summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2024-10-03rust: net::phy always define device_table in module_phy_driver macroFUJITA Tomonori
device_table in module_phy_driver macro is defined only when the driver is built as a module. So a PHY driver imports phy::DeviceId module in the following way then hits `unused import` warning when it's compiled as built-in: use kernel::net::phy::DeviceId; kernel::module_phy_driver! { drivers: [PhyQT2025], device_table: [ DeviceId::new_with_driver::<PhyQT2025>(), ], Put device_table in a const. It's not included in the kernel image if unused (when the driver is compiled as built-in), and the compiler doesn't complain. Signed-off-by: FUJITA Tomonori <fujita.tomonori@gmail.com> Reviewed-by: Alice Ryhl <aliceryhl@google.com> Link: https://patch.msgid.link/20240930134038.1309-1-fujita.tomonori@gmail.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-03net: phy: microchip_t1: Interrupt support for lan887xDivya Koppera
Add support for link up and link down interrupts in lan887x. Signed-off-by: Divya Koppera <divya.koppera@microchip.com> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Link: https://patch.msgid.link/20241001144421.6661-1-divya.koppera@microchip.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-03Merge branch 'ipv4-convert-ip_route_input_slow-and-its-callers-to-dscp_t'Jakub Kicinski
Guillaume Nault says: ==================== ipv4: Convert ip_route_input_slow() and its callers to dscp_t. Prepare ip_route_input_slow() and its call chain to future conversion of ->flowi4_tos. The ->flowi4_tos field of "struct flowi4" is used in many different places, which makes it hard to convert it from __u8 to dscp_t. In order to avoid a big patch updating all its users at once, this patch series gradually converts some users to dscp_t. Those users now set ->flowi4_tos from a dscp_t variable that is converted to __u8 using inet_dscp_to_dsfield(). When all users of ->flowi4_tos will use a dscp_t variable, converting that field to dscp_t will just be a matter of removing all the inet_dscp_to_dsfield() conversions. This series concentrates on ip_route_input_slow() and its direct and indirect callers. ==================== Link: https://patch.msgid.link/cover.1727807926.git.gnault@redhat.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-03ipv4: Convert ip_route_input_slow() to dscp_t.Guillaume Nault
Pass a dscp_t variable to ip_route_input_slow(), instead of a plain u8, to prevent accidental setting of ECN bits in ->flowi4_tos. Only ip_route_input_rcu() actually calls ip_route_input_slow(). Since it already has a dscp_t variable to pass as parameter, we only need to remove the inet_dscp_to_dsfield() conversion. Signed-off-by: Guillaume Nault <gnault@redhat.com> Reviewed-by: David Ahern <dsahern@kernel.org> Link: https://patch.msgid.link/d6bca5f87eea9e83a3861e6e05594cdd252583c9.1727807926.git.gnault@redhat.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-03ipv4: Convert ip_route_input_rcu() to dscp_t.Guillaume Nault
Pass a dscp_t variable to ip_route_input_rcu(), instead of a plain u8, to prevent accidental setting of ECN bits in ->flowi4_tos. Callers of ip_route_input_rcu() to consider are: * ip_route_input_noref(), which already has a dscp_t variable to pass as parameter. We just need to remove the inet_dscp_to_dsfield() conversion. * inet_rtm_getroute(), which receives a u8 from user space and needs to convert it with inet_dsfield_to_dscp(). Signed-off-by: Guillaume Nault <gnault@redhat.com> Reviewed-by: David Ahern <dsahern@kernel.org> Link: https://patch.msgid.link/c4dbb5aa9cbc79c4fcb317abbffa7c7156bc56a7.1727807926.git.gnault@redhat.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-03ipv4: Convert ip_route_input_noref() to dscp_t.Guillaume Nault
Pass a dscp_t variable to ip_route_input_noref(), instead of a plain u8, to prevent accidental setting of ECN bits in ->flowi4_tos. Callers of ip_route_input_noref() to consider are: * arp_process() in net/ipv4/arp.c. This function sets the tos parameter to 0, which is already a valid dscp_t value, so it doesn't need to be adjusted for the new prototype. * ip_route_input(), which already has a dscp_t variable to pass as parameter. We just need to remove the inet_dscp_to_dsfield() conversion. * ipvlan_l3_rcv(), bpf_lwt_input_reroute(), ip_expire(), ip_rcv_finish_core(), xfrm4_rcv_encap_finish() and xfrm4_rcv_encap(), which get the DSCP directly from IPv4 headers and can simply use the ip4h_dscp() helper. While there, declare the IPv4 header pointers as const in ipvlan_l3_rcv() and bpf_lwt_input_reroute(). Also, modify the declaration of ip_route_input_noref() in include/net/route.h so that it matches the prototype of its implementation in net/ipv4/route.c. Signed-off-by: Guillaume Nault <gnault@redhat.com> Reviewed-by: David Ahern <dsahern@kernel.org> Link: https://patch.msgid.link/a8a747bed452519c4d0cc06af32c7e7795d7b627.1727807926.git.gnault@redhat.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-03ipv4: Convert ip_route_input() to dscp_t.Guillaume Nault
Pass a dscp_t variable to ip_route_input(), instead of a plain u8, to prevent accidental setting of ECN bits in ->flowi4_tos. Callers of ip_route_input() to consider are: * input_action_end_dx4_finish() and input_action_end_dt4() in net/ipv6/seg6_local.c. These functions set the tos parameter to 0, which is already a valid dscp_t value, so they don't need to be adjusted for the new prototype. * icmp_route_lookup(), which already has a dscp_t variable to pass as parameter. We just need to remove the inet_dscp_to_dsfield() conversion. * br_nf_pre_routing_finish(), ip_options_rcv_srr() and ip4ip6_err(), which get the DSCP directly from IPv4 headers. Define a helper to read the .tos field of struct iphdr as dscp_t, so that these function don't have to do the conversion manually. While there, declare *iph as const in br_nf_pre_routing_finish(), declare its local variables in reverse-christmas-tree order and move the "err = ip_route_input()" assignment out of the conditional to avoid checkpatch warning. Signed-off-by: Guillaume Nault <gnault@redhat.com> Reviewed-by: David Ahern <dsahern@kernel.org> Link: https://patch.msgid.link/e9d40781d64d3d69f4c79ac8a008b8d67a033e8d.1727807926.git.gnault@redhat.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-03ipv4: Convert icmp_route_lookup() to dscp_t.Guillaume Nault
Pass a dscp_t variable to icmp_route_lookup(), instead of a plain u8, to prevent accidental setting of ECN bits in ->flowi4_tos. Rename that variable ("tos" -> "dscp") to make the intent clear. While there, reorganise the function parameters to fill up horizontal space. Signed-off-by: Guillaume Nault <gnault@redhat.com> Reviewed-by: David Ahern <dsahern@kernel.org> Link: https://patch.msgid.link/294fead85c6035bcdc5fcf9a6bb4ce8798c45ba1.1727807926.git.gnault@redhat.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-03ipv4: avoid quadratic behavior in FIB insertion of common addressAlexandre Ferrieux
Mix netns into all IPv4 FIB hashes to avoid massive collision when inserting the same address in many netns. Signed-off-by: Alexandre Ferrieux <alexandre.ferrieux@orange.com> Reviewed-by: Eric Dumazet <edumazet@google.com> Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com> Link: https://patch.msgid.link/20241001231438.3855035-1-alexandre.ferrieux@orange.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-03Merge branch 'ena-link-irqs-queues-and-napi-instances'Jakub Kicinski
Joe Damato says: ==================== ena: Link IRQs, queues, and NAPI instances This series uses the netdev-genl API to link IRQs and queues to NAPI IDs so that this information is queryable by user apps. This is particularly useful for epoll-based busy polling apps which rely on having access to the NAPI ID. I've tested these commits on an EC2 instance with an ENA NIC configured and have included test output in the commit messages for each patch showing how to query the information. I noted in the implementation that the driver requests an IRQ for management purposes which does not have an associated NAPI. I tried to take this into account in patch 1, but would appreciate if ENA maintainers can verify I did this correctly. v1: https://lore.kernel.org/all/20240930195617.37369-1-jdamato@fastly.com/ ==================== Link: https://patch.msgid.link/20241002001331.65444-1-jdamato@fastly.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-03ena: Link queues to NAPIsJoe Damato
Link queues to NAPIs using the netdev-genl API so this information is queryable. $ ./tools/net/ynl/cli.py --spec Documentation/netlink/specs/netdev.yaml \ --dump queue-get --json='{"ifindex": 2}' [{'id': 0, 'ifindex': 2, 'napi-id': 8201, 'type': 'rx'}, {'id': 1, 'ifindex': 2, 'napi-id': 8202, 'type': 'rx'}, {'id': 2, 'ifindex': 2, 'napi-id': 8203, 'type': 'rx'}, {'id': 3, 'ifindex': 2, 'napi-id': 8204, 'type': 'rx'}, {'id': 4, 'ifindex': 2, 'napi-id': 8205, 'type': 'rx'}, {'id': 5, 'ifindex': 2, 'napi-id': 8206, 'type': 'rx'}, {'id': 6, 'ifindex': 2, 'napi-id': 8207, 'type': 'rx'}, {'id': 7, 'ifindex': 2, 'napi-id': 8208, 'type': 'rx'}, {'id': 0, 'ifindex': 2, 'napi-id': 8201, 'type': 'tx'}, {'id': 1, 'ifindex': 2, 'napi-id': 8202, 'type': 'tx'}, {'id': 2, 'ifindex': 2, 'napi-id': 8203, 'type': 'tx'}, {'id': 3, 'ifindex': 2, 'napi-id': 8204, 'type': 'tx'}, {'id': 4, 'ifindex': 2, 'napi-id': 8205, 'type': 'tx'}, {'id': 5, 'ifindex': 2, 'napi-id': 8206, 'type': 'tx'}, {'id': 6, 'ifindex': 2, 'napi-id': 8207, 'type': 'tx'}, {'id': 7, 'ifindex': 2, 'napi-id': 8208, 'type': 'tx'}] Signed-off-by: Joe Damato <jdamato@fastly.com> Reviewed-by: David Arinzon <darinzon@amazon.com> Link: https://patch.msgid.link/20241002001331.65444-3-jdamato@fastly.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-03ena: Link IRQs to NAPI instancesJoe Damato
Link IRQs to NAPI instances with netif_napi_set_irq. This information can be queried with the netdev-genl API. Note that the ENA device appears to allocate an IRQ for management purposes which does not have a NAPI associated with it; this commit takes this into consideration to accurately construct a map between IRQs and NAPI instances. Compare the output of /proc/interrupts for my ena device with the output of netdev-genl after applying this patch: $ cat /proc/interrupts | grep enp55s0 | cut -f1 --delimiter=':' 94 95 96 97 98 99 100 101 $ ./tools/net/ynl/cli.py --spec Documentation/netlink/specs/netdev.yaml \ --dump napi-get --json='{"ifindex": 2}' [{'id': 8208, 'ifindex': 2, 'irq': 101}, {'id': 8207, 'ifindex': 2, 'irq': 100}, {'id': 8206, 'ifindex': 2, 'irq': 99}, {'id': 8205, 'ifindex': 2, 'irq': 98}, {'id': 8204, 'ifindex': 2, 'irq': 97}, {'id': 8203, 'ifindex': 2, 'irq': 96}, {'id': 8202, 'ifindex': 2, 'irq': 95}, {'id': 8201, 'ifindex': 2, 'irq': 94}] Signed-off-by: Joe Damato <jdamato@fastly.com> Reviewed-by: David Arinzon <darinzon@amazon.com> Link: https://patch.msgid.link/20241002001331.65444-2-jdamato@fastly.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-03Merge branch 'packing-various-improvements-and-kunit-tests'Jakub Kicinski
Jacob Keller says: ==================== packing: various improvements and KUnit tests This series contains a handful of improvements and fixes for the packing library, including the addition of KUnit tests. There are two major changes which might be considered bug fixes: 1) The library is updated to handle arbitrary buffer lengths, fixing undefined behavior when operating on buffers which are not a multiple of 4 bytes. 2) The behavior of QUIRK_MSB_ON_THE_RIGHT is fixed to match the intended behavior when operating on packings that are not byte aligned. These are not sent to net because no driver currently depends on this behavior. For (1), the existing users of the packing API all operate on buffers which are multiples of 4-bytes. For (2), no driver currently uses QUIRK_MSB_ON_THE_RIGHT. The incorrect behavior was found while writing KUnit tests. This series also includes a handful of minor cleanups from Vladimir, as well as a change to introduce a separated pack() and unpack() API. This API is not (yet) used by a driver, but is the first step in implementing pack_fields() and unpack_fields() which will be used in future changes for the ice driver and changes Vladimir has in progress for other drivers using the packing API. This series is part 1 of a 2-part series for implementing use of lib/packing in the ice driver. The 2nd part includes a new pack_fields() and unpack_fields() implementation inspired by the ice driver's existing bit packing code. It is built on top of the split pack() and unpack() code. Additionally, the KUnit tests are built on top of pack() and unpack(), based on original selftests written by Vladimir. Fitting the entire library changes and drivers changes into a single series exceeded the usual series limits. v1: https://lore.kernel.org/r/20240930-packing-kunit-tests-and-split-pack-unpack-v1-0-94b1f04aca85@intel.com ==================== Link: https://patch.msgid.link/20241002-packing-kunit-tests-and-split-pack-unpack-v2-0-8373e551eae3@intel.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-03lib: packing: use GENMASK() for box_maskVladimir Oltean
This is an u8, so using GENMASK_ULL() for unsigned long long is unnecessary. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Link: https://patch.msgid.link/20241002-packing-kunit-tests-and-split-pack-unpack-v2-10-8373e551eae3@intel.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-03lib: packing: use BITS_PER_BYTE instead of 8Vladimir Oltean
This helps clarify what the 8 is for. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com> Link: https://patch.msgid.link/20241002-packing-kunit-tests-and-split-pack-unpack-v2-9-8373e551eae3@intel.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-03lib: packing: fix QUIRK_MSB_ON_THE_RIGHT behaviorJacob Keller
The QUIRK_MSB_ON_THE_RIGHT quirk is intended to modify pack() and unpack() so that the most significant bit of each byte in the packed layout is on the right. The way the quirk is currently implemented is broken whenever the packing code packs or unpacks any value that is not exactly a full byte. The broken behavior can occur when packing any values smaller than one byte, when packing any value that is not exactly a whole number of bytes, or when the packing is not aligned to a byte boundary. This quirk is documented in the following way: 1. Normally (no quirks), we would do it like this: :: 63 62 61 60 59 58 57 56 55 54 53 52 51 50 49 48 47 46 45 44 43 42 41 40 39 38 37 36 35 34 33 32 7 6 5 4 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 3 2 1 0 <snip> 2. If QUIRK_MSB_ON_THE_RIGHT is set, we do it like this: :: 56 57 58 59 60 61 62 63 48 49 50 51 52 53 54 55 40 41 42 43 44 45 46 47 32 33 34 35 36 37 38 39 7 6 5 4 24 25 26 27 28 29 30 31 16 17 18 19 20 21 22 23 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 3 2 1 0 That is, QUIRK_MSB_ON_THE_RIGHT does not affect byte positioning, but inverts bit offsets inside a byte. Essentially, the mapping for physical bit offsets should be reserved for a given byte within the payload. This reversal should be fixed to the bytes in the packing layout. The logic to implement this quirk is handled within the adjust_for_msb_right_quirk() function. This function does not work properly when dealing with the bytes that contain only a partial amount of data. In particular, consider trying to pack or unpack the range 53-44. We should always be mapping the bits from the logical ordering to their physical ordering in the same way, regardless of what sequence of bits we are unpacking. This, we should grab the following logical bits: Logical: 55 54 53 52 51 50 49 48 47 45 44 43 42 41 40 39 ^ ^ ^ ^ ^ ^ ^ ^ ^ And pack them into the physical bits: Physical: 48 49 50 51 52 53 54 55 40 41 42 43 44 45 46 47 Logical: 48 49 50 51 52 53 44 45 46 47 ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ The current logic in adjust_for_msb_right_quirk is broken. I believe it is intending to map according to the following: Physical: 48 49 50 51 52 53 54 55 40 41 42 43 44 45 46 47 Logical: 48 49 50 51 52 53 44 45 46 47 ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ That is, it tries to keep the bits at the start and end of a packing together. This is wrong, as it makes the packing change what bit is being mapped to what based on which bits you're currently packing or unpacking. Worse, the actual calculations within adjust_for_msb_right_quirk don't make sense. Consider the case when packing the last byte of an unaligned packing. It might have a start bit of 7 and an end bit of 5. This would have a width of 3 bits. The new_start_bit will be calculated as the width - the box_end_bit - 1. This will underflow and produce a negative value, which will ultimate result in generating a new box_mask of all 0s. For any other values, the result of the calculations of the new_box_end_bit, new_box_start_bit, and the new box_mask will result in the exact same values for the box_end_bit, box_start_bit, and box_mask. This makes the calculations completely irrelevant. If box_end_bit is 0, and box_start_bit is 7, then the entire function of adjust_for_msb_right_quirk will boil down to just: *to_write = bitrev8(*to_write) The other adjustments are attempting (incorrectly) to keep the bits in the same place but just reversed. This is not the right behavior even if implemented correctly, as it leaves the mapping dependent on the bit values being packed or unpacked. Remove adjust_for_msb_right_quirk() and just use bitrev8 to reverse the byte order when interacting with the packed data. In particular, for packing, we need to reverse both the box_mask and the physical value being packed. This is done after shifting the value by box_end_bit so that the reversed mapping is always aligned to the physical buffer byte boundary. The box_mask is reversed as we're about to use it to clear any stale bits in the physical buffer at this block. For unpacking, we need to reverse the contents of the physical buffer *before* masking with the box_mask. This is critical, as the box_mask is a logical mask of the bit layout before handling the QUIRK_MSB_ON_THE_RIGHT. Add several new tests which cover this behavior. These tests will fail without the fix and pass afterwards. Note that no current drivers make use of QUIRK_MSB_ON_THE_RIGHT. I suspect this is why there have been no reports of this inconsistency before. Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com> Link: https://patch.msgid.link/20241002-packing-kunit-tests-and-split-pack-unpack-v2-8-8373e551eae3@intel.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-03lib: packing: add additional KUnit testsJacob Keller
While reviewing the initial KUnit tests for lib/packing, Przemek pointed out that the test values have duplicate bytes in the input sequence. In addition, I noticed that the unit tests pack and unpack on a byte boundary, instead of crossing bytes. Thus, we lack good coverage of the corner cases of the API. Add additional unit tests to cover packing and unpacking byte buffers which do not have duplicate bytes in the unpacked value, and which pack and unpack to an unaligned offset. A careful reviewer may note the lack tests for QUIRK_MSB_ON_THE_RIGHT. This is because I found issues with that quirk during test implementation. This quirk will be fixed and the tests will be included in a future change. Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com> Link: https://patch.msgid.link/20241002-packing-kunit-tests-and-split-pack-unpack-v2-7-8373e551eae3@intel.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-03lib: packing: add KUnit tests adapted from selftestsJacob Keller
Add 24 simple KUnit tests for the lib/packing.c pack() and unpack() APIs. The first 16 tests exercise all combinations of quirks with a simple magic number value on a 16-byte buffer. The remaining 8 tests cover non-multiple-of-4 buffer sizes. These tests were originally written by Vladimir as simple selftest functions. I adapted them to KUnit, refactoring them into a table driven approach. This will aid in adding additional tests in the future. Co-developed-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com> Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com> Tested-by: Vladimir Oltean <vladimir.oltean@nxp.com> Link: https://patch.msgid.link/20241002-packing-kunit-tests-and-split-pack-unpack-v2-6-8373e551eae3@intel.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-03lib: packing: duplicate pack() and unpack() implementationsVladimir Oltean
packing() is now used in some hot paths, and it would be good to get rid of some ifs and buts that depend on "op", to speed things up a little bit. With the main implementations now taking size_t endbit, we no longer have to check for negative values. Update the local integer variables to also be size_t to match. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com> Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com> Link: https://patch.msgid.link/20241002-packing-kunit-tests-and-split-pack-unpack-v2-5-8373e551eae3@intel.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-03lib: packing: add pack() and unpack() wrappers over packing()Vladimir Oltean
Geert Uytterhoeven described packing() as "really bad API" because of not being able to enforce const correctness. The same function is used both when "pbuf" is input and "uval" is output, as in the other way around. Create 2 wrapper functions where const correctness can be ensured. Do ugly type casts inside, to be able to reuse packing() as currently implemented - which will _not_ modify the input argument. Also, take the opportunity to change the type of startbit and endbit to size_t - an unsigned type - in these new function prototypes. When int, an extra check for negative values is necessary. Hopefully, when packing() goes away completely, that check can be dropped. My concern is that code which does rely on the conditional directionality of packing() is harder to refactor without blowing up in size. So it may take a while to completely eliminate packing(). But let's make alternatives available for those who do not need that. Link: https://lore.kernel.org/netdev/20210223112003.2223332-1-geert+renesas@glider.be/ Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com> Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com> Link: https://patch.msgid.link/20241002-packing-kunit-tests-and-split-pack-unpack-v2-4-8373e551eae3@intel.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-03lib: packing: remove kernel-doc from header fileVladimir Oltean
It is not necessary to have the kernel-doc duplicated both in the header and in the implementation. It is better to have it near the implementation of the function, since in C, a function can have N declarations, but only one definition. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com> Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com> Link: https://patch.msgid.link/20241002-packing-kunit-tests-and-split-pack-unpack-v2-3-8373e551eae3@intel.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-03lib: packing: adjust definitions and implementation for arbitrary buffer lengthsVladimir Oltean
Jacob Keller has a use case for packing() in the intel/ice networking driver, but it cannot be used as-is. Simply put, the API quirks for LSW32_IS_FIRST and LITTLE_ENDIAN are naively implemented with the undocumented assumption that the buffer length must be a multiple of 4. All calculations of group offsets and offsets of bytes within groups assume that this is the case. But in the ice case, this does not hold true. For example, packing into a buffer of 22 bytes would yield wrong results, but pretending it was a 24 byte buffer would work. Rather than requiring such hacks, and leaving a big question mark when it comes to discontinuities in the accessible bit fields of such buffer, we should extend the packing API to support this use case. It turns out that we can keep the design in terms of groups of 4 bytes, but also make it work if the total length is not a multiple of 4. Just like before, imagine the buffer as a big number, and its most significant bytes (the ones that would make up to a multiple of 4) are missing. Thus, with a big endian (no quirks) interpretation of the buffer, those most significant bytes would be absent from the beginning of the buffer, and with a LSW32_IS_FIRST interpretation, they would be absent from the end of the buffer. The LITTLE_ENDIAN quirk, in the packing() API world, only affects byte ordering within groups of 4. Thus, it does not change which bytes are missing. Only the significance of the remaining bytes within the (smaller) group. No change intended for buffer sizes which are multiples of 4. Tested with the sja1105 driver and with downstream unit tests. Link: https://lore.kernel.org/netdev/a0338310-e66c-497c-bc1f-a597e50aa3ff@intel.com/ Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Tested-by: Jacob Keller <jacob.e.keller@intel.com> Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com> Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com> Link: https://patch.msgid.link/20241002-packing-kunit-tests-and-split-pack-unpack-v2-2-8373e551eae3@intel.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-03lib: packing: refuse operating on bit indices which exceed size of bufferVladimir Oltean
While reworking the implementation, it became apparent that this check does not exist. There is no functional issue yet, because at call sites, "startbit" and "endbit" are always hardcoded to correct values, and never come from the user. Even with the upcoming support of arbitrary buffer lengths, the "startbit >= 8 * pbuflen" check will remain correct. This is because we intend to always interpret the packed buffer in a way that avoids discontinuities in the available bit indices. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Tested-by: Jacob Keller <jacob.e.keller@intel.com> Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com> Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com> Link: https://patch.msgid.link/20241002-packing-kunit-tests-and-split-pack-unpack-v2-1-8373e551eae3@intel.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-03Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/netJakub Kicinski
Cross-merge networking fixes after downstream PR. No conflicts and no adjacent changes. Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-03Merge tag 'net-6.12-rc2' of ↵Linus Torvalds
git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net Pull networking fixes from Paolo Abeni: "Including fixes from ieee802154, bluetooth and netfilter. Current release - regressions: - eth: mlx5: fix wrong reserved field in hca_cap_2 in mlx5_ifc - eth: am65-cpsw: fix forever loop in cleanup code Current release - new code bugs: - eth: mlx5: HWS, fixed double-free in error flow of creating SQ Previous releases - regressions: - core: avoid potential underflow in qdisc_pkt_len_init() with UFO - core: test for not too small csum_start in virtio_net_hdr_to_skb() - vrf: revert "vrf: remove unnecessary RCU-bh critical section" - bluetooth: - fix uaf in l2cap_connect - fix possible crash on mgmt_index_removed - dsa: improve shutdown sequence - eth: mlx5e: SHAMPO, fix overflow of hd_per_wq - eth: ip_gre: fix drops of small packets in ipgre_xmit Previous releases - always broken: - core: fix gso_features_check to check for both dev->gso_{ipv4_,}max_size - core: fix tcp fraglist segmentation after pull from frag_list - netfilter: nf_tables: prevent nf_skb_duplicated corruption - sctp: set sk_state back to CLOSED if autobind fails in sctp_listen_start - mac802154: fix potential RCU dereference issue in mac802154_scan_worker - eth: fec: restart PPS after link state change" * tag 'net-6.12-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (48 commits) sctp: set sk_state back to CLOSED if autobind fails in sctp_listen_start dt-bindings: net: xlnx,axi-ethernet: Add missing reg minItems doc: net: napi: Update documentation for napi_schedule_irqoff net/ncsi: Disable the ncsi work before freeing the associated structure net: phy: qt2025: Fix warning: unused import DeviceId gso: fix udp gso fraglist segmentation after pull from frag_list bridge: mcast: Fail MDB get request on empty entry vrf: revert "vrf: Remove unnecessary RCU-bh critical section" net: ethernet: ti: am65-cpsw: Fix forever loop in cleanup code net: phy: realtek: Check the index value in led_hw_control_get ppp: do not assume bh is held in ppp_channel_bridge_input() selftests: rds: move include.sh to TEST_FILES net: test for not too small csum_start in virtio_net_hdr_to_skb() net: gso: fix tcp fraglist segmentation after pull from frag_list ipv4: ip_gre: Fix drops of small packets in ipgre_xmit net: stmmac: dwmac4: extend timeout for VLAN Tag register busy bit check net: add more sanity checks to qdisc_pkt_len_init() net: avoid potential underflow in qdisc_pkt_len_init() with UFO net: ethernet: ti: cpsw_ale: Fix warning on some platforms net: microchip: Make FDMA config symbol invisible ...
2024-10-03Merge tag 'v6.12-rc1-ksmbd-fixes' of git://git.samba.org/ksmbdLinus Torvalds
Pull smb server fixes from Steve French: - small cleanup patches leveraging struct size to improve access bounds checking * tag 'v6.12-rc1-ksmbd-fixes' of git://git.samba.org/ksmbd: ksmbd: Use struct_size() to improve smb_direct_rdma_xmit() ksmbd: Annotate struct copychunk_ioctl_req with __counted_by_le() ksmbd: Use struct_size() to improve get_file_alternate_info()
2024-10-03Merge tag 'vfs-6.12-rc2.fixes.2' of ↵Linus Torvalds
git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs Pull vfs fixes from Christian Brauner: "vfs: - Ensure that iter_folioq_get_pages() advances to the next slot otherwise it will end up using the same folio with an out-of-bound offset. iomap: - Dont unshare delalloc extents which can't be reflinked, and thus can't be shared. - Constrain the file range passed to iomap_file_unshare() directly in iomap instead of requiring the callers to do it. netfs: - Use folioq_count instead of folioq_nr_slot to prevent an unitialized value warning in netfs_clear_buffer(). - Fix missing wakeup after issuing writes by scheduling the write collector only if all the subrequest queues are empty and thus no writes are pending. - Fix two minor documentation bugs" * tag 'vfs-6.12-rc2.fixes.2' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs: iomap: constrain the file range passed to iomap_file_unshare iomap: don't bother unsharing delalloc extents netfs: Fix missing wakeup after issuing writes Documentation: add missing folio_queue entry folio_queue: fix documentation netfs: Fix a KMSAN uninit-value error in netfs_clear_buffer iov_iter: fix advancing slot in iter_folioq_get_pages()
2024-10-03net: mana: Add get_link and get_link_ksettings in ethtoolErni Sri Satya Vennela
Add support for the ethtool get_link and get_link_ksettings operations. Display standard port information using ethtool. Before the change: $ethtool enP30832s1 > No data available After the change: $ethtool enP30832s1 > Settings for enP30832s1: Supported ports: [ ] Supported link modes: Not reported Supported pause frame use: No Supports auto-negotiation: No Supported FEC modes: Not reported Advertised link modes: Not reported Advertised pause frame use: No Advertised auto-negotiation: No Advertised FEC modes: Not reported Speed: Unknown! Duplex: Full Auto-negotiation: off Port: Other PHYAD: 0 Transceiver: internal Link detected: yes Signed-off-by: Erni Sri Satya Vennela <ernis@linux.microsoft.com> Reviewed-by: Haiyang Zhang <haiyangz@microsoft.com> Reviewed-by: Shradha Gupta <shradhagupta@linux.microsoft.com> Reviewed-by: Simon Horman <horms@kernel.org> Link: https://patch.msgid.link/1727674934-12130-1-git-send-email-ernis@linux.microsoft.com Signed-off-by: Paolo Abeni <pabeni@redhat.com>
2024-10-03sctp: set sk_state back to CLOSED if autobind fails in sctp_listen_startXin Long
In sctp_listen_start() invoked by sctp_inet_listen(), it should set the sk_state back to CLOSED if sctp_autobind() fails due to whatever reason. Otherwise, next time when calling sctp_inet_listen(), if sctp_sk(sk)->reuse is already set via setsockopt(SCTP_REUSE_PORT), sctp_sk(sk)->bind_hash will be dereferenced as sk_state is LISTENING, which causes a crash as bind_hash is NULL. KASAN: null-ptr-deref in range [0x0000000000000000-0x0000000000000007] RIP: 0010:sctp_inet_listen+0x7f0/0xa20 net/sctp/socket.c:8617 Call Trace: <TASK> __sys_listen_socket net/socket.c:1883 [inline] __sys_listen+0x1b7/0x230 net/socket.c:1894 __do_sys_listen net/socket.c:1902 [inline] Fixes: 5e8f3f703ae4 ("sctp: simplify sctp listening code") Reported-by: syzbot+f4e0f821e3a3b7cee51d@syzkaller.appspotmail.com Signed-off-by: Xin Long <lucien.xin@gmail.com> Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com> Link: https://patch.msgid.link/a93e655b3c153dc8945d7a812e6d8ab0d52b7aa0.1727729391.git.lucien.xin@gmail.com Signed-off-by: Paolo Abeni <pabeni@redhat.com>
2024-10-03dt-bindings: net: xlnx,axi-ethernet: Add missing reg minItemsRavikanth Tuniki
Add missing reg minItems as based on current binding document only ethernet MAC IO space is a supported configuration. There is a bug in schema, current examples contain 64-bit addressing as well as 32-bit addressing. The schema validation does pass incidentally considering one 64-bit reg address as two 32-bit reg address entries. If we change axi_ethernet_eth1 example node reg addressing to 32-bit schema validation reports: Documentation/devicetree/bindings/net/xlnx,axi-ethernet.example.dtb: ethernet@40000000: reg: [[1073741824, 262144]] is too short To fix it add missing reg minItems constraints and to make things clearer stick to 32-bit addressing in examples. Fixes: cbb1ca6d5f9a ("dt-bindings: net: xlnx,axi-ethernet: convert bindings document to yaml") Signed-off-by: Ravikanth Tuniki <ravikanth.tuniki@amd.com> Signed-off-by: Radhey Shyam Pandey <radhey.shyam.pandey@amd.com> Acked-by: Conor Dooley <conor.dooley@microchip.com> Link: https://patch.msgid.link/1727723615-2109795-1-git-send-email-radhey.shyam.pandey@amd.com Signed-off-by: Paolo Abeni <pabeni@redhat.com>
2024-10-03doc: net: napi: Update documentation for napi_schedule_irqoffSean Anderson
Since commit 8380c81d5c4f ("net: Treat __napi_schedule_irqoff() as __napi_schedule() on PREEMPT_RT"), napi_schedule_irqoff will do the right thing if IRQs are threaded. Therefore, there is no need to use IRQF_NO_THREAD. Signed-off-by: Sean Anderson <sean.anderson@linux.dev> Reviewed-by: Bagas Sanjaya <bagasdotme@gmail.com> Reviewed-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Link: https://patch.msgid.link/20240930153955.971657-1-sean.anderson@linux.dev Signed-off-by: Paolo Abeni <pabeni@redhat.com>
2024-10-03Merge tag 'nf-24-10-02' of ↵Paolo Abeni
git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf Pablo Neira Ayuso says: ==================== Netfilter fixes for net The following patchset contains Netfilter fixes for net: 1) Fix incorrect documentation in uapi/linux/netfilter/nf_tables.h regarding flowtable hooks, from Phil Sutter. 2) Fix nft_audit.sh selftests with newer nft binaries, due to different (valid) audit output, also from Phil. 3) Disable BH when duplicating packets via nf_dup infrastructure, otherwise race on nf_skb_duplicated for locally generated traffic. From Eric. 4) Missing return in callback of selftest C program, from zhang jiao. netfilter pull request 24-10-02 * tag 'nf-24-10-02' of git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf: selftests: netfilter: Add missing return value netfilter: nf_tables: prevent nf_skb_duplicated corruption selftests: netfilter: Fix nft_audit.sh for newer nft binaries netfilter: uapi: NFTA_FLOWTABLE_HOOK is NLA_NESTED ==================== Link: https://patch.msgid.link/20241002202421.1281311-1-pablo@netfilter.org Signed-off-by: Paolo Abeni <pabeni@redhat.com>
2024-10-03net: mana: Increase the DEF_RX_BUFFERS_PER_QUEUE to 1024Shradha Gupta
Through some experiments, we found out that increasing the default RX buffers count from 512 to 1024, gives slightly better throughput and significantly reduces the no_wqe_rx errs on the receiver side. Along with these, other parameters like cpu usage, retrans seg etc also show some improvement with 1024 value. Following are some snippets from the experiments ntttcp tests with 512 Rx buffers --------------------------------------- connections| throughput| no_wqe errs| --------------------------------------- 1 | 40.93Gbps | 123,211 | 16 | 180.15Gbps | 190,120 | 128 | 180.20Gbps | 173,508 | 256 | 180.27Gbps | 189,884 | ntttcp tests with 1024 Rx buffers --------------------------------------- connections| throughput| no_wqe errs| --------------------------------------- 1 | 44.22Gbps | 19,864 | 16 | 180.19Gbps | 4,430 | 128 | 180.21Gbps | 2,560 | 256 | 180.29Gbps | 1,529 | So, increasing the default RX buffers per queue count to 1024 Signed-off-by: Shradha Gupta <shradhagupta@linux.microsoft.com> Reviewed-by: Haiyang Zhang <haiyangz@microsoft.com> Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com> Link: https://patch.msgid.link/1727667875-29908-1-git-send-email-shradhagupta@linux.microsoft.com Signed-off-by: Paolo Abeni <pabeni@redhat.com>
2024-10-03selftests/net: Add missing va_end.zhang jiao
There is no va_end after va_copy, just add it. Signed-off-by: zhang jiao <zhangjiao2@cmss.chinamobile.com> Reviewed-by: Simon Horman <horms@kernel.org> Link: https://patch.msgid.link/20240927040050.7851-1-zhangjiao2@cmss.chinamobile.com Signed-off-by: Paolo Abeni <pabeni@redhat.com>
2024-10-03iomap: constrain the file range passed to iomap_file_unshareDarrick J. Wong
File contents can only be shared (i.e. reflinked) below EOF, so it makes no sense to try to unshare ranges beyond EOF. Constrain the file range parameters here so that we don't have to do that in the callers. Fixes: 5f4e5752a8a3 ("fs: add iomap_file_dirty") Signed-off-by: Darrick J. Wong <djwong@kernel.org> Link: https://lore.kernel.org/r/20241002150213.GC21853@frogsfrogsfrogs Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-10-03iomap: don't bother unsharing delalloc extentsDarrick J. Wong
If unshare encounters a delalloc reservation in the srcmap, that means that the file range isn't shared because delalloc reservations cannot be reflinked. Therefore, don't try to unshare them. Signed-off-by: Darrick J. Wong <djwong@kernel.org> Link: https://lore.kernel.org/r/20241002150040.GB21853@frogsfrogsfrogs Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-10-03net/ncsi: Disable the ncsi work before freeing the associated structureEddie James
The work function can run after the ncsi device is freed, resulting in use-after-free bugs or kernel panic. Fixes: 2d283bdd079c ("net/ncsi: Resource management") Signed-off-by: Eddie James <eajames@linux.ibm.com> Link: https://patch.msgid.link/20240925155523.1017097-1-eajames@linux.ibm.com Signed-off-by: Paolo Abeni <pabeni@redhat.com>
2024-10-02net: phy: qt2025: Fix warning: unused import DeviceIdFUJITA Tomonori
Fix the following warning when the driver is compiled as built-in: warning: unused import: `DeviceId` --> drivers/net/phy/qt2025.rs:18:5 | 18 | DeviceId, Driver, | ^^^^^^^^ | = note: `#[warn(unused_imports)]` on by default device_table in module_phy_driver macro is defined only when the driver is built as a module. Use phy::DeviceId in the macro instead of importing `DeviceId` since `phy` is always used. Fixes: fd3eaad826da ("net: phy: add Applied Micro QT2025 PHY driver") Reported-by: kernel test robot <lkp@intel.com> Closes: https://lore.kernel.org/oe-kbuild-all/202409190717.i135rfVo-lkp@intel.com/ Reviewed-by: Alice Ryhl <aliceryhl@google.com> Reviewed-by: Trevor Gross <tmgross@umich.edu> Signed-off-by: FUJITA Tomonori <fujita.tomonori@gmail.com> Reviewed-by: Fiona Behrens <me@kloenk.dev> Acked-by: Miguel Ojeda <ojeda@kernel.org> Link: https://patch.msgid.link/20240926121404.242092-1-fujita.tomonori@gmail.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-02Merge branch 'net-pcs-xpcs-cleanups-batch-1'Jakub Kicinski
Russell King says: ==================== net: pcs: xpcs: cleanups batch 1 First, sorry for the bland series subject - this is the first in a number of cleanup series to the XPCS driver. This series has some functional changes beyond merely cleanups, notably the first patch. This series starts off with a patch that moves the PCS reset from the xpcs_create*() family of calls to when phylink first configures the PHY. The motivation for this change is to get rid of the interface argument to the xpcs_create*() functions, which I see as unnecessary complexity. This patch should be tested on Wangxun and STMMAC drivers. Patch 2 removes the now unnecessary interface argument from the internal xpcs_create() and xpcs_init_iface() functions. With this, xpcs_init_iface() becomes a misnamed function, but patch 3 removes this function, moving its now meager contents to xpcs_create(). Patch 4 adds xpcs_destroy_pcs() and xpcs_create_pcs_mdiodev() functions which return and take a phylink_pcs, allowing SJA1105 and Wangxun drivers to be converted to using the phylink_pcs structure internally. Patches 5 through 8 convert both these drivers to that end. Patch 9 drops the interface argument from the remaining xpcs_create*() functions, addressing the only remaining caller of these functions, that being the STMMAC driver. As patch 7 removed the direct calls to the XPCS config/link-up functions, the last patch makes these functions static. ==================== Link: https://patch.msgid.link/ZvwdKIp3oYSenGdH@shell.armlinux.org.uk Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-02net: pcs: xpcs: make xpcs_do_config() and xpcs_link_up() internalRussell King (Oracle)
As nothing outside pcs-xpcs.c calls neither xpcs_do_config() nor xpcs_link_up(), remove their exports and prototypes. Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk> Link: https://patch.msgid.link/E1svfMv-005ZIv-2M@rmk-PC.armlinux.org.uk Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-02net: pcs: xpcs: drop interface argument from xpcs_create*()Russell King (Oracle)
The XPCS sub-driver no longer uses the "interface" argument to the xpcs_create_mdiodev() and xpcs_create_fwnode() functions. Remove this now unnecessary argument, updating the stmmac driver appropriately. Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk> Link: https://patch.msgid.link/E1svfMp-005ZIp-UX@rmk-PC.armlinux.org.uk Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-02net: dsa: sja1105: use phylink_pcs internallyRussell King (Oracle)
Use xpcs_create_pcs_mdiodev() to create the XPCS instance, storing and using the phylink_pcs pointer internally, rather than dw_xpcs. Use xpcs_destroy_pcs() to destroy the XPCS instance when we've finished with it. Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk> Link: https://patch.msgid.link/E1svfMk-005ZIj-R3@rmk-PC.armlinux.org.uk Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-02net: dsa: sja1105: call PCS config/link_up via pcs_ops structureRussell King (Oracle)
Call the PCS operations through the ops structure, which avoids needing to export xpcs internal functions. Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk> Link: https://patch.msgid.link/E1svfMf-005ZId-Mx@rmk-PC.armlinux.org.uk Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-02net: dsa: sja1105: simplify static configuration reloadRussell King (Oracle)
The static configuration reload saves the port speed in the static configuration tables by first converting it from the internal respresentation to the SPEED_xxx ethtool representation, and then converts it back to restore the setting. This is because sja1105_adjust_port_config() takes the speed as SPEED_xxx. However, this is unnecessarily complex. If we split sja1105_adjust_port_config() up, we can simply save and restore the mac[port].speed member in the static configuration tables. Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk> Link: https://patch.msgid.link/E1svfMa-005ZIX-If@rmk-PC.armlinux.org.uk Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-02net: wangxun: txgbe: use phylink_pcs internallyRussell King (Oracle)
Use xpcs_create_pcs_mdiodev() to create the XPCS instance, storing and using the phylink_pcs pointer internally, rather than dw_xpcs. Use xpcs_destroy_pcs() to destroy the XPCS instance when we've finished with it. Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Link: https://patch.msgid.link/E1svfMV-005ZIR-FE@rmk-PC.armlinux.org.uk Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-02net: pcs: xpcs: add xpcs_destroy_pcs() and xpcs_create_pcs_mdiodev()Russell King (Oracle)
Provide xpcs create/destroy functions that return and take a phylink_pcs pointer instead of an xpcs pointer. This will be used by drivers that have been converted to use phylink_pcs pointers internally, rather than dw_xpcs pointers. As xpcs_create_mdiodev() no longer makes use of its interface argument, pass PHY_INTERFACE_MODE_NA into xpcs_create_mdiodev() until it is removed later in the series. Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk> Link: https://patch.msgid.link/E1svfMQ-005ZIL-Bi@rmk-PC.armlinux.org.uk Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-02net: pcs: xpcs: get rid of xpcs_init_iface()Russell King (Oracle)
xpcs_init_iface() no longer does anything with the interface mode, and now merely does configuration related to the PMA ID. Move this back into xpcs_create() as it doesn't warrant being a separate function anymore. Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk> Link: https://patch.msgid.link/E1svfML-005ZIF-84@rmk-PC.armlinux.org.uk Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-02net: pcs: xpcs: drop interface argument from internal functionsRussell King (Oracle)
Now that we no longer use the "interface" argument when creating the XPCS sub-driver, remove it from xpcs_create() and xpcs_init_iface(). Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk> Link: https://patch.msgid.link/E1svfMG-005ZI9-3k@rmk-PC.armlinux.org.uk Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-02net: pcs: xpcs: move PCS reset to .pcs_pre_config()Russell King (Oracle)
Move the PCS reset to .pcs_pre_config() rather than at creation time, which means we call the reset function with the interface that we're actually going to be using to talk to the downstream device. Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com> Tested-by: Vladimir Oltean <vladimir.oltean@nxp.com> # sja1105 Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk> Tested-by: for them? Link: https://patch.msgid.link/E1svfMA-005ZI3-Va@rmk-PC.armlinux.org.uk Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-10-02gso: fix udp gso fraglist segmentation after pull from frag_listWillem de Bruijn
Detect gso fraglist skbs with corrupted geometry (see below) and pass these to skb_segment instead of skb_segment_list, as the first can segment them correctly. Valid SKB_GSO_FRAGLIST skbs - consist of two or more segments - the head_skb holds the protocol headers plus first gso_size - one or more frag_list skbs hold exactly one segment - all but the last must be gso_size Optional datapath hooks such as NAT and BPF (bpf_skb_pull_data) can modify these skbs, breaking these invariants. In extreme cases they pull all data into skb linear. For UDP, this causes a NULL ptr deref in __udpv4_gso_segment_list_csum at udp_hdr(seg->next)->dest. Detect invalid geometry due to pull, by checking head_skb size. Don't just drop, as this may blackhole a destination. Convert to be able to pass to regular skb_segment. Link: https://lore.kernel.org/netdev/20240428142913.18666-1-shiming.cheng@mediatek.com/ Fixes: 9fd1ff5d2ac7 ("udp: Support UDP fraglist GRO/GSO.") Signed-off-by: Willem de Bruijn <willemb@google.com> Cc: stable@vger.kernel.org Link: https://patch.msgid.link/20241001171752.107580-1-willemdebruijn.kernel@gmail.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
generated by cgit (git 2.34.1) at 2025-05-05 01:13:15 +0000