recv() returns 0 when the remote end of the connection is shut down, but
there is no such rule for send(). Remove the incorrect treatment of
send() returning 0 as an error. The remaining code will treat a send()
return value of 0 as a non-error condition and will continue to work
correctly.
Change-Id: Ia753b802d95428104a62d1acb13c5fa437add6b4
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/369299
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
If a connection is closed by the remote end, the JSON-RPC server could
potentially call poll() on the pollfd containing the invalid file
descriptor. Rework the logic so that the pollfd's fd field is set to a
negative value so that poll() will ignore it until the connection is
fully cleaned up by spdk_jsonrpc_server_conn_remove().
Also add handling for the potential error conditions returned by poll()
so that if something does go wrong, the server doesn't get stuck polling
a broken pollfd forever.
Change-Id: Id02e3a230740c8ffd513888cb0564891b3aca069
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/369285
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Make sure that we check any closed connections even if poll() indicates
that no sockets are ready, since closed sockets won't ever trigger
poll().
Change-Id: Ie543cc2848d07f3d8e22662cb22c3d0f5cf3d174
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/369292
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Move the per-connection send buffer into each request, and allow a
connection to have a queue of responses ready to be sent.
Change-Id: If6b2151691c4cd76f3cf7cde0cdd8f20cac77ceb
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/368470
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
This will be necessary when the lifetime of a request can exceed the
lifetime of the values stored in spdk_jsonrpc_server_conn.
Change-Id: Icd9772eb142e3f6ae69303aff1e12bc213f435a4
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/368455
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
This will greatly simplify the implementation of asynchronous requests,
and asynchronous requests also allow clients to submit multiple
overlapped requests, making batches unnecessary for executing multiple
RPCs at once. Additionally, our RPC client (scripts/rpc.py) does not
use batch requests.
Change-Id: I2529793c54b43acbacd934d82926aa32e286210c
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/368449
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
This will enable asynchronous request handling in a future patch, and it
also removes the need for the RPC handlers to know about request id and
the JSON-RPC rules about notification-only requests.
Change-Id: I25aaa8e48bff8d5594ffcccecb61842b1e31ec3c
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/368225
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
SPDK_COUNTOF works like sizeof, except it returns the number of elements
in an array instead of the number of bytes.
Change-Id: I38ff4dd3485ed9b630cc5660ff84851d0031911f
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
The SPDK_TRACELOG macro depends on a CONFIG setting (DEBUG), so it
should not be part of the public API.
Create a new include/spdk_internal directory for headers that should
only be used within SPDK, not exported for public use.
Change-Id: I39b90ce57da3270e735ba32210c4b3a3468c460b
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Explicitly include spdk.common.mk at the top of all lib Makefiles so
that CONFIG options and other predefined variables are set.
Change-Id: I1e560c294fe8242602e45191a280f4295533ae44
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
