summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2020-01-16net/mlx5: Add ignore level support fwd to table rulesPaul Blakey
If user sets ignore flow level flag on a rule, that rule can point to a flow table of any level, including those with levels equal or less than the level of the flow table it is added on. This with unamanged tables will be used to create a FDB chain/prio hierarchy much larger than currently supported level range. Signed-off-by: Paul Blakey <paulb@mellanox.com> Reviewed-by: Mark Bloch <markb@mellanox.com> Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
2020-01-16net/mlx5: fs_core: Introduce unmanaged flow tablesPaul Blakey
Currently, Most of the steering tree is statically declared ahead of time, with steering prios instances allocated for each fdb chain to assign max number of levels for each of them. This allows fs_core to manage the connections and levels of the flow tables hierarcy to prevent loops, but restricts us with the number of supported chains and priorities. Introduce unmananged flow tables, allowing the user to manage the flow table connections. A unamanged table is detached from the fs_core flow table hierarcy, and is only connected back to the hierarchy by explicit FTEs forward actions. This will be used together with firmware that supports ignoring the flow table levels to increase the number of supported chains and prios. Signed-off-by: Paul Blakey <paulb@mellanox.com> Reviewed-by: Roi Dayan <roid@mellanox.com> Reviewed-by: Oz Shlomo <ozsh@mellanox.com> Reviewed-by: Mark Bloch <markb@mellanox.com> Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
2020-01-16Merge branch 'mlx5-next' of ↵Saeed Mahameed
git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux This merge syncs with mlx5-next latest HW bits and layout updates for next features, in addition one patch that improves mlx5_create_auto_grouped_flow_table() API across all mlx5 users. * 'mlx5-next' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux: net/mlx5: Refactor mlx5_create_auto_grouped_flow_table net/mlx5e: Add discard counters per priority net/mlx5e: Expose FEC feilds and related capability bit net/mlx5: Add mlx5_ifc definitions for connection tracking support net/mlx5: Add copy header action struct layout net/mlx5: Expose resource dump register mapping net/mlx5: Add structures and defines for MIRC register net/mlx5: Read MCAM register groups 1 and 2 net/mlx5: Add structures layout for new MCAM access reg groups net/mlx5: Expose vDPA emulation device capabilities net/mlx5: Add Virtio Emulation related device capabilities Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
2020-01-16net/mlx5: Refactor mlx5_create_auto_grouped_flow_tablePaul Blakey
Refactor mlx5_create_auto_grouped_flow_table() to use ft_attr param which already carries the max_fte, prio and flags memebers, and is used the same in similar mlx5_create_flow_table() function. Signed-off-by: Paul Blakey <paulb@mellanox.com> Reviewed-by: Roi Dayan <roid@mellanox.com> Reviewed-by: Oz Shlomo <ozsh@mellanox.com> Reviewed-by: Mark Bloch <markb@mellanox.com> Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
2020-01-17software node: introduce CONFIG_KUNIT_DRIVER_PE_TESTAlan Maguire
Currently the property entry kunit tests are built if CONFIG_KUNIT=y. This will cause warnings when merged with the kunit tree that now supports tristate CONFIG_KUNIT. While the tests appear to compile as a module, we get a warning about missing module license. It's better to have a per-test suite CONFIG variable so that we can do selective building of kunit-based suites, and can also avoid merge issues like this. Fixes: c032ace71c29 ("software node: add basic tests for property entries") Reported-by: Stephen Rothwell <sfr@canb.auug.org.au> Reported-by: Randy Dunlap <rdunlap@infradead.org> Signed-off-by: Alan Maguire <alan.maguire@oracle.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2020-01-17Merge back cpuidle material for v5.6.Rafael J. Wysocki
2020-01-16PM: hibernate: fix crashes with init_on_free=1Alexander Potapenko
Upon resuming from hibernation, free pages may contain stale data from the kernel that initiated the resume. This breaks the invariant inflicted by init_on_free=1 that freed pages must be zeroed. To deal with this problem, make clear_free_pages() also clear the free pages when init_on_free is enabled. Fixes: 6471384af2a6 ("mm: security: introduce init_on_alloc=1 and init_on_free=1 boot options") Reported-by: Johannes Stezenbach <js@sig21.net> Signed-off-by: Alexander Potapenko <glider@google.com> Cc: 5.3+ <stable@vger.kernel.org> # 5.3+ Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2020-01-16net/mlx5e: Add discard counters per priorityAharon Landau
Add counters that count (per priority) the number of received packets that dropped due to lack of buffers on a physical port. If this counter is increasing, it implies that the adapter is congested and cannot absorb the traffic coming from the network. Signed-off-by: Aharon Landau <aharonl@mellanox.com> Reviewed-by: Tariq Toukan <tariqt@mellanox.com> Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
2020-01-16net/mlx5e: Expose FEC feilds and related capability bitAya Levin
Introduce 50G per lane FEC modes capability bit and newly supported fields in PPLM register which allow this configuration. Signed-off-by: Aya Levin <ayal@mellanox.com> Reviewed-by: Eran Ben Elisha <eranbe@mellanox.com> Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
2020-01-16net/mlx5: Add mlx5_ifc definitions for connection tracking supportPaul Blakey
Add the required hardware definitions to mlx5_ifc: ignore_flow_level, registers, copy_header, and fwd_and_modify cap. Signed-off-by: Paul Blakey <paulb@mellanox.com> Reviewed-by: Roi Dayan <roid@mellanox.com> Reviewed-by: Oz Sholomo <ozsh@mellanox.com> Reviewed-by: Mark Bloch <markb@mellanox.com> Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
2020-01-16net/mlx5: Add copy header action struct layoutHamdan Igbaria
Add definition for copy header action, copy action is used to copy header fields from source to destination. Signed-off-by: Hamdan Igbaria <hamdani@mellanox.com> Signed-off-by: Alex Vesker <valex@mellanox.com> Reviewed-by: Alex Vesker <valex@mellanox.com> Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
2020-01-16net/mlx5: Expose resource dump register mappingAya Levin
Add new register enumeration for resource dump. Add layout mapping for resource dump: access command and response. Signed-off-by: Aya Levin <ayal@mellanox.com> Reviewed-by: Moshe Shemesh <moshe@mellanox.com> Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
2020-01-16net/mlx5: Add structures and defines for MIRC registerEran Ben Elisha
Add needed structures, layouts and defines for MIRC (Management Image Re-activation Control) register. This structure will be used for the FSM reactivation flow in the downstream patches. Signed-off-by: Eran Ben Elisha <eranbe@mellanox.com> Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
2020-01-16net/mlx5: Read MCAM register groups 1 and 2Eran Ben Elisha
On load, Driver caches MCAM (Management Capabilities Mask Register) registers. in addition to the only MCAM register group (0) the driver already reads, here we add support for reading groups 1 and 2. Signed-off-by: Eran Ben Elisha <eranbe@mellanox.com> Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
2020-01-16net/mlx5: Add structures layout for new MCAM access reg groupsEran Ben Elisha
MCAM has 3 access_reg_groups (0-2). Defines data structures in order to read and parse access_reg_groups #1 and #2. Signed-off-by: Eran Ben Elisha <eranbe@mellanox.com> Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
2020-01-16selftests/bpf: Fix test_progs send_signal flakiness with nmi modeYonghong Song
Alexei observed that test_progs send_signal may fail if run with command line "./test_progs" and the tests will pass if just run "./test_progs -n 40". I observed similar issue with nmi subtest failure and added a delay 100 us in Commit ab8b7f0cb358 ("tools/bpf: Add self tests for bpf_send_signal_thread()") and the problem is gone for me. But the issue still exists in Alexei's testing environment. The current code uses sample_freq = 50 (50 events/second), which may not be enough. But if the sample_freq value is larger than sysctl kernel/perf_event_max_sample_rate, the perf_event_open syscall will fail. This patch changed nmi perf testing to use sample_period = 1, which means trying to sampling every event. This seems fixing the issue. Fixes: ab8b7f0cb358 ("tools/bpf: Add self tests for bpf_send_signal_thread()") Signed-off-by: Yonghong Song <yhs@fb.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Andrii Nakryiko <andriin@fb.com> Link: https://lore.kernel.org/bpf/20200116174004.1522812-1-yhs@fb.com
2020-01-16Merge tag 'timers-v5.5-rc6' of ↵Thomas Gleixner
https://git.linaro.org/people/daniel.lezcano/linux into timers/core Pull clocksource/events driver updates from Daniel Lezcano - Add suspend/resume for Hyper-V clocksource (Dexuan Cui) - Fix Kconfig indentation (Krzysztof Kozlowski) - Add SoC specific bindings for RZ/G2N (r8a774b1) (Biju Das) - Add Microchip PIT64B support (Claudiu Beznea) - Convert the cadence ttc driver to a platform driver to initialize later (Rajan Vaja) - Fix a memory leak when the initialization fails on bcm2835 (Colin Ian King) - Use the combo devm_platform_ioremap_resource() function for em-sti, ti-dm and switch to platform_get_irq (Yangtao Li) - Fix Exynos naming in the driver (Krzysztof Kozlowski) - Fix an uninitialized pointer access in ti-dm timer (Tony Lindgren) - Fix a sparse warning in microchip-pit64b (Claudiu Beznea) - Code reorg without functional changes for Hyper-V clocksource (Andrea Parri) - Decrease the Hyper-V clocksource rating in order to let the stable TSC to be selected instead (Andrea Parri)
2020-01-16selinux: do not allocate ancillary buffer on first loadOndrej Mosnacek
In security_load_policy(), we can defer allocating the newpolicydb ancillary array to after checking state->initialized, thereby avoiding the pointless allocation when loading policy the first time. Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> [PM: merged portions by hand] Reviewed-by: Stephen Smalley <sds@tycho.nsa.gov> Signed-off-by: Paul Moore <paul@paul-moore.com>
2020-01-16EDAC/mce_amd: Make fam_ops static globalBorislav Petkov
... and do not kmalloc a three-pointer struct. Which simplifies mce_amd_init() a bit. No functional changes. Signed-off-by: Borislav Petkov <bp@suse.de> Link: https://lkml.kernel.org/r/20200116163403.GF27148@zn.tnic
2020-01-16PM: suspend: Add sysfs attribute to control the "sync on suspend" behaviorJonas Meurer
The sysfs attribute `/sys/power/sync_on_suspend` controls, whether or not filesystems are synced by the kernel before system suspend. Congruously, the behaviour of build-time switch CONFIG_SUSPEND_SKIP_SYNC is slightly changed: It now defines the run-tim default for the new sysfs attribute `/sys/power/sync_on_suspend`. The run-time attribute is added because the existing corresponding build-time Kconfig flag for (`CONFIG_SUSPEND_SKIP_SYNC`) is not flexible enough. E.g. Linux distributions that provide pre-compiled kernels usually want to stick with the default (sync filesystems before suspend) but under special conditions this needs to be changed. One example for such a special condition is user-space handling of suspending block devices (e.g. using `cryptsetup luksSuspend` or `dmsetup suspend`) before system suspend. The Kernel trying to sync filesystems after the underlying block device already got suspended obviously leads to dead-locks. Be aware that you have to take care of the filesystem sync yourself before suspending the system in those scenarios. Signed-off-by: Jonas Meurer <jonas@freesources.org> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2020-01-16lib/vdso: Only read hrtimer_res when needed in __cvdso_clock_getres()Christophe Leroy
Only perform READ_ONCE(vd[CS_HRES_COARSE].hrtimer_res) for HRES and RAW clocks. Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Link: https://lore.kernel.org/r/7ac2f0d21652f95e2bbdfa6bd514ae6c7caf53ab.1579196675.git.christophe.leroy@c-s.fr
2020-01-16MIPS: vdso: Define BUILD_VDSO32 when building a 32bit kernelThomas Gleixner
The confinement of the 32bit specific VDSO functions missed to define BUILD_VDSO32 when building a 32bit MIPS kernel: arch/mips/vdso/vgettimeofday.c: In function __vdso_clock_gettime: arch/mips/vdso/vgettimeofday.c:17:9: error: implicit declaration of function __cvdso_clock_gettime32 arch/mips/vdso/vgettimeofday.c: In function __vdso_clock_getres: arch/mips/vdso/vgettimeofday.c:39:9: error: implicit declaration of function __cvdso_clock_getres_time32 Force the define for 32bit builds in the VDSO Makefile. Fixes: bf279849ad59 ("lib/vdso: Build 32 bit specific functions in the right context") Reported-by: kbuild test robot <lkp@intel.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: Paul Burton <paulburton@kernel.org> Link: https://lore.kernel.org/r/87d0bjfaqa.fsf@nanos.tec.linutronix.de
2020-01-16kselftest: Minimise dependency of get_size on C library interfacesSiddhesh Poyarekar
It was observed[1] on arm64 that __builtin_strlen led to an infinite loop in the get_size selftest. This is because __builtin_strlen (and other builtins) may sometimes result in a call to the C library function. The C library implementation of strlen uses an IFUNC resolver to load the most efficient strlen implementation for the underlying machine and hence has a PLT indirection even for static binaries. Because this binary avoids the C library startup routines, the PLT initialization never happens and hence the program gets stuck in an infinite loop. On x86_64 the __builtin_strlen just happens to expand inline and avoid the call but that is not always guaranteed. Further, while testing on x86_64 (Fedora 31), it was observed that the test also failed with a segfault inside write() because the generated code for the write function in glibc seems to access TLS before the syscall (probably due to the cancellation point check) and fails because TLS is not initialised. To mitigate these problems, this patch reduces the interface with the C library to just the syscall function. The syscall function still sets errno on failure, which is undesirable but for now it only affects cases where syscalls fail. [1] https://bugs.linaro.org/show_bug.cgi?id=5479 Signed-off-by: Siddhesh Poyarekar <siddhesh@gotplt.org> Reported-by: Masami Hiramatsu <masami.hiramatsu@linaro.org> Tested-by: Masami Hiramatsu <masami.hiramatsu@linaro.org> Reviewed-by: Tim Bird <tim.bird@sony.com> Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
2020-01-16selftests/livepatch: Remove unused local variable in set_ftrace_enabled()Miroslav Benes
set_ftrace_enabled() contains unused local variable "sysctl". Remove it. Signed-off-by: Miroslav Benes <mbenes@suse.cz> Reviewed-by: Kamalesh Babulal <kamalesh@linux.vnet.ibm.com> Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
2020-01-16selftests/livepatch: Replace set_dynamic_debug() with setup_config() in READMEMiroslav Benes
Commit 35c9e74cff4c ("selftests/livepatch: Make dynamic debug setup and restore generic") introduced setup_config() to set up the environment for each test. It superseded set_dynamic_debug(). README still mentions set_dynamic_debug(), so update it to setup_config() which should be used now in every test. Signed-off-by: Miroslav Benes <mbenes@suse.cz> Reviewed-by: Kamalesh Babulal <kamalesh@linux.vnet.ibm.com> Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
2020-01-16staging: wfx: update TODOJérôme Pouiller
Some work has been done :) Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com> Link: https://lore.kernel.org/r/20200115135338.14374-66-Jerome.Pouiller@silabs.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2020-01-16staging: wfx: simplify hif_multi_tx_confirm()Jérôme Pouiller
Usage of the "buf_loc" variable does not simplify the function. Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com> Link: https://lore.kernel.org/r/20200115135338.14374-65-Jerome.Pouiller@silabs.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2020-01-16staging: wfx: simplify wfx_tx_queue_get_num_queued()Jérôme Pouiller
wfx_tx_queue_get_num_queued() can take advantage of BIT() instead of maintaining one variable for a counter and another for a mask. In add, wfx_tx_queue_get_num_queued() has no real reason to return a size_t instead of an int. Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com> Link: https://lore.kernel.org/r/20200115135338.14374-64-Jerome.Pouiller@silabs.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2020-01-16staging: wfx: simplify hif_handle_tx_data()Jérôme Pouiller
Since enum action has now only two cases, it can be dropped. Then hif_handle_tx_data() can be simplified. Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com> Link: https://lore.kernel.org/r/20200115135338.14374-63-Jerome.Pouiller@silabs.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2020-01-16staging: wfx: remove check for interface stateJérôme Pouiller
Obviously, the value of wvif->state as no reason to be wrong. At least, if it the case, dropping the frame is probably not the bast thing to do. Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com> Link: https://lore.kernel.org/r/20200115135338.14374-62-Jerome.Pouiller@silabs.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2020-01-16staging: wfx: remove unused do_probeJérôme Pouiller
The identifier do_probe is unused since "staging: wfx: remove workaround to send probe requests" Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com> Link: https://lore.kernel.org/r/20200115135338.14374-61-Jerome.Pouiller@silabs.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2020-01-16staging: wfx: ensure that packet_id is uniqueJérôme Pouiller
In current code, packet_id is derivated from mac802.11 packet sequence number, but this number is only unique for a station. It is not sufficient. Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com> Link: https://lore.kernel.org/r/20200115135338.14374-60-Jerome.Pouiller@silabs.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2020-01-16staging: wfx: simplify wfx_ps_notify_sta()Jérôme Pouiller
wfx_ps_notify_sta() is too complex compared to the task it do. Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com> Link: https://lore.kernel.org/r/20200115135338.14374-59-Jerome.Pouiller@silabs.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2020-01-16staging: wfx: simplify handling of IEEE80211_TX_CTL_SEND_AFTER_DTIMJérôme Pouiller
When mac80211 ask for a frame to be sent after a DTIM, driver should: 1. Update TIM with multicast bit set (using update_ie). This function can be called whenever. 2. Keep buffered all frames marked "after dtim" 3. When it receive a suspend_resume indication (see wfx_suspend_resume_mc()), send all the buffered frames. This indication is sent by the firmware 4ms before the dtim. 4. If one of the frames returns status "REQUEUE", it means that the DTIM period was ended before to be able to send the frame. 5. When all the buffered frames were sent or if DTIM period was ended, driver should update the TIM with multicast bit reset. All the mess with the asynchronous works can be dropped. Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com> Link: https://lore.kernel.org/r/20200115135338.14374-58-Jerome.Pouiller@silabs.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2020-01-16staging: wfx: simplify wfx_suspend_resume_mc()Jérôme Pouiller
Indeed, it is not necessary to pass whole hif_ind_suspend_resume_tx to wfx_suspend_resume_mc(). In add, the structure hif_ind_suspend_resume_tx come from hardware API. It is not intended to be manipulated in upper layers of the driver. Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com> Link: https://lore.kernel.org/r/20200115135338.14374-57-Jerome.Pouiller@silabs.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2020-01-16staging: wfx: firmware never return PS status for stationsJérôme Pouiller
At the beginning, firmware could send suspend_resume indication to notify that a station wake up or sleep down. However, mac80211 already handles power save status of stations and this behavior has been removed from the firmware. So now, when suspend_resume indication is received, it is always to notify that a DTIM is about to be sent. So, it is possible to simply wfx_suspend_resume(). Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com> Link: https://lore.kernel.org/r/20200115135338.14374-56-Jerome.Pouiller@silabs.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2020-01-16staging: wfx: sta and dtimJérôme Pouiller
wfx_ps_notify() is called for any changes in the TIM. However, association ID 0 is a very special case that should be handled independently. Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com> Link: https://lore.kernel.org/r/20200115135338.14374-55-Jerome.Pouiller@silabs.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2020-01-16staging: wfx: pspoll_mask make no senseJérôme Pouiller
pspoll_mask is here to send data buffered in driver. But since station is marked buffered, TIM for this station is 1 and mac80211 will call sta_notify when a ps-poll is received. So pspoll_mask is useless and sta_alseep_mask is sufficient. Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com> Link: https://lore.kernel.org/r/20200115135338.14374-54-Jerome.Pouiller@silabs.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2020-01-16staging: wfx: replace wfx_tx_get_tid() with ieee80211_get_tid()Jérôme Pouiller
wfx_tx_get_tid() was used as a wrapper around ieee80211_get_tid(). It did sometime return WFX_MAX_TID to ask to upper layers to not include the frame in "buffered" counter. The objective of this behavior is not clear, but tests has shown that wfx_tx_get_tid() can be replaced by ieee80211_get_tid() without any regressions. BTW, it is not necessary to save the tid in tx_rpiv since it can be retrieved from the 802.11 header. Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com> Link: https://lore.kernel.org/r/20200115135338.14374-53-Jerome.Pouiller@silabs.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2020-01-16staging: wfx: check that no tx is pending before release staJérôme Pouiller
Just for sanity. Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com> Link: https://lore.kernel.org/r/20200115135338.14374-52-Jerome.Pouiller@silabs.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2020-01-16staging: wfx: simplify the link-id allocationJérôme Pouiller
The "link-id" is a slot number provided to the chip. A link-id is allocated to every station associated with the chip (mainly when the chip is in AP mode). It is more or less the same thing than the association ID, but it is limited to 14 values. Firmware uses the link-id to track the power save status of the stations. The current code try to associate a link-id as soon as data are exchanged with station. It is far easier to rely on sta_add() and sta_remove(). Until now the value WFX_LINK_ID_NO_ASSOC, was only used when no more link-id was available. Now, we also use this value for not-yet-associated stations (that was its primary behavior). Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com> Link: https://lore.kernel.org/r/20200115135338.14374-51-Jerome.Pouiller@silabs.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2020-01-16staging: wfx: simplify wfx_set_tim_impl()Jérôme Pouiller
Argument provided to wfx_set_tim_impl() is always wvif->aid0_bit_set and there is no reason to provide another argument. Also rename wfx_set_tim_impl() into wfx_update_tim() to reflect the new behavior. Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com> Link: https://lore.kernel.org/r/20200115135338.14374-50-Jerome.Pouiller@silabs.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2020-01-16staging: wfx: fix RCU usageJérôme Pouiller
Indeed, sta was used after call to rcu_unlock() Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com> Link: https://lore.kernel.org/r/20200115135338.14374-49-Jerome.Pouiller@silabs.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2020-01-16staging: wfx: fix bss_lossJérôme Pouiller
wfx_tx_confirm_cb() retrieves the station associated with a frame using the MAC address from the 802.11 header. In the other side wfx_tx() retrieves the station using sta field from the ieee80211_tx_control argument. In wfx_cqm_bssloss_sm(), wfx_tx() was called directly without valid sta field, but with a valid MAC address in 802.11 header. So there the processing of this packet was unbalanced and may produce weird bugs. Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com> Link: https://lore.kernel.org/r/20200115135338.14374-48-Jerome.Pouiller@silabs.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2020-01-16staging: wfx: relocate "buffered" information to sta_privJérôme Pouiller
It simplify the code if field buffered is hosted in the struct sta_priv instead of in the struct wfx_link_entry. More globally, struct wfx_link_entry has no real reasons to exist and should be dropped soon. Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com> Link: https://lore.kernel.org/r/20200115135338.14374-47-Jerome.Pouiller@silabs.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2020-01-16staging: wfx: remove handling of "early_data"Jérôme Pouiller
It seems that purpose of "early_data" was to prevent sending data to mac80211 before station was completely associated. It is a useless precaution. Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com> Link: https://lore.kernel.org/r/20200115135338.14374-46-Jerome.Pouiller@silabs.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2020-01-16staging: wfx: fix possible overflow on jiffies comparaisonJérôme Pouiller
MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit It is recommended to use function time_*() to compare jiffies. Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com> Link: https://lore.kernel.org/r/20200115135338.14374-45-Jerome.Pouiller@silabs.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2020-01-16staging: wfx: fix case where RTS threshold is 0Jérôme Pouiller
If RTS threshold is 0, it currently disables RTS. It should mean "enabled for every frames". Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com> Link: https://lore.kernel.org/r/20200115135338.14374-44-Jerome.Pouiller@silabs.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2020-01-16staging: wfx: do not update uapsd if not necessaryJérôme Pouiller
wfx_conf_tx() is called for each queue. On every call, the function updates UAPSD mask and PM mode for all queues. It is a pity since the UAPSD configuration very rarely changes and it makes exchanges between the host and the chip more difficult to track. This patch avoid to update UAPSD and Power Mode in most usual cases. Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com> Link: https://lore.kernel.org/r/20200115135338.14374-43-Jerome.Pouiller@silabs.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2020-01-16staging: wfx: with multiple vifs, force PS only if channels differsJérôme Pouiller
When multiple vif are in use (the only supported configuration is one station and one AP), the driver force power save flag on station. This behavior allows the station to leave the station channel and make its business on AP channel. However, this has a big impact on station performances (especially since only legacy PS is supported). When both vifs use the same channel, it is not necessary to keep this restriction. This greatly improve station performances. Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com> Link: https://lore.kernel.org/r/20200115135338.14374-42-Jerome.Pouiller@silabs.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
generated by cgit (git 2.34.1) at 2025-09-20 03:34:38 +0000