psock_snd.sh in net from ubuntu_kernel_selftests ADT failure with focal/groovy/hirsute/impish

[Kleber Sacilotto de Souza]

[Impact]
The expected to fail 'raw gso min size - 1' test in psock_snd.sh is
not failing as expected, and thus this test will fail with:

  raw gso min size - 1 (expected to fail)
  tx: 1524
  rx: 1472
  OK

This is because after commit 7c6d2ec ("net: be more gentle about silly
gso requests coming from user"), we relaxed the min gso_size check in
virtio_net_hdr_to_skb().

So when a packet which is smaller then the gso_size, GSO for this
packet will not be set, the packet will be send/recv successfully.

[Fix]
* cfba3fb68960b4 selftests/net: remove min gso test in packet_snd

This patch can be cherry-picked into all affected kernels.

I decided to skip those 5.8 variants, as the current cycle
(2021-09-06) might be the last cycle for them.

[Test]
Run this psock_snd.sh script, this test will no longer exist and the
script can proceed with the next test.

[Where problems could occur]
Removing this broken test should not have any negative impact to our
tests. We will need to update our hints later.

[Original Bug Report]
Testing failed on focal/linux 5.4.0-44.48:
    amd64: https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-focal/focal/amd64/l/linux/20200812_171444_31971@/log.gz
    arm64: https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-focal/focal/arm64/l/linux/20200812_210509_145fd@/log.gz
    ppc64el: https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-focal/focal/ppc64el/l/linux/20200812_165855_1dabd@/log.gz
    s390x: https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-focal/focal/s390x/l/linux/20200812_153600_9c7cb@/log.gz

psock_snd.sh output:

16:28:30 DEBUG| [stdout] # selftests: net: psock_snd.sh
16:28:31 DEBUG| [stdout] # dgram
16:28:31 DEBUG| [stdout] # tx: 128
16:28:31 DEBUG| [stdout] # rx: 142
16:28:31 DEBUG| [stdout] # rx: 100
16:28:31 DEBUG| [stdout] # OK
16:28:31 DEBUG| [stdout] #
16:28:31 DEBUG| [stdout] # dgram bind
16:28:31 DEBUG| [stdout] # tx: 128
16:28:31 DEBUG| [stdout] # rx: 142
16:28:31 DEBUG| [stdout] # rx: 100
16:28:31 DEBUG| [stdout] # OK
16:28:31 DEBUG| [stdout] #
16:28:31 DEBUG| [stdout] # raw
16:28:31 DEBUG| [stdout] # tx: 142
16:28:31 DEBUG| [stdout] # rx: 142
16:28:31 DEBUG| [stdout] # rx: 100
16:28:31 DEBUG| [stdout] # OK
16:28:31 DEBUG| [stdout] #
16:28:31 DEBUG| [stdout] # raw bind
16:28:31 DEBUG| [stdout] # tx: 142
16:28:31 DEBUG| [stdout] # rx: 142
16:28:31 DEBUG| [stdout] # rx: 100
16:28:31 DEBUG| [stdout] # OK
16:28:31 DEBUG| [stdout] #
16:28:31 DEBUG| [stdout] # raw qdisc bypass
16:28:31 DEBUG| [stdout] # tx: 142
16:28:31 DEBUG| [stdout] # rx: 142
16:28:31 DEBUG| [stdout] # rx: 100
16:28:31 DEBUG| [stdout] # OK
16:28:31 DEBUG| [stdout] #
16:28:31 DEBUG| [stdout] # raw vlan
16:28:31 DEBUG| [stdout] # tx: 146
16:28:31 DEBUG| [stdout] # rx: 100
16:28:31 DEBUG| [stdout] # OK
16:28:31 DEBUG| [stdout] #
16:28:31 DEBUG| [stdout] # raw vnet hdr
16:28:31 DEBUG| [stdout] # tx: 152
16:28:31 DEBUG| [stdout] # rx: 142
16:28:31 DEBUG| [stdout] # rx: 100
16:28:31 DEBUG| [stdout] # OK
16:28:31 DEBUG| [stdout] #
16:28:31 DEBUG| [stdout] # raw csum_off
16:28:31 DEBUG| [stdout] # tx: 152
16:28:31 DEBUG| [stdout] # rx: 142
16:28:31 DEBUG| [stdout] # rx: 100
16:28:31 DEBUG| [stdout] # OK
16:28:31 DEBUG| [stdout] #
16:28:31 DEBUG| [stdout] # raw csum_off with bad offset (fails)
16:28:31 DEBUG| [stdout] # ./psock_snd: write: Invalid argument
16:28:31 DEBUG| [stdout] # raw min size
16:28:31 DEBUG| [stdout] # tx: 42
16:28:31 DEBUG| [stdout] # rx: 0
16:28:31 DEBUG| [stdout] # OK
16:28:31 DEBUG| [stdout] #
16:28:31 DEBUG| [stdout] # raw mtu size
16:28:31 DEBUG| [stdout] # tx: 1514
16:28:31 DEBUG| [stdout] # rx: 1472
16:28:31 DEBUG| [stdout] # OK
16:28:31 DEBUG| [stdout] #
16:28:31 DEBUG| [stdout] # raw mtu size + 1 (fails)
16:28:31 DEBUG| [stdout] # ./psock_snd: write: Message too long
16:28:31 DEBUG| [stdout] # raw vlan mtu size + 1 (fails)
16:28:31 DEBUG| [stdout] # ./psock_snd: write: Message too long
16:28:32 DEBUG| [stdout] # dgram mtu size
16:28:32 DEBUG| [stdout] # tx: 1500
16:28:32 DEBUG| [stdout] # rx: 1472
16:28:32 DEBUG| [stdout] # OK
16:28:32 DEBUG| [stdout] #
16:28:32 DEBUG| [stdout] # dgram mtu size + 1 (fails)
16:28:32 DEBUG| [stdout] # ./psock_snd: write: Message too long
16:28:32 DEBUG| [stdout] # raw truncate hlen (fails: does not arrive)
16:28:32 DEBUG| [stdout] # tx: 14
16:28:32 DEBUG| [stdout] # ./psock_snd: recv: Resource temporarily unavailable
16:28:32 DEBUG| [stdout] # raw truncate hlen - 1 (fails: EINVAL)
16:28:32 DEBUG| [stdout] # ./psock_snd: write: Invalid argument
16:28:32 DEBUG| [stdout] # raw gso min size
16:28:32 DEBUG| [stdout] # tx: 1525
16:28:32 DEBUG| [stdout] # rx: 1473
16:28:32 DEBUG| [stdout] # OK
16:28:32 DEBUG| [stdout] #
16:28:32 DEBUG| [stdout] # raw gso min size - 1 (fails)
16:28:32 DEBUG| [stdout] # tx: 1524
16:28:32 DEBUG| [stdout] # rx: 1472
16:28:32 DEBUG| [stdout] # OK
16:28:32 DEBUG| [stdout] #
16:28:32 DEBUG| [stdout] not ok 22 selftests: net: psock_snd.sh # exit=1

This failure is similar to bug 1884234, however, this test didn't fail on ADT with focal/linux before 5.4.0-44.48.

[Ubuntu Kernel Bot (ubuntu-kernel-bot)]

This bug is missing log files that will aid in diagnosing the problem. While running an Ubuntu kernel (not a mainline or third-party kernel) please enter the following command in a terminal window:

apport-collect 1892213

and then change the status of the bug to 'Confirmed'.

If, due to the nature of the issue you have encountered, you are unable to run this command, please add a comment stating that fact and change the bug status to 'Confirmed'.

This change has been made by an automated script, maintained by the Ubuntu Kernel Team.

[Kleber Sacilotto de Souza (kleber-souza)]

I was able to reproduce it with mainline kernel 5.9.0-050900rc1-generic.

The testcase "raw gso min size - 1 (fails)" is failing because what it executes ("./in_netns.sh ./psock_snd -v -c -g -l 1472") is returning success when it expects a failure.

So I'm wondering if psock_snd.sh is making some assumption which is not valid anymore.

[Kleber Sacilotto de Souza (kleber-souza)]

I have reverted the following two commits from Ubuntu-5.4.0-44.48 and the test is now successfully:

commit 78aedad3a6b9264a27f30d2c1cf50e6f94d22530
Author: Eric Dumazet <email address hidden>
Date: Thu May 28 14:57:47 2020 -0700

    net: be more gentle about silly gso requests coming from user

    BugLink: https://bugs.launchpad.net/bugs/1883184

    [ Upstream commit 7c6d2ecbda83150b2036a2b36b21381ad4667762 ]

commit 3bfeea9b0b203e14acb2c9ba122da08e3a6eb93c
Author: Willem de Bruijn <email address hidden>
Date: Mon May 25 15:07:40 2020 -0400

    net: check untrusted gso_size at kernel entry

    BugLink: https://bugs.launchpad.net/bugs/1883184

    [ Upstream commit 6dd912f82680761d8fb6b1bb274a69d4c7010988 ]

The description from 7c6d2ecbda really seems to match the change that's causing the testcase to not fail anymore.

[Kelsey Steele (kelsey-steele)]

Failed on Bionic/oracle : 5.4.0-1026.26~18.04.1 : amd64

[Po-Hsu Lin (cypressyew)]

Found on 5.4.0-1030.32~18.04.1 - oracle

[Kelsey Steele (kelsey-steele)]

Found on Groovy/linux 5.8.0-31.33

[Ian May (ian-may)]

Found on Focal/aws 5.4.0-1042.44

[Guilherme G. Piccoli (gpiccoli)]

Found on Focal/kvm cycle sru-20210412

[Krzysztof Kozlowski (krzk)]

Found on focal/linux-gke 5.4.0-1043.45

[Guilherme G. Piccoli (gpiccoli)]

Observed in F/KVM, cycle sru-20210510.

[Guilherme G. Piccoli (gpiccoli)]

Observed on G/KVM, cycle sru-20210510 .

[Guilherme G. Piccoli (gpiccoli)]

Observed on H/KVM, cycle sru-20210510 .

[Kleber Sacilotto de Souza (kleber-souza)]

Seen with focal/linux 5.4.0-75.84.

[Guilherme G. Piccoli (gpiccoli)]

Observed on G/KVM, cycle sru-20210531 .

[Kleber Sacilotto de Souza (kleber-souza)]

Seen with focal/linux: 5.4.0-76.85.

[Kleber Sacilotto de Souza (kleber-souza)]

Issue found with bionic/linux-hwe-5.4 5.4.0-79.88~18.04.1.

[Guilherme G. Piccoli (gpiccoli)]

Observed this on B/oracle-5.4, cycle sru-20210621 .

[Po-Hsu Lin (cypressyew)]

https://lists.ubuntu.com/archives/kernel-team/2021-September/123898.html

[Ubuntu Kernel Bot (ubuntu-kernel-bot)]

This bug is awaiting verification that the kernel in -proposed solves the problem. Please test the kernel and update this bug with the results. If the problem is solved, change the tag 'verification-needed-focal' to 'verification-done-focal'. If the problem still exists, change the tag 'verification-needed-focal' to 'verification-failed-focal'.

If verification is not done by 5 working days from today, this fix will be dropped from the source code, and this bug will be closed.

See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Thank you!

[Po-Hsu Lin (cypressyew)]

Verified on F-oem-5.10.0-1047.49

ok 22 selftests: net: psock_snd.sh

[Ubuntu Kernel Bot (ubuntu-kernel-bot)]

This bug is awaiting verification that the linux/5.11.0-38.42 kernel in -proposed solves the problem. Please test the kernel and update this bug with the results. If the problem is solved, change the tag 'verification-needed-hirsute' to 'verification-done-hirsute'. If the problem still exists, change the tag 'verification-needed-hirsute' to 'verification-failed-hirsute'.

If verification is not done by 5 working days from today, this fix will be dropped from the source code, and this bug will be closed.

See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Thank you!

[Launchpad Janitor (janitor)]

This bug was fixed in the package linux-oem-5.13 - 5.13.0-1014.18

---------------
linux-oem-5.13 (5.13.0-1014.18) focal; urgency=medium

  * focal/linux-oem-5.13: 5.13.0-1014.18 -proposed tracker (LP: #1944429)

  * CVE-2021-40490
    - ext4: fix race writing to an inline_data file while its xattrs are changing

  * CVE-2021-41073
    - io_uring: ensure symmetry in handling iter types in loop_rw_iter()

 -- Timo Aaltonen <email address hidden> Wed, 22 Sep 2021 18:34:24 +0300

[Launchpad Janitor (janitor)]

This bug was fixed in the package linux-oem-5.10 - 5.10.0-1049.51

---------------
linux-oem-5.10 (5.10.0-1049.51) focal; urgency=medium

  * focal/linux-oem-5.10: 5.10.0-1049.50 -proposed tracker (LP: #1944209)

  * e1000e extremly slow (LP: #1930754)
    - SAUCE: e1000e: Separate TGP board type from SPT
    - SAUCE: e1000e: Fixing packet loss issues on new platforms

  * CVE-2021-41073
    - io_uring: ensure symmetry in handling iter types in loop_rw_iter()

 -- Chia-Lin Kao (AceLan) <email address hidden> Mon, 27 Sep 2021 18:33:36 +0800

[Stefan Bader (smb)]

Looking at the latest ADT run [1] this subtest is no longer failing.

[1] https://autopkgtest.ubuntu.com/results/autopkgtest-hirsute/hirsute/amd64/l/linux/20210928_124449_9b7e9@/log.gz

[Launchpad Janitor (janitor)]

This bug was fixed in the package linux - 5.13.0-19.19

---------------
linux (5.13.0-19.19) impish; urgency=medium

  * impish/linux: 5.13.0-19.19 -proposed tracker (LP: #1946337)

  * impish:linux-aws 5.13 panic during systemd autotest (LP: #1946001)
    - [Config] disable KFENCE

 -- Andrea Righi <email address hidden> Thu, 07 Oct 2021 11:09:51 +0200

[Kleber Sacilotto de Souza (kleber-souza)]

The latest ADT results for bionic/linux-hwe-5.4 (5.4.0-89.100~18.04.1) show that psock_snd.sh is no longer failing.

https://autopkgtest.ubuntu.com/results/autopkgtest-bionic/bionic/amd64/l/linux-hwe-5.4/20211002_141345_704e6@/log.gz

[Launchpad Janitor (janitor)]

This bug was fixed in the package linux - 5.11.0-38.42

---------------
linux (5.11.0-38.42) hirsute; urgency=medium

  * hirsute/linux: 5.11.0-38.42 -proposed tracker (LP: #1944863)

  * Packaging resync (LP: #1786013)
    - debian/dkms-versions -- update from kernel-versions (main/2021.09.27)

  * Wobbly graphics on built-in display since linux-image-5.11.0-22-generic
    (LP: #1936708)
    - drm/i915/dp: Use max params for panels < eDP 1.4

  * Hirsute update: upstream stable patchset 2021-09-22 (LP: #1944610)
    - net: qrtr: fix another OOB Read in qrtr_endpoint_post
    - bpf: Fix ringbuf helper function compatibility
    - bpf: Fix NULL pointer dereference in bpf_get_local_storage() helper
    - ASoC: rt5682: Adjust headset volume button threshold
    - ASoC: component: Remove misplaced prefix handling in pin control functions
    - ARC: Fix CONFIG_STACKDEPOT
    - netfilter: conntrack: collect all entries in one cycle
    - once: Fix panic when module unload
    - blk-iocost: fix lockdep warning on blkcg->lock
    - ovl: fix uninitialized pointer read in ovl_lookup_real_one()
    - net: mscc: Fix non-GPL export of regmap APIs
    - can: usb: esd_usb2: esd_usb2_rx_event(): fix the interchange of the CAN RX
      and TX error counters
    - ceph: correctly handle releasing an embedded cap flush
    - riscv: Ensure the value of FP registers in the core dump file is up to date
    - Revert "btrfs: compression: don't try to compress if we don't have enough
      pages"
    - drm/amdgpu: Cancel delayed work when GFXOFF is disabled
    - Revert "USB: serial: ch341: fix character loss at high transfer rates"
    - USB: serial: option: add new VID/PID to support Fibocom FG150
    - usb: renesas-xhci: Prefer firmware loading on unknown ROM state
    - usb: dwc3: gadget: Fix dwc3_calc_trbs_left()
    - usb: dwc3: gadget: Stop EP0 transfers during pullup disable
    - scsi: core: Fix hang of freezing queue between blocking and running device
    - RDMA/bnxt_re: Add missing spin lock initialization
    - IB/hfi1: Fix possible null-pointer dereference in _extend_sdma_tx_descs()
    - RDMA/bnxt_re: Remove unpaired rtnl unlock in bnxt_re_dev_init()
    - ice: do not abort devlink info if board identifier can't be found
    - net: usb: pegasus: fixes of set_register(s) return value evaluation;
    - igc: Use num_tx_queues when iterating over tx_ring queue
    - e1000e: Fix the max snoop/no-snoop latency for 10M
    - RDMA/efa: Free IRQ vectors on error flow
    - ip_gre: add validation for csum_start
    - xgene-v2: Fix a resource leak in the error handling path of 'xge_probe()'
    - net: marvell: fix MVNETA_TX_IN_PRGRS bit number
    - net/sched: ets: fix crash when flipping from 'strict' to 'quantum'
    - ipv6: use siphash in rt6_exception_hash()
    - ipv4: use siphash instead of Jenkins in fnhe_hashfun()
    - cxgb4: dont touch blocked freelist bitmap after free
    - rtnetlink: Return correct error on changing device netns
    - net: hns3: clear hardware resource when loading driver
    - net: hns3: add waiting time before cmdq memory is released
    - net: hns3: fix duplicate node in VLAN list
    - net: hns3: fix get wrong pfc_en when query P...

[Launchpad Janitor (janitor)]

This bug was fixed in the package linux - 5.4.0-89.100

---------------
linux (5.4.0-89.100) focal; urgency=medium

  * focal/linux: 5.4.0-89.100 -proposed tracker (LP: #1944901)

  * Packaging resync (LP: #1786013)
    - debian/dkms-versions -- update from kernel-versions (main/2021.09.27)

  * ext4 journal recovery fails w/ data=journal + mmap (LP: #1847340)
    - jbd2: introduce/export functions
      jbd2_journal_submit|finish_inode_data_buffers()
    - jbd2, ext4, ocfs2: introduce/use journal callbacks
      j_submit|finish_inode_data_buffers()
    - ext4: data=journal: fixes for ext4_page_mkwrite()
    - ext4: data=journal: write-protect pages on j_submit_inode_data_buffers()
    - ext4: fix mmap write protection for data=journal mode

  * CVE-2021-40490
    - ext4: fix race writing to an inline_data file while its xattrs are changing

  * Obsolete patch "UBUNTU: SAUCE: ext4: fix directory index node split
    corruption" (LP: #1942902)
    - Revert "UBUNTU: SAUCE: ext4: fix directory index node split corruption"

  * psock_snd.sh in net from ubuntu_kernel_selftests ADT failure with
    focal/groovy/hirsute/impish (LP: #1892213)
    - selftests/net: remove min gso test in packet_snd

  * Focal update: v5.4.143 upstream stable release (LP: #1944212)
    - ext4: fix EXT4_MAX_LOGICAL_BLOCK macro
    - x86/fpu: Make init_fpstate correct with optimized XSAVE
    - ath: Use safer key clearing with key cache entries
    - ath9k: Clear key cache explicitly on disabling hardware
    - ath: Export ath_hw_keysetmac()
    - ath: Modify ath_key_delete() to not need full key entry
    - ath9k: Postpone key cache entry deletion for TXQ frames reference it
    - mtd: cfi_cmdset_0002: fix crash when erasing/writing AMD cards
    - media: zr364xx: propagate errors from zr364xx_start_readpipe()
    - media: zr364xx: fix memory leaks in probe()
    - media: drivers/media/usb: fix memory leak in zr364xx_probe
    - USB: core: Avoid WARNings for 0-length descriptor requests
    - dmaengine: xilinx_dma: Fix read-after-free bug when terminating transfers
    - dmaengine: usb-dmac: Fix PM reference leak in usb_dmac_probe()
    - ARM: dts: am43x-epos-evm: Reduce i2c0 bus speed for tps65218
    - dmaengine: of-dma: router_xlate to return -EPROBE_DEFER if controller is not
      yet available
    - scsi: megaraid_mm: Fix end of loop tests for list_for_each_entry()
    - scsi: scsi_dh_rdac: Avoid crash during rdac_bus_attach()
    - scsi: core: Avoid printing an error if target_alloc() returns -ENXIO
    - scsi: core: Fix capacity set to zero after offlinining device
    - ARM: dts: nomadik: Fix up interrupt controller node names
    - net: usb: lan78xx: don't modify phy_device state concurrently
    - drm/amd/display: Fix Dynamic bpp issue with 8K30 with Navi 1X
    - Bluetooth: hidp: use correct wait queue when removing ctrl_wait
    - iommu: Check if group is NULL before remove device
    - cpufreq: armada-37xx: forbid cpufreq for 1.2 GHz variant
    - dccp: add do-while-0 stubs for dccp_pr_debug macros
    - virtio: Protect vqs list access
    - vhost: Fix the calculation in vhost_overflow()
    - bpf: Clear zext_dst of dead insns
    - bnxt: don't lock the tx ...