| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Curently, if I have no ipmi_lib_path, ipmid crashes:
$ ./ipmid
Registering handler:[/usr/lib/host-ipmid/H��dH�%(]
ERROR opening [/usr/lib/host-ipmid/H��dH�%(]: /usr/lib/host-ipmid/H��dH�%(: cannot open shared object file: No such file or directory
*** Error in `./ipmid': munmap_chunk(): invalid pointer: 0x0000000000401eb6 ***
Aborted (core dumped)
We need to check for the error case of a scandir().
Signed-off-by: Jeremy Kerr <jk@ozlabs.org>
|
| |
|
|
|
|
|
|
|
|
|
| |
With the updated HostIpmi dbus interface, netfns are not packed with
LUNs, and the CC is separate from the data. This means we no longer
have to define NETFN_* as arbitrarily shifted, and we have the added
bonus of replying to messages where lun != 0.
We may want to separate the CC returned from the handlers in future too.
Signed-off-by: Jeremy Kerr <jk@ozlabs.org>
|
| |
|
|
|
|
|
|
|
|
|
|
| |
Currently, the IPMI daemon will only receive messages from a specific
(arbitrary) object path that the debug daemon used.
Instead, we should be listening to any signal that matches our interface
(org.openbmc.HostIpmi.ReceivedMessage). When replying to the message,
we should be using the signal sender info to properly address the
reply method.
Signed-off-by: Jeremy Kerr <jk@ozlabs.org>
|
| |
|
|
|
|
| |
support for hex values with cli argument
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
|
| |
|
|
| |
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
|
| |
|
|
|
|
|
| |
Fixed bug where ipmi sensor readings > offset 7 where not asserted
Added testcases for code
Kept sensors 35,37-3A hardcoded until dbus side decides if they
want to represent them.
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
