Commit handles possible int *flow via user input. atol()/make_int() calls
got replaced and, at least, 'errno' is being checked after strtol() calls
while reading data from file.
Reported-by: Duncan Idaho
Commit fixes fd leaks in 'lib/ipmi_ekananlyzer.c' and 'lib/ipmi_fwum.c'.
Commit fixes file descriptor leaks in ipmi_ekanalyzer_fru_file2structure(),
source 'lib/ipmi_ekanalyzer.c', and in KfwumGetFileSize(),
source 'lib/ipmi_fwum.c'.
Reported-by: dcb
Commit fixes corruption in 'lan alert print' output. This bug comes from two
consecutive calls to get_lan_param_select() which returns pointer to struct. In
the end, the second call would over-write data from the first one, as 'ptype'
and 'paddr' were pointing at the same address.
Thanks to Rob Swindell for logging this bug and testing the patch.
Provides configuration script, systemd unit and sysconfig defaults.
On hosts running net-snmp and PEF alerting on BMC:
1. Redirect SNMP get requests to BMC (Service Processor).
2. Redirect SNMP Traps from BMC (Service Processor) to the Hosts' configured Trap sink.
Commit fixes incorrect array index in get_lan_param_select(). The param value
matches the cmd in the lan_param struct, but it's not the same as the
ipmi_lan_params[] array index. Here is an example of what I mean. This problem
becomes apparent with the OEM extensions which use values 192+.
Commit for Jeff Bastian
Commit fixes segfault in 'set' sub-command when no IPMI interface is
present/available. This is caused by 'intf->session' being used without check
whether it's NULL or not which leads to NULL reference in printf().
Fixed the hpm upgstatus command to print the results of the command
without having to specify -v to enable verbose operation.
It's not clear that this command is even useful to the user at all.
Perhaps it should be removed entirely.
Commit for Dan Gora
Update the help menu for the ipmitool hpm command. Fix the formatting
and explain better the options.
Changed the option 'upgrade <file> all' to 'upgrade <file> force'.
This makes it clearer that the user is overriding the version check
that 'upgrade <file>' performs.
Commit for Dan Gora
Added a new field to show the Deferred firmware version value as part
of the 'hpm check' subcommand.
The deferred firmware version is the version of the firmware which has
been loaded with the 'hpm upgrade' command, but not yet activated.
Commit for Dan Gora
Fixed a bug where the rolling percentage display could corrupt the
version line if the first call to HpmDisplayUpgrade resulted in a
non-zero percentage.
This was produced by performing:
ipmitool hpm upgrade hpm1fru.img
where hpm1fru.img is a small (820 byte) file.
We now call HpmDisplayUpgrade before we transfer any data to ensure
that we get 0% printed before the transfer starts.
Commit for Dan Gora
Added a '^' character to the ID field which will be set if the
component can be upgraded with the specified file with the 'hpm check
<file>' subcommand.
Commit for Dan Gora
Update HpmfwupgPreUpgradeCheck to check if the Aux version is also
different when it compares the Image version to the active and
rollback versions.
Fixed the indenting and style of HpmfwupgPreUpgradeCheck to make it
more readable.
Commit for Dan Gora
Commit squashes white spaces in 'lib/ipmi_dcmi.c'. Tabulators are used for
indentation now. I've tried to clean up comments, long lines, trailing white
spaces and tabs, code formatting. However, more work is needed especially in
ipmi_dcmi_main() where indentation level is reaching one too many.
No functional changes.
gethostbyname() can't handle IPv6. Despite this fact it seems worth of
nothing to check whether gethostbyname() really returned AF_INET or
not. And that's what attached patch does, resp. in case IPv6 is
returned then ipmitool shoud/will terminate.
