You can subscribe to this list here.
| 2000 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
(14) |
Nov
(315) |
Dec
(298) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2001 |
Jan
(254) |
Feb
(467) |
Mar
(430) |
Apr
(345) |
May
(406) |
Jun
(336) |
Jul
(313) |
Aug
(265) |
Sep
(433) |
Oct
(462) |
Nov
(387) |
Dec
(232) |
| 2002 |
Jan
(352) |
Feb
(556) |
Mar
(463) |
Apr
(500) |
May
(557) |
Jun
(337) |
Jul
(317) |
Aug
(279) |
Sep
(273) |
Oct
(354) |
Nov
(267) |
Dec
(347) |
| 2003 |
Jan
(351) |
Feb
(445) |
Mar
(520) |
Apr
(665) |
May
(499) |
Jun
(393) |
Jul
(304) |
Aug
(425) |
Sep
(262) |
Oct
(329) |
Nov
(220) |
Dec
(174) |
| 2004 |
Jan
(365) |
Feb
(479) |
Mar
(515) |
Apr
(522) |
May
(214) |
Jun
(471) |
Jul
(292) |
Aug
(341) |
Sep
(243) |
Oct
(446) |
Nov
(294) |
Dec
(147) |
| 2005 |
Jan
(171) |
Feb
(209) |
Mar
(218) |
Apr
(321) |
May
(233) |
Jun
(534) |
Jul
(268) |
Aug
(345) |
Sep
(498) |
Oct
(557) |
Nov
(459) |
Dec
(238) |
| 2006 |
Jan
(288) |
Feb
(180) |
Mar
(151) |
Apr
(113) |
May
(164) |
Jun
(277) |
Jul
(160) |
Aug
(383) |
Sep
(221) |
Oct
(404) |
Nov
(358) |
Dec
(163) |
| 2007 |
Jan
(293) |
Feb
(175) |
Mar
(202) |
Apr
(155) |
May
(427) |
Jun
(484) |
Jul
(414) |
Aug
(125) |
Sep
(131) |
Oct
(160) |
Nov
(79) |
Dec
(70) |
| 2008 |
Jan
(133) |
Feb
(115) |
Mar
(158) |
Apr
(194) |
May
(197) |
Jun
(230) |
Jul
(146) |
Aug
(68) |
Sep
(93) |
Oct
(53) |
Nov
(95) |
Dec
(69) |
| 2009 |
Jan
(81) |
Feb
(162) |
Mar
(215) |
Apr
(216) |
May
(78) |
Jun
(131) |
Jul
(61) |
Aug
(176) |
Sep
(127) |
Oct
(28) |
Nov
(83) |
Dec
(94) |
| 2010 |
Jan
(100) |
Feb
(187) |
Mar
(320) |
Apr
(161) |
May
(194) |
Jun
(142) |
Jul
(129) |
Aug
(139) |
Sep
(239) |
Oct
(202) |
Nov
(139) |
Dec
(196) |
| 2011 |
Jan
(195) |
Feb
(191) |
Mar
(201) |
Apr
(127) |
May
(84) |
Jun
(126) |
Jul
(101) |
Aug
(237) |
Sep
(123) |
Oct
(104) |
Nov
(197) |
Dec
(114) |
| 2012 |
Jan
(65) |
Feb
(85) |
Mar
(129) |
Apr
(84) |
May
(94) |
Jun
(83) |
Jul
(89) |
Aug
(85) |
Sep
(89) |
Oct
(73) |
Nov
(34) |
Dec
(38) |
| 2013 |
Jan
(89) |
Feb
(30) |
Mar
(25) |
Apr
(18) |
May
(20) |
Jun
(45) |
Jul
(74) |
Aug
(37) |
Sep
(72) |
Oct
(30) |
Nov
(67) |
Dec
(24) |
| 2014 |
Jan
(23) |
Feb
(16) |
Mar
(40) |
Apr
(37) |
May
(12) |
Jun
(18) |
Jul
(30) |
Aug
(26) |
Sep
(24) |
Oct
(32) |
Nov
(15) |
Dec
(33) |
| 2015 |
Jan
(15) |
Feb
(45) |
Mar
(21) |
Apr
(24) |
May
(22) |
Jun
(7) |
Jul
(57) |
Aug
(17) |
Sep
(16) |
Oct
(3) |
Nov
(8) |
Dec
(13) |
| 2016 |
Jan
(7) |
Feb
(14) |
Mar
(40) |
Apr
(8) |
May
(10) |
Jun
(6) |
Jul
(8) |
Aug
(10) |
Sep
(19) |
Oct
(20) |
Nov
(45) |
Dec
(10) |
| 2017 |
Jan
(10) |
Feb
(12) |
Mar
(3) |
Apr
(17) |
May
(41) |
Jun
(21) |
Jul
(13) |
Aug
(13) |
Sep
(7) |
Oct
(23) |
Nov
(10) |
Dec
(23) |
| 2018 |
Jan
(45) |
Feb
(3) |
Mar
(57) |
Apr
(107) |
May
(173) |
Jun
(47) |
Jul
(28) |
Aug
(26) |
Sep
(38) |
Oct
(56) |
Nov
(22) |
Dec
(11) |
| 2019 |
Jan
(37) |
Feb
(8) |
Mar
(7) |
Apr
(29) |
May
(32) |
Jun
(5) |
Jul
(21) |
Aug
(31) |
Sep
(38) |
Oct
(8) |
Nov
(13) |
Dec
(10) |
| 2020 |
Jan
(9) |
Feb
(33) |
Mar
(14) |
Apr
(4) |
May
(16) |
Jun
(11) |
Jul
(14) |
Aug
(50) |
Sep
(24) |
Oct
(3) |
Nov
(14) |
Dec
(13) |
| 2021 |
Jan
(18) |
Feb
(15) |
Mar
(12) |
Apr
(9) |
May
(9) |
Jun
(8) |
Jul
(6) |
Aug
(7) |
Sep
(26) |
Oct
(17) |
Nov
(6) |
Dec
(2) |
| 2022 |
Jan
(3) |
Feb
(11) |
Mar
(7) |
Apr
(15) |
May
(5) |
Jun
(4) |
Jul
(29) |
Aug
(6) |
Sep
(7) |
Oct
|
Nov
(4) |
Dec
(1) |
| 2023 |
Jan
|
Feb
|
Mar
|
Apr
(10) |
May
(3) |
Jun
(5) |
Jul
(3) |
Aug
(10) |
Sep
(10) |
Oct
(7) |
Nov
(2) |
Dec
(4) |
| 2024 |
Jan
(22) |
Feb
(5) |
Mar
(11) |
Apr
(20) |
May
(16) |
Jun
(9) |
Jul
(14) |
Aug
(5) |
Sep
(7) |
Oct
(4) |
Nov
(3) |
Dec
|
| 2025 |
Jan
(6) |
Feb
(6) |
Mar
(14) |
Apr
(2) |
May
|
Jun
(2) |
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
(9) |
| 2026 |
Jan
(7) |
Feb
(2) |
Mar
(3) |
Apr
|
May
|
Jun
(7) |
Jul
|
Aug
(2) |
Sep
(1) |
Oct
(1) |
Nov
(13) |
Dec
|
|
From: Stuart H. <st...@sp...> - 2026-11-16 18:53:38
|
On 2026/11/16 08:57, Bart Van Assche wrote: > On 11/16/26 7:04 AM, Stuart Henderson wrote: > > On 2026/11/16 06:48, Wes Hardaker wrote: > > > Stuart Henderson <st...@sp...> writes: > > > > > > > - The stricter OID_LENGTH macro using __builtin_types_compatible_p > > > > in 589e9c5 results in build failures when used from C++ code. > > > > > > What we should do, IMHO, is not not use any C++ code at all anywhere. > > > It would be best to rewrite the macro to use C compatible conventions > > > entirely. > > > > > > If there is something truly beneficial that helps C++ code (or > > > compilers), then the ifdef you mentioned is better than nothing. > > > > > > We do know that many people do use non-C++ compilers, so this is another > > > show stopper in my eyes. > > > > Other way round, it works with a C compiler but not C++. > > The OID_LENGTH() macro on the master branch should again support both C > and C++. Further feedback is welcome. > > Thanks, > > Bart. Thank you. I didn't run into any other problems (built various things including sane-backends, php, lldpd, monitoring-plugins, zabbix, hplip, nut, collectd, asterisk with SNMP support). |
|
From: Wes H. <har...@us...> - 2026-11-16 17:45:27
|
Bart Van Assche <bva...@ac...> writes: > The OID_LENGTH() macro on the master branch should again support both C > and C++. Further feedback is welcome. Thank you!!! -- Wes Hardaker Please mail all replies to net...@li... |
|
From: Bart V. A. <bva...@ac...> - 2026-11-16 16:57:21
|
On 11/16/26 7:04 AM, Stuart Henderson wrote: > On 2026/11/16 06:48, Wes Hardaker wrote: >> Stuart Henderson <st...@sp...> writes: >> >>> - The stricter OID_LENGTH macro using __builtin_types_compatible_p >>> in 589e9c5 results in build failures when used from C++ code. >> >> What we should do, IMHO, is not not use any C++ code at all anywhere. >> It would be best to rewrite the macro to use C compatible conventions >> entirely. >> >> If there is something truly beneficial that helps C++ code (or >> compilers), then the ifdef you mentioned is better than nothing. >> >> We do know that many people do use non-C++ compilers, so this is another >> show stopper in my eyes. > > Other way round, it works with a C compiler but not C++. The OID_LENGTH() macro on the master branch should again support both C and C++. Further feedback is welcome. Thanks, Bart. |
|
From: Stuart H. <st...@sp...> - 2026-11-16 15:04:42
|
On 2026/11/16 06:48, Wes Hardaker wrote: > Stuart Henderson <st...@sp...> writes: > > > - The stricter OID_LENGTH macro using __builtin_types_compatible_p > > in 589e9c5 results in build failures when used from C++ code. > > What we should do, IMHO, is not not use any C++ code at all anywhere. > It would be best to rewrite the macro to use C compatible conventions > entirely. > > If there is something truly beneficial that helps C++ code (or > compilers), then the ifdef you mentioned is better than nothing. > > We do know that many people do use non-C++ compilers, so this is another > show stopper in my eyes. Other way round, it works with a C compiler but not C++. |
|
From: Wes H. <har...@us...> - 2026-11-16 14:48:42
|
Stuart Henderson <st...@sp...> writes: > - The stricter OID_LENGTH macro using __builtin_types_compatible_p > in 589e9c5 results in build failures when used from C++ code. What we should do, IMHO, is not not use any C++ code at all anywhere. It would be best to rewrite the macro to use C compatible conventions entirely. If there is something truly beneficial that helps C++ code (or compilers), then the ifdef you mentioned is better than nothing. We do know that many people do use non-C++ compilers, so this is another show stopper in my eyes. -- Wes Hardaker Please mail all replies to net...@li... |
|
From: Stuart H. <st...@sp...> - 2026-11-16 13:54:42
|
On 2026/11/16 09:49, Stuart Henderson wrote: > On 2026/11/15 14:11, Bart Van Assche wrote: > > On 11/15/26 9:39 AM, Wes Hardaker wrote: > > > Bart Van Assche <bva...@ac...> writes: > > > > > > > That's unfortunate. > > > > > > I suspect a read-only failed to build bug is a big enough one to warrant > > > another RC with the fix in it. Specifically, a good number of people do > > > build with --read-only. > > > > Stuart, the read-only build should have been fixed on the master branch. > > Please help with retesting the master branch. > > Confirmed, read-only + standard builds are now working for me. I've now run builds of ports which depend on net-snmp and ran into two issues: - snmp_sess_error() changed from void * to struct session_list * in 2f22405, I noticed because NUT failed to build - I think it's their error and they should have used snmp_error not snmp_sess_error in https://github.com/networkupstools/nut/blame/master/drivers/snmp-ups.c#L1792 - if so then that's not a problem for Net-SNMP to deal with, however while investigating I found that the netsnmp_sess_api.3 manpage still uses void *. - The stricter OID_LENGTH macro using __builtin_types_compatible_p in 589e9c5 results in build failures when used from C++ code. https://gcc.gnu.org/onlinedocs/gcc/Other-Builtins.html says that the builtin is only available from C. Should the conditional have && !defined(__cplusplus) ? Seen here: https://github.com/willamowius/gnugk/blob/d0e15e12a20a0c769ab5d999bb2cf6f70d7e2772/snmp.cxx#L87 clang: snmp.cxx:87:11: error: expected '(' for function-style cast or type construction 87 | trapOID[ OID_LENGTH(trapOID) - 1 ] = trapNumber; | ^~~~~~~~~~~~~~~~~~~ /usr/local/include/net-snmp/library/asn1.h:65:56: note: expanded from macro 'OID_LENGTH' 65 | sizeof(int[-__builtin_types_compatible_p(typeof(x), typeof(&(x)[0]))])) | ~~~~~~~~~^ gcc: snmp.cxx: In function 'void SendNetSNMPTrap(unsigned int, SNMPLevel, SNMPGroup, const PString&)': /usr/local/include/net-snmp/library/asn1.h:65:18: error: '__builtin_types_compatible_p' was not declared in this scope 65 | sizeof(int[-__builtin_types_compatible_p(typeof(x), typeof(&(x)[0]))])) |
|
From: Stuart H. <st...@sp...> - 2026-11-16 09:49:50
|
On 2026/11/15 14:11, Bart Van Assche wrote: > On 11/15/26 9:39 AM, Wes Hardaker wrote: > > Bart Van Assche <bva...@ac...> writes: > > > > > That's unfortunate. > > > > I suspect a read-only failed to build bug is a big enough one to warrant > > another RC with the fix in it. Specifically, a good number of people do > > build with --read-only. > > Stuart, the read-only build should have been fixed on the master branch. > Please help with retesting the master branch. Confirmed, read-only + standard builds are now working for me. |
|
From: Bart V. A. <bva...@ac...> - 2026-11-15 22:11:32
|
On 11/15/26 9:39 AM, Wes Hardaker wrote: > Bart Van Assche <bva...@ac...> writes: > >> That's unfortunate. > > I suspect a read-only failed to build bug is a big enough one to warrant > another RC with the fix in it. Specifically, a good number of people do > build with --read-only. Stuart, the read-only build should have been fixed on the master branch. Please help with retesting the master branch. Thanks, Bart. |
|
From: Wes H. <har...@us...> - 2026-11-15 17:39:34
|
Bart Van Assche <bva...@ac...> writes: > That's unfortunate. I suspect a read-only failed to build bug is a big enough one to warrant another RC with the fix in it. Specifically, a good number of people do build with --read-only. -- Wes Hardaker Please mail all replies to net...@li... |
|
From: Bart V. A. <bva...@ac...> - 2026-11-11 17:32:25
|
On 11/11/26 7:08 AM, Stuart Henderson wrote: > At least with clang, build does error out. I have no opinion whether > it's important or not though, so no problem as far as I'm concerned > to wait until after 5.10. > > cc -MMD -MP -MF snmp_api.d -MT snmp_api.lo -I../include -I. -I../snmplib -D_GNU_SOURCE -D_ALL_SOURCE -D_THREAD_SAFE -D__EXTENSIONS__ -O2 -pipe -g -DNETSNMP_ENABLE_IPV6 -fno-strict-aliasing -DNETSNMP_REMOVE_U64 -O2 -pipe -g -Uopenbsd8 -Dopenbsd8=openbsd8 -DNO_LOCALE_NUMERIC -DNO_LOCALE_COLLATE -DNO_LOCALE_NUMERIC -DNO_LOCALE_COLLATE -DNO_LOCALE_MONETARY -DNO_LOCALE_TIME -DNO_LOCALE_MESSAGES -DLIBC_HANDLES_MISMATCHED_CTYPE -fno-strict-aliasing -fno-delete-null-pointer-checks -pipe -fstack-protector-strong -I/usr/local/include -I/usr/libdata/perl5/amd64-openbsd/CORE -c snmp_api.c -fPIC -DPIC -o .libs/snmp_api.o > snmp_api.c:3076:9: error: fallthrough annotation does not directly precede switch label > 3076 | NETSNMP_FALLTHROUGH; > | ^ > ../include/net-snmp/net-snmp-config.h:2449:33: note: expanded from macro 'NETSNMP_FALLTHROUGH' > 2449 | # define NETSNMP_FALLTHROUGH __attribute__((__fallthrough__)) > | ^ > 1 error generated. That's unfortunate. The changes I think that are needed to fix the compiler warnings for read-only mode are available here: https://github.com/bvanassche/net-snmp/commits/master/ Bart. |
|
From: Stuart H. <st...@sp...> - 2026-11-11 15:08:59
|
On 2026/11/11 07:01, Bart Van Assche wrote:
> On 11/11/26 3:28 AM, Stuart Henderson wrote:
> > Probably not a show-stopper, but this fixes --enable-read-only build:
> >
> > Index: snmplib/snmp_api.c
> > --- snmplib/snmp_api.c.orig
> > +++ snmplib/snmp_api.c
> > @@ -3069,11 +3069,12 @@ _snmp_build(u_char ** pkt, size_t * pkt_len, size_t *
> > switch (pdu->command) {
> > case SNMP_MSG_RESPONSE:
> > netsnmp_assert(0 == (pdu->flags & UCD_MSG_FLAG_EXPECT_RESPONSE));
> > -#ifndef NETSNMP_NOTIFY_ONLY
> > +#if !defined(NETSNMP_NOTIFY_ONLY) && !defined(NETSNMP_NO_WRITE_SUPPORT)
> > NETSNMP_FALLTHROUGH;
> > +#endif
> > +#ifndef NETSNMP_NOTIFY_ONLY
> > case SNMP_MSG_GET:
> > case SNMP_MSG_GETNEXT:
> > - NETSNMP_FALLTHROUGH;
> > #endif /* ! NETSNMP_NOTIFY_ONLY */
> > #ifndef NETSNMP_NO_WRITE_SUPPORT
> > case SNMP_MSG_SET:
>
> Let's keep any compiler warning fixes for after the 5.10 release.
>
> Additionally, more changes are required than the above to fix the
> compiler warnings for the read-only build.
>
> Thanks,
>
> Bart.
At least with clang, build does error out. I have no opinion whether
it's important or not though, so no problem as far as I'm concerned
to wait until after 5.10.
cc -MMD -MP -MF snmp_api.d -MT snmp_api.lo -I../include -I. -I../snmplib -D_GNU_SOURCE -D_ALL_SOURCE -D_THREAD_SAFE -D__EXTENSIONS__ -O2 -pipe -g -DNETSNMP_ENABLE_IPV6 -fno-strict-aliasing -DNETSNMP_REMOVE_U64 -O2 -pipe -g -Uopenbsd8 -Dopenbsd8=openbsd8 -DNO_LOCALE_NUMERIC -DNO_LOCALE_COLLATE -DNO_LOCALE_NUMERIC -DNO_LOCALE_COLLATE -DNO_LOCALE_MONETARY -DNO_LOCALE_TIME -DNO_LOCALE_MESSAGES -DLIBC_HANDLES_MISMATCHED_CTYPE -fno-strict-aliasing -fno-delete-null-pointer-checks -pipe -fstack-protector-strong -I/usr/local/include -I/usr/libdata/perl5/amd64-openbsd/CORE -c snmp_api.c -fPIC -DPIC -o .libs/snmp_api.o
snmp_api.c:3076:9: error: fallthrough annotation does not directly precede switch label
3076 | NETSNMP_FALLTHROUGH;
| ^
../include/net-snmp/net-snmp-config.h:2449:33: note: expanded from macro 'NETSNMP_FALLTHROUGH'
2449 | # define NETSNMP_FALLTHROUGH __attribute__((__fallthrough__))
| ^
1 error generated.
|
|
From: Stuart H. <st...@sp...> - 2026-11-11 11:47:34
|
On 2026/11/07 10:34, Wes Hardaker via Net-snmp-coders wrote: > > We're starting down the final stretch of releasing Net-SNMP 5.10. The > first (and maybe last?) release-candidate version is available here: > > https://sourceforge.net/projects/net-snmp/files/net-snmp/5.10-pre-releases/ > > At this point only security fixes or show-stopper patches that have been > agreed to by multiple developers will be included in future release > candidates. If no future patches seem needed, then we'll publish 5.10 > in a few weeks after enough people have had the chance to test it. Probably not a show-stopper, but this fixes --enable-read-only build: Index: snmplib/snmp_api.c --- snmplib/snmp_api.c.orig +++ snmplib/snmp_api.c @@ -3069,11 +3069,12 @@ _snmp_build(u_char ** pkt, size_t * pkt_len, size_t * switch (pdu->command) { case SNMP_MSG_RESPONSE: netsnmp_assert(0 == (pdu->flags & UCD_MSG_FLAG_EXPECT_RESPONSE)); -#ifndef NETSNMP_NOTIFY_ONLY +#if !defined(NETSNMP_NOTIFY_ONLY) && !defined(NETSNMP_NO_WRITE_SUPPORT) NETSNMP_FALLTHROUGH; +#endif +#ifndef NETSNMP_NOTIFY_ONLY case SNMP_MSG_GET: case SNMP_MSG_GETNEXT: - NETSNMP_FALLTHROUGH; #endif /* ! NETSNMP_NOTIFY_ONLY */ #ifndef NETSNMP_NO_WRITE_SUPPORT case SNMP_MSG_SET: |
|
From: Wes H. <har...@us...> - 2026-11-07 18:34:59
|
We're starting down the final stretch of releasing Net-SNMP 5.10. The first (and maybe last?) release-candidate version is available here: https://sourceforge.net/projects/net-snmp/files/net-snmp/5.10-pre-releases/ At this point only security fixes or show-stopper patches that have been agreed to by multiple developers will be included in future release candidates. If no future patches seem needed, then we'll publish 5.10 in a few weeks after enough people have had the chance to test it. -- Wes Hardaker Please mail all replies to net...@li... |
|
From: Wes H. <har...@us...> - 2026-10-15 06:13:09
|
We have published Net-SNMP 5.10.pre3 at: https://sourceforge.net/projects/net-snmp/files/net-snmp/5.10-pre-releases/ Here is the relevant section of the NEWS file: *5.10.pre3*: snmpd: - SECURITY: Fix a SMUX denial of service caused by blocking peer handshakes (CVE-2026-89147) - SECURITY: Fix newline injection in persistent configuration storage (CWE-93) - AgentX: Enforce session ownership on unregistration and validate registration ranges/OID lengths - Log listening addresses and ports for snmpd and AgentX at startup - Remove global return_buf[] from the public Net-SNMP API - Numerous crash, memory leak, and bounds-check fixes across AgentX, VACM, USM, trap conversion, and MIB modules - IF-MIB: Fix ifLastChange timestamps and linkUpDown traps on BSD, anchor include_ifmib_iface_prefix regexes, and read stored ifXTable config at startup - UCD-SNMP-MIB: Fix dskTable crash when includeAllDisks is set - HOST-RESOURCES-MIB: Fix crashes in macOS disk/partition/software tables - DISMAN-PING-MIB: Fix IPv6 ping support and timeout loops snmptrapd: - Treat setgid() and setuid() failures as fatal errors when dropping privileges - Log listening addresses and ports at startup - Fix TLS-TCP session polling and malformed SNMPv1 trap crashes snmplib: - Make the TLS-TCP transport fully asynchronous and non-blocking, and derive TLSTM securityName from the verified certificate chain - Port DES support to the OpenSSL EVP API and load OpenSSL 3.0+ legacy/default providers - Add per-peer connection limits and frame timeouts for TCP/Unix stream transports, and enable TCP_NODELAY - Add support for IPV6_RECVPKTINFO - Split snmp_sess_poll() into snmp_poll() and snmp_sess_poll() - Significantly speed up MIB parsing for large MIB trees - Update credentials when re-adding existing USM users from a session - Fix priority range filtering in -L* logging options - Numerous memory leak, buffer overflow, and crash fixes across ASN.1, MIB parsing, transports, and security models perl & python: - python: Add DontCheckOrdering and Cc session options to walk() - perl: Fix Windows Winsock compatibility and select() busy-loops - Numerous memory leak, buffer overflow, and crash fixes in both bindings building: - Generate compiler dependency files dynamically during the build (removing static Makefile.depend files; add --disable-dependency-tracking) - Fix 64-bit Windows SOCKET handling and add Visual Studio 2026 support - Install FreeBSD 15 and 16 system headers -- Wes Hardaker Please mail all replies to net...@li... |
|
From: Simon C. <sim...@mp...> - 2026-09-16 18:56:24
|
FYI, Best, Simon name "net-snmp" layer "networking-layer" version "5.9.4" id "CVE-2025-68615" status "Unpatched" link "https://nvd.nist.gov/vuln/detail/CVE-2025-68615" summary "net-snmp is a SNMP application library, tools and daemon. Prior to versions 5.9.5 and 5.10.pre2, a specially crafted packet to an net-snmp snmptrapd daemon can cause a buffer overflow and the daemon to crash. This issue has been patched in versions 5.9.5 and 5.10.pre2." scorev2 "0.0" scorev3 "9.8" scorev4 "0.0" modified "2026-06-17T09:59:21.000" vector "NETWORK" vectorString "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H" detail "version-in-range" description "May need backporting (fixed from 5.9.5)" id "CVE-2026-89147" status "Unpatched" link "https://nvd.nist.gov/vuln/detail/CVE-2026-89147" summary "Net-SNMP through 5.9.5.2 contains a denial of service vulnerability in the SMUX module where smux_accept() performs an unauthenticated blocking read without timeout on newly accepted connections. An unauthenticated remote client can connect to the SMUX listener and send no data, causing the single-threaded snmpd main loop to block indefinitely and suspend all SNMP processing." scorev2 "0.0" scorev3 "7.5" scorev4 "8.7" modified "2026-09-11T11:16:58.000" vector "NETWORK" vectorString "CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X" detail "version-in-range" description "May need backporting (fixed from >5.9.5.2)" |
|
From: Wes H. <har...@us...> - 2026-08-07 16:42:22
|
John Dill via Net-snmp-coders <net...@li...> writes: > I have a question about whether anyone uses net-snmp in the embedded > domain. It definitely has been used in those contexts yes. > The issue is that importing the entirety of openssl and net-snmp is > likely too large resource wise for the hardware to manage, so I'm > curious if anyone has ideas for creating a compact "library" with a > smaller footprint of features. You should be aware that you can specify --with-openssl="internal" which uses the most minimal extracts of (an older) openssl to support SNMPv3 with USM. You won't get DTLS though, but it doesn't sound like you need that. Also see the --enable-minimalist and --enable-mini-agent flags. -- Wes Hardaker Please mail all replies to net...@li... |
|
From: John D. <joh...@sa...> - 2026-08-06 21:40:35
|
Hi, I have a question about whether anyone uses net-snmp in the embedded domain. We have a large source tree for running software on an embedded platform that is based on a solaris cross-compiler for powerpc. Development is Ada but with C bindings to some lower level APIs. The issue is that importing the entirety of openssl and net-snmp is likely too large resource wise for the hardware to manage, so I'm curious if anyone has ideas for creating a compact "library" with a smaller footprint of features. I do not need to support the entire SNMP ecosystem, just SNMPv3 with MD5/SHA plus DES/AES with object identifiers and PDU creation. I'm not sure I need traps or the like, just Set/Get family of functions may be sufficient. I know that custom library based on a subset of openssl (or libtomcrypt) might get us a compact set of crypto functions, but net-snmp is pretty big. I have built it for Windows and can use the tool suite with openssl support in a hosted environment. I've thought about some kind of function trace to develop a profile of what functions we would need using net-snmp Windows installation to demo the devices we're attempting to integrate into our platform. Not sure if anyone is in the same kind of embedded domain (with constrained processor/memory requirements) that I'm in and needs to add a base level of SNMPv3 support and could provide some tips. Thanks, John Dill ________________________________ ATTENTION: The information contained in this message is intended for the addressee only and may contain Confidential / Controlled Unclassified Information (CUI). If you are not the intended addressee, please delete this message and notify the sender; you should not copy or distribute this message or disclose its contents to anyone. Any views or opinions expressed in this message are those of the individual(s) and not necessarily of the organization. No guarantee is implied that this message or any attachment is virus free or has not been intercepted and amended. |
|
From: Paul S <net...@co...> - 2026-06-17 13:33:51
|
SNMPv3 bulkwalk()/get() fail when using async/callback AND explicit
EngineID (Issue 3)
I'm having issues retrieving data via SNMPv3 when using async/callback
AND explicitly specifying the EngineID.
(This is separate to the EngineID discovery blocking issue I've logged
elsewhere)
FAIL:
If I make an asynchronous bulkwalk() request, explicity specifying the
EngineID, I receive an immediate empty callback. Debugging shows the
first SNMP request/reply packet successfully returned complete with
device data, but this is never returned to the user and there is no
further SNMP traffic or callbacks.
SUCCESS:
If I make a Non-asynchronous bulkwalk() request explicity specifying
the EngineID, the requested table is returned once polled.
SUCCESS:
If I make an asynchronous bulkwalk() request, leaving the EngineID to
be automagically discovered, the requested table is returned via
callback (VarList of VarBinds).
Similarly with get()
FAIL:
If I make an asynchronous get() request, explicity specifying the
EngineID, I get a segmentation fault.
SUCCESS:
If I make a Non-asynchronous get() request explicity specifying the
EngineID, the requested MIB value is returned.
SUCCESS:
If I make an asynchronous get() request, leaving the EngineID to be
automagically discovered, the correct VarBind result is returned via
callback.
If feel this indicates that I'm correctly handling both callbacks and
the setting of the EngineID, but used together it surfaces some bug.
Anyone any workaround such that I can use async AND explicit EngineID?
Any advice gratefully received. Further details below.
Background:
I use async/callbacks because of the number of devices I'm regularly
polling (thousands)
I'm looking to explicitly specify the EngineIDs because:
1) It reduces SNMPv3 overhead as I already have the EngineIDs available
2) It removes the blocking during the automagic EngineID discovery
that occurs in bulkwalk(), get(), etc. when a device is unavailable.
I potentially have a couple of dozen inaccessible devices at any time,
and these block the running of my MainLoop() against accessible
devices. This causes:
2a) Timeout of devices that are accessible as the inaccessible device
blocking of MainLoop() affects SNMP data processing of accessible
devices.
2b) Inaccessible devices whose automagic EngineID discovery fails do
not release their SNMP::Session. This cannot be closed manually and
the UDP socket is not returned to the OS until Perl exits.
Test script which fails - (use only async OR EngineID and this should work):
#!/usr/bin/perl
use warnings;
use strict;
use SNMP;
$SNMP::auto_init_mib=0;
$SNMP::use_numeric=1;
$SNMP::debugging=2;
$SNMP::verbose=2;
$SNMP::non_increasing=1;
use Time::HiRes qw(gettimeofday);
use POSIX qw(strftime);
my $generic_dev = { dev=> {
DestHost => '172.19.99.99',
RemotePort => 161,
Version => '3',
SecName => "RO-SECNAME",
SecLevel => "authPriv",
AuthProto => "SHA",
AuthPass => 'AUTHPASS',
PrivProto => "AES",
PrivPass => 'PRIVPASS',
#SecEngineId => '', #Autodiscover
SecEngineId => '800012345678901234567890', #Explicit EngineID
previously acquired
ContextEngineId => '',
Retries => 3,
Timeout => 1000000,
}};
my $varlist = snmpwalk($generic_dev);
logMe('Result count (non-async): '. scalar @{$varlist->[0]})
if((ref($varlist) eq 'ARRAY') && ($varlist->[0])); #Non-async
result count
die "Cannot do bulkwalk: $generic_dev->{ErrorStr}
($generic_dev->{ErrorNum})\n" if $generic_dev->{ErrorNum};
&SNMP::MainLoop();
exit;
sub snmpwalk {
my $tr=shift;
if(!defined($tr->{sess})) {
$tr->{sess} = SNMP::Session->new( %{$tr->{dev}} ) or logMe("$0:
impossible to obtain a new work v3 SNMP::Session; $!");
}
my $oid='.1.3.6.1.2.1.2.2.1.2'; #ifName table
my @arr=( [ $oid ] );
my $vars = new SNMP::VarList(@arr);
logMe('v3 calling bulkwalk() - ' . $tr->{dev}->{DestHost});
#my $varlist = $tr->{sess}->bulkwalk(0, 25, $vars); #Non-async
my $varlist = $tr->{sess}->bulkwalk(0, 25, $vars, [ \&callback,
$tr, $vars ]); #Async
logMe('v3 returned bulkwalk() - ' . $tr->{dev}->{DestHost});
return $varlist;
}
sub callback { #Should be called on data returned, or (no response) timeout
my $tr=shift;
my $vars=shift;
my $vals=shift;
logMe('Callback for ' . $tr->{dev}->{DestHost});
logMe('Result count (async): '. scalar @{$vals->[0]})
if($vals->[0]); #Async result count
}
sub logMe {
my ($t, $nsec) = gettimeofday;
my @t = localtime $t;
printf("%s %s\n", POSIX::strftime('%T.', @t) . $nsec, shift);
}
Environment:
Ubuntu 22.04.5 LTS
libsnmp-base/jammy-updates,jammy-security,now 5.9.1+dfsg-1ubuntu2.9
libsnmp40/jammy-updates,jammy-security,now 5.9.1+dfsg-1ubuntu2.9 amd64
[installed,automatic]
libsnmp-perl/jammy-updates,jammy-security,now 5.9.1+dfsg-1ubuntu2.9
amd64 [installed]
perl5 (revision 5 version 34 subversion 0)
Many thanks,
Paul S.
|
|
From: Paul S <net...@co...> - 2026-06-15 13:17:33
|
Quoting Paul S <net...@co...>: > > Quoting Wes Hardaker <har...@us...>: > >> Paul S <net...@co...> writes: >> >>> From packet analysis, I see that when bulkwalk() is called, the SNMP >>> library first queries that device's EngineID if this is not already >>> cached. >> >> SNMPv3 does have a larger start up cost because of this while it does >> engine synchronization. It would surprise me that bulkwalk is the only >> function with this problem. Are you sure the same symptoms don't occur >> with other types of requests (walks or even gets)? > > Actually, I'm not sure as I'm only using bulkwalk(), but I have no > doubt that engine synchronisation must happen for walks/gets if its not > already cached. > I have now tested this further. Yes, both bulkwalk() and get() block until the EngineID is obtained, or the device times-out. Issue No.1: This blocking means that I cannot call MainLoop(), which means the hundreds of other devices whose SNMP is currently being handled by MainLoop() are suspended until unavailable devices are timed-out. This is having a major knock-on to my ability to run a regular poll-loop should ever a number of devices be unavailable. Issue No.2: get() and bulkwalk() differ in the event of a failing to auto-obtain an EngineID: Upon timeout, get() performs a callback with an empty varlist, and the session is closeable by setting undef. This releases the UDP socket, which can be seen as freed by the OS. bulkwalk() never performs the callback in the event it times-out obtaining the EngineID. Subsequent setting such a session as undef does NOT release the UDP socket. In an environment where devices are polled in a loop, inaccessible devices will build up a large number of "stuck" sessions and UDP sockets, ultimately leading to OS resource starvation. I am unable to see anything in SNMP.pm that can point me in a direction to look next - any suggestions? Is this something handled with the XS or the library itself? Are there any workarounds to Issues 1 and 2 above, or would a code snippet help to demo the issue? Many thanks, Paul |
|
From: Paul S <net...@co...> - 2026-06-11 13:59:32
|
Quoting Wes Hardaker <har...@us...>: > Paul S <net...@co...> writes: > >> From packet analysis, I see that when bulkwalk() is called, the SNMP >> library first queries that device's EngineID if this is not already >> cached. > > SNMPv3 does have a larger start up cost because of this while it does > engine synchronization. It would surprise me that bulkwalk is the only > function with this problem. Are you sure the same symptoms don't occur > with other types of requests (walks or even gets)? Actually, I'm not sure as I'm only using bulkwalk(), but I have no doubt that engine synchronisation must happen for walks/gets if its not already cached. However, it's not the engine synchronization itself that's the issue - it's the fact that the engine synchonisation is not handled within SNMP::MainLoop(). Thus the attempted synchronisation of one unavailable device blocks the asynchronous SNMP processing of hundreds of other devices until that single device times-out. There may be a number of devices unavailable, and all of these will block asynchonous polling each time they are retried and found to be unavailable. As an alternative immediate solution I tried to handle the engine synchonrisation in a separate process, and pass the EngineID via SecEngineId in SNMP::Session->new() in my code. If I specify 0 as the EngineID (default, as above), we have the overhead of the blocking EngineId request (a working device returns Report 33, usmStatsUnknownEngineIDs increments) If I specify the correct EngineID, The device replies Report 31, usmStatsNotInTimeWindows increments, which implies I have the EngineID correct, but bulkwalk() itself then fails to complete - callback is immediately called with empty data. If I specify an incorrect EngineID the device returns Report 33, usmStatsUnknownEngineIDs increments, and this attempt is retried before timing-out and the callback called with empty data. This does not block. So, specifying a non-zero SecEngineId avoids blocking of bulkwalk as there is no EngineID synchonrisation happening in the background. This would solve my issue except it seems bulkwalk() fails even when a correct EngineID is specified. Any ideas how I might work around this? Many thanks, Paul -- Paul S |
|
From: Wes H. <har...@us...> - 2026-06-11 12:46:46
|
Paul S <net...@co...> writes: > From packet analysis, I see that when bulkwalk() is called, the SNMP > library first queries that device's EngineID if this is not already > cached. SNMPv3 does have a larger start up cost because of this while it does engine synchronization. It would surprise me that bulkwalk is the only function with this problem. Are you sure the same symptoms don't occur with other types of requests (walks or even gets)? -- Wes Hardaker Please mail all replies to net...@li... |
|
From: Paul S <net...@co...> - 2026-06-09 16:37:11
|
Hi, Issue: SNMPv3 bulkwalk()/callback blocking during background EngineID queries I have been successfully using Perl SNMP package/Net-SNMP library for years polling 3000+ SNMPv2. At any one time, there are parallel queries to around 200 devices, and the non-blocking+callback solution works perfectly. I am currently running Ubuntu 22.04.5 LTS, Perl 5.34.0, SNMP.pm 5.0901 (5.9.1+dfsg-1ubuntu2.9 Perl module libsnmp-base modules) However, the devices are now reconfigured for SNMPv3, and whilst my code still works fine, it runs very slowly when there are a number of devices unavailable (say 1%) at startup. (Such as when a remote site is offline/many devices under maintenance). My code consists of a loop with SNMP::MainLoop(), which periodically fires off various device bulkwalk()s with callback refs, (setting up a SNMP session beforehand, if none exists). The callbacks process the returned data, and undef the session, if necessary. I can provide a bare minimum code to demonstrate the issue. From packet analysis, I see that when bulkwalk() is called, the SNMP library first queries that device's EngineID if this is not already cached. Despite using callbacks, bulkwalk() unexpectedly blocks until either the EngineID is returned and cached, or until the request to the device for the EngineID times out. If the device is unavailable, this blocks the running of SNMP::MainLoop() for up to the device's timeout/retry duration. Given that some devices are on some very slow links, the timeout to an unavailable device may be longer than that of a device that is part way through a large table bulkwalk(). When SNMP::MainLoop() is eventually called, the latter device may find itself timed-out, through no fault of its own - as any bulkwalk()'s blocking duration will add to their timeout. To compound this, I appear to receive no callback in the event an unavailable device times-out (no EngineID received). I am also unable to close(undef) such a session, and the OS-allocated UDP session socket remains until the Perl program exits. I periodically retry unavailable devices to see if they have become available (I have no other way to tell), which blocks all other devices being polled if the device is still unavailable, and eats up a UDP socket on each occasion. At any point in time, for various reasons, I may have ~30 devices that don't respond. Each device is set to 3 retries of 2 seconds each, and each device will of course block sequentially (as its call to bulkwalk() blocks). I attempt to poll each of my devices every 5 minutes, and this means that the 30 devices block for around 3 minutes of that 5 minutes, and eat 360 UDP sockets per hour (or 8640 sockets per day). Whilst I could implement some form of exponential back-off for polling, this doesn't escape from the fact I can only back-off for a short maximum period, merely slowing the issue. As UDP sockets are chewed-up, I need to restart the process to release them anyway, which just restarts the issue (and not withstanding the fact that ALL devices will then need to re-cache their EngineID's, and each of these devices will block sequentially on their first bulkwalk() ). Should a single device's EngineID request block like this (or is there a way to make it non-blocking)? Even if I decrease the retries and timeouts, I'm unlikely to go below 1 second each, which will still impact around 30 seconds in each poll period. Is there a way to ensure that failed EngineID requests result in a callback, such that I can close and release the UDP socket back to the OS? This aside, I have to say the library has been rock-solid for me for decades. Very many thanks indeed. Regards, Paul S |
|
From: Craig S. <cs...@dr...> - 2026-06-04 09:59:11
|
On Thu, 4 Jun 2026 at 18:30, Pushpa Thimmaiah <pus...@gm...> wrote: > I am seeing following message . > "systemstats_linux: unexpected header length in /proc/net/snmp." > I am using net-snmp 5.9.4 and linuxkernel v6.6. Could you please share > upgrade net-snmp is clean fix or applying patch from latest to v5.9.4 is > valid? > Hi, For Debian Trixie we added this patch[1] which is based upon this one[2] - Craig 1: https://salsa.debian.org/debian/net-snmp/-/blob/trixie/debian/patches/linux_systemstats?ref_type=heads 2: https://github.com/net-snmp/net-snmp/commit/49d60ba57f4b462df7dc5fd5b38b4425dab0982c |
Hi All, I am seeing following message . "systemstats_linux: unexpected header length in /proc/net/snmp." I am using net-snmp 5.9.4 and linuxkernel v6.6. Could you please share upgrade net-snmp is clean fix or applying patch from latest to v5.9.4 is valid? Thank you, Pushpa.T |
|
From: Pushpa T. <pus...@gm...> - 2026-03-27 04:46:10
|
Hi Wes, Thank you for the response. I am using net-snmp-5.9 on my device, md5+AES192/AES256 works fine. However, I have heard that this combination is invalid. So wanted to check here if anything like invalid combinations. Thanks, Pushpa.T On Fri, Mar 27, 2026 at 2:59 AM Wes Hardaker <har...@us...> wrote: > Pushpa Thimmaiah <pus...@gm...> writes: > > > I have compiled net-snmp-5.9.4 with blumenthal-aes and need > clarification for the following. > > 1. Does combination MD5 and AES192/AES256 invalid? > > MD5 doesn't generate a long enough key to be used with the longer-key > AES algorithms. I honestly forget whether or not the algorithm in 3414 > (which we implement) extends key sizes if the hash algorithm produces > something too short. I might go look later, but you could check > yourself. > > (if the tools don't work, then that's the other clue :-/ ) > > -- > Wes Hardaker > Please mail all replies to net...@li... > |