kernel/linux.git/drivers/nvme/host/fabrics.h, branch v6.2

nvme: implement In-Band authentication

2022-08-02T23:14:49+00:00

Implement NVMe-oF In-Band authentication according to NVMe TPAR 8006. This patch adds two new fabric options 'dhchap_secret' to specify the pre-shared key (in ASCII respresentation according to NVMe 2.0 section 8.13.5.8 'Secret representation') and 'dhchap_ctrl_secret' to specify the pre-shared controller key for bi-directional authentication of both the host and the controller. Re-authentication can be triggered by writing the PSK into the new controller sysfs attribute 'dhchap_secret' or 'dhchap_ctrl_secret'. Signed-off-by: Hannes Reinecke Reviewed-by: Sagi Grimberg Signed-off-by: Christoph Hellwig [axboe: fold in clang build fix] Signed-off-by: Jens Axboe

nvme-fabrics: add a request timeout helper

2022-05-16T06:07:25+00:00

The RDAMA and TCP transport both complete the timed out request in the same manner and hence code is duplicated. Add and use the helper nvmf_complete_timed_out_request() to remove the duplicate code. Signed-off-by: Chaitanya Kulkarni Reviewed-by: Sagi Grimberg Signed-off-by: Christoph Hellwig

nvme-fabrics: fix state check in nvmf_ctlr_matches_baseopts()

2022-02-03T06:30:57+00:00

Controller deletion/reset, immediately followed by or concurrent with a reconnect, is hard failing the connect attempt resulting in a complete loss of connectivity to the controller. In the connect request, fabrics looks for an existing controller with the same address components and aborts the connect if a controller already exists and the duplicate connect option isn't set. The match routine filters out controllers that are dead or dying, so they don't interfere with the new connect request. When NVME_CTRL_DELETING_NOIO was added, it missed updating the state filters in the nvmf_ctlr_matches_baseopts() routine. Thus, when in this new state, it's seen as a live controller and fails the connect request. Correct by adding the DELETING_NIO state to the match checks. Fixes: ecca390e8056 ("nvme: fix deadlock in disconnect during scan_work and/or ana_work") Cc: # v5.7+ Signed-off-by: Uday Shankar Reviewed-by: James Smart Reviewed-by: Sagi Grimberg Signed-off-by: Christoph Hellwig

nvme: display correct subsystem NQN

2021-10-20T17:16:02+00:00

With discovery controllers supporting unique subsystem NQNs the actual subsystem NQN might be different from that one passed in via the connect args. So add a helper to display the resulting subsystem NQN. Signed-off-by: Hannes Reinecke Reviewed-by: Chaitanya Kulkarni Signed-off-by: Christoph Hellwig

nvme: Add connect option 'discovery'

2021-10-20T17:16:02+00:00

Add a connect option 'discovery' to specify that the connection should be made to a discovery controller, not a normal I/O controller. With discovery controllers supporting unique subsystem NQNs we cannot easily distinguish by the subsystem NQN if this should be a discovery connection, but we need this information to blank out options not supported by discovery controllers. Signed-off-by: Hannes Reinecke Reviewed-by: Chaitanya Kulkarni Signed-off-by: Christoph Hellwig

nvme: use blk_execute_rq() for passthrough commands

2021-06-30T21:35:38+00:00

The generic blk_execute_rq() knows how to handle polled completions. Use that instead of implementing an nvme specific handler. Signed-off-by: Keith Busch Reviewed-by: Christoph Hellwig Reviewed-by: Chaitanya Kulkarni Link: https://lore.kernel.org/r/20210610214437.641245-3-kbusch@kernel.org Signed-off-by: Jens Axboe

nvme-tcp: allow selecting the network interface for connections

2021-06-03T07:29:24+00:00

In our application, we need a way to force TCP connections to go out a specific IP interface instead of letting Linux select the interface based on the routing tables. Add the 'host-iface' option to allow specifying the interface to use. When the option host-iface is specified, the driver uses the specified interface to set the option SO_BINDTODEVICE on the TCP socket before connecting. This new option is needed in addtion to the existing host-traddr for the following reasons: Specifying an IP interface by its associated IP address is less intuitive than specifying the actual interface name and, in some cases, simply doesn't work. That's because the association between interfaces and IP addresses is not predictable. IP addresses can be changed or can change by themselves over time (e.g. DHCP). Interface names are predictable [1] and will persist over time. Consider the following configuration. 1: lo: mtu 65536 qdisc noqueue state ... link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 inet 100.0.0.100/24 scope global lo valid_lft forever preferred_lft forever 2: enp0s3: mtu 1500 qdisc ... link/ether 08:00:27:21:65:ec brd ff:ff:ff:ff:ff:ff inet 100.0.0.100/24 scope global enp0s3 valid_lft forever preferred_lft forever 3: enp0s8: mtu 1500 qdisc ... link/ether 08:00:27:4f:95:5c brd ff:ff:ff:ff:ff:ff inet 100.0.0.100/24 scope global enp0s8 valid_lft forever preferred_lft forever The above is a VM that I configured with the same IP address (100.0.0.100) on all interfaces. Doing a reverse lookup to identify the unique interface associated with 100.0.0.100 does not work here. And this is why the option host_iface is required. I understand that the above config does not represent a standard host system, but I'm using this to prove a point: "We can never know how users will configure their systems". By te way, The above configuration is perfectly fine by Linux. The current TCP implementation for host_traddr performs a bind()-before-connect(). This is a common construct to set the source IP address on a TCP socket before connecting. This has no effect on how Linux selects the interface for the connection. That's because Linux uses the Weak End System model as described in RFC1122 [2]. On the other hand, setting the Source IP Address has benefits and should be supported by linux-nvme. In fact, setting the Source IP Address is a mandatory FedGov requirement (e.g. connection to a RADIUS/TACACS+ server). Consider the following configuration. $ ip addr list dev enp0s8 3: enp0s8: mtu 1500 qdisc ... link/ether 08:00:27:4f:95:5c brd ff:ff:ff:ff:ff:ff inet 192.168.56.101/24 brd 192.168.56.255 scope global enp0s8 valid_lft 426sec preferred_lft 426sec inet 192.168.56.102/24 scope global secondary enp0s8 valid_lft forever preferred_lft forever inet 192.168.56.103/24 scope global secondary enp0s8 valid_lft forever preferred_lft forever inet 192.168.56.104/24 scope global secondary enp0s8 valid_lft forever preferred_lft forever Here we can see that several addresses are associated with interface enp0s8. By default, Linux always selects the default IP address, 192.168.56.101, as the source address when connecting over interface enp0s8. Some users, however, want the ability to specify a different source address (e.g., 192.168.56.102, 192.168.56.103, ...). The option host_traddr can be used as-is to perform this function. In conclusion, I believe that we need 2 options for TCP connections. One that can be used to specify an interface (host-iface). And one that can be used to set the source address (host-traddr). Users should be allowed to use one or the other, or both, or none. Of course, the documentation for host_traddr will need some clarification. It should state that when used for TCP connection, this option only sets the source address. And the documentation for host_iface should say that this option is only available for TCP connections. References: [1] https://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames/ [2] https://tools.ietf.org/html/rfc1122 Tested both IPv4 and IPv6 connections. Signed-off-by: Martin Belanger Reviewed-by: Sagi Grimberg Reviewed-by: Hannes Reinecke Signed-off-by: Christoph Hellwig

nvme: move the fabrics queue ready check routines to core

2021-05-04T07:35:49+00:00

queue_rq() in pci only checks if the dispatched queue (nvmeq) is ready, e.g. not being suspended. Since nvme_alloc_admin_tags() in reset flow restarts the admin queue, users are able to submit admin commands to a controller before reset_work() completes. Commands submitted under this condition may interfere with commands that performs identify, IO queue setup in reset_work(), and may result in a hang described in the following patch. As seen in the fabrics, user commands are prevented from being executed under inproper controller states. We may reuse this logic to maintain a clear admin queue during reset_work(). Signed-off-by: Tao Chiu Signed-off-by: Cody Wong Reviewed-by: Leon Chien Reviewed-by: Keith Busch Signed-off-by: Christoph Hellwig

nvme-fabrics: only reserve a single tag

2021-03-18T04:38:48+00:00

Fabrics drivers currently reserve two tags on the admin queue. But given that the connect command is only run on a freshly created queue or after all commands have been force aborted we only need to reserve a single tag. Signed-off-by: Christoph Hellwig Reviewed-by: Sagi Grimberg Reviewed-by: Chaitanya Kulkarni Reviewed-by: Hannes Reinecke Reviewed-by: Daniel Wagner

nvme-fabrics: reject I/O to offline device

2020-12-01T19:36:37+00:00

Commands get stuck while Host NVMe-oF controller is in reconnect state. The controller enters into reconnect state when it loses connection with the target. It tries to reconnect every 10 seconds (default) until a successful reconnect or until the reconnect time-out is reached. The default reconnect time out is 10 minutes. Applications are expecting commands to complete with success or error within a certain timeout (30 seconds by default). The NVMe host is enforcing that timeout while it is connected, but during reconnect the timeout is not enforced and commands may get stuck for a long period or even forever. To fix this long delay due to the default timeout, introduce new "fast_io_fail_tmo" session parameter. The timeout is measured in seconds from the controller reconnect and any command beyond that timeout is rejected. The new parameter value may be passed during 'connect'. The default value of -1 means no timeout (similar to current behavior). Signed-off-by: Victor Gladkov Signed-off-by: Chaitanya Kulkarni Reviewed-by: Hannes Reinecke Reviewed-by: Sagi Grimberg Reviewed-by: Chao Leng Signed-off-by: Christoph Hellwig