aboutsummaryrefslogtreecommitdiff
path: root/drivers
AgeCommit message (Collapse)AuthorFilesLines
2023-06-21wifi: iwlwifi: don't load old firmware for BzJohannes Berg1-1/+1
This is a future product, don't try to load ancient firmware images for it, they don't exist anyway. Signed-off-by: Johannes Berg <[email protected]> Signed-off-by: Gregory Greenman <[email protected]> Link: https://lore.kernel.org/r/20230621130443.a6961592f258.Ib7afecd46b1963164481c2acf35d2582691ef0bc@changeid Signed-off-by: Johannes Berg <[email protected]>
2023-06-21wifi: iwlwifi: don't load old firmware for ScJohannes Berg1-1/+1
This is a future product, don't try to load ancient firmware images for it, they don't exist anyway. Signed-off-by: Johannes Berg <[email protected]> Signed-off-by: Gregory Greenman <[email protected]> Link: https://lore.kernel.org/r/20230621130443.a15e7bf936cb.I68c3c71fda62c837e4da885a42471bf772ac1202@changeid Signed-off-by: Johannes Berg <[email protected]>
2023-06-21wifi: iwlwifi: give Sc devices their own familyJohannes Berg2-2/+3
They're not the same as Bz or any prior ones, and there's already one place in the driver that would erroneously assign a workaround to A-step Sc devices if they're just treated as a version of Bz. Fix that. Signed-off-by: Johannes Berg <[email protected]> Signed-off-by: Gregory Greenman <[email protected]> Link: https://lore.kernel.org/r/20230621130443.e98272ddb808.If18577b2393f631d1bfaa931287cae106fa32438@changeid Signed-off-by: Johannes Berg <[email protected]>
2023-06-21wifi: iwlwifi: split 22000.c into multiple filesJohannes Berg7-928/+1206
Split the configuration list in 22000.c into four new files, per new device family, so we don't have this huge unusable file. Yes, this duplicates a few small things, but that's still much better than what we have now. Signed-off-by: Johannes Berg <[email protected]> Signed-off-by: Gregory Greenman <[email protected]> Link: https://lore.kernel.org/r/20230621130443.7543603b2ee7.Ia8dd54216d341ef1ddc0531f2c9aa30d30536a5d@changeid Signed-off-by: Johannes Berg <[email protected]>
2023-06-21wifi: iwlwifi: Add support for new CNVi (SC)Mukesh Sisodiya3-1/+137
Add support for the new Integrated Connectivity (CNVi) and Companion RF (CRF) versions and their combinations to handle new devices. Signed-off-by: Mukesh Sisodiya <[email protected]> Signed-off-by: Gregory Greenman <[email protected]> Link: https://lore.kernel.org/r/20230621130443.716fd707e847.I34f6ffd61e3210c926868a3e961b16d1742bba29@changeid Signed-off-by: Johannes Berg <[email protected]>
2023-06-21wifi: iwlwifi: Add support for new Bz versionMukesh Sisodiya3-6/+75
Add support for the new version of the Bz CNVI device. Signed-off-by: Mukesh Sisodiya <[email protected]> Signed-off-by: Johannes Berg <[email protected]> Signed-off-by: Gregory Greenman <[email protected]> Link: https://lore.kernel.org/r/20230621130443.82d436d5f346.I0154c202c5d895cb002a2b7c827b9536e81a84b5@changeid Signed-off-by: Johannes Berg <[email protected]>
2023-06-21wifi: iwlwifi: mvm: Add support for scan version 17Ilan Peer2-42/+99
Add support for scan request command version 17, which supports specifying the maximal EIRP PSD value that can be used for probe request transmission on a given channel. Signed-off-by: Ilan Peer <[email protected]> Signed-off-by: Gregory Greenman <[email protected]> Link: https://lore.kernel.org/r/20230620125813.0a41c847d450.I0c9b45cc3eb39d44c75d3bdca84f0a91fdad1fa1@changeid Signed-off-by: Johannes Berg <[email protected]>
2023-06-21wifi: iwlwifi: mvm: adjust skip-over-dtim in D3Johannes Berg1-4/+3
The current formula can skip both too much and not enough time, given the +1 (where the comment about firmware is wrong). Adjust the formula accordingly. Signed-off-by: Johannes Berg <[email protected]> Signed-off-by: Gregory Greenman <[email protected]> Link: https://lore.kernel.org/r/20230620125813.32406b6828ae.I88c315b85f7c56ac6109f84580b95a3dd104ff6c@changeid Signed-off-by: Johannes Berg <[email protected]>
2023-06-21wifi: iwlwifi: mvm: check only affected linksJohannes Berg1-1/+2
When hostapd starts up, it may start up with only one link while the other is still scanning for overlapping BSSes. A station might start to connect at this point, but we run into this warning instead. Since there's no need to check for _all_ links, restrict the check to just the affected links that the STA will be using. Fixes: 57974a55d995 ("wifi: iwlwifi: mvm: refactor iwl_mvm_mac_sta_state_common()") Reported-by: Miri Korenblit <[email protected]> Signed-off-by: Johannes Berg <[email protected]> Signed-off-by: Gregory Greenman <[email protected]> Link: https://lore.kernel.org/r/20230620125813.c3d5a006ec21.Ib4715381f598f4c18d67cd9598ebd5cdbe7d2b09@changeid Signed-off-by: Johannes Berg <[email protected]>
2023-06-21wifi: iwlwifi: mvm: avoid baid size integer overflowJohannes Berg1-1/+1
Roee reported various hard-to-debug crashes with pings in EHT aggregation scenarios. Enabling KASAN showed that we access the BAID allocation out of bounds, and looking at the code a bit shows that since the reorder buffer entry (struct iwl_mvm_reorder_buf_entry) is 128 bytes if debug such as lockdep is enabled, then staring from an agg size 512 we overflow the size calculation, and allocate a much smaller structure than we should, causing slab corruption once we initialize this. Fix this by simply using u32 instead of u16. Reported-by: Roee Goldfiner <[email protected]> Signed-off-by: Johannes Berg <[email protected]> Signed-off-by: Gregory Greenman <[email protected]> Link: https://lore.kernel.org/r/20230620125813.f428c856030d.I2c2bb808e945adb71bc15f5b2bac2d8957ea90eb@changeid Signed-off-by: Johannes Berg <[email protected]>
2023-06-21wifi: iwlwifi: mvm: indicate HW decrypt for beacon protectionJohannes Berg1-3/+6
We've already done the 'decryption' here, so tell mac80211 it need not do it again. Fixes: b1fdc2505abc ("iwlwifi: mvm: advertise BIGTK client support if available") Signed-off-by: Johannes Berg <[email protected]> Signed-off-by: Gregory Greenman <[email protected]> Link: https://lore.kernel.org/r/20230620125813.a50cf68fbf2e.Ieceacbe3789d81ea02ae085ad8d1f8813a33c31b@changeid Signed-off-by: Johannes Berg <[email protected]>
2023-06-21wifi: iwlwifi: mvm: support new flush_sta methodJohannes Berg3-0/+28
For iwlwifi this is simple to implement, and on newer hardware it's an improvement since we have per-station queues. Signed-off-by: Johannes Berg <[email protected]> Signed-off-by: Gregory Greenman <[email protected]> Link: https://lore.kernel.org/r/20230620125813.a1f8ec20b727.I48594b708b41aa55dc2b8c3d346b4412ad3a5ba3@changeid Signed-off-by: Johannes Berg <[email protected]>
2023-06-21wifi: iwlwifi: Add support for new PCI IdMukesh Sisodiya1-0/+2
Add support for the PCI Id 51F1 without IMR support. Signed-off-by: Mukesh Sisodiya <[email protected]> Signed-off-by: Gregory Greenman <[email protected]> Link: https://lore.kernel.org/r/20230620125813.9800e652e789.Ic06a085832ac3f988c8ef07d856c8e281563295d@changeid Signed-off-by: Johannes Berg <[email protected]>
2023-06-21wifi: iwlwifi: mvm: send LARI configuration earlierBenjamin Berg1-1/+2
Sending the LARI configuration may trigger calibration, which can have undesired side effects. Move the command to be send earlier (before the phy contexts are registered) to avoid unintended side effects. Signed-off-by: Benjamin Berg <[email protected]> Signed-off-by: Gregory Greenman <[email protected]> Link: https://lore.kernel.org/r/20230620125813.80742497eb3f.I3e599a796290082e6d331ea495a5591d55de4726@changeid Signed-off-by: Johannes Berg <[email protected]>
2023-06-21wifi: iwlwifi: remove disable_dummy_notificationJohannes Berg2-5/+1
This struct member is read-only, so can never change away from the default value of zero. Remove the code that's in an if on the value, since it's effectively dead code. Signed-off-by: Johannes Berg <[email protected]> Signed-off-by: Gregory Greenman <[email protected]> Link: https://lore.kernel.org/r/20230620125813.e7c96d0aa805.I5b158ce15e48393d2896c0bff9f644d983f0e92d@changeid Signed-off-by: Johannes Berg <[email protected]>
2023-06-21wifi: iwlwifi: limit EHT capabilities based on PCIe link speedJohannes Berg3-4/+24
If a discrete NIC is connected to a PCIe link hat isn't at least Gen3 (8.0 GT/s), then we cannot sustain 320 MHz traffic, so remove that from EHT capabilities in that case. While at it, also move setting 320 MHz beamformee to the right place in the code so it's not set while not supporting 320 MHz. Signed-off-by: Johannes Berg <[email protected]> Signed-off-by: Gregory Greenman <[email protected]> Link: https://lore.kernel.org/r/20230620125813.b77a1574a0a7.Id4120c161fb7df6dedc70d5f3e3829e9117b8cb1@changeid Signed-off-by: Johannes Berg <[email protected]>
2023-06-21wifi: iwlwifi: mvm: add EHT A-MPDU size exponent supportJohannes Berg1-9/+18
Add support for reading the EHT MAC capabilities A-MPDU size exponent field, as indicated by the draft spec. Also clarify the existing code a bit and add comments so it's clearer to understand what's going on here. Signed-off-by: Johannes Berg <[email protected]> Signed-off-by: Gregory Greenman <[email protected]> Link: https://lore.kernel.org/r/20230620125813.c5e00045d90f.I7520787fca8f8430a564adedf975d069ad8c5417@changeid Signed-off-by: Johannes Berg <[email protected]>
2023-06-21wifi: iwlwifi: mvm: use min_t() for agg_sizeJohannes Berg1-3/+2
We can use min_t() for the agg_size and avoid spelling out the (firmware) limit twice. Signed-off-by: Johannes Berg <[email protected]> Signed-off-by: Gregory Greenman <[email protected]> Link: https://lore.kernel.org/r/20230620125813.211768036c1f.I78b7eea32eaae20cc9f32869aa3f42814634ce9a@changeid Signed-off-by: Johannes Berg <[email protected]>
2023-06-21wifi: iwlwifi: mvm: use EHT maximum MPDU length on 2.4 GHzJohannes Berg1-2/+13
On 2.4 GHz there's no VHT, so EHT defines its own bits for the maximum MPDU length. Use them when telling firmware about the maximum. Signed-off-by: Johannes Berg <[email protected]> Signed-off-by: Gregory Greenman <[email protected]> Link: https://lore.kernel.org/r/20230620125813.fd5322bb48a4.Ic471045f83229ceaacce25edcf992d3ce2c75de5@changeid Signed-off-by: Johannes Berg <[email protected]>
2023-06-21wifi: iwlwifi: nvm: handle EHT/320 MHz regulatory flagJohannes Berg1-0/+7
Handle the regulatory EHT/320 MHz flags from firmware just like any other flags before it. Signed-off-by: Johannes Berg <[email protected]> Signed-off-by: Gregory Greenman <[email protected]> Link: https://lore.kernel.org/r/20230620125813.2c5e886c08f3.Ibc5c27d973d0590e2dea1f50435f9cf3ba8c2c09@changeid Signed-off-by: Johannes Berg <[email protected]>
2023-06-21wifi: iwlwifi: mvm: make iwl_mvm_set_fw_mu_edca_params mld awareEmmanuel Grumbach3-7/+7
We need to work on the right link there. Signed-off-by: Emmanuel Grumbach <[email protected]> Signed-off-by: Gregory Greenman <[email protected]> Link: https://lore.kernel.org/r/20230620125813.8762a90e8857.Ic5b8e96140a449fd1ed7008907d67fc36fe98506@changeid Signed-off-by: Johannes Berg <[email protected]>
2023-06-21wifi: iwlwifi: fw: send marker cmd before suspend cmdMiri Korenblit4-28/+58
This is needed to sync the times in the FW and driver logs Signed-off-by: Miri Korenblit <[email protected]> Signed-off-by: Gregory Greenman <[email protected]> Link: https://lore.kernel.org/r/20230620125813.9c84322c41b5.Id13816b3ece103f88514a7523b22bb2b9dcc8ab7@changeid Signed-off-by: Johannes Berg <[email protected]>
2023-06-21wifi: iwlwifi: mvm: check the right csa_activeEmmanuel Grumbach1-1/+4
When the firmware says that the channel switch is happening, we check that we know about that switch by checking the csa_active bit. Until now, we checked the bss_conf from the vif instead of taking the bss_conf of the link. Fix that. Signed-off-by: Emmanuel Grumbach <[email protected]> Signed-off-by: Gregory Greenman <[email protected]> Link: https://lore.kernel.org/r/20230620125813.63f835a4f578.I0bb2a231e4da506b7c751dc23a428558f9ecfa75@changeid Signed-off-by: Johannes Berg <[email protected]>
2023-06-21wifi: iwlwifi: pcie: add size assertionsJohannes Berg2-1/+11
Ensure that the TX command scratch fits into the buffer provided by the first TB. It does, of course, but add some build-time validations in case we touch this code. Signed-off-by: Johannes Berg <[email protected]> Signed-off-by: Gregory Greenman <[email protected]> Link: https://lore.kernel.org/r/20230620125813.8f54f2990b92.If19a038dfd633d4601e3d44dd0ff678bc0a851e9@changeid Signed-off-by: Johannes Berg <[email protected]>
2023-06-21wifi: iwlwifi: pcie: refactor RB status size calculationJohannes Berg1-10/+14
We have three places doing this check, and even in slightly different ways (with/without an intermediate). Refactor that to a new small inline function. Signed-off-by: Johannes Berg <[email protected]> Signed-off-by: Gregory Greenman <[email protected]> Link: https://lore.kernel.org/r/20230620125813.f3e87ddd5bce.Ifefba753043b68c394590a35bc6914a0f6497fd3@changeid Signed-off-by: Johannes Berg <[email protected]>
2023-06-21wifi: p54: Add missing MODULE_FIRMWARE macroJuerg Haefliger1-0/+1
Add the missing MODULE_FIRMWARE macro for "3826.eeprom". Signed-off-by: Juerg Haefliger <[email protected]> Acked-by: Christian Lamparter <[email protected]> Signed-off-by: Kalle Valo <[email protected]> Link: https://lore.kernel.org/r/[email protected]
2023-06-21wifi: rtw89: use struct to parse firmware headerPing-Ke Shih3-80/+87
A firmware contains basic header, sections and optional dynamic header. Define them by a struct, so it will be easier to understand the layout, and also simply access these elements. Signed-off-by: Ping-Ke Shih <[email protected]> Signed-off-by: Kalle Valo <[email protected]> Link: https://lore.kernel.org/r/[email protected]
2023-06-21wifi: rtw89: TX power stuffs replace confusing naming of _max with _numZong-Zhe Yang8-43/+43
Some old declarations about TX power stuffs were named with confusing `_max`. But, they mean "the number of". So we change them to be named with `_num`. (No logic is changed.) Signed-off-by: Zong-Zhe Yang <[email protected]> Signed-off-by: Ping-Ke Shih <[email protected]> Signed-off-by: Kalle Valo <[email protected]> Link: https://lore.kernel.org/r/[email protected]
2023-06-21wifi: rtw89: 8851b: configure to force 1 TX power valuePing-Ke Shih2-0/+25
RTL8851B is a chip with only single RF path, and it must use 1 TX power value for transmission, so force 1 TX power value to prevent hardware logic gets wrong TX power values randomly in certain samples. Signed-off-by: Ping-Ke Shih <[email protected]> Signed-off-by: Kalle Valo <[email protected]> Link: https://lore.kernel.org/r/[email protected]
2023-06-21wifi: rtw89: 8851b: rfk: update IQK to version 0x8Ping-Ke Shih2-54/+63
The main change is to adjust RX calibration groups from {0,1,2,3} to {0,2} in 5 GHz, so reduce elements from 4 to 2, and use index to iterate them. Meanwhile, always do RX narrowband calibration (ID_NBRXK) for each group. NCTL is used to assist IQK, so also update NCTL to 0x6 along with internal tag HALRF_029_00_103. Signed-off-by: Ping-Ke Shih <[email protected]> Signed-off-by: Kalle Valo <[email protected]> Link: https://lore.kernel.org/r/[email protected]
2023-06-21wifi: rtw89: 8851b: rfk: add LCK trackPing-Ke Shih4-1/+69
LCK is short for LC Tank calibration. To keep RF performance, do this calibration if difference of thermal value is over a threshold. Signed-off-by: Ping-Ke Shih <[email protected]> Signed-off-by: Kalle Valo <[email protected]> Link: https://lore.kernel.org/r/[email protected]
2023-06-21wifi: rtw89: 8851b: update TX power tables to R28Zong-Zhe Yang1-110/+110
Update 8851B TX power tables to RF version R28. TX power tables' changes: * TX power limit and TX power shape: update 5 GHz configurations for FCC and IC Signed-off-by: Zong-Zhe Yang <[email protected]> Signed-off-by: Ping-Ke Shih <[email protected]> Signed-off-by: Kalle Valo <[email protected]> Link: https://lore.kernel.org/r/[email protected]
2023-06-21wifi: rtw89: 8851b: update RF radio A parameters to R28Ping-Ke Shih1-6/+22
Update 8851b radio A parameters to R28 along with internal HALRF_029_00_103 Signed-off-by: Ping-Ke Shih <[email protected]> Signed-off-by: Kalle Valo <[email protected]> Link: https://lore.kernel.org/r/[email protected]
2023-06-21wifi: rtw88: fix not entering PS mode after AP stopsPo-Hao Huang5-0/+35
Without this patch, firmware only track beacons for port 0 and since we will always start AP on port 0, this results in misbehavior of power saving mode on other ports after AP stops. The "default port" H2C command is used to notify which port should firmware track. Update the correct settings to firmware so power saving mode can work properly. Signed-off-by: Po-Hao Huang <[email protected]> Signed-off-by: Ping-Ke Shih <[email protected]> Signed-off-by: Kalle Valo <[email protected]> Link: https://lore.kernel.org/r/[email protected]
2023-06-21wifi: rtw88: refine register based H2C commandPo-Hao Huang2-0/+56
Since register based H2C commands don't need endian conversion. Introduce a new API that don't do conversion and send it directly. New caller are expected to encode with cpu order and gradually replace the old ones. Signed-off-by: Po-Hao Huang <[email protected]> Signed-off-by: Ping-Ke Shih <[email protected]> Signed-off-by: Kalle Valo <[email protected]> Link: https://lore.kernel.org/r/[email protected]
2023-06-21wifi: rtw88: Stop high queue during scanPo-Hao Huang2-2/+6
When traversing channel list, TX in high queue should be disabled along with beacon function, so packets won't be sent to incorrect channels. Signed-off-by: Po-Hao Huang <[email protected]> Signed-off-by: Ping-Ke Shih <[email protected]> Signed-off-by: Kalle Valo <[email protected]> Link: https://lore.kernel.org/r/[email protected]
2023-06-21wifi: rtw88: Skip high queue in hci_flushPo-Hao Huang1-2/+3
The flush period may not always intersect with DTIM and when that happens, an error log "timed out to flush pci TX ring[6]" is shown. Bypass this since hardware will do proper transmission on the next DTIM period for broadcast/multicast packets in high queue. Signed-off-by: Po-Hao Huang <[email protected]> Signed-off-by: Ping-Ke Shih <[email protected]> Signed-off-by: Kalle Valo <[email protected]> Link: https://lore.kernel.org/r/[email protected]
2023-06-21wifi: rtw88: Fix AP mode incorrect DTIM behaviorPo-Hao Huang5-1/+13
Broadcast and multicast packets in high queue should be transmitted all at once during DTIM. But without proper settings, hardware fails to recognize that there are multiple packets and fetches only one. Fix this by signaling hardware with more data bit set when there are packets in the high queue. Signed-off-by: Po-Hao Huang <[email protected]> Signed-off-by: Ping-Ke Shih <[email protected]> Signed-off-by: Kalle Valo <[email protected]> Link: https://lore.kernel.org/r/[email protected]
2023-06-21wifi: rtw88: use struct instead of macros to set TX descPo-Hao Huang4-113/+106
Remove macros that set TX descriptors. Use struct and le32_encode_bits() with mask definitions. Signed-off-by: Po-Hao Huang <[email protected]> Signed-off-by: Ping-Ke Shih <[email protected]> Signed-off-by: Kalle Valo <[email protected]> Link: https://lore.kernel.org/r/[email protected]
2023-06-21wifi: rtw88: process VO packets without workqueue to avoid PTK rekey failedChih-Kang Chang3-3/+14
In the wpa_supplicant rekey flow, it sends an EAPOL packet 4/4 through nl80211_tx_control_port() and triggers wake_tx_queue() in the driver. Then, it sends nl80211_new_key() to configure a new key in mac80211. However, in wake_tx_queue(), a workqueue is used to process the tx packet, which might cause the driver to process the EAPOL packet later than nl80211_new_key(). As a result, the EAPOL 4/4 packet is dropped by mac80211 due to the rekey configuration being finished. The EAPOL packets belongs to VO packets that need high priority. Therefore, we process VO packets directly without workqueue to ensure that packets can process immediately. VO is normally used by voice application that is low traffic load and low latency, that doesn't affect user experience. We test iperf with VO packets(iperf3 -P4 -u -b 10000M -S 0xdf) before after TX throughput 162M 162M ping RTT 3.8ms 3.7ms Signed-off-by: Chih-Kang Chang <[email protected]> Signed-off-by: Ping-Ke Shih <[email protected]> Signed-off-by: Kalle Valo <[email protected]> Link: https://lore.kernel.org/r/[email protected]
2023-06-21wifi: rtw88: Fix action frame transmission fail before associationPo-Hao Huang1-2/+1
For combo chips, antennas were controlled by bluetooth only during power on. If WiFi wish to do transmission, notification to the coexistence module are required. Previously we only do this before authentication. To allow transmission before auth, such as management TX, now we start the initiation of coexistence earlier so antennas are shared between WiFi and bluetooth after set_channel(), and frames could then be sent. Signed-off-by: Po-Hao Huang <[email protected]> Signed-off-by: Ping-Ke Shih <[email protected]> Signed-off-by: Kalle Valo <[email protected]> Link: https://lore.kernel.org/r/[email protected]
2023-06-21Revert "virtio-blk: support completion batching for the IRQ path"Michael S. Tsirkin1-45/+37
This reverts commit 07b679f70d73483930e8d3c293942416d9cd5c13. This change appears to have broken things... We now see applications hanging during disk accesses. e.g. multi-port virtio-blk device running in h/w (FPGA) Host running a simple 'fio' test. [global] thread=1 direct=1 ioengine=libaio norandommap=1 group_reporting=1 bs=4K rw=read iodepth=128 runtime=1 numjobs=4 time_based [job0] filename=/dev/vda [job1] filename=/dev/vdb [job2] filename=/dev/vdc ... [job15] filename=/dev/vdp i.e. 16 disks; 4 queues per disk; simple burst of 4KB reads This is repeatedly run in a loop. After a few, normally <10 seconds, fio hangs. With 64 queues (16 disks), failure occurs within a few seconds; with 8 queues (2 disks) it may take ~hour before hanging. Last message: fio-3.19 Starting 8 threads Jobs: 1 (f=1): [_(7),R(1)][68.3%][eta 03h:11m:06s] I think this means at the end of the run 1 queue was left incomplete. 'diskstats' (run while fio is hung) shows no outstanding transactions. e.g. $ cat /proc/diskstats ... 252 0 vda 1843140071 0 14745120568 712568645 0 0 0 0 0 3117947 712568645 0 0 0 0 0 0 252 16 vdb 1816291511 0 14530332088 704905623 0 0 0 0 0 3117711 704905623 0 0 0 0 0 0 ... Other stats (in the h/w, and added to the virtio-blk driver ([a]virtio_queue_rq(), [b]virtblk_handle_req(), [c]virtblk_request_done()) all agree, and show every request had a completion, and that virtblk_request_done() never gets called. e.g. PF= 0 vq=0 1 2 3 [a]request_count - 839416590 813148916 105586179 84988123 [b]completion1_count - 839416590 813148916 105586179 84988123 [c]completion2_count - 0 0 0 0 PF= 1 vq=0 1 2 3 [a]request_count - 823335887 812516140 104582672 75856549 [b]completion1_count - 823335887 812516140 104582672 75856549 [c]completion2_count - 0 0 0 0 i.e. the issue is after the virtio-blk driver. This change was introduced in kernel 6.3.0. I am seeing this using 6.3.3. If I run with an earlier kernel (5.15), it does not occur. If I make a simple patch to the 6.3.3 virtio-blk driver, to skip the blk_mq_add_to_batch()call, it does not fail. e.g. kernel 5.15 - this is OK virtio_blk.c,virtblk_done() [irq handler] if (likely(!blk_should_fake_timeout(req->q))) { blk_mq_complete_request(req); } kernel 6.3.3 - this fails virtio_blk.c,virtblk_handle_req() [irq handler] if (likely(!blk_should_fake_timeout(req->q))) { if (!blk_mq_complete_request_remote(req)) { if (!blk_mq_add_to_batch(req, iob, virtblk_vbr_status(vbr), virtblk_complete_batch)) { virtblk_request_done(req); //this never gets called... so blk_mq_add_to_batch() must always succeed } } } If I do, kernel 6.3.3 - this is OK virtio_blk.c,virtblk_handle_req() [irq handler] if (likely(!blk_should_fake_timeout(req->q))) { if (!blk_mq_complete_request_remote(req)) { virtblk_request_done(req); //force this here... if (!blk_mq_add_to_batch(req, iob, virtblk_vbr_status(vbr), virtblk_complete_batch)) { virtblk_request_done(req); //this never gets called... so blk_mq_add_to_batch() must always succeed } } } Perhaps you might like to fix/test/revert this change... Martin Reported-by: kernel test robot <[email protected]> Closes: https://lore.kernel.org/oe-kbuild-all/[email protected]/ Cc: Suwan Kim <[email protected]> Tested-by: [email protected] Reported-by: "Roberts, Martin" <[email protected]> Message-Id: <336455b4f630f329380a8f53ee8cad3868764d5c.1686295549.git.mst@redhat.com> Signed-off-by: Michael S. Tsirkin <[email protected]>
2023-06-20net: stmmac: dwmac-qcom-ethqos: add support for emac4 on sa8775p platformsBartosz Golaszewski1-14/+51
sa8775p uses EMAC version 4, add the relevant defines, rename the has_emac3 switch to has_emac_ge_3 (has emac greater-or-equal than 3) and add the new compatible. Signed-off-by: Bartosz Golaszewski <[email protected]> Signed-off-by: Jakub Kicinski <[email protected]>
2023-06-20net: stmmac: add new switch to struct plat_stmmacenet_dataBartosz Golaszewski1-1/+1
On some platforms, the PCS can be integrated in the MAC so the driver will not see any PCS link activity. Add a switch that allows the platform drivers to let the core code know. Signed-off-by: Bartosz Golaszewski <[email protected]> Reviewed-by: Jose Abreu <[email protected]> Signed-off-by: Jakub Kicinski <[email protected]>
2023-06-20net: stmmac: dwmac-qcom-ethqos: add support for SGMIIBartosz Golaszewski1-0/+37
On sa8775p the MAC is connected to the external PHY over SGMII so add support for it to the driver. Signed-off-by: Bartosz Golaszewski <[email protected]> Signed-off-by: Jakub Kicinski <[email protected]>
2023-06-20net: stmmac: dwmac-qcom-ethqos: prepare the driver for more PHY modesBartosz Golaszewski1-5/+26
In preparation for supporting SGMII, let's make the code a bit more generic. Add a new callback for MAC configuration so that we can assign a different variant of it in the future. Signed-off-by: Bartosz Golaszewski <[email protected]> Reviewed-by: Andrew Halaney <[email protected]> Signed-off-by: Jakub Kicinski <[email protected]>
2023-06-20net: stmmac: dwmac-qcom-ethqos: add support for the phyaux clockBartosz Golaszewski1-15/+16
On sa8775p, the EMAC revision is 4 and we use SGMII instead of RGMII. There's no "rgmii" clock but there's a fourth clock under a different name: "phyaux". Add a new field to the chip data struct that specifies the link clock name. Default to "rgmii" for backward compatibility. Signed-off-by: Bartosz Golaszewski <[email protected]> Signed-off-by: Jakub Kicinski <[email protected]>
2023-06-20net: stmmac: dwmac-qcom-ethqos: add support for the optional serdes phyBartosz Golaszewski1-0/+37
On sa8775p platforms, there's a SGMII SerDes PHY between the MAC and external PHY that we need to enable and configure. Signed-off-by: Bartosz Golaszewski <[email protected]> Signed-off-by: Jakub Kicinski <[email protected]>
2023-06-20net: stmmac: dwmac-qcom-ethqos: remove stray spaceBartosz Golaszewski1-1/+1
There's an unnecessary space in the rgmii_updatel() function, remove it. Signed-off-by: Bartosz Golaszewski <[email protected]> Reviewed-by: Andrew Halaney <[email protected]> Signed-off-by: Jakub Kicinski <[email protected]>
2023-06-20net: stmmac: dwmac-qcom-ethqos: add a newline between headersBartosz Golaszewski1-0/+1
Typically we use a newline between global and local headers so add it here as well. Signed-off-by: Bartosz Golaszewski <[email protected]> Reviewed-by: Andrew Halaney <[email protected]> Signed-off-by: Jakub Kicinski <[email protected]>
Powered by cgit, Gruvboxed by Blaster4385.