Felix Fietkau [Tue, 25 Jul 2023 19:06:03 +0000 (21:06 +0200)]
wifi: mt76: mt7603: fix tx filter/flush function
Setting MT_TX_ABORT does not abort any transmission for a wtbl index on its
own. Instead, it modifies the behavior of a queue flush to make it selectively
flush packets for a particular wtbl index.
Adjust powersave filtering to make use of this in order to avoid running into
unnecessary timeouts while flushing
Felix Fietkau [Tue, 25 Jul 2023 10:37:59 +0000 (12:37 +0200)]
wifi: mt76: mt7603: fix beacon interval after disabling a single vif
When disabling beacons on a vif, intval is 0. Ensure that dev->mt76.beacon_int
is not overwritten in this case, so that beacons continue to work for other
interfaces.
Rafał Miłecki [Mon, 8 May 2023 15:58:20 +0000 (17:58 +0200)]
dt-bindings: mt76: support pointing to EEPROM using NVMEM cell
All kind of calibration data should be described as NVMEM cells of NVMEM
devices. That is more generic solution than "mediatek,mtd-eeprom" which
is MTD specific.
Add support for EEPROM NVMEM cells and deprecate existing MTD-based
property.
Cc: Christian Marangi <ansuelsmth@gmail.com> Signed-off-by: Rafał Miłecki <rafal@milecki.pl> Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Signed-off-by: Felix Fietkau <nbd@nbd.name>
Christian Marangi [Sat, 8 Jul 2023 18:29:36 +0000 (20:29 +0200)]
wifi: mt76: add support for providing eeprom in nvmem cells
Add support for providing eeprom in nvmem cells by adding nvmem cell as
an alternative source for mt76_get_of_eeprom().
Nvmem cells will follow standard nvmem cell definition and needs to be
called 'eeprom' to be correctly identified.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com> Reviewed-by: Daniel Golle <daniel@makrotopia.org> Tested-by: Daniel Golle <daniel@makrotopia.org> Signed-off-by: Felix Fietkau <nbd@nbd.name>
Christian Marangi [Sat, 8 Jul 2023 18:29:35 +0000 (20:29 +0200)]
wifi: mt76: split get_of_eeprom in subfunction
In preparation for NVMEM support, split get_of_eeprom() in subfunction
to tidy the code and facilitate the addition of alternative method to
get eeprom data. No behaviour change intended.
While at it also drop OF ifdef checks as OF have stubs and calling
of_get_property would result in the same error returned.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com> Reviewed-by: Daniel Golle <daniel@makrotopia.org> Tested-by: Daniel Golle <daniel@makrotopia.org> Signed-off-by: Felix Fietkau <nbd@nbd.name>
Lorenzo Bianconi [Thu, 22 Jun 2023 16:50:32 +0000 (18:50 +0200)]
wifi: mt76: connac: add connac3 mac library
Introduce connac3_mac in mt76_connac library to reuse mac code shared
between WiFi7 chipsets. So far connac3 library contains just radiotap
parsing code.
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org> Signed-off-by: Felix Fietkau <nbd@nbd.name>
Lorenzo Bianconi [Thu, 22 Jun 2023 16:50:29 +0000 (18:50 +0200)]
mt76: connac: move more mt7921/mt7915 mac shared code in connac lib
Move the following routines in mt76-connac lib since they are shared
between mt7915 and mt7921:
- mt76_connac2_tx_check_aggr
- mt76_connac2_txwi_free
- mt76_connac2_tx_token_put
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org> Signed-off-by: Felix Fietkau <nbd@nbd.name>
StanleyYP Wang [Mon, 5 Jun 2023 15:21:38 +0000 (23:21 +0800)]
wifi: mt76: mt7996: use correct phy for background radar event
If driver directly uses the band_idx reported from the radar event to
access mt76_phy array, it will get the wrong phy for background radar.
Fix this by adjusting the statement.
Fixes: 98686cd21624 ("wifi: mt76: mt7996: add driver for MediaTek Wi-Fi 7 (802.11be) devices") Signed-off-by: StanleyYP Wang <StanleyYP.Wang@mediatek.com> Signed-off-by: Shayne Chen <shayne.chen@mediatek.com> Signed-off-by: Felix Fietkau <nbd@nbd.name>
Peter Chiu [Mon, 5 Jun 2023 15:21:37 +0000 (23:21 +0800)]
wifi: mt76: connac: add support to set ifs time by mcu command
There's a race between driver and fw on some tx/rx control registers
when setting ifs, which will cause accidental hw queue pause problems.
Avoid this by setting ifs time with bss_info mcu command.
Reviewed-by: Shayne Chen <shayne.chen@mediatek.com> Signed-off-by: Peter Chiu <chui-hao.chiu@mediatek.com> Signed-off-by: Shayne Chen <shayne.chen@mediatek.com> Signed-off-by: Felix Fietkau <nbd@nbd.name>
Peter Chiu [Mon, 5 Jun 2023 15:21:34 +0000 (23:21 +0800)]
wifi: mt76: mt7996: fix bss wlan_idx when sending bss_info command
The bmc_tx_wlan_idx should be the wlan_idx of the current bss rather
than peer AP's wlan_idx, otherwise there will appear some frame
decryption problems on station mode.
Fixes: 98686cd21624 ("wifi: mt76: mt7996: add driver for MediaTek Wi-Fi 7 (802.11be) devices") Reviewed-by: Shayne Chen <shayne.chen@mediatek.com> Signed-off-by: Peter Chiu <chui-hao.chiu@mediatek.com> Signed-off-by: Shayne Chen <shayne.chen@mediatek.com> Signed-off-by: Felix Fietkau <nbd@nbd.name>
Peter Chiu [Mon, 5 Jun 2023 15:21:33 +0000 (23:21 +0800)]
wifi: mt76: connac: add support for dsp firmware download
Add FW_START_WORKING_PDA_DSP for the indication of starting DSP
firmware download, which is for phy-related control.
The firmware is transparent to the driver, but it's necessary for the
firmware download process.
Reviewed-by: Shayne Chen <shayne.chen@mediatek.com> Signed-off-by: Peter Chiu <chui-hao.chiu@mediatek.com> Signed-off-by: Shayne Chen <shayne.chen@mediatek.com> Signed-off-by: Felix Fietkau <nbd@nbd.name>
Rany Hany [Sun, 28 May 2023 18:04:28 +0000 (21:04 +0300)]
wifi: mt76: mt7915: fix command timeout in AP stop period
Due to AP stop improperly, mt7915 driver would face random command timeout
by chip fw problem. Migrate AP start/stop process to .start_ap/.stop_ap and
congiure BSS network settings in both hooks.
The new flow is shown below.
* AP start
.start_ap()
configure BSS network resource
set BSS to connected state
.bss_info_changed()
enable fw beacon offload
* AP stop
.bss_info_changed()
disable fw beacon offload (skip this command)
.stop_ap()
set BSS to disconnected state (beacon offload disabled automatically)
destroy BSS network resource
Based on "mt76: mt7921: fix command timeout in AP stop period"
Signed-off-by: Rany Hany <rany_hany@riseup.net> Signed-off-by: Felix Fietkau <nbd@nbd.name>
Ming Yen Hsieh [Thu, 18 May 2023 14:08:14 +0000 (22:08 +0800)]
wifi: mt76: mt7921: fix non-PSC channel scan fail
Due to the scan command may only request legacy bands and PSC channel
in 6GHz band, we are unable to scan the APs on non-PSC channel in this
case. Enable WIPHY_FLAG_SPLIT_SCAN_6GHZ to support non-PSC channel
(obtained during scan on legacy bands) in 6GHz scan request.
Fixes: 50ac15a511e3 ("mt76: mt7921: add 6GHz support") Signed-off-by: Ming Yen Hsieh <mingyen.hsieh@mediatek.com> Signed-off-by: Deren Wu <deren.wu@mediatek.com> Signed-off-by: Felix Fietkau <nbd@nbd.name>
Deren Wu [Wed, 17 May 2023 09:18:24 +0000 (17:18 +0800)]
wifi: mt76: mt7921: fix skb leak by txs missing in AMSDU
txs may be dropped if the frame is aggregated in AMSDU. When the problem
shows up, some SKBs would be hold in driver to cause network stopped
temporarily. Even if the problem can be recovered by txs timeout handling,
mt7921 still need to disable txs in AMSDU to avoid this issue.
Cc: stable@vger.kernel.org Fixes: 163f4d22c118 ("mt76: mt7921: add MAC support") Reviewed-by: Shayne Chen <shayne.chen@mediatek.com> Signed-off-by: Deren Wu <deren.wu@mediatek.com> Reviewed-by: Simon Horman <simon.horman@corigine.com> Signed-off-by: Felix Fietkau <nbd@nbd.name>
Alexander Couzens [Mon, 15 May 2023 13:27:12 +0000 (15:27 +0200)]
wifi: mt76: mt7915: add support for MT7981
Add support for the MediaTek MT7981 SoC which is similar to the MT7986
but with a newer IP cores and only 2x ARM Cortex-A53 instead of 4x.
Unlike MT7986 the MT7981 can only connect a single wireless frontend,
usually MT7976 is used for DBDC.
Signed-off-by: Alexander Couzens <lynxis@fe80.eu> Signed-off-by: Daniel Golle <daniel@makrotopia.org> Reviewed-by: Simon Horman <simon.horman@corigine.com> Signed-off-by: Felix Fietkau <nbd@nbd.name>
Peter Chiu [Wed, 10 May 2023 04:53:19 +0000 (12:53 +0800)]
wifi: mt76: report non-binding skb tx rate when WED is active
AQL relies on accurate tx rate reporting, so add this patch for
non-binding skb rate when WED is enabled.
Co-developed-by: Ryder Lee <ryder.lee@mediatek.com> Signed-off-by: Ryder Lee <ryder.lee@mediatek.com> Signed-off-by: Peter Chiu <chui-hao.chiu@mediatek.com> Signed-off-by: Felix Fietkau <nbd@nbd.name>
Peter Chiu [Wed, 10 May 2023 04:53:18 +0000 (12:53 +0800)]
wifi: mt76: mt7915: rework tx bytes counting when WED is active
Concurrent binding/non-binding skbs could be handled anywhere which leads
to mixed byte counting, so switch to use PPDU TxS reporting regardless Tx
paths when WED is active.
Fixes: 43eaa3689507 ("wifi: mt76: add PPDU based TxS support for WED device") Co-developed-by: Ryder Lee <ryder.lee@mediatek.com> Signed-off-by: Ryder Lee <ryder.lee@mediatek.com> Signed-off-by: Peter Chiu <chui-hao.chiu@mediatek.com> Signed-off-by: Felix Fietkau <nbd@nbd.name>
Peter Chiu [Wed, 10 May 2023 04:53:17 +0000 (12:53 +0800)]
wifi: mt76: mt7915: rework tx packets counting when WED is active
PPDU TxS can only report MPDU count whereas mac80211 requires MSDU scale
(NL80211_STA_INFO_TX_PACKETS), so switch to get MSDU counts from WA
statistic.
Note that mt7915 WA firmware only counts tx_packet for WED path, so driver
needs to take care of host path additionally.
Fixes: 43eaa3689507 ("wifi: mt76: add PPDU based TxS support for WED device") Co-developed-by: Ryder Lee <ryder.lee@mediatek.com> Signed-off-by: Ryder Lee <ryder.lee@mediatek.com> Signed-off-by: Peter Chiu <chui-hao.chiu@mediatek.com> Signed-off-by: Felix Fietkau <nbd@nbd.name>
Ryder Lee [Wed, 10 May 2023 04:53:16 +0000 (12:53 +0800)]
wifi: mt76: mt7915: report tx retries/failed counts for non-WED path
Get missing tx retries/failed counts from txfree done events and report
them via mt7915_sta_statistics().
Co-developed-by: Peter Chiu <chui-hao.chiu@mediatek.com> Signed-off-by: Peter Chiu <chui-hao.chiu@mediatek.com> Signed-off-by: Ryder Lee <ryder.lee@mediatek.com> Signed-off-by: Felix Fietkau <nbd@nbd.name>
StanleyYP Wang [Tue, 9 May 2023 03:11:57 +0000 (11:11 +0800)]
wifi: mt76: mt7915: fix background radar event being blocked
The background radar uses MT_RX_SEL2 as its band indication, so fix it.
Fixes: 7a53eecd5c87 (wifi: mt76: mt7915: check the correctness of event data) Signed-off-by: StanleyYP Wang <StanleyYP.Wang@mediatek.com> Signed-off-by: Ryder Lee <ryder.lee@mediatek.com> Signed-off-by: Felix Fietkau <nbd@nbd.name>
Ryder Lee [Tue, 2 May 2023 03:10:06 +0000 (11:10 +0800)]
wifi: mt76: mt7615: enable BSS_CHANGED_MU_GROUPS support
The Group ID Management frame is an Action frame of category VHT.
It is transmitted by the AP to assign or change the user position
of a STA for one or more group IDs.
Also, sniffer can use a given group id to monitor STA that belongs
to that group.
Notify underlying driver of changes via BSS_CHANGED_MU_GROUPS.
Signed-off-by: Ryder Lee <ryder.lee@mediatek.com> Signed-off-by: Felix Fietkau <nbd@nbd.name>
Ryder Lee [Tue, 2 May 2023 03:10:05 +0000 (11:10 +0800)]
wifi: mt76: mt7996: enable BSS_CHANGED_MU_GROUPS support
The Group ID Management frame is an Action frame of category VHT.
It is transmitted by the AP to assign or change the user position
of a STA for one or more group IDs.
Also, sniffer can use a given group id to monitor STA that belongs
to that group.
Signed-off-by: Ryder Lee <ryder.lee@mediatek.com> Signed-off-by: Felix Fietkau <nbd@nbd.name>
Ryder Lee [Wed, 26 Apr 2023 23:05:15 +0000 (07:05 +0800)]
wifi: mt76: mt7996: fix header translation logic
When header translation failure is indicated, the hardware will insert
an extra 2-byte field containing the data length after the protocol
type field. This happens either when the LLC-SNAP pattern did not match,
or if a VLAN header was detected.
The previous commit accidentally breaks the logic, so reverts back.
Fixes: 27db47ab1f47 (wifi: mt76: mt7996: enable mesh HW amsdu/de-amsdu support) Signed-off-by: Ryder Lee <ryder.lee@mediatek.com> Signed-off-by: Felix Fietkau <nbd@nbd.name>
Matt Whitlock [Thu, 20 Apr 2023 19:24:51 +0000 (15:24 -0400)]
mt76: mt7921: don't assume adequate headroom for SDIO headers
mt7921_usb_sdio_tx_prepare_skb() calls mt7921_usb_sdio_write_txwi() and
mt7921_skb_add_usb_sdio_hdr(), both of which blindly assume that
adequate headroom will be available in the passed skb. This assumption
typically is satisfied when the skb was allocated in the net core for
transmission via the mt7921 netdev (although even that is only an
optimization and is not strictly guaranteed), but the assumption is
sometimes not satisfied when the skb originated in the receive path of
another netdev and was passed through to the mt7921, such as by the
bridge layer. Blindly prepending bytes to an skb is always wrong.
This commit introduces a call to skb_cow_head() before the call to
mt7921_usb_sdio_write_txwi() in mt7921_usb_sdio_tx_prepare_skb() to
ensure that at least MT_SDIO_TXD_SIZE + MT_SDIO_HDR_SIZE bytes can be
pushed onto the skb.
Without this fix, I can trivially cause kernel panics by bridging an
MT7921AU-based USB 802.11ax interface with an Ethernet interface on an
Intel Atom-based x86 system using its onboard RTL8169 PCI Ethernet
adapter and also on an ARM-based Raspberry Pi 1 using its onboard
SMSC9512 USB Ethernet adapter. Note that the panics do not occur in
every system configuration, as they occur only if the receiving netdev
leaves less headroom in its received skbs than the mt7921 needs for its
SDIO headers.
Here is an example stack trace of this panic on Raspberry Pi OS Lite
2023-02-21 running kernel 6.1.24+ [1]:
skb_panic from skb_push+0x44/0x48
skb_push from mt7921_usb_sdio_tx_prepare_skb+0xd4/0x190 [mt7921_common]
mt7921_usb_sdio_tx_prepare_skb [mt7921_common] from mt76u_tx_queue_skb+0x94/0x1d0 [mt76_usb]
mt76u_tx_queue_skb [mt76_usb] from __mt76_tx_queue_skb+0x4c/0xc8 [mt76]
__mt76_tx_queue_skb [mt76] from mt76_txq_schedule.part.0+0x13c/0x398 [mt76]
mt76_txq_schedule.part.0 [mt76] from mt76_txq_schedule_all+0x24/0x30 [mt76]
mt76_txq_schedule_all [mt76] from mt7921_tx_worker+0x58/0xf4 [mt7921_common]
mt7921_tx_worker [mt7921_common] from __mt76_worker_fn+0x9c/0xec [mt76]
__mt76_worker_fn [mt76] from kthread+0xbc/0xe0
kthread from ret_from_fork+0x14/0x34
After this fix, bridging the mt7921 interface works fine on both of my
previously problematic systems.
Johannes Wiesboeck [Thu, 6 Jul 2023 09:34:36 +0000 (11:34 +0200)]
wifi: mwifiex: Set WIPHY_FLAG_NETNS_OK flag
Allow moving the wiphy device for mwififex to a non-inital network
namespace. Many wireless drivers enable this flag implicitly by
initializing through the generic ieee80211_alloc_hw_nm function. As
mwifiex does not utilize this function WIPHY_FLAG_NETNS_OK must be set
explicitly.
Moving to network namespace and working connection tested using
wpa_supplicant from a network namespace as well as from a container on
GyroidOS. The hardware was the Toradex Apalis i.MX8QM Board using the
Azurewave AW-CM276NF wireless module.
Dmitry Antipov [Thu, 29 Jun 2023 08:51:01 +0000 (11:51 +0300)]
wifi: mwifiex: fix fortify warning
When compiling with gcc 13.1 and CONFIG_FORTIFY_SOURCE=y,
I've noticed the following:
In function ‘fortify_memcpy_chk’,
inlined from ‘mwifiex_construct_tdls_action_frame’ at drivers/net/wireless/marvell/mwifiex/tdls.c:765:3,
inlined from ‘mwifiex_send_tdls_action_frame’ at drivers/net/wireless/marvell/mwifiex/tdls.c:856:6:
./include/linux/fortify-string.h:529:25: warning: call to ‘__read_overflow2_field’
declared with attribute warning: detected read beyond size of field (2nd parameter);
maybe use struct_group()? [-Wattribute-warning]
529 | __read_overflow2_field(q_size_field, size);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
and it happens because the fortification logic interprets this
as an attempt to overread 1-byte 'u.action.category' member of
'struct ieee80211_mgmt'. To silence this warning, it's enough
to pass an address of 'u.action' itself instead of an address
of its first member.
This also fixes an improper usage of 'sizeof()'. Since 'skb' is
extended with 'sizeof(mgmt->u.action.u.tdls_discover_resp) + 1'
bytes (where 1 is actually 'sizeof(mgmt->u.action.category)'),
I assume that the same number of bytes should be copied.
Suggested-by: Brian Norris <briannorris@chromium.org> Signed-off-by: Dmitry Antipov <dmantipov@yandex.ru> Reviewed-by: Brian Norris <briannorris@chromium.org> Signed-off-by: Kalle Valo <kvalo@kernel.org> Link: https://lore.kernel.org/r/20230629085115.180499-2-dmantipov@yandex.ru
wifi: rtw89: debug: Fix error handling in rtw89_debug_priv_btc_manual_set()
If there is a failure during kstrtobool_from_user()
rtw89_debug_priv_btc_manual_set should return a negative error code
instead of returning the count directly.
Fix this bug by returning an error code instead of a count after
a failed call of the function "kstrtobool_from_user". Moreover
I omitted the label "out" with this source code correction.
Dmitry Antipov [Wed, 28 Jun 2023 07:23:18 +0000 (10:23 +0300)]
wifi: rtw88: simplify vif iterators
Since all iterators called by 'rtw_iterate_vifs()' never uses
'mac' argument, it may be omitted, and 'struct rtw_vifs_entry'
may be simplified accordingly.
Dmitry Antipov [Wed, 28 Jun 2023 07:23:17 +0000 (10:23 +0300)]
wifi: rtw88: remove unused USB bulkout size set
Drop no longer used 'bulkout_size' of 'struct rtw_usb' as well
as related macros from usb.h and leftovers in 'rtw_usb_parse()'.
This follows commit 462c8db6a011 ("wifi: rtw88: usb: drop
now unnecessary URB size check").
Dmitry Antipov [Wed, 28 Jun 2023 07:23:16 +0000 (10:23 +0300)]
wifi: rtw88: remove unused and set but unused leftovers
Drop unused and set but unused 'last_push' of 'struct rtw_txq',
'wireless_set' of 'struct rtw_sta_info', 'usb_txagg_num' of
'struct rtw_usb' and 'n' of 'struct rx_usb_ctrl_block', unused
definition of 'struct rtw_timer_list', adjust related code.
Dmitry Antipov [Wed, 28 Jun 2023 07:23:15 +0000 (10:23 +0300)]
wifi: rtw88: delete timer and free skb queue when unloading
Fix possible crash and memory leak on driver unload by deleting
TX purge timer and freeing C2H queue in 'rtw_core_deinit()',
shrink critical section in the latter by freeing COEX queue
out of TX report lock scope.
Eric Dumazet [Thu, 20 Jul 2023 11:09:01 +0000 (11:09 +0000)]
ipv6: remove hard coded limitation on ipv6_pinfo
IPv6 inet sockets are supposed to have a "struct ipv6_pinfo"
field at the end of their definition, so that inet6_sk_generic()
can derive from socket size the offset of the "struct ipv6_pinfo".
This is very fragile, and prevents adding bigger alignment
in sockets, because inet6_sk_generic() does not work
if the compiler adds padding after the ipv6_pinfo component.
We are currently working on a patch series to reorganize
TCP structures for better data locality and found issues
similar to the one fixed in commit f5d547676ca0
("tcp: fix tcp_inet6_sk() for 32bit kernels")
Alternative would be to force an alignment on "struct ipv6_pinfo",
greater or equal to __alignof__(any ipv6 sock) to ensure there is
no padding. This does not look great.
v2: fix typo in mptcp_proto_v6_init() (Paolo)
Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Chao Wu <wwchao@google.com> Cc: Wei Wang <weiwan@google.com> Cc: Coco Li <lixiaoyan@google.com> Cc: YiFei Zhu <zhuyifei@google.com> Reviewed-by: Simon Horman <simon.horman@corigine.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Patrick Rohr [Wed, 19 Jul 2023 14:52:13 +0000 (07:52 -0700)]
net: add sysctl accept_ra_min_rtr_lft
This change adds a new sysctl accept_ra_min_rtr_lft to specify the
minimum acceptable router lifetime in an RA. If the received RA router
lifetime is less than the configured value (and not 0), the RA is
ignored.
This is useful for mobile devices, whose battery life can be impacted
by networks that configure RAs with a short lifetime. On such networks,
the device should never gain IPv6 provisioning and should attempt to
drop RAs via hardware offload, if available.
Signed-off-by: Patrick Rohr <prohr@google.com> Cc: Maciej Żenczykowski <maze@google.com> Cc: Lorenzo Colitti <lorenzo@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
`strncpy` is deprecated for use on NUL-terminated destination strings [1].
Even call sites utilizing length-bounded destination buffers should
switch over to using `strtomem` or `strtomem_pad`. In this case,
however, the compiler is unable to determine the size of the `data`
buffer which renders `strtomem` unusable. Due to this, `strscpy`
should be used.
It should be noted that most call sites already zero-initialize the
destination buffer. However, I've opted to use `strscpy_pad` to maintain
the same exact behavior that `strncpy` produced (zero-padded tail up to
`len`).
Link: https://github.com/KSPP/linux/issues/90 Reviewed-by: Nick Desaulniers <ndesaulniers@google.com> Reviewed-by: Kees Cook <keescook@chromium.org> Signed-off-by: Justin Stitt <justinstitt@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
====================
Process connector bug fixes & enhancements
Oracle DB is trying to solve a performance overhead problem it has been
facing for the past 10 years and using this patch series, we can fix this
issue.
Oracle DB runs on a large scale with 100000s of short lived processes,
starting up and exiting quickly. A process monitoring DB daemon which
tracks and cleans up after processes that have died without a proper exit
needs notifications only when a process died with a non-zero exit code
(which should be rare).
Due to the pmon architecture, which is distributed, each process is
independent and has minimal interaction with pmon. Hence fd based
solutions to track a process's spawning and exit cannot be used. Pmon
needs to detect the abnormal death of a process so it can cleanup after.
Currently it resorts to checking /proc every few seconds. Other methods
we tried like using system call to reduce the above overhead were not
accepted upstream.
With this change, we add event based filtering to proc connector module
so that DB can only listen to the events it is interested in. A new
event type PROC_EVENT_NONZERO_EXIT is added, which is only sent by kernel
to a listening application when any process exiting has a non-zero exit
status.
This change will give Oracle DB substantial performance savings - it takes
50ms to scan about 8K PIDs in /proc, about 500ms for 100K PIDs. DB does
this check every 3 secs, so over an hour we save 10secs for 100K PIDs.
With this, a client can register to listen for only exit or fork or a mix or
all of the events. This greatly enhances performance - currently, we
need to listen to all events, and there are 9 different types of events.
For eg. handling 3 types of events - 8K-forks + 8K-exits + 8K-execs takes
200ms, whereas handling 2 types - 8K-forks + 8K-exits takes about 150ms,
and handling just one type - 8K exits takes about 70ms.
Measuring the time using pidfds for monitoring 8K process exits took 4
times longer - 200ms, as compared to 70ms using only exit notifications
of proc connector. Hence, we cannot use pidfd for our use case.
This kind of a new event could also be useful to other applications like
Google's lmkd daemon, which needs a killed process's exit notification.
This patch series is organized as follows -
Patch 1 : Needed for patch 3 to work.
Patch 2 : Needed for patch 3 to work.
Patch 3 : Fixes some bugs in proc connector, details in the patch.
Patch 4 : Adds event based filtering for performance enhancements.
Patch 5 : Allow non-root users access to proc connector events.
Patch 6 : Selftest code for proc connector.
v9->v10 changes:
- Rebased to net-next, re-compiled and re-tested.
v8->v9 changes:
- Added sha1 ("title") of reversed patch as suggested by Eric Dumazet.
v7->v8 changes:
- Fixed an issue pointed by Liam Howlett in v7.
v5->v6 changes:
- Incorporated Liam Howlett's comments
- Removed FILTER define from proc_filter.c and added a "-f" run-time
option to run new filter code.
- Made proc_filter.c a selftest in tools/testing/selftests/connector
v4->v5 changes:
- Change the cover letter
- Fix a small issue in proc_filter.c
v3->v4 changes:
- Fix comments by Jakub Kicinski to incorporate root access changes
within bind call of connector
v2->v3 changes:
- Fix comments by Jakub Kicinski to separate netlink (patch 2) (after
layering) from connector fixes (patch 3).
- Minor fixes suggested by Jakub.
- Add new multicast group level permissions check at netlink layer.
Split this into netlink & connector layers (patches 6 & 7)
v1->v2 changes:
- Fix comments by Jakub Kicinski to keep layering within netlink and
update kdocs.
- Move non-root users access patch last in series so remaining patches
can go in first.
v->v1 changes:
- Changed commit log in patch 4 as suggested by Christian Brauner
- Changed patch 4 to make more fine grained access to non-root users
- Fixed warning in cn_proc.c, Reported-by: kernel test robot <lkp@intel.com>
- Fixed some existing warnings in cn_proc.c
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Run as ./proc_filter -f to run new filter code. Run without "-f" to run
usual proc connector code without the new filtering code.
Signed-off-by: Anjali Kulkarni <anjali.k.kulkarni@oracle.com> Reviewed-by: Liam R. Howlett <Liam.Howlett@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net>
There were a couple of reasons for not allowing non-root users access
initially - one is there was some point no proper receive buffer
management in place for netlink multicast. But that should be long
fixed. See link below for more context.
Second is that some of the messages may contain data that is root only. But
this should be handled with a finer granularity, which is being done at the
protocol layer. The only problematic protocols are nf_queue and the
firewall netlink. Hence, this restriction for non-root access was relaxed
for NETLINK_ROUTE initially:
https://lore.kernel.org/all/20020612013101.A22399@wotan.suse.de/
This restriction has also been removed for following protocols:
NETLINK_KOBJECT_UEVENT, NETLINK_AUDIT, NETLINK_SOCK_DIAG,
NETLINK_GENERIC, NETLINK_SELINUX.
Since process connector messages are not sensitive (process fork, exit
notifications etc.), and anyone can read /proc data, we can allow non-root
access here. However, since process event notification is not the only
consumer of NETLINK_CONNECTOR, we can make this change even more
fine grained than the protocol level, by checking for multicast group
within the protocol.
Allow non-root access for NETLINK_CONNECTOR via NL_CFG_F_NONROOT_RECV
but add new bind function cn_bind(), which allows non-root access only
for CN_IDX_PROC multicast group.
Signed-off-by: Anjali Kulkarni <anjali.k.kulkarni@oracle.com> Reviewed-by: Liam R. Howlett <Liam.Howlett@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net>
This patch adds the capability to filter messages sent by the proc
connector on the event type supplied in the message from the client
to the connector. The client can register to listen for an event type
given in struct proc_input.
This event based filteting will greatly enhance performance - handling
8K exits takes about 70ms, whereas 8K-forks + 8K-exits takes about 150ms
& handling 8K-forks + 8K-exits + 8K-execs takes 200ms. There are currently
9 different types of events, and we need to listen to all of them. Also,
measuring the time using pidfds for monitoring 8K process exits took
much longer - 200ms, as compared to 70ms using only exit notifications of
proc connector.
We also add a new event type - PROC_EVENT_NONZERO_EXIT, which is
only sent by kernel to a listening application when any process exiting,
has a non-zero exit status. This will help the clients like Oracle DB,
where a monitoring process wants notfications for non-zero process exits
so it can cleanup after them.
This kind of a new event could also be useful to other applications like
Google's lmkd daemon, which needs a killed process's exit notification.
The patch takes care that existing clients using old mechanism of not
sending the event type work without any changes.
cn_filter function checks to see if the event type being notified via
proc connector matches the event type requested by client, before
sending(matches) or dropping(does not match) a packet.
Signed-off-by: Anjali Kulkarni <anjali.k.kulkarni@oracle.com> Reviewed-by: Liam R. Howlett <Liam.Howlett@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net>
The current proc connector code has the foll. bugs - if there are more
than one listeners for the proc connector messages, and one of them
deregisters for listening using PROC_CN_MCAST_IGNORE, they will still get
all proc connector messages, as long as there is another listener.
Another issue is if one client calls PROC_CN_MCAST_LISTEN, and another one
calls PROC_CN_MCAST_IGNORE, then both will end up not getting any messages.
This patch adds filtering and drops packet if client has sent
PROC_CN_MCAST_IGNORE. This data is stored in the client socket's
sk_user_data. In addition, we only increment or decrement
proc_event_num_listeners once per client. This fixes the above issues.
cn_release is the release function added for NETLINK_CONNECTOR. It uses
the newly added netlink_release function added to netlink_sock. It will
free sk_user_data.
Signed-off-by: Anjali Kulkarni <anjali.k.kulkarni@oracle.com> Reviewed-by: Liam R. Howlett <Liam.Howlett@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net>
A new function netlink_release is added in netlink_sock to store the
protocol's release function. This is called when the socket is deleted.
This can be supplied by the protocol via the release function in
netlink_kernel_cfg. This is being added for the NETLINK_CONNECTOR
protocol, so it can free it's data when socket is deleted.
Signed-off-by: Anjali Kulkarni <anjali.k.kulkarni@oracle.com> Reviewed-by: Liam R. Howlett <Liam.Howlett@oracle.com> Acked-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
netlink: Reverse the patch which removed filtering
To use filtering at the connector & cn_proc layers, we need to enable
filtering in the netlink layer. This reverses the patch which removed
netlink filtering - commit ID for that patch: 549017aa1bb7 (netlink: remove netlink_broadcast_filtered).
Signed-off-by: Anjali Kulkarni <anjali.k.kulkarni@oracle.com> Reviewed-by: Liam R. Howlett <Liam.Howlett@oracle.com> Acked-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
page_pool_return_page() is a historic artefact from before
recycling of pages attached to skbs was supported. Theoretical
uses for it may be thought up but in practice all existing
users can be converted to use skb_mark_for_recycle() instead.
This code was previously posted as part of the memory provider RFC.
https://lore.kernel.org/all/20230707183935.997267-1-kuba@kernel.org/
====================
Jakub Kicinski [Thu, 20 Jul 2023 01:04:09 +0000 (18:04 -0700)]
net: page_pool: merge page_pool_release_page() with page_pool_return_page()
Now that page_pool_release_page() is not exported we can
merge it with page_pool_return_page(). I believe that
the "Do not replace this with page_pool_return_page()"
comment was there in case page_pool_return_page() was
not inlined, to avoid two function calls.
Acked-by: Jesper Dangaard Brouer <hawk@kernel.org> Reviewed-by: Yunsheng Lin <linyunsheng@huawei.com> Link: https://lore.kernel.org/r/20230720010409.1967072-5-kuba@kernel.org Reviewed-by: Alexander Lobakin <aleksander.lobakin@intel.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Jakub Kicinski [Thu, 20 Jul 2023 01:04:08 +0000 (18:04 -0700)]
net: page_pool: hide page_pool_release_page()
There seems to be no user calling page_pool_release_page()
for legit reasons, all the users simply haven't been converted
to skb-based recycling, yet. Previous changes converted them.
Update the docs, and unexport the function.
Jakub Kicinski [Thu, 20 Jul 2023 01:04:06 +0000 (18:04 -0700)]
eth: tsnep: let page recycling happen with skbs
tsnep builds an skb with napi_build_skb() and then calls
page_pool_release_page() for the page in which that skb's
head sits. Use recycling instead, recycling of heads works
just fine.
genetlink: add explicit ordering break check for split ops
Currently, if cmd in the split ops array is of lower value than the
previous one, genl_validate_ops() continues to do the checks as if
the values are equal. This may result in non-obvious WARN_ON() hit in
these check.
Instead, check the incorrect ordering explicitly and put a WARN_ON()
in case it is broken.
David S. Miller [Fri, 21 Jul 2023 08:55:54 +0000 (09:55 +0100)]
Merge branch 'octeontx2-pf-round-robin-sched'
Hariprasad Kelam says:
====================
octeontx2-pf: support Round Robin scheduling
octeontx2 and CN10K silicons support Round Robin scheduling. When multiple
traffic flows reach transmit level with the same priority, with Round Robin
scheduling traffic flow with the highest quantum value is picked. With this
support, the user can add multiple classes with the same priority and
different quantum in htb offload.
This series of patches adds support for the same.
Patch1: implement transmit schedular allocation algorithm as preparation
for support round robin scheduling.
Patch2: Allow quantum parameter in HTB offload mode.
Patch3: extends octeontx2 htb offload support for Round Robin scheduling
Patch4: extend QOS documentation for Round Robin scheduling
Hariprasad Kelam (1):
docs: octeontx2: extend documentation for Round Robin scheduling
Naveen Mamindlapalli (3):
octeontx2-pf: implement transmit schedular allocation algorithm
sch_htb: Allow HTB quantum parameter in offload mode
octeontx2-pf: htb offload support for Round Robin scheduling
---
v4 * update classid values in documentation.
v3 * 1. update QOS documentation for round robin scheduling
2. added out of bound checks for quantum parameter
v2 * change data type of otx2_index_used to reduce size of structure
otx2_qos_cfg
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
octeontx2-pf: htb offload support for Round Robin scheduling
When multiple traffic flows reach Transmit level with the same
priority, with Round robin scheduling traffic flow with the highest
quantum value is picked. With this support, the user can add multiple
classes with the same priority and different quantum. This patch
does necessary changes to support the same.
Signed-off-by: Naveen Mamindlapalli <naveenm@marvell.com> Signed-off-by: Hariprasad Kelam <hkelam@marvell.com> Signed-off-by: David S. Miller <davem@davemloft.net>
sch_htb: Allow HTB quantum parameter in offload mode
The current implementation of HTB offload returns the EINVAL error for
quantum parameter. This patch removes the error returning checks for
'quantum' parameter and populates its value to tc_htb_qopt_offload
structure such that driver can use the same.
Add quantum parameter check in mlx5 driver, as mlx5 devices are not capable
of supporting the quantum parameter when htb offload is used. Report error
if quantum parameter is set to a non-default value.
Signed-off-by: Naveen Mamindlapalli <naveenm@marvell.com> Signed-off-by: Hariprasad Kelam <hkelam@marvell.com> Signed-off-by: David S. Miller <davem@davemloft.net>
projects / users / jedix / linux-maple.git / log