- 11 Mar, 2024 40 commits
-
-
Geliang Tang authored
To unify the output formats of all test scripts, this patch adds four more helpers: mptcp_lib_pr_ok() mptcp_lib_pr_skip() mptcp_lib_pr_fail() mptcp_lib_pr_info() to print out [ OK ], [SKIP], [FAIL] and 'INFO: ' with colors. Use them in all scripts to print the "ok/skip/fail/info' using the same 'format'. Having colors helps to quickly identify issues when looking at a long list of output logs and results. Note that now all print the same keywords, which was not the case before, but it is good to uniform that. Signed-off-by:
Geliang Tang <tanggeliang@kylinos.cn> Reviewed-by:
Matthieu Baerts (NGI0) <matttbe@kernel.org> Signed-off-by:
Jakub Kicinski <kuba@kernel.org>
-
Geliang Tang authored
This patch uses addition assignment operator (+=) to append strings instead of duplicating the variable name in mptcp_connect.sh and mptcp_join.sh. This can make the statements shorter. Note: in mptcp_connect.sh, add a local variable extra in do_transfer to save the various extra warning logs, using += to append it. And add a new variable tc_info to save various tc info, also using += to append it. This can make the code more readable and prepare for the next commit. Signed-off-by:
-
Geliang Tang authored
This patch adds a new helper mptcp_lib_print_title(), a wrapper of mptcp_lib_inc_test_counter() and mptcp_lib_pr_title_counter(), to print out test counter in each test result and increase the counter. Use this helper to print out test counters for every tests in diag.sh, mptcp_connect.sh, mptcp_sockopt.sh, pm_netlink.sh, simult_flows.sh, and userspace_pm.sh. diag.sh: 01 no msk on netns creation [ ok ] 02 listen match for dport 10000 [ ok ] 03 listen match for sport 10000 [ ok ] 04 listen match for saddr and sport [ ok ] 05 all listen sockets [ ok ] mptcp_connect.sh: 01 New MPTCP socket can be blocked via sysctl [ OK ] 02 Validating network environment with pings [ OK ] INFO: Using loss of 0.85% delay 31 ms reorder .. with delay 7ms on ns3eth4 03 ns1 MPTCP -> ns1 (10.0.1.1:10000 ) MPTCP (duration 69ms) [ OK ] 04 ns1 MPTCP -> ns1 (10.0.1.1:10001 ) TCP (duration 20ms) [ OK ] 05 ns1 TCP -> ns1 (10.0.1.1:10002 ) MPTCP (duration 16ms) [ OK ] mptcp_sockopt.sh: 01 Transfer v4 [ OK ] 02 Mark v4 [ OK ] 03 Transfer v6 [ OK ] 04 Mark v6 [ OK ] 05 SOL_MPTCP sockopt v4 [ OK ] pm_netlink.sh: 01 defaults addr list [ OK ] 02 simple add/get addr [ OK ] 03 dump addrs [ OK ] 04 simple del addr [ OK ] 05 dump addrs after del [ OK ] simult_flows.sh: 01 balanced bwidth 7391 max 8456 [ OK ] 02 balanced bwidth - reverse direction 7403 max 8456 [ OK ] 03 balanced bwidth with unbalanced delay 7429 max 8456 [ OK ] 04 balanced bwidth with unbalanced delay - reverse ... 7485 max 8456 [ OK ] 05 unbalanced bwidth 7549 max 8456 [ OK ] userspace_pm.sh: 01 Created network namespaces ns1, ns2 [ OK ] INFO: Make connections 02 Established IPv4 MPTCP Connection ns2 => ns1 [ OK ] 03 Established IPv6 MPTCP Connection ns2 => ns1 [ OK ] INFO: Announce tests 04 ADD_ADDR 10.0.2.2 (ns2) => ns1, invalid token [ OK ] 05 ADD_ADDR id:67 10.0.2.2 (ns2) => ns1, reuse port [ OK ] Having test counters helps to quickly identify issues when looking at a long list of output logs and results. Signed-off-by:
-
Geliang Tang authored
This patch adds a new variable MPTCP_LIB_TEST_FORMAT as the test title printing format. Also add a helper mptcp_lib_print_title() to use this format to print the test title with test counters. They are used in mptcp_join.sh first. Each MPTCP selftest is having subtests, and it helps to give them a number to quickly identify them. This can be managed by mptcp_lib.sh, reusing what has been done here. The following commit will use these new helpers in the other tests. Signed-off-by:
-
Geliang Tang authored
Variable TEST_COUNT are used in mptcp_connect.sh and mptcp_join.sh as test counters, which are initialized to 0, while variable test_cnt are used in diag.sh and simult_flows.sh, which are initialized to 1. To maintain consistency, this patch renames them all as MPTCP_LIB_TEST_COUNTER, initializes it to 1, and exports it into mptcp_lib.sh. Signed-off-by:
-
Geliang Tang authored
Only total test results are printed out in mptcp_sockopt.sh: PASS: all packets had packet mark set PASS: SOL_MPTCP getsockopt has expected information PASS: TCP_INQ cmsg/ioctl -t tcp PASS: TCP_INQ cmsg/ioctl -6 -t tcp PASS: TCP_INQ cmsg/ioctl -r tcp PASS: TCP_INQ cmsg/ioctl -6 -r tcp PASS: TCP_INQ cmsg/ioctl -r tcp -t tcp They mismatch with the test results: ok 1 - mptcp_sockopt: mark ipv4 ok 2 - mptcp_sockopt: transfer ipv4 ok 3 - mptcp_sockopt: mark ipv6 ok 4 - mptcp_sockopt: transfer ipv6 ok 5 - mptcp_sockopt: sockopt v4 ok 6 - mptcp_sockopt: sockopt v6 ok 7 - mptcp_sockopt: TCP_INQ: -t tcp ok 8 - mptcp_sockopt: TCP_INQ: -6 -t tcp ok 9 - mptcp_sockopt: TCP_INQ: -r tcp ok 10 - mptcp_sockopt: TCP_INQ: -6 -r tcp ok 11 - mptcp_sockopt: TCP_INQ: -r tcp -t tcp 'mptcp_sockopt.sh' now display more detailed results + why (what you had in a former patch from v6, merged here). It no longer displays 'PASS:', because it is duplicated info now that the detailed are displayed: Transfer v4 [ OK ] Mark v4 [ OK ] Transfer v6 [ OK ] Mark v6 [ OK ] SOL_MPTCP sockopt v4 [ OK ] SOL_MPTCP sockopt v6 [ OK ] TCP_INQ cmsg/ioctl -t tcp [ OK ] TCP_INQ cmsg/ioctl -6 -t tcp [ OK ] TCP_INQ cmsg/ioctl -r tcp [ OK ] TCP_INQ cmsg/ioctl -6 -r tcp [ OK ] TCP_INQ cmsg/ioctl -r tcp -t tcp [ OK ] Also fix the TAP output: ok 1 - mptcp_sockopt: transfer ipv4 ok 2 - mptcp_sockopt: mark ipv4 ok 3 - mptcp_sockopt: transfer ipv6 ok 4 - mptcp_sockopt: mark ipv6 ok 5 - mptcp_sockopt: sockopt v4 ok 6 - mptcp_sockopt: sockopt v6 ok 7 - mptcp_sockopt: TCP_INQ: -t tcp ok 8 - mptcp_sockopt: TCP_INQ: -6 -t tcp ok 9 - mptcp_sockopt: TCP_INQ: -r tcp ok 10 - mptcp_sockopt: TCP_INQ: -6 -r tcp ok 11 - mptcp_sockopt: TCP_INQ: -r tcp -t tcp Signed-off-by:
-
Geliang Tang authored
The first [ OK ] in the output of mptcp_connect.sh misaligns with the others: New MPTCP socket can be blocked via sysctl [ OK ] INFO: validating network environment with pings INFO: Using loss of 0.85% delay 16 ms reorder 95% 70% with delay 4ms on ns1 MPTCP -> ns1 (10.0.1.1:10000 ) MPTCP (duration 184ms) [ OK ] ns1 MPTCP -> ns1 (10.0.1.1:10001 ) TCP (duration 50ms) [ OK ] ns1 TCP -> ns1 (10.0.1.1:10002 ) MPTCP (duration 55ms) [ OK ] This patch aligns them by using 69 chars to display the first two lines, and 50 chars for the other. Since 19 chars are used to display duration time. Also print out a [ OK ] at the end of the 2nd line for consistency. Signed-off-by:
-
Geliang Tang authored
This patch adds a new dedicated counter 'port' instead of TEST_COUNT to increase port numbers in mptcp_connect.sh. This can avoid outputting discontinuous test counters. Signed-off-by:
-
Geliang Tang authored
Some error messages are printed to stderr while the others are printed to 'stdout'. As part of the unification, this patch drop "1>&2" to let all errors messages are printed to 'stdout'. Signed-off-by:
-
Uwe Kleine-König authored
The .remove() callback for a platform driver returns an int which makes many driver authors wrongly assume it's possible to do error handling by returning an error code. However the value returned is ignored (apart from emitting a warning) and this typically results in resource leaks. To improve here there is a quest to make the remove callback return void. In the first step of this quest all drivers are converted to .remove_new(), which already returns void. Eventually after all drivers are converted, .remove_new() will be renamed to .remove(). Trivially convert this driver from always returning zero in the remove callback to the void returning variant. Signed-off-by:
Uwe Kleine-König <u.kleine-koenig@pengutronix.de> Acked-by:
Herve Codina <herve.codina@bootlin.com> Link: https://lore.kernel.org/r/9684419fd714cc489a3ef36d838d3717bb6aec6d.1709886922.git.u.kleine-koenig@pengutronix.deSigned-off-by:
-
Jakub Kicinski authored
Petr Machata says: ==================== mlxsw: Support for nexthop group statistics ECMP is a fundamental component in L3 designs. However, it's fragile. Many factors influence whether an ECMP group will operate as intended: hash policy (i.e. the set of fields that contribute to ECMP hash calculation), neighbor validity, hash seed (which might lead to polarization) or the type of ECMP group used (hash-threshold or resilient). At the same time, collection of statistics that would help an operator determine that the group performs as desired, is difficult. Support for nexthop group statistics and their HW collection has been introduced recently. In this patch set, add HW stats collection support to mlxsw. This patchset progresses as follows: - Patches #1 and #2 add nexthop IDs to notifiers. - Patches #3 and #4 are code-shaping. - Patches #5, #6 and #7 adjust the flow counter code. - Patches #8 and #9 add HW nexthop counters. - Patch #10 adjusts the HW counter code to allow sharing the same counter for several resilient group buckets with the same NH ID. - Patch #11 adds a selftest. ==================== Link: https://lore.kernel.org/r/cover.1709901020.git.petrm@nvidia.comSigned-off-by:
-
Petr Machata authored
Add to lib.sh support for fetching NH stats, and a new library, router_mpath_nh_lib.sh, with the common code for testing NH stats. Use the latter from router_mpath_nh.sh and router_mpath_nh_res.sh. The test works by sending traffic through a NH group, and checking that the reported values correspond to what the link that ultimately receives the traffic reports having seen. Signed-off-by:
Petr Machata <petrm@nvidia.com> Link: https://lore.kernel.org/r/2a424c54062a5f1efd13b9ec5b2b0e29c6af2574.1709901020.git.petrm@nvidia.comSigned-off-by:
-
Petr Machata authored
For resilient groups, we can reuse the same counter for all the buckets that share the same nexthop. Keep a reference count per counter, and keep all these counters in a per-next hop group xarray, which serves as a NHID->counter cache. If a counter is already present for a given NHID, just take a reference and use the same counter. Signed-off-by:
Ido Schimmel <idosch@nvidia.com> Link: https://lore.kernel.org/r/cdd00084533fc83ac5917562f54642f008205bf3.1709901020.git.petrm@nvidia.comSigned-off-by:
-
Petr Machata authored
When hw_stats is set on a group, install nexthop counters on members of a group. Counter allocation request is moved from nexthop object initialization to the update code. The previous placement made sense: when the counters are enabled by dpipe, the counters are installed to all existing nexthops and all nexthops created from then on get them. For the finer-grained nexthop group statistics, this is unsuitable. The existing placement was kept for the IPv4 and IPv6 nexthops. Resilient group replacement emits a pre_replace notification, and then any bucket_replace notifications if there were any replacements at all. If the group is balanced and the nexthop composition of the replaced group didn't change, there will be no such notifiers. Therefore hook to the pre_replace notifier and mark all buckets for update, to un/install the counters. When reporting deltas for resilient groups, use the nexthop ID that we stored in a previous patch to look up to which nexthop a bucket contributes. Co-developed-by:
-
Petr Machata authored
The core interfaces for collecting per-NH statistics are built around nexthops even for resilient groups. Because mlxsw models each bucket as a nexthop, the core next hop that a given bucket contributes to needs to be looked up. In order to be able to match the two up, we need to track nexthop ID for members of group nexthop objects. For simplicity, do it for all nexthop objects, not just group members. Signed-off-by:
-
Petr Machata authored
The next patch will add the ability to share nexthop counters among mlxsw nexthops backed by the same core nexthop. To have a place to store reference count, the counter should be kept in a dedicated structure. In this patch, introduce the structure together with the related helpers, sans the refcount, which comes in the next patch. Signed-off-by:
-
Petr Machata authored
mlxsw_sp_nexthop_counter_disable() decays to a nop when called on a disabled counter, but mlxsw_sp_nexthop_counter_enable() can't similarly be called on an enabled counter. This would be useful in the following patches. Add the missing condition. Signed-off-by:
-
Petr Machata authored
For the report_delta-like interface like a previous patch has added for collection of NH group statistics, it's easiest to read the counter and have the HW clear it right away. Thus, change mlxsw_sp_flow_counter_get() to take a bool indicating whether this should be done. Signed-off-by:
-
Petr Machata authored
In order to be able to diagnose failures in counter allocation, have the function mlxsw_sp_nexthop_counter_enable() return an error code. Signed-off-by:
-
Petr Machata authored
The function mlxsw_sp_nexthop_counter_alloc() doesn't directly allocate anything, and mlxsw_sp_nexthop_counter_free() doesn't directly free. For the following patches, we will need names for functions that actually do those things. Therefore rename to mlxsw_sp_nexthop_counter_enable() and mlxsw_sp_nexthop_counter_disable() to free up the namespace. Signed-off-by:
-
Petr Machata authored
When sending the notifications to collect NH statistics for resilient groups, the driver will need to know the nexthop IDs in individual buckets to look up the right counter. To that end, move the nexthop ID from struct nh_notifier_grp_entry_info to nh_notifier_single_info. Signed-off-by:
-
Petr Machata authored
The NEXTHOP_EVENT_RES_TABLE_PRE_REPLACE notifier currently keeps the group ID unset. That makes it impossible to look up the group for which the notifier is intended. This is not an issue at the moment, because the only client is netdevsim, and that just so that it veto replacements, which is a static property not tied to a particular group. But for any practical use, the ID is necessary. Set it. Signed-off-by:
-
Eric Dumazet authored
Move gro_find_receive_by_type() and gro_find_complete_by_type() to include/net/gro.h where they belong. Also use _NET_GRO_H instead of _NET_IPV6_GRO_H to protect include/net/gro.h from multiple inclusions. Signed-off-by:
Eric Dumazet <edumazet@google.com> Reviewed-by:
Simon Horman <horms@kernel.org> Link: https://lore.kernel.org/r/20240308102230.296224-1-edumazet@google.comSigned-off-by:
-
Matthew Wood authored
Add a space (' ') prefix to every userdata line to match docs for dev-kmsg. To account for this extra character in each userdata entry, reduce userdata entry names (directory name) from 54 characters to 53. According to the dev-kmsg docs, a space is used for subsequent lines to mark them as continuation lines. > A line starting with ' ', is a continuation line, adding > key/value pairs to the log message, which provide the machine > readable context of the message, for reliable processing in > userspace. Testing for this patch:: cd /sys/kernel/config/netconsole && mkdir cmdline0 cd cmdline0 mkdir userdata/test && echo "hello" > userdata/test/value mkdir userdata/test2 && echo "hello2" > userdata/test2/value echo "message" > /dev/kmsg Outputs:: 6.8.0-rc5-virtme,12,493,231373579,-;message test=hello test2=hello2 And I confirmed all testing works as expected from the original patchset Fixes: df03f830 ("net: netconsole: cache userdata formatted string in netconsole_target") Signed-off-by:Matthew Wood <thepacketgeek@gmail.com> Reviewed-by:
Breno Leitao <leitao@debian.org> Link: https://lore.kernel.org/r/20240308002525.248672-1-thepacketgeek@gmail.comSigned-off-by:
-
Heiner Kallweit authored
Switch to new function phy_support_eee. This allows to simplify the code because data->tx_lpi_enabled is now populated by phy_ethtool_get_eee(). Signed-off-by:
Heiner Kallweit <hkallweit1@gmail.com> Link: https://lore.kernel.org/r/92462328-5c9b-4d82-9ce4-ea974cda4900@gmail.comSigned-off-by:
-
Heiner Kallweit authored
Handling case err == 0 in the other branch allows to simplify the code. In addition I assume in "err & phydev->eee_cfg.tx_lpi_enabled" it should have been a logical and operator. It works as expected also with the bitwise and, but using a bitwise and with a bool value looks ugly to me. Signed-off-by:
-
Andy Shevchenko authored
of_gpio.h is deprecated and subject to remove. The driver doesn't use it, simply remove the unused header. Signed-off-by:
Andy Shevchenko <andriy.shevchenko@linux.intel.com> Link: https://lore.kernel.org/r/20240307122346.3677534-1-andriy.shevchenko@linux.intel.comSigned-off-by:
-
Justin Swartz authored
Disable LEDs just before resetting the MT7530 to avoid situations where the ESW_P4_LED_0 and ESW_P3_LED_0 pin states may cause an unintended external crystal frequency to be selected. The HT_XTAL_FSEL (External Crystal Frequency Selection) field of HWTRAP (the Hardware Trap register) stores a 2-bit value that represents the state of the ESW_P4_LED_0 and ESW_P4_LED_0 pins (seemingly) sampled just after the MT7530 has been reset, as: ESW_P4_LED_0 ESW_P3_LED_0 Frequency ----------------------------------------- 0 1 20MHz 1 0 40MHz 1 1 25MHz The value of HT_XTAL_FSEL is bootstrapped by pulling ESW_P4_LED_0 and ESW_P3_LED_0 up or down accordingly, but: if a 40MHz crystal has been selected and the ESW_P3_LED_0 pin is high during reset, or a 20MHz crystal has been selected and the ESW_P4_LED_0 pin is high during reset, then the value of HT_XTAL_FSEL will indicate that a 25MHz crystal is present. By default, the state of the LED pins is PHY controlled to reflect the link state. To illustrate, if a board has: 5 ports with active low LED control, and HT_XTAL_FSEL bootstrapped for 40MHz. When the MT7530 is powered up without any external connection, only the LED associated with Port 3 is illuminated as ESW_P3_LED_0 is low. In this state, directly after mt7530_setup()'s reset is performed, the HWTRAP register (0x7800) reflects the intended HT_XTAL_FSEL (HWTRAP bits 10:9) of 40MHz: mt7530-mdio mdio-bus:1f: mt7530_read: 00007800 == 00007dcf >>> bin(0x7dcf >> 9 & 0b11) '0b10' But if a cable is connected to Port 3 and the link is active before mt7530_setup()'s reset takes place, then HT_XTAL_FSEL seems to be set for 25MHz: mt7530-mdio mdio-bus:1f: mt7530_read: 00007800 == 00007fcf >>> bin(0x7fcf >> 9 & 0b11) '0b11' Once HT_XTAL_FSEL reflects 25MHz, none of the ports are functional until the MT7621 (or MT7530 itself) is reset. By disabling the LED pins just before reset, the chance of an unintended HT_XTAL_FSEL value is reduced. Signed-off-by:Justin Swartz <justin.swartz@risingedge.co.za> Link: https://lore.kernel.org/r/20240305043952.21590-1-justin.swartz@risingedge.co.zaSigned-off-by:
-
Andy Shevchenko authored
of_gpio.h is deprecated and subject to remove. The driver doesn't use it, simply remove the unused header. Signed-off-by:
-
Ricardo B. Marliere authored
Since commit 43a7206b ("driver core: class: make class_register() take a const *"), the driver core allows for struct class to be in read-only memory, so move the ptp_class structure to be declared at build time placing it into read-only memory, instead of having to be dynamically allocated at boot time. Suggested-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by:
Ricardo B. Marliere <ricardo@marliere.net> Reviewed-by:
-
Hangbin Liu authored
ynl-gen-c.py supports check unterminated-ok, but the yaml schemas don't have this key. Add this to the yaml files. Signed-off-by:
Hangbin Liu <liuhangbin@gmail.com> Link: https://lore.kernel.org/r/20240308081239.3281710-1-liuhangbin@gmail.comSigned-off-by:
-
Hangbin Liu authored
Support using pre-defined values in checks so we don't need to use hard code number for the string, binary length. e.g. we have a definition like #define TEAM_STRING_MAX_LEN 32 Which defined in yaml like: definitions: - name: string-max-len type: const value: 32 It can be used in the attribute-sets like attribute-sets: - name: attr-option name-prefix: team-attr-option- attributes: - name: name type: string checks: len: string-max-len With this patch it will be converted to [TEAM_ATTR_OPTION_NAME] = { .type = NLA_STRING, .len = TEAM_STRING_MAX_LEN, } Signed-off-by: -
Mina Almasry authored
The check is duplicated in 2 places, factor it out into a common helper. Signed-off-by:
Mina Almasry <almasrymina@google.com> Reviewed-by:
Yunsheng Lin <linyunsheng@huawei.com> Reviewed-by:
Ilias Apalodimas <ilias.apalodimas@linaro.org> Link: https://lore.kernel.org/r/20240308204500.1112858-1-almasrymina@google.comSigned-off-by:
-
David S. Miller authored
Jason Xing says: ==================== annotate data-races around sysctl_tcp_wmem[0] Adding simple READ_ONCE() can avoid reading the sysctl knob meanwhile someone is trying to change it. ==================== Signed-off-by:David S. Miller <davem@davemloft.net>
-
Jason Xing authored
When reading wmem[0], it could be changed concurrently without READ_ONCE() protection. So add one annotation here. Signed-off-by:
Jason Xing <kernelxing@tencent.com> Reviewed-by:
-
Jason Xing authored
It's possible that writer and the reader can manipulate the same sysctl knob concurrently. Using READ_ONCE() to prevent reading an old value. Signed-off-by:
Mat Martineau <martineau@kernel.org> Signed-off-by:
-
Jakub Kicinski authored
Running the page-pool sample on production machines under moderate networking load shows recycling rate higher than 100%: $ page-pool eth0[2] page pools: 14 (zombies: 0) refs: 89088 bytes: 364904448 (refs: 0 bytes: 0) recycling: 100.3% (alloc: 1392:2290247724 recycle: 469289484:1828235386) Note that outstanding refs (89088) == slow alloc * cache size (1392 * 64) which means this machine is recycling page pool pages perfectly, not a single page has been released. The extra 0.3% is because sample ignores allocations from the ptr_ring. Treat those the same as alloc_fast, the ring vs cache alloc is already captured accurately enough by recycling stats. With the fix: $ page-pool eth0[2] page pools: 14 (zombies: 0) refs: 89088 bytes: 364904448 (refs: 0 bytes: 0) recycling: 100.0% (alloc: 1392:2331141604 recycle: 473625579:1857460661) Signed-off-by: -
Eric Dumazet authored
After commits ca065d0c ("udp: no longer use SLAB_DESTROY_BY_RCU") and 7ae215d2 ("bpf: Don't refcount LISTEN sockets in sk_assign()") UDP early demux no longer need to grab a refcount on the UDP socket. This save two atomic operations per incoming packet for connected sockets. Signed-off-by:
Paolo Abeni <pabeni@redhat.com> Signed-off-by:
-
David S. Miller authored
Gavrilov Ilia says: ==================== fix incorrect parameter validation in the *_get_sockopt() functions This v2 series fix incorrent parameter validation in *_get_sockopt() functions in several places. version 2 changes: - reword the patch description - add two patches for net/kcm and net/x25 ==================== Signed-off-by: -
Gavrilov Ilia authored
The 'len' variable can't be negative when assigned the result of 'min_t' because all 'min_t' parameters are cast to unsigned int, and then the minimum one is chosen. To fix the logic, check 'len' as read from 'optlen', where the types of relevant variables are (signed) int. Fixes: 1da177e4 ("Linux-2.6.12-rc2") Signed-off-by:
Gavrilov Ilia <Ilia.Gavrilov@infotecs.ru> Signed-off-by:
-
