Adding support for asynchronous file transfer #404
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Signed-off-by: Kamalkumar Patel <[email protected]>
Patel-Kamalkumar
force-pushed
the
AIO
branch
5 times, most recently
from
fc3db41
to
dac7bd9
Compare
Adding support for Nonblocking socket so, pldm will not blocked during data transfer between host and BMC. in previous case pldm was blocked if DMA transfer blocked in middle of operation. Signed-off-by: Kamalkumar Patel <[email protected]>
Patel-Kamalkumar
force-pushed
the
AIO
branch
2 times, most recently
from
c85d7d1
to
d44ef4e
Compare
sagisin
reviewed
May 30, 2023
sagisin
reviewed
May 30, 2023
Patel-Kamalkumar
changed the title
Patel-Kamalkumar
requested review from
sagisin,
manojkiraneda,
jaypadath,
Pavithrab7 and
ArchanaKakani
- File transfer between BMC and Host is synchronous mechanism. - File transfer interrupted because of many reasons, so if file transfer interrupted between BMC and Host then PLDM is ending up in the hung state and not able to respond to the upcoming new requests. User has to do work around to restart the PLDM to get work again. - I have added asynchronous file transfer support using eventloop mechanism and non-blocking socket communication, so file transfer will be taken care by event loop and PLDM is free to receive another request from the Host. - If file transfer will be stuck or interrupted during asynchronous transfer then time-out occurs and file transfer will be aborted and same status communicated to the Host.
| try | ||
| { | ||
| handler = getHandlerByType(fileType, fileHandle); | ||
| handler = getSharedHandlerByType(fileType, fileHandle); |
There was a problem hiding this comment.
Why are we using shared pointer? Once the file handler is created, we are supposed to save it and resuse it.
Can we use same gethandlerByType function to return unique pointer and then move the pointer to functionPtr?
There was a problem hiding this comment.
when we call any function of read/write for DMA transfer then first time it's create event loop and return from that function and delete unique_ptr because of limited scope. we can't move unique_ptr inside event loop because it's delete any point of time and we don't know, and while creating even loop need to pass function pointer which will be called after file transfer completed and do post operation. so as per our requirement shared_ptr is best to use.
Commit the marker lid validate errors for the min-mif level and access key expiration so that it generates an SRC when we hit the error during code update. Signed-off-by: Sagar Srinivas <[email protected]>
This commit adds change to move the fetching of activation property inside the inband code update path. As it is a no-op for out of band updates. Signed-off-by: Sagar Srinivas <[email protected]>
The ibm,rainier1s-4u json had missing nvme slots and were not modelled by PLDM. This commit adds the same. Fixes: 561341 Change-Id: I0559f1761a6ec223fa62aa14c283885d3aa37e49 Signed-off-by: Pavithra Barithaya <[email protected]>
This commit adds 5 second dbus timeout value to dbus calls with no reply. Previously a commit was merged that included 5 seconds dbus timeout value for all dbus call with reply. This is added to make sure PLDM does not wait for more than 5 seconds when a dbus call is made with or without reply if no response is received from other daemon or host. The 5 second timeout value is overridden to 10 seconds in meta-ibm layer. Signed-off-by: [email protected] <[email protected]>
When IBM's hypervisor receives an event log (PEL) using the file API, it may end up discarding it, returning the reason the file was discarded in the status argument of the file ack command. It can do this for two reasons, either it doesn't have any more room to store event logs, or the event log was malformed. In these cases, call the HostReject D-Bus method on the PEL daemon so it can handle it. The code must convert the value of the status returned into the proper D-Bus enum value. Without this call, the operating system was missing event logs when there were floods of them. Tested: API was called when system was flooded with event logs, allowing the PEL daemon to resend them as space opens up. Change-Id: I9d3acf4e5c5f885934ef96e89ddb93dd8f9c10a3 Signed-off-by: Matt Spinler <[email protected]>
As part of adding new features there were traces that are still using standard ostream pattern of logging. As PLDM repository migrated to structured logging hence migrating the new log traces to follow the same practice. Additionally minor refinement of traces is also handled as a part of this commit, listed below - 1. Correction in type of log (ex- error or info) 2. Correction in spelling of log message 3. Removing logs from test file. Signed-off-by: Riya Dixit <[email protected]>
Patel-Kamalkumar
force-pushed
the
AIO
branch
3 times, most recently
from
39d7dc5
to
2bb2365
Compare
- File transfer between BMC and Host is synchronous mechanism. - File transfer interrupted because of many reasons, so if file transfer interrupted between BMC and Host then PLDM is ending up in the hung state and not able to respond to the upcoming new requests. User has to do work around to restart the PLDM to get work again. - I have added asynchronous file transfer support using eventloop mechanism and non-blocking socket communication, so file transfer will be taken care by event loop and PLDM is free to receive another request from the Host. - If file transfer will be stuck or interrupted during asynchronous transfer then time-out occurs and file transfer will be aborted and same status communicated to the Host.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Adding support for asynchronous file transfer
file transfer interrupted between BMC and Host then PLDM is
ending up in the hung state and not able to respond to the
upcoming new requests. User has to do work around to restart
the PLDM to get work again.
eventloop mechanism and non-blocking socket communication,
so file transfer will be taken care by event loop and PLDM
is free to receive another request from the Host.
asynchronous transfer then time-out occurs and file transfer
will be aborted and same
status communicated to the Host.