| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
| |
automake does not seem to include noinst_*_MANS in the dist tarball.
Signed-off-by: Mike Gilbert <floppym@gentoo.org>
|
|\
| |
| | |
Fixes for #411 and #413
|
| |
| |
| |
| |
| |
| | |
Adding curly braces to fix the code.
Signed-off-by: Eivind Naess <eivnaes@yahoo.com>
|
| |
| |
| |
| |
| |
| | |
Fixes issue #411.
Signed-off-by: Eivind Naess <eivnaes@yahoo.com>
|
|/
|
|
| |
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
|
|
|
|
|
|
| |
History from the 2.4.x series is moved to Changes-2.4.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
|
|
|
|
|
|
|
| |
Use unsigned constants to avoid integer overflows when shifting, and
because xmit_accm[] is an array of unsigned values.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
|
|
|
|
|
|
|
|
|
|
|
|
| |
In rc_avpair_insert, if the list element "p" is non-NULL but not
actually in the list "a", we can end up with this_node being NULL and
being dereferenced.
By changing the while test to this_node->next we avoid having
this_node being NULL; the loop will terminate when this_node == p or
this_node->next == NULL, which is what we want.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
|
|
|
|
| |
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
|
|
|
|
|
|
|
|
| |
They aren't used (pppdump only needs decompression), and removing the
unused code avoids getting reports from automated tools about possible
errors in the unused code.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
|
|
|
|
|
|
| |
* CI: add automatic tests, enable ASAN build&test
Signed-off-by: Ilya Shipitsin <chipitsine@gmail.com>
|
|\
| |
| | |
Additional fixes for various configure options (less frequently used) that broke
|
| |
| |
| |
| |
| |
| | |
Correcting the call to DesEncrypt() where needed.
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This change fixes the build when
- ./configure is run with --disable-plugins
- ./configure is run with --disable-peap --disable-eaptls
--disable-microsoft-extensions
The latter disables the MPPE encryption too, but <pppd/crypto.h> is
still needed.
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
|
| |
| |
| |
| |
| |
| |
| | |
You can't reference variables that aren't included in the compile.
Add #ifdef PPP_WITH_IPV6CP in the get_notifier_by_type function.
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
|
| |
| |
| |
| |
| | |
Allow distributions to specify a default CA path. Fix for github issue #405.
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
|
|\ \
| | |
| | | |
Removes autotools include guards for stdbool.h and friends
|
| | |
| | |
| | |
| | | |
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
|
| |/
| |
| |
| | |
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
|
|/
|
|
|
|
|
| |
(#401)
Signed-off-by: Wei Xing <skyxwwalker@gmail.com>
Co-authored-by: Wei Xing <skyxwwalker@gmail.com>
|
|
|
|
|
|
|
|
| |
Running Valgrind memcheck tool on chat.c while communicating with a
modem found additional memory leaks. These fixes plumb those.
Signed-off-by: Robert Bartel <r.bartel@gmx.net>
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
|
|
|
|
|
|
|
|
|
|
|
|
| |
There were two issues here, the report_buffer is too small to hold the
value, and accessing the memory outside its bounds. The following fixes
was made:
- Expand the size of report_buffer to 4096 from 256, this is to account
for handling of really long GSM USSD report strings
- Make sure to not to access memory outside the bounds of the buffer
Signed-off-by: Robert Bartel <r.bartel@gmx.net>
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Fix for coverity issue 436265, we should cap copy to size of destination buffer
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
* Fix for coverity issue 436262, llv6_ntoa() returns a pointer to a buffer that can be up to 64 bytes long; likely not a problem, but this will quiet coverity
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
* Fix for coverity issue 436251, not freeing path in the normal flow of the code
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
* Fixing coverity issue #436258, Digest maybe uninitialized in some paths of this code
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
* Fix for coverity issue 436254, forgot to free 's' before returning from the function?
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
* Fixing coverity issue #436251, memory leak in put_string() function
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
* Fixing coverity issue 436215, should copy at most sizeof(devname) bytes
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
* Fixing coverity issue #436203, if no authentication (or no accounting) server was found, we still need to free the allocated local instance
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
* Fixing coverity issue #436171, use of uninitialized variable
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
* Use of signed vs unsigned variable in printf for MD4Update
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
* Fixing coverity issue #436182, fixing possible buffer overrun in handling of PW_CLASS attribute
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
* Fixing coverity issue #436156
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
* Compile errors
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
[paulus@ozlabs.org - Squashed to avoid breaking bisection]
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
|
|\
| |
| | |
Few more changes before 2.5.0 release
|
| |
| |
| |
| | |
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
|
| |
| |
| |
| | |
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
|
| |
| |
| |
| | |
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
|
| |
| |
| |
| |
| |
| | |
additions by Paul Mackerras
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
|
| |
| |
| |
| | |
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
|
| |
| |
| | |
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
|
|/
|
|
|
|
|
|
|
|
|
|
| |
* Add 'const' parameter to input arguments in crypto_ms.*
* Round ZPasswordHash buffers up to 24 bytes, as the DES MakeKey() function
accesses ZPasswordHash[21]
Closes github issue #392
[paulus@ozlabs.org - tidied up headline and commit message]
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
2.5.0 (#379)
This commit does several things, being a squash-and-merge of a series
of changes; squashed in order not to break bisection.
* Clean up pppd.h, moving declarations that should only be accessed by
pppd code (not by users of pppd) to a new pppd-private.h. Also,
other parts of pppd.h were moved to multilink.h, chap.h, eap.h,
eui64.h, and a new options.h.
* Provide an API for access to data that is needed by plugins (in no
particular order):
- ifname
- ifunit
- remote_name
- remote_number
- peer_authname
- status (now called "code" internally)
- phase
- doing_multilink
- multilink_master
- idle_time_limit
- link_connect_time
- max_connect_time
- link_stats
- ipparam
- hostname
- got_sigterm
- got_sigusr2
- got_sighup
- session_number
- maxoctets
- maxoctets_dir
- debug
- persist
- devnam
- modem
- peer_authname
- sync_serial
* Update the version number to 2.5.0.
* Detect availability of stddef.h and stdarg.h.
* Rename some headers:
- pppcrypt.c/h to crypto_ms.c/h
- ppp-crypto.c/h to crypto.c/h
- ppp-crypto-priv.h to crypto-priv.h
- chap-new.c/h to chap.c/h
* Remove chap-md5.h, crypto-priv.h, eap-tls.h, etc. from the list of
header files to be installed.
* Provide typedefs for the hook functions.
* Provide a typedef for the "phase" variable.
* Provide a typedef for the link statistics array.
* Remove the option_t typedef.
* Rename the following functions by adding a "ppp_" prefix (with the
intention that these are a "public" API for use by plugins):
- option_error
- add_options
- int_option,
- options_from_file
- script_setenv
- bad_ip_adrs,
- netif_get/set_mtu (renamed to ppp_get/set_mtu)
- get_time
- timeout
- untimeout
- safe_fork
- sys_close
- set_session_number
- update_link_stats (renamed to ppp_get_link_stats)
- add_notifier (renamed to ppp_add_notify)
- remove_notifier (renamed to ppp_del_notify)
- generic_[dis]establish_ppp (to ppp_generic_[dis]establish)
* Rename ppp_devnam to ppp_devname.
* Rename ppp_available() to ppp_check_kernel_support().
* Use unsigned char instead of u_char, unsigned short instead of
u_short, uint32_t instead of u_int32_t.
* Add const to some declarations
* Update comments
* Change the interface for notifiers to use an enum to identify which
notifier is to be modified.
* Provide an API for getting the path to a file, with an enum to
identify different types of file.
* Link plugins with the -DPLUGIN flag
[paulus@ozlabs.org - wrote commit message]
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
|
|
|
| |
Signed-off-by: Dianne Skoll <dianne@skoll.ca>
|
|
|
| |
Signed-off-by: lilinjie <lilinjie@uniontech.com>
|
|\
| |
| | |
radius: distinguish between User-Request and Admin-Reset.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
For the purposes of our definition:
User-Request - remote side hanging up.
Admin-Reset - local side hanging up.
Reasoning is that typically radius will be used to authentication
dial-in users, so if the pppd gets killed locally, that's not the User
(client) requesting hangup, but rather the local administrator (be that
a manual kill, or as a result of a CoA/Disconnect).
Signed-off-by: Jaco Kroon <jaco@uls.co.za>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
If the lcp-echo-adaptive option is specified, it means that seeing
received traffic on the link is considered to be an indication that
the link is working. Hence, this resets the count of missing LCP
echo-replies to 0 when traffic is seen. Without this, occasional
echo failures interspersed with link traffic can accumulate and end up
causing a disconnection even when the link is working correctly.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This improves the way that signals are handled in chat.
First, signal handlers should not be calling functions which are not
async-signal-safe; doing so incurs the possibility of deadlock. Thus
we can't call fatal() in signal handlers; instead we set 'fatalsig',
which functions both as a flag and as an indication of which signal
occurred, and check that at various points (basically after any
operation which might block) using the new function checksigs().
Secondly, using sigaction rather than signal() means that we can
control whether calls such as read() get restarted after a signal, and
whether the signal disposition gets reset when the signal is
delivered. That simplifies sigalrm(); we no longer need to
re-register the handler, and we don't need the kludge of setting stdin
to non-blocking mode in order to get the read() in get_char() to
return.
This also removes a #ifdef ultrix since ultrix is no longer supported.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
|
|\ \
| | |
| | |
| | |
| | |
| | |
| | | |
Fix libsystemd detection
Merged despite bogus signoff in the commits being merged, since the changes are trivial.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
|
| | |
| | |
| | |
| | |
| | |
| | | |
And update package indexes, otherwise apt-get might not find the package.
Signed-off-by: Chocobo1 <Chocobo1@users.noreply.github.com>
|
|/ /
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
1. Use uppercase for `prefix` parameter
`SYSTEMD_CFLAGS` is used elsewhere so `prefix` cannot be lowercase.
https://autotools.info/pkgconfig/pkg_check_modules.html
2. The module name should be `libsystemd`
Previously it will result in the following compile error when building pppd/auth.c:
> /usr/bin/ld: pppd-auth.o: undefined reference to symbol 'sd_notify@@LIBSYSTEMD_209'
> /usr/bin/ld: /usr/lib/libsystemd.so.0: error adding symbols: DSO missing from command line
This is due to missing `-lsystemd-daemon` flag which is provided by `libsystemd-daemon-devel`
package on Debian or `systemd-libs` on ArchLinux. And the proper .pc file in the package is
`libsystemd` not `systemd`.
https://stackoverflow.com/a/38303241
Signed-off-by: Chocobo1 <Chocobo1@users.noreply.github.com>
|
|\ \
| | |
| | | |
pppd: Retry registering interface when on rtnetlink -EBUSY error
|
| |/
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Due to workaround in kernel module ppp_generic.ko in function
ppp_nl_newlink(), kernel may return -EBUSY error to prevent possible
mutex deadlock. In this case userspace needs to retry its request.
Proper way would be to fix kernel module to order requests and mutex
locking, so prevent deadlock in kernel and so never return this error to
userspace. Until it happens we need retry code in userspace.
Signed-off-by: Pali Rohár <pali@kernel.org>
|
|\ \
| | |
| | | |
Add option to show all options (show-options)
|
| | |
| | |
| | |
| | |
| | |
| | | |
text to include copyright and package name from autotools.
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
|
|\ \ \
| |_|/
|/| | |
Create an new API that allows us to abstract the use of the standard crypto functions
|
| | |
| | |
| | |
| | |
| | |
| | | |
incarnation DES w.r.t. RFC2759
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
|
| | |
| | |
| | |
| | |
| | |
| | | |
Avoid using these variables in function descriptors.
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
|
| |/
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This re-introduces the missing DES encryption functions copied from Openssl 3.0 project. Incorporates a new API for performing MD4/MD5/SHA and encryption using DES-ECB mode.
Unit tests are included for respective digest/encryption functions using this new API. With this change, you can pass configure --without-openssl to use the internally provided functions. If you do have openssl, then it will default to use these functions. This also provides a framework to allow other vendors to provide crypto.
This closes #333, partially addresses #242 (except the pkcs11 engine support). Word has it that openssl is working on support for this, and the libp11 / opensc project are inclined not to support this.
Signed-off-by: Eivind Næss <eivnaes@yahoo.com>
|