aboutsummaryrefslogtreecommitdiff
AgeCommit message (Collapse)AuthorFilesLines
2022-09-05net: ipa: simplify gsi_channel_trans_last()Alex Elder1-26/+7
Using a little logic we can simplify gsi_channel_trans_last(). The first condition in that function looks like this: if (trans_info->allocated_id != trans_info->free_id) And if that's false, we proceed to the next one: if (trans_info->committed_id != trans_info->allocated_id) Failure of the first test implies: trans_info->allocated_id == trans_info->free_id And therefore, the second one can be rewritten this way: if (trans_info->committed_id != trans_info->free_id) Substituting free_id for allocated_id and committed_id can also be done in the code blocks executed when these conditions yield true. The net result is that all three blocks for TX endpoints can be consolidated into just one. The two blocks of code at the end of that function (used for both TX and RX channels) can be similarly consolidated into a single block. Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-09-05net: ipa: use IDs exclusively for last transactionAlex Elder1-15/+27
Always use transaction IDs when finding the "last" transaction to await when quiescing a channel. This basically extends what was done in the previous patch to all other transaction state IDs. As a result we are no longer updating any transaction lists inside gsi_channel_trans_last(), so there's no need to take the spinlock. Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-09-05net: ipa: use IDs for last allocated transactionAlex Elder1-3/+9
Use the allocated and free transaction IDs to determine whether the "last" transaction used for quiescing a channel is in allocated state. The last allocated transaction that has not been committed (if any) immediately precedes the first free transaction in the transaction array. Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-09-05net: ipa: rework last transaction determinationAlex Elder3-36/+28
When quiescing a channel, we find the "last" transaction, which is the latest one to have been allocated. (New transaction allocation will have been prevented by the time this is called.) Currently we do this by looking for the first non-empty transaction list in each state, then return the last entry from that last. Instead, determine the last entry in each list (if any) and return that entry if found. Temporarily (locally) introduce list_last_entry_or_null() as a helper for this, mirroring list_first_entry_or_null(). This macro definition will be removed by an upcoming patch. Remove the temporary warnings added by the previous commit. Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-09-05r8169: use devm_clk_get_optional_enabled() to simplify the codeHeiner Kallweit1-34/+3
Now that we have devm_clk_get_optional_enabled(), we don't have to open-code it. Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-09-05r8169: remove comment about apparently non-existing chip versionsHeiner Kallweit1-6/+0
It's not clear where these entries came from, and as I wrote in the comment: Not even Realtek's r8101 driver knows these chip id's. So remove the comment. Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-09-05r8169: merge handling of chip versions 12 and 17 (RTL8168B)Heiner Kallweit3-8/+1
It's not clear why XID's 380 and 381..387 ever got different chip version id's. VER_12 and VER_17 are handled exactly the same. Therefore merge handling under the VER_17 umbrella. Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-09-05Merge branch 'altera-tse-phylink'David S. Miller11-447/+536
Maxime Chevallier says: ==================== net: altera: tse: phylink conversion This is V4 of a series converting the Altera TSE driver to phylink, introducing a new PCS driver along the way. The Altera TSE can be built with a SGMII/1000BaseX PCS, allowing to use SFP ports with this MAC, which is the end goal of adding phylink support and a proper PCS driver. The PCS itself can either be mapped in the MAC's register space, in that case, it's accessed through 32 bits registers, with the higher 16 bits always 0. Alternatively, it can sit on its own register space, exposing 16 bits registers, some of which ressemble the standard PHY registers. To tackle that rework, several things needs updating, starting by the DT binding, since we add support for a new register range for the PCS. Hence, the first patch of the series is a conversion to YAML of the existing binding. Then, patch 2 does a bit of simple cleanup to the TSE driver, using nice reverse xmas tree definitions. Patch 3 adds the actual PCS driver, as a standalone driver. Some future series will then reuse that PCS driver from the dwmac-socfpga driver, which implements support for this exact PCS too, allowing to share the code nicely. Patch 4 is then a phylink conversion of the altera_tse driver, to use this new PCS driver. Finally, patch 5 updates the newly converted DT binding to support the pcs register range. This series contains bits and pieces for this conversion, please tell me if you want me to send it as individual patches. V4 Changes: - Add missing MODULE_* macros to the TSE PCS driver V3 Changes: - YAML binding conversion changes and PCS addition changes thanks to Krzysztof's reviews V2 Changes : - Fixed the binding after the YAML conversion - Added a pcs_validate() callback - Introduced a comment to justify a soft reset for the PCS ==================== Signed-off-by: David S. Miller <davem@davemloft.net>
2022-09-05dt-bindings: net: altera: tse: add an optional pcs register rangeMaxime Chevallier1-0/+27
Some implementations of the TSE have their PCS as an external bloc, exposed at its own register range. Document this, and add a new example showing a case using the pcs and the new phylink conversion to connect an sfp port to a TSE mac. Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-09-05net: altera: tse: convert to phylinkMaxime Chevallier4-314/+141
Convert the Altera Triple Speed Ethernet Controller to phylink. This controller supports MII, GMII and RGMII with its MAC, and SGMII + 1000BaseX through a small embedded PCS. The PCS itself has a register set very similar to what is found in a typical 802.3 ethernet PHY, but this register set memory-mapped instead of lying on an mdio bus. Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-09-05net: pcs: add new PCS driver for altera TSE PCSMaxime Chevallier5-0/+206
The Altera Triple Speed Ethernet has a SGMII/1000BaseC PCS that can be integrated in several ways. It can either be part of the TSE MAC's address space, accessed through 32 bits accesses on the mapped mdio device 0, or through a dedicated 16 bits register set. This driver allows using the TSE PCS outside of altera TSE's driver, since it can be used standalone by other MACs. Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-09-05net: altera: tse: cosmetic change to use reverse xmas tree orderingMaxime Chevallier2-22/+23
Make the driver code cleaner through a strictly cosmetic change, using he reverse xmas tree variable declaration ordering. Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-09-05dt-bindings: net: Convert Altera TSE bindings to yamlMaxime Chevallier2-113/+141
Convert the bindings for the Altera Triple-Speed Ethernet to yaml. Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-09-04Merge tag 'wireless-next-2022-09-03' of ↵David S. Miller71-7097/+26678
git://git.kernel.org/pub/scm/linux/kernel/git/wireless/wireless-next Johannes Berg says: ==================== drivers - rtw89: large update across the map, e.g. coex, pci(e), etc. - ath9k: uninit memory read fix - ath10k: small peer map fix and a WCN3990 device fix - wfx: underflow stack - the "change MAC address while IFF_UP" change from James we discussed - more MLO work, including a set of fixes for the previous code, now that we have more code we can exercise it more - prevent some features with MLO that aren't ready yet (AP_VLAN and 4-address connections) ==================== Signed-off-by: David S. Miller <davem@davemloft.net>
2022-09-03wifi: mac80211_hwsim: fix multi-channel handling in netlink RXJohannes Berg1-22/+4
In netlink RX, now that we can actually have multiple channel contexts for MLO, things don't work well as we only keep a single pointer, and then on link switching we might NULL it, and hit the return if the channel is NULL. However, we already use mac80211_hwsim_tx_iter() which deals with all this, so remove the test and adjust the remaining code a bit. This then means we no longer use the chanctx pointer, so remove it as well. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2022-09-03wifi: mac80211: call drv_sta_state() under sdata_lock() in reconfigJohannes Berg1-15/+13
Currently, other paths calling drv_sta_state() hold the mutex and therefore drivers can assume that, and look at links with that protection. Fix that for the reconfig path as well; to do it more easily use ieee80211_reconfig_stations() for the AP/AP_VLAN station reconfig as well. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2022-09-03wifi: nl80211: add MLD address to assoc BSS entriesJohannes Berg2-2/+6
Add an MLD address attribute to BSS entries that the interface is currently associated with to help userspace figure out what's going on. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2022-09-03wifi: mac80211: mlme: refactor QoS settings codeJohannes Berg2-13/+27
Refactor the code to apply QoS settings to the driver so we can call it on link switch. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2022-09-03wifi: mac80211_hwsim: warn on invalid link addressJohannes Berg1-0/+2
Catch the bugs fixed in mac80211 by the previous commits and warn if an invalid address is added (or removed). Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2022-09-03wifi: mac80211: fix double SW scan stopJohannes Berg1-1/+1
When we stop a not-yet-started scan, we erroneously call into the driver, causing a sequence of sw_scan_start() followed by sw_scan_complete() twice. This will cause a warning in hwsim with next in line commit that validates the address passed to wmediumd/virtio. Fix this by doing the calls only if we were actually scanning. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2022-09-03wifi: mac80211: mlme: assign link address correctlyJohannes Berg1-3/+2
Right now, we assign the link address only after we add the link to the driver, which is quite obviously wrong. It happens to work in many cases because it gets updated immediately, and then link_conf updates may update it, but it's clearly not really right. Set the link address during ieee80211_mgd_setup_link() so it's set before telling the driver about the link. Fixes: 81151ce462e5 ("wifi: mac80211: support MLO authentication/association with one link") Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2022-09-03wifi: mac80211: move link code to a new fileJohannes Berg4-254/+274
We probably should've done that originally, we already have about 300 lines of code there, and will add more. Move all the link code we wrote to a new file. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2022-09-03wifi: mac80211_hwsim: refactor RX a bitJohannes Berg1-11/+15
Refactor some common RX functionality between the netlink and non-netlink paths, adding the special hwsim TLV (if compiled) also in the netlink path. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2022-09-03wifi: mac80211_hwsim: check STA magic in change_sta_linksJohannes Berg1-0/+2
Just as an additional check that mac80211 isn't doing anything strange, add a check of the STA magic (which gets assigned when the station is added, and cleared when the station is removed). Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2022-09-03wifi: mac80211: remove unused arg to ieee80211_chandef_eht_operJohannes Berg3-6/+4
We don't need the sdata argument, and it doesn't make any sense for a direct conversion from one value to another, so just remove the argument Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2022-09-03wifi: mac80211_hwsim: remove multicast workaroundJohannes Berg1-6/+1
Now that we have proper multicast TX in mac80211, there's no longer a need to fake something here. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2022-09-03wifi: nl80211: remove redundant err variableJinpeng Cui1-4/+1
Return value from rdev_set_mcast_rate() directly instead of taking this in another redundant variable. Reported-by: Zeal Robot <zealci@zte.com.cn> Signed-off-by: Jinpeng Cui <cui.jinpeng2@zte.com.cn> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2022-09-03wifi: mac80211: Support POWERED_ADDR_CHANGE featureJames Prestwood2-2/+71
Adds support in mac80211 for NL80211_EXT_FEATURE_POWERED_ADDR_CHANGE. The motivation behind this functionality is to fix limitations of address randomization on frequencies which are disallowed in world roaming. The way things work now, if a client wants to randomize their address per-connection it must power down the device, change the MAC, and power back up. Here lies a problem since powering down the device may result in frequencies being disabled (until the regdom is set). If the desired BSS is on one such frequency the client is unable to connect once the phy is powered again. For mac80211 based devices changing the MAC while powered is possible but currently disallowed (-EBUSY). This patch adds some logic to allow a MAC change while powered by removing the interface, changing the MAC, and adding it again. mac80211 will advertise support for this feature so userspace can determine the best course of action e.g. disallow address randomization on certain frequencies if not supported. There are certain limitations put on this which simplify the logic: - No active connection - No offchannel work, including scanning. Signed-off-by: James Prestwood <prestwoj@gmail.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2022-09-03wifi: nl80211: Add POWERED_ADDR_CHANGE featureJames Prestwood1-0/+9
Add a new extended feature bit signifying that the wireless hardware supports changing the MAC address while the underlying net_device is powered. Note that this has a different meaning from IFF_LIVE_ADDR_CHANGE as additional restrictions might be imposed by the hardware, such as: - No connection is active on this interface, carrier is off - No scan is in progress - No offchannel operations are in progress Signed-off-by: James Prestwood <prestwoj@gmail.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2022-09-03wifi: mac80211: prevent 4-addr use on MLDsJohannes Berg2-0/+11
We haven't tried this yet, and it's not very likely to work well right now, so for now disable 4-addr use on interfaces that are MLDs. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Link: https://lore.kernel.org/r/20220902161143.f2e4cc2efaa1.I5924e8fb44a2d098b676f5711b36bbc1b1bd68e2@changeid Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2022-09-03wifi: mac80211: prevent VLANs on MLDsJohannes Berg1-0/+5
Do not allow VLANs to be added to AP interfaces that are MLDs, this isn't going to work because the link structs aren't propagated to the VLAN interfaces yet. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Link: https://lore.kernel.org/r/20220902161144.8c88531146e9.If2ef9a3b138d4f16ed2fda91c852da156bdf5e4d@changeid Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2022-09-03Merge branch 'net_sched-redundant-resource-cleanups'David S. Miller2-44/+17
Zhengchao Shao says: ==================== net: sched: remove redundant resource cleanup when init() fails qdisc_create() calls .init() to initialize qdisc. If the initialization fails, qdisc_create() will call .destroy() to release resources. ==================== Signed-off-by: David S. Miller <davem@davemloft.net>
2022-09-03net: sched: htb: remove redundant resource cleanup in htb_init()Zhengchao Shao1-27/+9
If htb_init() fails, qdisc_create() invokes htb_destroy() to clear resources. Therefore, remove redundant resource cleanup in htb_init(). Signed-off-by: Zhengchao Shao <shaozhengchao@huawei.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-09-03net: sched: fq_codel: remove redundant resource cleanup in fq_codel_init()Zhengchao Shao1-17/+8
If fq_codel_init() fails, qdisc_create() invokes fq_codel_destroy() to clear resources. Therefore, remove redundant resource cleanup in fq_codel_init(). Signed-off-by: Zhengchao Shao <shaozhengchao@huawei.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-09-03net: fec: add stop mode support for imx8 platformWei Fang2-0/+39
The current driver support stop mode by calling machine api. The patch add dts support to set GPR register for stop request. imx8mq enter stop/exit stop mode by setting GPR bit, which can be accessed by A core. imx8qm enter stop/exit stop mode by calling IMX_SC ipc APIs that communicate with M core ipc service, and the M core set the related GPR bit at last. Signed-off-by: Wei Fang <wei.fang@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-09-03r8152: Add MAC passthrough support for Lenovo Travel HubAndré Apitzsch1-0/+2
The Lenovo USB-C Travel Hub supports MAC passthrough. Signed-off-by: André Apitzsch <git@apitzsch.eu> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-09-03net/ipv4: Use __DECLARE_FLEX_ARRAY() helperGustavo A. R. Silva1-14/+6
We now have a cleaner way to keep compatibility with user-space (a.k.a. not breaking it) when we need to keep in place a one-element array (for its use in user-space) together with a flexible-array member (for its use in kernel-space) without making it hard to read at the source level. This is through the use of the new __DECLARE_FLEX_ARRAY() helper macro. The size and memory layout of the structure is preserved after the changes. See below. Before changes: $ pahole -C ip_msfilter net/ipv4/igmp.o struct ip_msfilter { union { struct { __be32 imsf_multiaddr_aux; /* 0 4 */ __be32 imsf_interface_aux; /* 4 4 */ __u32 imsf_fmode_aux; /* 8 4 */ __u32 imsf_numsrc_aux; /* 12 4 */ __be32 imsf_slist[1]; /* 16 4 */ }; /* 0 20 */ struct { __be32 imsf_multiaddr; /* 0 4 */ __be32 imsf_interface; /* 4 4 */ __u32 imsf_fmode; /* 8 4 */ __u32 imsf_numsrc; /* 12 4 */ __be32 imsf_slist_flex[0]; /* 16 0 */ }; /* 0 16 */ }; /* 0 20 */ /* size: 20, cachelines: 1, members: 1 */ /* last cacheline: 20 bytes */ }; After changes: $ pahole -C ip_msfilter net/ipv4/igmp.o struct ip_msfilter { __be32 imsf_multiaddr; /* 0 4 */ __be32 imsf_interface; /* 4 4 */ __u32 imsf_fmode; /* 8 4 */ __u32 imsf_numsrc; /* 12 4 */ union { __be32 imsf_slist[1]; /* 16 4 */ struct { struct { } __empty_imsf_slist_flex; /* 16 0 */ __be32 imsf_slist_flex[0]; /* 16 0 */ }; /* 16 0 */ }; /* 16 4 */ /* size: 20, cachelines: 1, members: 5 */ /* last cacheline: 20 bytes */ }; In the past, we had to duplicate the whole original structure within a union, and update the names of all the members. Now, we just need to declare the flexible-array member to be used in kernel-space through the __DECLARE_FLEX_ARRAY() helper together with the one-element array, within a union. This makes the source code more clean and easier to read. Link: https://github.com/KSPP/linux/issues/193 Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org> Reviewed-by: Kees Cook <keescook@chromium.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-09-02net/sched: cls_api: remove redundant 0 check in tcf_qevent_init()Zhengchao Shao1-3/+0
tcf_qevent_parse_block_index() never returns a zero block_index. Therefore, it is unnecessary to check the value of block_index in tcf_qevent_init(). Signed-off-by: Zhengchao Shao <shaozhengchao@huawei.com> Link: https://lore.kernel.org/r/20220901011617.14105-1-shaozhengchao@huawei.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-09-02net: lantiq_etop: Fix return type for implementation of ndo_start_xmitGUO Zihua1-1/+1
Since Linux now supports CFI, it will be a good idea to fix mismatched return type for implementation of hooks. Otherwise this might get cought out by CFI and cause a panic. ltq_etop_tx() would return either NETDEV_TX_BUSY or NETDEV_TX_OK, so change the return type to netdev_tx_t directly. Signed-off-by: GUO Zihua <guozihua@huawei.com> Link: https://lore.kernel.org/r/20220902081521.59867-1-guozihua@huawei.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-09-02net: sunplus: Fix return type for implementation of ndo_start_xmitGUO Zihua1-1/+2
Since Linux now supports CFI, it will be a good idea to fix mismatched return type for implementation of hooks. Otherwise this might get cought out by CFI and cause a panic. spl2sw_ethernet_start_xmit() would return either NETDEV_TX_BUSY or NETDEV_TX_OK, so change the return type to netdev_tx_t directly. Signed-off-by: GUO Zihua <guozihua@huawei.com> Link: https://lore.kernel.org/r/20220902081550.60095-1-guozihua@huawei.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-09-02net: xscale: Fix return type for implementation of ndo_start_xmitGUO Zihua1-1/+1
Since Linux now supports CFI, it will be a good idea to fix mismatched return type for implementation of hooks. Otherwise this might get cought out by CFI and cause a panic. eth_xmit() would return either NETDEV_TX_BUSY or NETDEV_TX_OK, so change the return type to netdev_tx_t directly. Signed-off-by: GUO Zihua <guozihua@huawei.com> Link: https://lore.kernel.org/r/20220902081612.60405-1-guozihua@huawei.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-09-02net: broadcom: Fix return type for implementation ofGUO Zihua1-1/+1
Since Linux now supports CFI, it will be a good idea to fix mismatched return type for implementation of hooks. Otherwise this might get cought out by CFI and cause a panic. bcm4908_enet_start_xmit() would return either NETDEV_TX_BUSY or NETDEV_TX_OK, so change the return type to netdev_tx_t directly. Signed-off-by: GUO Zihua <guozihua@huawei.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Link: https://lore.kernel.org/r/20220902075407.52358-1-guozihua@huawei.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-09-02net: ieee802154: Fix compilation error when ↵Gal Pressman1-4/+2
CONFIG_IEEE802154_NL802154_EXPERIMENTAL is disabled When CONFIG_IEEE802154_NL802154_EXPERIMENTAL is disabled, NL802154_CMD_DEL_SEC_LEVEL is undefined and results in a compilation error: net/ieee802154/nl802154.c:2503:19: error: 'NL802154_CMD_DEL_SEC_LEVEL' undeclared here (not in a function); did you mean 'NL802154_CMD_SET_CCA_ED_LEVEL'? 2503 | .resv_start_op = NL802154_CMD_DEL_SEC_LEVEL + 1, | ^~~~~~~~~~~~~~~~~~~~~~~~~~ | NL802154_CMD_SET_CCA_ED_LEVEL Unhide the experimental commands, having them defined in an enum makes no difference. Fixes: 9c5d03d36251 ("genetlink: start to validate reserved header bytes") Signed-off-by: Gal Pressman <gal@nvidia.com> Acked-by: Stefan Schmidt <stefan@datenfreihafen.org> Tested-by: Sudip Mukherjee <sudipm.mukherjee@gmail.com> Link: https://lore.kernel.org/r/20220902030620.2737091-1-kuba@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-09-02net: remove netif_tx_napi_add()Jakub Kicinski1-2/+0
All callers are now gone. Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-09-02net: bql: add more documentationEric Dumazet1-6/+26
Add some documentation for netdev_tx_sent_queue() and netdev_tx_completed_queue() Stating that netdev_tx_completed_queue() must be called once per TX completion round is apparently not obvious for everybody. Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-09-02Merge branch 'net-ipa-transaction-state-IDs'David S. Miller2-14/+84
Alex Elder says: ==================== net: ipa: use IDs to track transaction state This series is the first of three groups of changes that simplify the way the IPA driver tracks the state of its transactions. Each GSI channel has a fixed number of transactions allocated at initialization time. The number allocated matches the number of TREs in the transfer ring associated with the channel. This is because the transfer ring limits the number of transfers that can ever be underway, and in the worst case, each transaction represents a single TRE. Transactions go through various states during their lifetime. Currently a set of lists keeps track of which transactions are in each state. Initially, all transactions are free. An allocated transaction is placed on the allocated list. Once an allocated transaction is committed, it is moved from the allocated to the committed list. When a committed transaction is sent to hardware (via a doorbell) it is moved to the pending list. When hardware signals that some work has completed, transactions are moved to the completed list. Finally, when a completed transaction is polled it's moved to the polled list before being removed when it becomes free. Changing a transaction's state thus normally involves manipulating two lists, and to prevent corruption a spinlock is held while the lists are updated. Transactions move through their states in a well-defined sequence though, and they do so strictly in order. So transaction 0 is always allocated before transaction 1; transaction 0 is always committed before transaction 1; and so on, through completion, polling, and becoming free. Because of this, it's sufficient to just keep track of which transaction is the first in each state. The rest of the transactions in a given state can be derived from the first transaction in an "adjacent" state. As a result, we can track the state of all transactions with a set of indexes, and can update these without the need for a spinlock. This first group of patches just defines the set of indexes that will be used for this new way of tracking transaction state. Two more groups of patches will follow. I've broken the 17 patches into these three groups to facilitate review. ==================== Signed-off-by: David S. Miller <davem@davemloft.net>
2022-09-02net: ipa: track polled transactions with an IDAlex Elder2-16/+24
Add a transaction ID to track the first element in the transaction array that has been polled. Advance the ID when we are releasing a transaction. Temporarily add warnings that verify that the first polled transaction tracked by the ID matches the first element on the polled list, both when polling and freeing. Remove the temporary warnings added by the previous commit. Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-09-02net: ipa: track completed transactions with an IDAlex Elder2-17/+24
Add a transaction ID field to track the first element in the transaction array that has completed but has not yet been polled. Advance the ID when we are processing a transaction in the NAPI polling loop (where completed transactions become polled). Temporarily add warnings that verify that the first completed transaction tracked by the ID matches the first element on the completed list, both when pending and completing. Remove the temporary warnings added by the previous commit. Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-09-02net: ipa: track pending transactions with an IDAlex Elder2-18/+29
Add a transaction ID field to track the first element in the transaction array that is pending (sent to hardware) but not yet complete. Advance the ID when a completion event for a channel indicates that transactions have completed. Temporarily add warnings that verify that the first pending transaction tracked by the ID matches the first element on the pending list, both when pending and completing, as well as when resetting the channel. Remove the temporary warnings added by the previous commit. Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-09-02net: ipa: track committed transactions with an IDAlex Elder2-17/+29
Add a transaction ID field to track the first element in a channel's transaction array that has been committed, but not yet passed to the hardware. Advance the ID when the hardware is notified via doorbell that TREs from a transaction are ready for consumption. Temporarily add warnings that verify that the first committed transaction tracked by the ID matches the first element on the committed list, both when committing and pending (at doorbell). Remove the temporary warnings added by the previous commit. Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Powered by cgit, Gruvboxed by Blaster4385.