netdev - Re: Kernel warning on 5.15.187+ caused: net/sched: Always pass notifications when child class becomes empty

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <779ce04d-2053-4196-b989-f801720e65bc@hauke-m.de>
Date: Thu, 7 Aug 2025 19:34:29 +0200
From: Hauke Mehrtens <hauke@...ke-m.de>
To: Jamal Hadi Salim <jhs@...atatu.com>, Lion Ackermann <nnamrec@...il.com>
Cc: netdev@...r.kernel.org, stable@...r.kernel.org
Subject: Re: Kernel warning on 5.15.187+ caused: net/sched: Always pass
 notifications when child class becomes empty

On 8/6/25 14:48, Jamal Hadi Salim wrote:
> On Wed, Aug 6, 2025 at 2:35 AM Lion Ackermann <nnamrec@...il.com> wrote:
>>
>> Hi,
>>
>> On 8/5/25 10:28 PM, Hauke Mehrtens wrote:
>>> Hi,
>>>
>>> We see a kernel warning with kernel 5.15.187 and later in OpenWrt 23.05.
>>>
>>> When I revert this patch it is gone:
>>> commit e269f29e9395527bc00c213c6b15da04ebb35070
>>> Author: Lion Ackermann <nnamrec@...il.com>
>>> Date:   Mon Jun 30 15:27:30 2025 +0200
>>>      net/sched: Always pass notifications when child class becomes empty
>>>      [ Upstream commit 103406b38c600fec1fe375a77b27d87e314aea09 ]
>>>
>>> Link: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v5.15.189&id=e269f29e9395527bc00c213c6b15da04ebb35070
>>>
>>> Hauke
>>>
>>>
>>> Warning:
>>> [   90.429594] ------------[ cut here ]------------
>>> [   90.429929] WARNING: CPU: 0 PID: 3667 at net/sched/sch_htb.c:609 0xffff800000d1b5dc
>>> [   90.430561] Modules linked in: pppoe ppp_async nft_fib_inet nf_flow_table_ipv6 nf_flow_table_ipv4 nf_flow_table_inet pppox ppp_generic nft_reject_ipv6 nft_reject_ipv4 nft_reject_inet nft_reject nft_redir nft_quota nft_objref nft_numgen nft_nat nft_masq nft_log nft_limit nft_hash nft_flow_offload nft_fib_ipv6 nft_fib_ipv4 nft_fib nft_ct nft_counter nft_compat nft_chain_nat nf_tables nf_nat nf_flow_table nf_conntrack iptable_mangle iptable_filter ipt_REJECT ipt_ECN ip_tables xt_time xt_tcpudp xt_tcpmss xt_statistic xt_multiport xt_mark xt_mac xt_limit xt_length xt_hl xt_ecn xt_dscp xt_comment xt_TCPMSS xt_LOG xt_HL xt_DSCP xt_CLASSIFY x_tables smsc slhc sfp sch_cake ravb nfnetlink nf_reject_ipv6 nf_reject_ipv4 nf_log_syslog nf_defrag_ipv6 nf_defrag_ipv4 mdio_i2c mdio_gpio mdio_bitbang libcrc32c e1000e atlantic sch_tbf sch_ingress sch_htb sch_hfsc em_u32 cls_u32 cls_route cls_matchall cls_fw cls_flow cls_basic act_skbedit act_mirred act_gact gpio_pca953x i2c_mux_pca954x
>>> i2c_mux i2c_dev
>>> [   90.431030]  sp805_wdt dwmac_sun8i dwmac_rk dwmac_imx nicvf nicpf thunder_bgx thunder_xcv dwmac_generic stmmac_platform stmmac rvu_nicvf rvu_nicpf rvu_af rvu_mbox mvpp2 mvneta vmxnet3 fec fsl_enetc fsl_enetc_mdio fsl_enetc_ierb fsl_dpaa2_eth ifb mdio_thunder mdio_cavium mdio_bcm_unimac xgmac_mdio pcs_lynx fsl_mc_dpio genet nls_utf8 nls_iso8859_1 nls_cp437 pcs_xpcs marvell10g marvell macsec sha512_generic seqiv jitterentropy_rng drbg hmac rtc_rx8025 vfat fat ptp realtek broadcom bcm_phy_lib aquantia hwmon crc_ccitt pps_core phylink mii
>>> [   90.435728] CPU: 0 PID: 3667 Comm: tc Not tainted 5.15.189 #0
>>> [   90.436003] Hardware name: linux,dummy-virt (DT)
>>> [   90.436306] pstate: 60400005 (nZCv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--)
>>> [   90.436578] pc : 0xffff800000d1b5dc
>>> [   90.436736] lr : qdisc_tree_reduce_backlog+0xa0/0x120
>>> [   90.436930] sp : ffff80000b73b6d0
>>> [   90.437053] x29: ffff80000b73b6d0 x28: ffff000003d81030 x27: ffff0000033bf800
>>> [   90.437335] x26: ffff0000033bf800 x25: ffff80000b73b8e0 x24: ffff000003677400
>>> [   90.437576] x23: ffff800000d1f5e0 x22: 0000000000000000 x21: 0000000000000000
>>> [   90.437816] x20: 0000000000000000 x19: ffff0000034f6800 x18: 0000000000000000
>>> [   90.438054] x17: 0000000000000000 x16: 0000000000000000 x15: 0000ffffc7ffd278
>>> [   90.438299] x14: 000122a60001216d x13: 0001204400011f1c x12: 00011df300011cca
>>> [   90.438548] x11: 00011b9100011a68 x10: 0001194000011817 x9 : 0000000000000000
>>> [   90.438794] x8 : ffff000003677e00 x7 : 0000000000000000 x6 : 0000000000000000
>>> [   90.439037] x5 : ffff0000029df000 x4 : 0000000000000003 x3 : ffff000003684000
>>> [   90.439277] x2 : ffff800000d1b5a0 x1 : 0000000000000000 x0 : ffff000003c64140
>>> [   90.439629] Call trace:
>>> [   90.439859]  0xffff800000d1b5dc
>>> [   90.440021]  qdisc_tree_reduce_backlog+0xa0/0x120
>>> [   90.440184]  0xffff800000d1dd8c
>>> [   90.440302]  tc_ctl_tclass+0x15c/0x460
>>> [   90.440440]  rtnetlink_rcv_msg+0x1b8/0x320
>>> [   90.440583]  netlink_rcv_skb+0x5c/0x130
>>> [   90.440715]  rtnetlink_rcv+0x18/0x2c
>>> [   90.440841]  netlink_unicast+0x214/0x310
>>> [   90.440975]  netlink_sendmsg+0x1a0/0x3dc
>>> [   90.441120]  ____sys_sendmsg+0x1b8/0x220
>>> [   90.441257]  ___sys_sendmsg+0x84/0xf0
>>> [   90.441385]  __sys_sendmsg+0x48/0xb0
>>> [   90.441509]  __arm64_sys_sendmsg+0x24/0x30
>>> [   90.441647]  invoke_syscall.constprop.0+0x5c/0x110
>>> [   90.441811]  do_el0_svc+0x6c/0x150
>>> [   90.441929]  el0_svc+0x28/0xc0
>>> [   90.442039]  el0t_64_sync_handler+0xe8/0x114
>>> [   90.442186]  el0t_64_sync+0x1a4/0x1a8
>>> [   90.442408] ---[ end trace b27ff72c6be40920 ]---
>>>
>>
>> It looks like some of the "make class ops idempotent" patches are
>> missing on this stable tree (the patch depends on them)?
>>
>>      sch_htb: make htb_deactivate() idempotent
>>      sch_qfq: make qfq_qlen_notify() idempotent
>>      sch_hfsc: make hfsc_qlen_notify() idempotent
>>      sch_drr: make drr_qlen_notify() idempotent
> 
> I think you are right - I believe the following patches somehow are
> missing in 5.15.187:
> https://lore.kernel.org/netdev/20250707210801.372995-1-victor@mojatatu.com/
> https://lore.kernel.org/netdev/20250403211033.166059-1-xiyou.wangcong@gmail.com/
> 
> 5.15.189 seems to have Victor's patch but for some reason only one of
> several in Cong's patches was ever merged in stable, basically:
> https://lore.kernel.org/netdev/20250403211033.166059-6-xiyou.wangcong@gmail.com/
> 
> cheers,
> jamal

Hi,

Thanks for the help.

I cherry picked these patches on top of 5.15.189 and then I do not see 
the warning any more:
https://git.kernel.org/linus/5ba8b837b522d7051ef81bacf3d95383ff8edce5
https://git.kernel.org/linus/df008598b3a00be02a8051fde89ca0fbc416bd55
https://git.kernel.org/linus/51eb3b65544c9efd6a1026889ee5fb5aa62da3bb
https://git.kernel.org/linus/55f9eca4bfe30a15d8656f915922e8c98b7f0728

I will ask Greg to add them to stable 5.15 too.

Hauke