Commit moves one #ifdef upwards in the code, so resources are not wasted
pointlessly. There is no need to allocate memory for variables/structs and open
interface, if we're about to quit immediately because LAN+ support is not
enabled.
Error message got changed as well and it gets printed on STDERR where it
belongs!
Commit changes return value of ipmi_fru_get_adjust_size_from_buffer() from void
to int in order to determine whether call was successful or wasn't.
ipmi_fru_upg_ekeying() got re-work as well.
Changes to ipmi_fru_get_adjust_size_from_buffer():
* CHUNK_SIZE removed
* code formatting
* return value void -> int
Changes to ipmi_fru_upg_ekeying():
* check whether pFilename is NULL
* check return values of called functions
* re-work code - return early on error instead of nested if()
* added (error) messages
Commit fixes memory leanks in functions ipmi_fru_set_field_string() and
ipmi_fru_set_field_string_rebuild() in 'lib/ipmi_fru.c'.
Note: it seems to me like these for() cycles do nothing.
Reported-by: Ales Ledvinka
Commit changes function ipmi_lcd_get_platform_model_name() to return integer
instead of void to signal potentional problems/issues.
Reported-by: Ales Ledvinka
Commit changes are related to ipmi_ldc_set_lock() in 'lib/ipmi_delloem.c':
* use 'else if' instead of 'if' since it's the same given the context and
probably better
* use 'rc' variable, since it's there, rather than 'return'
Note: ipmi_lcd_set_lock() would use some re-write.
Reported-by: Ales Ledvinka
Commit changes evaluation order of 'oemtype' which is an input parameter to
ipmi_oem_setup() in 'lib/ipmi_oem.c'. Make the check whether 'oemtype' is NULL
or not one of the first things we do rather than blindly use 'oemtype' in
strncmp() and then check if it isn't NULL. Yes, this is a potential crash
waiting to happen.
Reported-by: Ales Ledvinka
Commit fixes NULL dereference in 'lib/ipmi_sdr.c' by checking whether Sensor is
type of 'full' or 'compact'.
Fix by: Jim Mank
Reported by: Ales Ledvinka
ID: 3601106 - # ipmitool sensor get NACname; output incorrect/inconsistent
After a lengthy discussion with Jim, I'm removing free() of sdr_record_list in
ipmi_sensor_get() as this is freed in 'lib/ipmi_main.c' on/prior to ipmitool's
exit.
Commit re-works output of # ipmitool sensor get; command. It removes calls to
sensor specific functions with call to ipmi_sdr_print_listentry(). This way,
output is unified with the output of '-v sdr list' and actually has corrent
information. Well, at least errors are gone and no information seems to be
missing.
Commit fixes help and return codes as well.
Commit fixes crash in ipmi_print_sensor_info() in 'lib/ipmi_dcmi.c' due to NULL
reference. 'rec' is correctly checked whether NULL or not, but it's used
immediately despite of the outcome of the NULL test.
Also, ipmi_sdr_print_rawentry() is used instead of specific print function.
ID: 3597782 - 'lib/ipmi_mc' - sysinfo_param() has two consecutive returns
ID: 3597781 - 'mc getsysinfo|setsysinfo' help has typos
Commit changes/fixes issues described in tickets #3598203, #3597782 and
#3597781.
Typos in help text of 'getsysinfo'/'setsysinfo' got fixed and function
printf_sysinfo_usage() has been created.
Consecutive return in function sysinfo_param() has been removed.
Some code has been re-worked:
* ipmi_getsysinfo() -> ipmi_mc_getsysinfo(),
* ipmi_setsysinfo() -> ipmi_mc_setsysinfo();
* ipmi_sysinfo_main() extended for 'is_set' parameter to signal whether it's
a get/set operation
* return codes changed
* code styling clean-up
* a bit of code clean-up in ipmi_sysinfo_main(), ipmi_mc_getsysinfo(),
ipmi_mc_setsysinfo()
Commit fixes possible NULL reference in str2*() functions in 'lib/helper.c'.
Although it's less likely to happen in ipmitool's daily life, it's still
possible.
Let's make sure it doesn't happen.
Commit fixes output printed by # ipmitool sdr info; about 'SDR Repository Alloc
info supported' due to incorrect member name; probably because of copy-paste
error.
Commit fixes return code of # ipmitool sdr list|elist; when invalid
input/unknown command is given.
Return code is > 0 now and error message is printed.
