"); >
Am 01.02.23 um 15:05 schrieb Stefan Metzmacher: > Am 31.01.23 um 19:28 schrieb David Howells: >> When the cifs client is talking to the ksmbd server by RDMA and the ksmbd >> server has "smb3 encryption = yes" in its config file, the normal PDU >> stream is encrypted, but the directly-delivered data isn't in the stream >> (and isn't encrypted), but is rather delivered by DDP/RDMA packets (at >> least with IWarp). >> >> Currently, the direct delivery fails with: >> >> buf can not contain only a part of read data >> WARNING: CPU: 0 PID: 4619 at fs/cifs/smb2ops.c:4731 handle_read_data+0x393/0x405 >> ... >> RIP: 0010:handle_read_data+0x393/0x405 >> ... >> smb3_handle_read_data+0x30/0x37 >> receive_encrypted_standard+0x141/0x224 >> cifs_demultiplex_thread+0x21a/0x63b >> kthread+0xe7/0xef >> ret_from_fork+0x22/0x30 >> >> The problem apparently stemming from the fact that it's trying to manage >> the decryption, but the data isn't in the smallbuf, the bigbuf or the page >> array). >> >> This can be fixed simply by inserting an extra case into handle_read_data() >> that checks to see if use_rdma_mr is true, and if it is, just setting >> rdata->got_bytes to the length of data delivered and allowing normal >> continuation. >> >> This can be seen in an IWarp packet trace. With the upstream code, it does >> a DDP/RDMA packet, which produces the warning above and then retries, >> retrieving the data inline, spread across several SMBDirect messages that >> get glued together into a single PDU. With the patch applied, only the >> DDP/RDMA packet is seen. >> >> Note that this doesn't happen if the server isn't told to encrypt stuff and >> it does also happen with softRoCE. >> >> Signed-off-by: David Howells <dhowells@redhat.com> >> cc: Steve French <smfrench@gmail.com> >> cc: Tom Talpey <tom@talpey.com> >> cc: Long Li <longli@microsoft.com> >> cc: Namjae Jeon <linkinjeon@kernel.org> >> cc: Stefan Metzmacher <metze@samba.org> >> cc: linux-cifs@vger.kernel.org >> >> Link: https://lore.kernel.org/r/166855224228.1998592.2212551359609792175.stgit@warthog.procyon.org.uk/ # v1 >> --- >> fs/cifs/smb2ops.c | 3 +++ >> 1 file changed, 3 insertions(+) >> >> diff --git a/fs/cifs/smb2ops.c b/fs/cifs/smb2ops.c >> index cea578a45ed8..73b66ac86abf 100644 >> --- a/fs/cifs/smb2ops.c >> +++ b/fs/cifs/smb2ops.c >> @@ -4733,6 +4733,9 @@ handle_read_data(struct TCP_Server_Info *server, struct mid_q_entry *mid, >> if (length < 0) >> return length; >> rdata->got_bytes = data_len; >> + } else if (use_rdma_mr) { >> + /* The data was delivered directly by RDMA. */ >> + rdata->got_bytes = data_len; > > I actually don't understand why this would only be a problem with encryption. > > I guess there's much more needed and data_offset should most likely be > ignored completely in the rdma offload case. So I'd guess its just luck > that we don't trigger the below warning/error more often. I guess it might be related to smb3_handle_read_data passing server->pdu_size to handle_read_data(), while server->pdu_size is the outer size of the transform and not the size of the decrypted pdu. Maybe receive_encrypted_standard() needs to reset server->pdu_size during this: if (mid_entry && mid_entry->handle) ret = mid_entry->handle(server, mid_entry); else ret = cifs_handle_standard(server, mid_entry); But this code is so overly complex, that's way to hard to understand...
diff --git a/fs/cifs/smb2ops.c b/fs/cifs/smb2ops.c index cea578a45ed8..73b66ac86abf 100644 --- a/fs/cifs/smb2ops.c +++ b/fs/cifs/smb2ops.c @@ -4733,6 +4733,9 @@ handle_read_data(struct TCP_Server_Info *server, struct mid_q_entry *mid, if (length < 0) return length; rdata->got_bytes = data_len; + } else if (use_rdma_mr) { + /* The data was delivered directly by RDMA. */ + rdata->got_bytes = data_len; } else { /* read response payload cannot be in both buf and pages */ WARN_ONCE(1, "buf can not contain only a part of read data");
When the cifs client is talking to the ksmbd server by RDMA and the ksmbd server has "smb3 encryption = yes" in its config file, the normal PDU stream is encrypted, but the directly-delivered data isn't in the stream (and isn't encrypted), but is rather delivered by DDP/RDMA packets (at least with IWarp). Currently, the direct delivery fails with: buf can not contain only a part of read data WARNING: CPU: 0 PID: 4619 at fs/cifs/smb2ops.c:4731 handle_read_data+0x393/0x405 ... RIP: 0010:handle_read_data+0x393/0x405 ... smb3_handle_read_data+0x30/0x37 receive_encrypted_standard+0x141/0x224 cifs_demultiplex_thread+0x21a/0x63b kthread+0xe7/0xef ret_from_fork+0x22/0x30 The problem apparently stemming from the fact that it's trying to manage the decryption, but the data isn't in the smallbuf, the bigbuf or the page array). This can be fixed simply by inserting an extra case into handle_read_data() that checks to see if use_rdma_mr is true, and if it is, just setting rdata->got_bytes to the length of data delivered and allowing normal continuation. This can be seen in an IWarp packet trace. With the upstream code, it does a DDP/RDMA packet, which produces the warning above and then retries, retrieving the data inline, spread across several SMBDirect messages that get glued together into a single PDU. With the patch applied, only the DDP/RDMA packet is seen. Note that this doesn't happen if the server isn't told to encrypt stuff and it does also happen with softRoCE. Signed-off-by: David Howells <dhowells@redhat.com> cc: Steve French <smfrench@gmail.com> cc: Tom Talpey <tom@talpey.com> cc: Long Li <longli@microsoft.com> cc: Namjae Jeon <linkinjeon@kernel.org> cc: Stefan Metzmacher <metze@samba.org> cc: linux-cifs@vger.kernel.org Link: https://lore.kernel.org/r/166855224228.1998592.2212551359609792175.stgit@warthog.procyon.org.uk/ # v1 --- fs/cifs/smb2ops.c | 3 +++ 1 file changed, 3 insertions(+)