- 12 Apr, 2022 7 commits
-
-
Haiyang Zhang authored
Handle XDP_REDIRECT action in netvsc driver. Also, transparently pass ndo_xdp_xmit to VF when available. Signed-off-by:
Haiyang Zhang <haiyangz@microsoft.com> Link: https://lore.kernel.org/r/1649362894-20077-1-git-send-email-haiyangz@microsoft.comSigned-off-by:
Jakub Kicinski <kuba@kernel.org>
-
Jakub Kicinski authored
Guillaume Nault says: ==================== ipv4: Convert several tos fields to dscp_t Continue the work started with commit a410a0cf ("ipv6: Define dscp_t and stop taking ECN bits into account in fib6-rules") and convert more structure fields and variables to dscp_t. This series focuses on struct fib_rt_info, struct fib_entry_notifier_info and their users (networking drivers). The purpose of dscp_t is to ensure that ECN bits don't influence IP route lookups. It does so by ensuring that dscp_t variables have the ECN bits cleared. Notes: * This series is entirely about type annotation and isn't supposed to have any user visible effect. * The first two patches have to introduce a few dsfield <-> dscp conversions in the affected drivers, but those are then removed when converting the internal driver structures (patches 3-5). In the end, drivers don't have to handle any conversion. ==================== Link: https://lore.kernel.org/r/cover.1649445279.git.gnault@redhat.comSigned-off-by:
-
Guillaume Nault authored
Use the new dscp_t type to replace the kern_tos field of struct prestera_kern_fib_cache. This ensures ECN bits are ignored and makes it compatible with the dscp fields of struct fib_entry_notifier_info and struct fib_rt_info. This also allows sparse to flag potential incorrect uses of DSCP and ECN bits. Signed-off-by:
Guillaume Nault <gnault@redhat.com> Reviewed-by:
Yevhen Orlov <yevhen.orlov@plvision.eu> Signed-off-by:
-
Guillaume Nault authored
Use the new dscp_t type to replace the tos field of struct mlxsw_sp_fib4_entry. This ensures ECN bits are ignored and makes it compatible with the dscp fields of fib_entry_notifier_info and fib_rt_info. This also allows sparse to flag potential incorrect uses of DSCP and ECN bits. Signed-off-by:
Ido Schimmel <idosch@nvidia.com> Signed-off-by:
-
Guillaume Nault authored
Use the new dscp_t type to replace the tos field of struct nsim_fib4_rt. This ensures ECN bits are ignored and makes it compatible with the dscp fields of struct fib_entry_notifier_info and struct fib_rt_info. This also allows sparse to flag potential incorrect uses of DSCP and ECN bits. Signed-off-by:
David Ahern <dsahern@kernel.org> Signed-off-by:
-
Guillaume Nault authored
Use the new dscp_t type to replace the tos field of struct fib_entry_notifier_info. This ensures ECN bits are ignored and makes it compatible with the dscp field of struct fib_rt_info. This also allows sparse to flag potential incorrect uses of DSCP and ECN bits. Signed-off-by:
-
Guillaume Nault authored
Use the new dscp_t type to replace the tos field of struct fib_rt_info. This ensures ECN bits are ignored and makes it compatible with the fa_dscp field of struct fib_alias. This also allows sparse to flag potential incorrect uses of DSCP and ECN bits. Signed-off-by:
-
- 11 Apr, 2022 21 commits
-
-
Grygorii Strashko authored
Since commit 1771afd4 ("net: cpsw: avoid alignment faults by taking NET_IP_ALIGN into account") the TI CPSW driver was switched to use correct define CPSW_HEADROOM_NA to avoid alignment faults, but there are two places left where CPSW_HEADROOM is still used (without causing issues). Hence, completely drop CPSW_HEADROOM define and use CPSW_HEADROOM_NA everywhere to avoid further mistakes in code. Signed-off-by:
Grygorii Strashko <grygorii.strashko@ti.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
David S. Miller authored
Mat Martineau says: ==================== mptcp: Miscellaneous changes for 5.19 Four separate groups of patches here: Patch 1 optimizes flag checking when releasing mptcp socket locks. Patches 2 and 3 update the packet scheduler when subflow priorities change. Patch 4 adds some pernet helper functions for MPTCP. Patches 5-8 add diag support for MPTCP listeners, including a selftest. ==================== Signed-off-by: -
Florian Westphal authored
Check dumping of mptcp listener sockets: 1. filter by dport should not return any results 2. filter by sport should return listen sk 3. filter by saddr+sport should return listen sk 4. no filter should return listen sk Signed-off-by:
Florian Westphal <fw@strlen.de> Signed-off-by:
Mat Martineau <mathew.j.martineau@linux.intel.com> Signed-off-by:
-
Florian Westphal authored
makes 'ss -Ml' show mptcp listen sockets. Iterate over the tcp listen sockets and pick those that have mptcp ulp info attached. mptcp_diag_get_info() is modified to prefer msk->first for mptcp sockets in listen state. This reports accurate number for recv and send queue (pending / max connection backlog counters). Sample output: ss -Mil State Recv-Q Send-Q Local Address:Port Peer Address:Port LISTEN 0 20 127.0.0.1:12000 0.0.0.0:* subflows_max:2 Signed-off-by: -
Florian Westphal authored
Problem is that listener iteration would call this from atomic context so this locking is not allowed. One way is to drop locks before calling the helper, but afaics the lock isn't really needed, all values are fetched via READ_ONCE(). Signed-off-by:
-
Florian Westphal authored
Raw access to cb->arg[] is deprecated, use a context structure. Signed-off-by:
-
Geliang Tang authored
This patch adds two pm_nl_pernet related helpers, named pm_nl_get_pernet() and pm_nl_get_pernet_from_msk() to get pm_nl_pernet from 'net' or 'msk'. Use these helpers instead of using net_generic() directly. Suggested-by:
Geliang Tang <geliang.tang@suse.com> Signed-off-by:
-
Paolo Abeni authored
Similar to the previous patch, for priority changes requested by the local PM. Reported-and-suggested-by:
Davide Caratti <dcaratti@redhat.com> Fixes: 06706542 ("mptcp: add the outgoing MP_PRIO support") Signed-off-by:
Paolo Abeni <pabeni@redhat.com> Signed-off-by:
-
Paolo Abeni authored
When an incoming MP_PRIO option changes the backup status of any subflow, we need to reset the packet scheduler status, or the next send could keep using the previously selected subflow, without taking in account the new priorities. Reported-by:
-
Paolo Abeni authored
The mptcp release callback checks several flags in atomic context, but only MPTCP_CLEAN_UNA can be up frequently. Reorganize the code to avoid multiple conditionals in the most common scenarios. Additional clarify a related comment. Signed-off-by:
-
git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf-nextDavid S. Miller authored
Pablo Neira Ayuso says: ==================== Netfilter updates for net-next The following patchset contains Netfilter updates for net-next: 1) Replace unnecessary list_for_each_entry_continue() in nf_tables, from Jakob Koschel. 2) Add struct nf_conntrack_net_ecache to conntrack event cache and use it, from Florian Westphal. 3) Refactor ctnetlink_dump_list(), also from Florian. 4) Bump module reference counter on cttimeout object addition/removal, from Florian. 5) Consolidate nf_log MAC printer, from Phil Sutter. 6) Add basic logging support for unknown ethertype, from Phil Sutter. 7) Consolidate check for sysctl nf_log_all_netns toggle, also from Phil. 8) Replace hardcode value in nft_bitwise, from Jeremy Sowden. 9) Rename BASIC-like goto tags in nft_bitwise to more meaningful names, also from Jeremy. 10) nft_fib support for reverse path filtering with policy-based routing on iif. Extend selftests to cover for this new usecase, from Florian. ==================== Signed-off-by: -
Florian Westphal authored
Its now possible to use fib expression in the forward chain (where both the input and output interfaces are known). Add a simple test case for this. Signed-off-by:
Pablo Neira Ayuso <pablo@netfilter.org>
-
Pablo Neira Ayuso authored
If policy-based routing using the iif selector is used, then the fib expression fails to look up for the reverse path from the prerouting hook because the input interface cannot be inferred. In order to support this scenario, extend the fib expression to allow to use after the route lookup, from the forward hook. This patch also adds support for the input hook for usability reasons. Since the prerouting hook cannot be used for the scenario described above, users need two rules: one for the forward chain and another rule for the input chain to check for the reverse path check for locally targeted traffic. Acked-by:
-
Lv Ruyi authored
There are some spelling mistakes in the comments for macro. Fix it. Reported-by:
Zeal Robot <zealci@zte.com.cn> Signed-off-by:
Lv Ruyi <lv.ruyi@zte.com.cn> Signed-off-by:
-
Felix Fietkau authored
Descriptor fields are little-endian Fixes: 804775df ("net: ethernet: mtk_eth_soc: add support for Wireless Ethernet Dispatch (WED)") Reported-by:
kernel test robot <lkp@intel.com> Signed-off-by:
Felix Fietkau <nbd@nbd.name> Signed-off-by:
-
Yang Yingliang authored
If syscon_regmap_lookup_by_phandle() fails, it never return NULL pointer, change the check to IS_ERR(). Fixes: 804775df ("net: ethernet: mtk_eth_soc: add support for Wireless Ethernet Dispatch (WED)") Reported-by:
Hulk Robot <hulkci@huawei.com> Signed-off-by:
Yang Yingliang <yangyingliang@huawei.com> Signed-off-by:
-
David S. Miller authored
Menglong Dong says: ==================== net: icmp: add skb drop reasons to icmp In the commit c504e5c2 ("net: skb: introduce kfree_skb_reason()"), we added the support of reporting the reasons of skb drops to kfree_skb tracepoint. And in this series patches, reasons for skb drops are added to ICMP protocol. In order to report the reasons of skb drops in 'sock_queue_rcv_skb()', the function 'sock_queue_rcv_skb_reason()' is introduced in the 1th patch, which is used in the 3th patch. As David Ahern suggested, the reasons for skb drops should be more general and not be code based. Therefore, in the 2th patch, SKB_DROP_REASON_PTYPE_ABSENT is renamed to SKB_DROP_REASON_UNHANDLED_PROTO, which is used for the cases of no L3 protocol handler, no L4 protocol handler, version extensions, etc. In the 3th patch, we introduce the new function __ping_queue_rcv_skb() to report drop reasons by its return value and keep the return value of ping_queue_rcv_skb() still. In the 4th patch, we make ICMP message handler functions return drop reasons, which means we change the return type of 'handler()' in 'struct icmp_control' from 'bool' to 'enum skb_drop_reason'. This changed its original intention, as 'false' means failure, but 'SKB_NOT_DROPPED_YET', which is 0, means success now. Therefore, we have to change all usages of these handler. Following "handler" functions are involved: icmp_unreach() icmp_redirect() icmp_echo() icmp_timestamp() icmp_discard() And following drop reasons are added(what they mean can be see in the document for them): SKB_DROP_REASON_ICMP_CSUM SKB_DROP_REASON_INVALID_PROTO The reason 'INVALID_PROTO' is introduced for the case that the packet doesn't follow rfc 1122 and is dropped. I think this reason is different from the 'UNHANDLED_PROTO', as the 'UNHANDLED_PROTO' means the packet is fine, and it is just not supported. This is not a common case, and I believe we can locate the problem from the data in the packet. For now, this 'INVALID_PROTO' is used for the icmp broadcasts with wrong types. Maybe there should be a document file for these reasons. For example, list all the case that causes the 'INVALID_PROTO' drop reason. Therefore, users can locate their problems according to the document. Changes since v4: - rename SKB_DROP_REASON_RFC_1122 to SKB_DROP_REASON_INVALID_PROTO Changes since v3: - rename SKB_DROP_REASON_PTYPE_ABSENT to SKB_DROP_REASON_UNHANDLED_PROTO in the 2th patch - fix the return value problem of ping_queue_rcv_skb() in the 3th patch - remove SKB_DROP_REASON_ICMP_TYPE and SKB_DROP_REASON_ICMP_BROADCAST and introduce the SKB_DROP_REASON_RFC_1122 in the 4th patch Changes since v2: - fix aliegnment problem in the 2th patch Changes since v1: - introduce __ping_queue_rcv_skb() instead of change the return value of ping_queue_rcv_skb() in the 2th patch, as Paolo suggested ==================== Signed-off-by:
-
Menglong Dong authored
Replace kfree_skb() used in icmp_rcv() and icmpv6_rcv() with kfree_skb_reason(). In order to get the reasons of the skb drops after icmp message handle, we change the return type of 'handler()' in 'struct icmp_control' from 'bool' to 'enum skb_drop_reason'. This may change its original intention, as 'false' means failure, but 'SKB_NOT_DROPPED_YET' means success now. Therefore, all 'handler' and the call of them need to be handled. Following 'handler' functions are involved: icmp_unreach() icmp_redirect() icmp_echo() icmp_timestamp() icmp_discard() And following new drop reasons are added: SKB_DROP_REASON_ICMP_CSUM SKB_DROP_REASON_INVALID_PROTO The reason 'INVALID_PROTO' is introduced for the case that the packet doesn't follow rfc 1122 and is dropped. This is not a common case, and I believe we can locate the problem from the data in the packet. For now, this 'INVALID_PROTO' is used for the icmp broadcasts with wrong types. Maybe there should be a document file for these reasons. For example, list all the case that causes the 'UNHANDLED_PROTO' and 'INVALID_PROTO' drop reason. Therefore, users can locate their problems according to the document. Reviewed-by:
Hao Peng <flyingpeng@tencent.com> Reviewed-by:
Jiang Biao <benbjiang@tencent.com> Signed-off-by:
Menglong Dong <imagedong@tencent.com> Reviewed-by:
-
Menglong Dong authored
In order to avoid to change the return value of ping_queue_rcv_skb(), introduce the function __ping_queue_rcv_skb(), which is able to report the reasons of skb drop as its return value, as Paolo suggested. Meanwhile, make ping_queue_rcv_skb() a simple call to __ping_queue_rcv_skb(). The kfree_skb() and sock_queue_rcv_skb() used in ping_queue_rcv_skb() are replaced with kfree_skb_reason() and sock_queue_rcv_skb_reason() now. Reviewed-by:
-
Menglong Dong authored
As David Ahern suggested, the reasons for skb drops should be more general and not be code based. Therefore, rename SKB_DROP_REASON_PTYPE_ABSENT to SKB_DROP_REASON_UNHANDLED_PROTO, which is used for the cases of no L3 protocol handler, no L4 protocol handler, version extensions, etc. From previous discussion, now we have the aim to make these reasons more abstract and users based, avoiding code based. Signed-off-by:
-
Menglong Dong authored
In order to report the reasons of skb drops in 'sock_queue_rcv_skb()', introduce the function 'sock_queue_rcv_skb_reason()'. As the return value of 'sock_queue_rcv_skb()' is used as the error code, we can't make it as drop reason and have to pass extra output argument. 'sock_queue_rcv_skb()' is used in many places, so we can't change it directly. Introduce the new function 'sock_queue_rcv_skb_reason()' and make 'sock_queue_rcv_skb()' an inline call to it. Reviewed-by:
-
- 10 Apr, 2022 12 commits
-
-
David S. Miller authored
Jakub Kicinski says: ==================== tls: rx: random refactoring part 2 TLS Rx refactoring. Part 2 of 3. This one focusing on the main loop. A couple of features to follow. ====================
-
Jakub Kicinski authored
The current invese logic is harder to follow (and adds extra tests to the fast path). We have to enumerate all cases which need to keep the skb before consuming it. It's simpler to jump out of the full record flow as we detect those cases. This makes it clear that partial consumption and peek can only reach end of the function thru the !zc case so move the code up there. Signed-off-by:
-
Jakub Kicinski authored
Whatever we do in the loop the skb should not remain on as ctx->recv_pkt afterwards. We can clear that pointer and restart strparser earlier. This adds overhead of extra linking and unlinking to rx_list but that's not large (upcoming change will switch to unlocked skb list operations). Signed-off-by:
-
Jakub Kicinski authored
tls_sw_advance_skb() always consumes the skb at the end of the loop. To fall here the following must be true: !async && !is_peek && !retain_skb retain_skb => !zc && rxm->full_len > len # but non-full record implies !zc, so above can be simplified as retain_skb => rxm->full_len > len !async && !is_peek && !(rxm->full_len > len) !async && !is_peek && rxm->full_len <= len tls_sw_advance_skb() returns false if len < rxm->full_len which can't be true given conditions above. Signed-off-by: -
Jakub Kicinski authored
Most of the conditions deciding if zero-copy can be used do not change throughout the iterations, so pre-calculate them. Signed-off-by:
-
Jakub Kicinski authored
We track both if the last record was handled by async crypto and how many records were async. This is not necessary. We implicitly assume once crypto goes async it will stay that way, otherwise we'd reorder records. So just track if we're in async mode, the exact number of records is not necessary. This change also forces us into "async" mode more consistently in case crypto ever decided to interleave async and sync. Signed-off-by:
-
Jakub Kicinski authored
tls_sw_advance_skb() caters to the async case when skb argument is NULL. In that case it simply unpauses the strparser. These are surprising semantics to a person reading the code, and result in higher LoC, so inline the __strp_unpause and only call tls_sw_advance_skb() when we actually move past an skb. Signed-off-by:
-
Jakub Kicinski authored
cmsg can be filled in during rx_list processing or normal receive. Consolidate the code. We don't need to keep the boolean to track if the cmsg was created. 0 is an invalid content type. Signed-off-by:
-
Jakub Kicinski authored
Since we are protected from async completions by decrypt_compl_lock we can drop the async_notify and reinit the completion before we start waiting. Signed-off-by:
-
Jakub Kicinski authored
We pass zc as a pointer to bool a few functions down as an in/out argument. This is error prone since C will happily evalue a pointer as a boolean (IOW forgetting *zc and writing zc leads to loss of developer time..). Wrap the arguments into a structure. Signed-off-by:
-
Jakub Kicinski authored
We plumb pointer to chunk all the way to the decryption method. It's set to the length of the text when decrypt_skb_update() returns. I think the code is written this way because original TLS implementation passed &chunk to zerocopy_from_iter() and this was carried forward as the code gotten more complex, without any refactoring. The fix for peek() introduced a new variable - to_decrypt which for all practical purposes is what chunk is going to get set to. Spare ourselves the pointer passing, use to_decrypt. Use this opportunity to clean things up a little further. Note that chunk / to_decrypt was mostly needed for the async path, since the sync path would access rxm->full_len (decryption transforms full_len from record size to text size). Use the right source of truth more explicitly. We have three cases: - async - it's TLS 1.2 only, so chunk == to_decrypt, but we need the min() because to_decrypt is a whole record and we don't want to underflow len. Note that we can't handle partial record by falling back to sync as it would introduce reordering against records in flight. - zc - again, TLS 1.2 only for now, so chunk == to_decrypt, we don't do zc if len < to_decrypt, no need to check again. - normal - it already handles chunk > len, we can factor out the assignment to rxm->full_len and share it with zc. Signed-off-by: -
Jakub Kicinski authored
sk is unused, remove it to make it clear the function doesn't poke at the socket. size_used is always 0 on input and @length on success. Signed-off-by:
-
