1. 19 Apr, 2024 33 commits
  2. 18 Apr, 2024 7 commits
    • Jakub Kicinski's avatar
      selftests: net: set the exit code correctly in Python tests · 4fa6bd4b
      Jakub Kicinski authored
      Test cases need to exit with non-zero status if they failed,
      we currently don't do that:
      
        # KTAP version 1
        # 1..3
        # # At /root/ksft-net-drv/drivers/net/./ping.py line 18:
        # # Check failed 1 != 2
        # not ok 1 ping.test_v4
        # ok 2 ping.test_v6
        # ok 3 ping.test_tcp
        # # Totals: pass:2 fail:1 xfail:0 xpass:0 skip:0 error:0
        ok 1 selftests: drivers/net: ping.py
        ^^^^
      
      It's a bit tempting to make the exit part of ksft_run(),
      but that only works well for very trivial setups. We can
      revisit this later, if people forget to call ksft_exit().
      
      Link: https://lore.kernel.org/r/20240417231146.2435572-3-kuba@kernel.orgSigned-off-by: default avatarJakub Kicinski <kuba@kernel.org>
      4fa6bd4b
    • Jakub Kicinski's avatar
      selftests: net: fix counting totals when some checks fail · 655614ea
      Jakub Kicinski authored
      Totals currently only pay attention to exceptions, if check fails
      (say ksft_eq()) the test case will be counted as pass:
      
        # At /ksft/drivers/net/./ping.py line 18:
        # Check failed 1 != 2
        not ok 1 ping.test_v4
        ok 2 ping.test_v6
        ok 3 ping.test_tcp
        # Totals: pass:3 fail:0 xfail:0 xpass:0 skip:0 error:0
                  ^^^^^^^^^^^^^
      
      Pay attention to the result.
      
      Fixes: b86761ff ("selftests: net: add scaffolding for Netlink tests in Python")
      Link: https://lore.kernel.org/r/20240417231146.2435572-2-kuba@kernel.orgSigned-off-by: default avatarJakub Kicinski <kuba@kernel.org>
      655614ea
    • Jakub Kicinski's avatar
      Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net · 41e3ddb2
      Jakub Kicinski authored
      Cross-merge networking fixes after downstream PR.
      
      Conflicts:
      
      include/trace/events/rpcgss.h
        386f4a73 ("trace: events: cleanup deprecated strncpy uses")
        a4833e3a ("SUNRPC: Fix rpcgss_context trace event acceptor field")
      
      Adjacent changes:
      
      drivers/net/ethernet/intel/ice/ice_tc_lib.c
        2cca35f5 ("ice: Fix checking for unsupported keys on non-tunnel device")
        784feaa6 ("ice: Add support for PFCP hardware offload in switchdev")
      Signed-off-by: default avatarJakub Kicinski <kuba@kernel.org>
      41e3ddb2
    • Linus Torvalds's avatar
      Merge tag 'net-6.9-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net · 7586c850
      Linus Torvalds authored
      Pull networking fixes from Jakub Kicinski:
       "A little calmer than usual, probably just the timing of sub-tree PRs.
      
        Including fixes from netfilter.
      
        Current release - regressions:
      
         - inet: bring NLM_DONE out to a separate recv() again, fix user space
           which assumes multiple recv()s will happen and gets blocked forever
      
         - drv: mlx5:
             - restore mistakenly dropped parts in register devlink flow
             - use channel mdev reference instead of global mdev instance for
               coalescing
             - acquire RTNL lock before RQs/SQs activation/deactivation
      
        Previous releases - regressions:
      
         - net: change maximum number of UDP segments to 128, fix virtio
           compatibility with Windows peers
      
         - usb: ax88179_178a: avoid writing the mac address before first
           reading
      
        Previous releases - always broken:
      
         - sched: fix mirred deadlock on device recursion
      
         - netfilter:
             - br_netfilter: skip conntrack input hook for promisc packets
             - fixes removal of duplicate elements in the pipapo set backend
             - various fixes for abort paths and error handling
      
         - af_unix: don't peek OOB data without MSG_OOB
      
         - drv: flower: fix fragment flags handling in multiple drivers
      
         - drv: ravb: fix jumbo frames and packet stats accounting
      
        Misc:
      
         - kselftest_harness: fix Clang warning about zero-length format
      
         - tun: limit printing rate when illegal packet received by tun dev"
      
      * tag 'net-6.9-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (46 commits)
        net: ethernet: ti: am65-cpsw-nuss: cleanup DMA Channels before using them
        net: usb: ax88179_178a: avoid writing the mac address before first reading
        net: ravb: Fix RX byte accounting for jumbo packets
        net: ravb: Fix GbEth jumbo packet RX checksum handling
        net: ravb: Allow RX loop to move past DMA mapping errors
        net: ravb: Count packets instead of descriptors in R-Car RX path
        net: ethernet: mtk_eth_soc: fix WED + wifi reset
        net:usb:qmi_wwan: support Rolling modules
        selftests: kselftest_harness: fix Clang warning about zero-length format
        net/sched: Fix mirred deadlock on device recursion
        netfilter: nf_tables: fix memleak in map from abort path
        netfilter: nf_tables: restore set elements when delete set fails
        netfilter: nf_tables: missing iterator type in lookup walk
        s390/ism: Properly fix receive message buffer allocation
        net: dsa: mt7530: fix port mirroring for MT7988 SoC switch
        net: dsa: mt7530: fix mirroring frames received on local port
        tun: limit printing rate when illegal packet received by tun dev
        ice: Fix checking for unsupported keys on non-tunnel device
        ice: tc: allow zero flags in parsing tc flower
        ice: tc: check src_vsi in case of traffic from VF
        ...
      7586c850
    • Linus Torvalds's avatar
      Merge tag 'gpio-fixes-for-v6.9-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux · 360a348f
      Linus Torvalds authored
      Pull gpio fixes from Bartosz Golaszewski:
      
       - use -ENOTSUPP consistently in Intel GPIO drivers
      
       - don't include dt-bindings headers in gpio-swnode code
      
       - add missing of device table to gpio-lpc32xx and fix autoloading
      
      * tag 'gpio-fixes-for-v6.9-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux:
        gpiolib: swnode: Remove wrong header inclusion
        gpio: lpc32xx: fix module autoloading
        gpio: crystalcove: Use -ENOTSUPP consistently
        gpio: wcove: Use -ENOTSUPP consistently
      360a348f
    • Siddharth Vadapalli's avatar
      net: ethernet: ti: am65-cpsw-nuss: cleanup DMA Channels before using them · c24cd679
      Siddharth Vadapalli authored
      The TX and RX DMA Channels used by the driver to exchange data with CPSW
      are not guaranteed to be in a clean state during driver initialization.
      The Bootloader could have used the same DMA Channels without cleaning them
      up in the event of failure. Thus, reset and disable the DMA Channels to
      ensure that they are in a clean state before using them.
      
      Fixes: 93a76530 ("net: ethernet: ti: introduce am65x/j721e gigabit eth subsystem driver")
      Reported-by: default avatarSchuyler Patton <spatton@ti.com>
      Signed-off-by: default avatarSiddharth Vadapalli <s-vadapalli@ti.com>
      Reviewed-by: default avatarRoger Quadros <rogerq@kernel.org>
      Link: https://lore.kernel.org/r/20240417095425.2253876-1-s-vadapalli@ti.comSigned-off-by: default avatarJakub Kicinski <kuba@kernel.org>
      c24cd679
    • Jose Ignacio Tornos Martinez's avatar
      net: usb: ax88179_178a: avoid writing the mac address before first reading · 56f78615
      Jose Ignacio Tornos Martinez authored
      After the commit d2689b6a ("net: usb: ax88179_178a: avoid two
      consecutive device resets"), reset operation, in which the default mac
      address from the device is read, is not executed from bind operation and
      the random address, that is pregenerated just in case, is direclty written
      the first time in the device, so the default one from the device is not
      even read. This writing is not dangerous because is volatile and the
      default mac address is not missed.
      
      In order to avoid this and keep the simplification to have only one
      reset and reduce the delays, restore the reset from bind operation and
      remove the reset that is commanded from open operation. The behavior is
      the same but everything is ready for usbnet_probe.
      
      Tested with ASIX AX88179 USB Gigabit Ethernet devices.
      Restore the old behavior for the rest of possible devices because I don't
      have the hardware to test.
      
      cc: stable@vger.kernel.org # 6.6+
      Fixes: d2689b6a ("net: usb: ax88179_178a: avoid two consecutive device resets")
      Reported-by: default avatarJarkko Palviainen <jarkko.palviainen@gmail.com>
      Signed-off-by: default avatarJose Ignacio Tornos Martinez <jtornosm@redhat.com>
      Link: https://lore.kernel.org/r/20240417085524.219532-1-jtornosm@redhat.comSigned-off-by: default avatarJakub Kicinski <kuba@kernel.org>
      56f78615