Move #ifdef from the inside to the outside of assert()'s
to get rid of this compiler warning:
warning: embedding a directive within macro arguments has undefined
behavior [-Wembedded-directive]
Signed-off-by: Alexander Amelkin <alexander@amelkin.msk.ru>
Refactor plugin ipmi_intf structure initialization for some
plugins to stop the compiler complaining like this:
warning: use of GNU old-style field designator
extension [-Wgnu-designator]
Signed-off-by: Alexander Amelkin <alexander@amelkin.msk.ru>
GitHub Actions macOS environment now has Xcode 12.0.1. [1]
Xcode 12 Clang enables -Werror=implicit-function-declaration by
default [2] and breaks the ipmitool build on macOS. [3]
[1] https://github.com/actions/virtual-environments/issues/1712
[2] https://developer.apple.com/documentation/xcode-release-notes/xcode-12-release-notes
"Clang now reports an error when you use a function without an
explicit declaration when building C or Objective-C code for macOS
(-Werror=implicit-function-declaration flag is on)."
[3] https://github.com/ipmitool/ipmitool/pull/252/checks?check_run_id=1389463134
"checking for ld used by gcc... /Applications/Xcode_12.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/ld
<snip>
free.c:66:13: error: implicit declaration of function 'getuid' is invalid in C99 [-Werror,-Wimplicit-function-declaration]
if (getuid() != 0) {
^
free.c:170:26: error: implicit declaration of function 'ipmi_get_oem' is invalid in C99 [-Werror,-Wimplicit-function-declaration]
intf->manufacturer_id = ipmi_get_oem(intf);"
^
Signed-off-by: Bing-Hua Wang <binghuawang@ami.com>
Some of the header guard names in ipmitool duplicated with the ones in
freeipmi.
* ipmitool/ipmi_fru.h and freeipmi/fru/ipmi-fru.h both uses IPMI_FRU_H
* ipmitool/ipmi_sdr.h and freeipmi/sdr/ipmi-sdr.h both uses IPMI_SDR_H
* ipmitool/ipmi_sel.h and freeipmi/sel/ipmi-sel.h both uses IPMI_SEL_H
This is problematic as including the 1st will prevent the 2nd from being
included, leading to the loss of any declarations, inline definitions,
or other "#includes" in the 2nd header.
For example, including ipmitool/ipmi_sel.h and freeipmi/freeipmi.h fails
to build.
$ cat test.c
#include <ipmitool/ipmi_sel.h>
#include <freeipmi/freeipmi.h>
$ gcc test.c -I${HOME}/src/oss/ipmitool/include
In file included from /usr/include/freeipmi/freeipmi.h:90,
from test.c:2:
/usr/include/freeipmi/sdr/oem/ipmi-sdr-oem-intel-node-manager.h:48:44: error: unknown type name 'ipmi_sdr_ctx_t'; did you mean 'ipmi_fru_ctx_t'?
48 | int ipmi_sdr_oem_parse_intel_node_manager (ipmi_sdr_ctx_t ctx,
| ^~~~~~~~~~~~~~
| ipmi_fru_ctx_t
Remove header guards and use "#pragma once" instead. In this way, we
don't have to do manual management to avoid name clashes.
* src/plugins/lan/md5.h is left as is being an external header
originally.
* src/plugins/imb/imbapi.h is a convoluted header consisting of multiple
header guards. Let's just add "#pragma once" and leave header guards
as is for now.
Signed-off-by: Bing-Hua Wang <binghuawang@ami.com>
The following commands is used on Linux to convert line endings to LF.
git config core.autocrlf input
git add --renormalize .
git config --unset core.autocrlf
It turns out there was only ipmi_kontronoem.h not using LF.
Signed-off-by: Bing-Hua Wang <binghuawang@ami.com>
This patch has been added to fix "error: comparison between signed
and unsigned integer expressions [-Werror=sign-compare]"
Signed-off-by: Shitalkumar Gandhi <shitalkumar.gandhi@seagate.com>
When building an RPM, --with-kerneldir is passed to ./configure. This adds
the include path /lib/modules/<kversion>/build/include that may break the
build but is not present when building with "make" so useless.
For instance while building ipmi_sel.c with kernel 5.3.7-301.fc31.x86_64 :
/lib/modules/5.3.7-301.fc31.x86_64/build/include/linux/stddef.h:11:2: error:
expected identifier before numeric constant
11 | false = 0,
| ^~~~~
Signed-off-by: Gilles Buloz <gilles.buloz@kontron.com>
imbapi.c:1393:52: warning: ISO C does not support ‘__FUNCTION__’ predefined identifier [-Wpedantic]
lprintf(LOG_DEBUG, "%s: munmap(0x%x,%d) success", __FUNCTION__,
Signed-off-by: Heiko Thiery <heiko.thiery@gmail.com>
- Drop a useless `extern` definition
- Remove static variable `errorCount` from header, move it to
the c file where it is actually used if needed
Related to ipmitool/ipmitool#220
Signed-off-by: Alexander Amelkin <alexander@amelkin.msk.ru>
Fix calculation of the buffer size for decoded 6-bit ASCII
strings. Previously the program could allocate too a short buffer
that caused buffer overflows and segmentation fault crashes on
certain FRU contents.
Signed-off-by: Alexander Amelkin <alexander@amelkin.msk.ru>
Restore using strncmp() for "options=" and similar
substrings.
Resolvesipmitool/ipmitool#223
Signed-off-by: Alexander Amelkin <alexander@amelkin.msk.ru>
- Initialize the log at the start of ipmi_main() to allow for proper
logging at the start;
- Remove the unused log_level_get() function;
- Update log_level_set() to take verbosity instead of log level
(default verbosity is 0, which is LOG_NOTICE log level),
use the function to update log level as `-v` is encountered
in command line;
- Move IANA PEN list debugging to verbosity 6. The list is too long
to see it in each debug output of verbosity 5 that is used for
debugging lan/lanplus packets;
- For verbosity >= 2 (that is `-vv`) add the ipmitool version
information at the start.
Signed-off-by: Alexander Amelkin <alexander@amelkin.msk.ru>
Simplified target "rpm" of Makefile, and removed rpmrc and rpmmacros to use
default settings and those from the spec file such as Distribution that has
been changed to "GitHub Build" instead of "Sourceforge Build".
Signed-off-by: Gilles Buloz <gilles.buloz@kontron.com>
The command "make rpm" was failing for the following reasons :
- setting the build directory to ./rpmbuild/ instead of default
~/rpmbuild/ was ignored.
- errors in the spec file : wrong dates and characters in the
changelog, and unpackaged files under misc/*
- debug package built even if unneeded but failing to build because
of missing requisites.
Using "rpmbuild -tb ipmitool-*.tar.gz" instead of 'make rpm" builds
under ~/rpmbuild/ but also fails because of the last two reasons.
Signed-off-by: Gilles Buloz <gilles.buloz@kontron.com>
The documentation now contains instructions on how to build on
Windows. The instructions have been verified and added to
a continous integration workflow with the previous commit,
so Windows buildability will from now on be automatically verified.
Resolves: ipmitool/ipmitool#148 (ipmitool for win with FreeIPMI)
ipmitool/ipmitool#200 (windows build error)
ipmitool/ipmitool#201 (windows build error)
ipmitool/ipmitool#203 (no lanplus on windows)
Signed-off-by: Alexander Amelkin <alexander@amelkin.msk.ru>
Update according to the new GitHub CI build configuration.
Add information regarding prerequisites for macOS X 10.15.
Signed-off-by: Alexander Amelkin <alexander@amelkin.msk.ru>
Neither of these systems have the `-t` option for `install`.
For Linux the option is not necessary and is only needed
if the target directory is specified before the source file(s).
Also, macOS produces strange files with -e suffix for man
pages when they are processed with sed using AC_CONFIG_FILES().
Move prefix expansion to another point in code to avoid these
unneeded intermediate files produced by configure.
Resolves: ipmitool/ipmitool#206
Signed-off-by: Alexander Amelkin <alexander@amelkin.msk.ru>
The `make distcheck` was failing because of some files deleted twice
and some intermediate files not deleted at all.
Signed-off-by: Alexander Amelkin <alexander@amelkin.msk.ru>
Remove a bashims from configure to make build compatible
with systems without bash.
Resolvesipmitool/ipmitool#205
Signed-off-by: Alexander Amelkin <alexander@amelkin.msk.ru>
Unify the comparison idioms use.
Always use `if(!strcmp())` for "if string equals"
and `if(strcmp())` for "if string is not equal".
Never use `== 0` and `!= 0` with `strcmp()`.
Minor reformatting of the code immediately surrounding the
refactored lines.
Resolvesipmitool/ipmitool#104
Signed-off-by: Alexander Amelkin <alexander@amelkin.msk.ru>
Reduce code duplication by extracting option names, types,
and value ranges into a separate structure, and rewriting
the option parsing code without mixing it with the data.
Signed-off-by: Alexander Amelkin <alexander@amelkin.msk.ru>
Clean up use of strcmp/strncmp/strncasecmp for command line arguments.
Never use anything but `strcmp()` unless absolutely neccessary.
Partialy resolvesipmitool/ipmitool#104
The current behavior:
- correct date format is not accepted by "sel time set".
- incorrect date format that looks correct is accepted, but time is not
set correctly.
- commands like
ipmitool sel time set "11/22/2013 trash"
are accepted.
Final part of the fixes for CVE-2020-5208, see
https://github.com/ipmitool/ipmitool/security/advisories/GHSA-g659-9qxw-p7cp
9 variants of stack buffer overflow when parsing `id_string` field of
SDR records returned from `CMD_GET_SDR` command.
SDR record structs have an `id_code` field, and an `id_string` `char`
array.
The length of `id_string` is calculated as `(id_code & 0x1f) + 1`,
which can be larger than expected 16 characters (if `id_code = 0xff`,
then length will be `(0xff & 0x1f) + 1 = 32`).
In numerous places, this can cause stack buffer overflow when copying
into fixed buffer of size `17` bytes from this calculated length.
Partial fix for CVE-2020-5208, see
https://github.com/ipmitool/ipmitool/security/advisories/GHSA-g659-9qxw-p7cp
The `ipmi_spd_print_fru` function has a similar issue as the one fixed
by the previous commit in `read_fru_area_section`. An initial request is
made to get the `fru.size`, which is used as the size for the allocation
of `spd_data`. Inside a loop, further requests are performed to get the
copy sizes which are not checked before being used as the size for a
copy into the buffer.
Partial fix for CVE-2020-5208, see
https://github.com/ipmitool/ipmitool/security/advisories/GHSA-g659-9qxw-p7cp
The `read_fru_area_section` function only performs size validation of
requested read size, and falsely assumes that the IPMI message will not
respond with more than the requested amount of data; it uses the
unvalidated response size to copy into `frubuf`. If the response is
larger than the request, this can result in overflowing the buffer.
The same issue affects the `read_fru_area` function.
Fix ipmitool not writing the last block of boot mailbox data if the
block is shorter than 3 bytes.
Signed-off-by: Ivan Mikhaylov <fr0st61te@gmail.com>
Signed-off-by: Alexander Amelkin <alexander@amelkin.msk.ru>
