Re: Bug#806130: linux: netlink_sock_destruct() crash in netns

[Pablo Neira Ayuso <pablo@netfilter.org>]

On Thu, Nov 26, 2015 at 10:14:58PM +0000, Ben Hutchings wrote:
> On Tue, 2015-11-24 at 16:49 +0100, Arturo Borrero Gonzalez wrote:
> > Source: linux
> > Version: 4.2.6-1
> > Severity: normal
> > 
> > Dear Maintainer,
> > 
> > thanks for your work in the linux kernel.
> > 
> > I'm experimenting reports like this:
> > 
> > [mar nov 24 16:38:16 2015] ------------[ cut here ]------------
> > [mar nov 24 16:38:16 2015] WARNING: CPU: 1 PID: 22224 at /build/linux-CrHvZ_/linux-4.2.6/net/netlink/af_netlink.c:946 netlink_sock_destruct+0x103/0x140()
> > [mar nov 24 16:38:16 2015] Modules linked in: nft_nat nft_chain_nat_ipv4 nf_nat_ipv4 nf_nat nft_log fuse uas usb_storage nft_counter nf_conntrack_ipv6 nf_defrag_ipv6 nf_conntrack_ipv4 nf_defrag_ipv4 nft_ct nf_conntrack nft_meta nft_hash nft_rbtree nfnetlink_log nf_tables_inet nf_tables_ipv6 nf_tables_ipv4 nf_tables nfnetlink nouveau snd_hda_codec_idt snd_hda_codec_generic mxm_wmi wmi video ttm drm_kms_helper drm iTCO_wdt i2c_algo_bit iTCO_vendor_support coretemp kvm_intel lpc_ich kvm snd_hda_intel mfd_core i2c_i801 snd_hda_codec evdev snd_hda_core pcspkr psmouse serio_raw snd_hwdep snd_pcm sg rng_core tpm_tis tpm 8250_fintek shpchp button snd_timer snd soundcore acpi_cpufreq processor thermal_sys parport_pc ppdev lp parport autofs4 ext4 crc16 mbcache jbd2 dm_mod sr_mod cdrom sd_mod ata_generic floppy ata_piix e100
> > [mar nov 24 16:38:16 2015]  mii libata scsi_mod ehci_pci uhci_hcd ehci_hcd usbcore usb_common
> > [mar nov 24 16:38:16 2015] CPU: 1 PID: 22224 Comm: bash Tainted: G        W       4.2.0-1-amd64 #1 Debian 4.2.6-1
> [...]
> > I can reproduce the issue:
> > 
> >  % sudo ip netns add test
> >  % sudo ip netns exec test nft -f test.nft
> >  % sudo ip netns del test
> > 
> > Find attached the test.nft file.

sk       Eth Pid    Groups   Rmem     Wmem     Dump     Locks
Drops     Inode
ffff8800cbb11000 0   0      00000000 0        0        0 2        0 19855   
ffff88011504f800 4   0      00000000 0        0        0 2        0 19877   
ffff8801153f7000 9   0      00000000 0        0        0 2        0 19876   
ffff880030c23000 10  0      00000000 0        0        0 2        0 19858   
ffff8800cbabb800 12  0      00000000 -4544    0        0 2        0 19890   

wmem is underflowing.

Could you check if this also happening in older kernel versions?

No need to test under netns, you can just run nft -f and check
/proc/net/netlink.

Let me know, thanks.
--
To unsubscribe from this list: send the line "unsubscribe netfilter-devel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html