- 04 Nov, 2022 4 commits
-
-
Colin Ian King authored
Variable stop_report_cnt is being set or incremented but is never being used for anything meaningful. The variable and code relating to it's use is redundant and can be removed. Signed-off-by:
Colin Ian King <colin.i.king@gmail.com> Acked-by:
Ping-Ke Shih <pkshih@realtek.com> Signed-off-by:
Kalle Valo <kvalo@kernel.org> Link: https://lore.kernel.org/r/20221031155637.871164-1-colin.i.king@gmail.com
-
Linus Walleij authored
We currently register the BCMA core even if the GPIO portions fail. There is no reason for this: the GPIO should register just fine, if it fails the BCMA driver should fail. We already gracefully handle the case where the GPIO driver is not compiled in. Signed-off-by:
Linus Walleij <linus.walleij@linaro.org> Signed-off-by:
-
Linus Walleij authored
The <linux/bcma/bcma_driver_chipcommon.h> is including the legacy header <linux/gpio.h> to obtain struct gpio_chip. Instead, include <linux/gpio/driver.h> where this struct is defined. It turns out that the brcm80211 brcmsmac depends on this to bring in the symbol gpio_is_valid(). The driver looks up the BCMA parent GPIO driver and checks that this succeeds, but then it goes on to use the deprecated GPIO call gpio_is_valid() to check the consistency of the .base member of the BCMA GPIO struct. The whole check can be dropped because the bcma_gpio is initialized in the declarations: struct gpio_chip *bcma_gpio = &cc_drv->gpio; And this can never be NULL. Cc: Jonas Gorski <jonas.gorski@gmail.com> Acked-by:
Arend van Spriel <arend.vanspriel@broadcom.com> Signed-off-by:
-
Minsuk Kang authored
This patch fixes a shift-out-of-bounds in brcmfmac that occurs in BIT(chiprev) when a 'chiprev' provided by the device is too large. It should also not be equal to or greater than BITS_PER_TYPE(u32) as we do bitwise AND with a u32 variable and BIT(chiprev). The patch adds a check that makes the function return NULL if that is the case. Note that the NULL case is later handled by the bus-specific caller, brcmf_usb_probe_cb() or brcmf_usb_reset_resume(), for example. Found by a modified version of syzkaller. UBSAN: shift-out-of-bounds in drivers/net/wireless/broadcom/brcm80211/brcmfmac/firmware.c shift exponent 151055786 is too large for 64-bit type 'long unsigned int' CPU: 0 PID: 1885 Comm: kworker/0:2 Tainted: G O 5.14.0+ #132 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.1-0-ga5cab58e9a3f-prebuilt.qemu.org 04/01/2014 Workqueue: usb_hub_wq hub_event Call Trace: dump_stack_lvl+0x57/0x7d ubsan_epilogue+0x5/0x40 __ubsan_handle_shift_out_of_bounds.cold+0x53/0xdb ? lock_chain_count+0x20/0x20 brcmf_fw_alloc_request.cold+0x19/0x3ea ? brcmf_fw_get_firmwares+0x250/0x250 ? brcmf_usb_ioctl_resp_wait+0x1a7/0x1f0 brcmf_usb_get_fwname+0x114/0x1a0 ? brcmf_usb_reset_resume+0x120/0x120 ? number+0x6c4/0x9a0 brcmf_c_process_clm_blob+0x168/0x590 ? put_dec+0x90/0x90 ? enable_ptr_key_workfn+0x20/0x20 ? brcmf_common_pd_remove+0x50/0x50 ? rcu_read_lock_sched_held+0xa1/0xd0 brcmf_c_preinit_dcmds+0x673/0xc40 ? brcmf_c_set_joinpref_default+0x100/0x100 ? rcu_read_lock_sched_held+0xa1/0xd0 ? rcu_read_lock_bh_held+0xb0/0xb0 ? lock_acquire+0x19d/0x4e0 ? find_held_lock+0x2d/0x110 ? brcmf_usb_deq+0x1cc/0x260 ? mark_held_locks+0x9f/0xe0 ? lockdep_hardirqs_on_prepare+0x273/0x3e0 ? _raw_spin_unlock_irqrestore+0x47/0x50 ? trace_hardirqs_on+0x1c/0x120 ? brcmf_usb_deq+0x1a7/0x260 ? brcmf_usb_rx_fill_all+0x5a/0xf0 brcmf_attach+0x246/0xd40 ? wiphy_new_nm+0x1476/0x1d50 ? kmemdup+0x30/0x40 brcmf_usb_probe+0x12de/0x1690 ? brcmf_usbdev_qinit.constprop.0+0x470/0x470 usb_probe_interface+0x25f/0x710 really_probe+0x1be/0xa90 __driver_probe_device+0x2ab/0x460 ? usb_match_id.part.0+0x88/0xc0 driver_probe_device+0x49/0x120 __device_attach_driver+0x18a/0x250 ? driver_allows_async_probing+0x120/0x120 bus_for_each_drv+0x123/0x1a0 ? bus_rescan_devices+0x20/0x20 ? lockdep_hardirqs_on_prepare+0x273/0x3e0 ? trace_hardirqs_on+0x1c/0x120 __device_attach+0x207/0x330 ? device_bind_driver+0xb0/0xb0 ? kobject_uevent_env+0x230/0x12c0 bus_probe_device+0x1a2/0x260 device_add+0xa61/0x1ce0 ? __mutex_unlock_slowpath+0xe7/0x660 ? __fw_devlink_link_to_suppliers+0x550/0x550 usb_set_configuration+0x984/0x1770 ? kernfs_create_link+0x175/0x230 usb_generic_driver_probe+0x69/0x90 usb_probe_device+0x9c/0x220 really_probe+0x1be/0xa90 __driver_probe_device+0x2ab/0x460 driver_probe_device+0x49/0x120 __device_attach_driver+0x18a/0x250 ? driver_allows_async_probing+0x120/0x120 bus_for_each_drv+0x123/0x1a0 ? bus_rescan_devices+0x20/0x20 ? lockdep_hardirqs_on_prepare+0x273/0x3e0 ? trace_hardirqs_on+0x1c/0x120 __device_attach+0x207/0x330 ? device_bind_driver+0xb0/0xb0 ? kobject_uevent_env+0x230/0x12c0 bus_probe_device+0x1a2/0x260 device_add+0xa61/0x1ce0 ? __fw_devlink_link_to_suppliers+0x550/0x550 usb_new_device.cold+0x463/0xf66 ? hub_disconnect+0x400/0x400 ? _raw_spin_unlock_irq+0x24/0x30 hub_event+0x10d5/0x3330 ? hub_port_debounce+0x280/0x280 ? __lock_acquire+0x1671/0x5790 ? wq_calc_node_cpumask+0x170/0x2a0 ? lock_release+0x640/0x640 ? rcu_read_lock_sched_held+0xa1/0xd0 ? rcu_read_lock_bh_held+0xb0/0xb0 ? lockdep_hardirqs_on_prepare+0x273/0x3e0 process_one_work+0x873/0x13e0 ? lock_release+0x640/0x640 ? pwq_dec_nr_in_flight+0x320/0x320 ? rwlock_bug.part.0+0x90/0x90 worker_thread+0x8b/0xd10 ? __kthread_parkme+0xd9/0x1d0 ? process_one_work+0x13e0/0x13e0 kthread+0x379/0x450 ? _raw_spin_unlock_irq+0x24/0x30 ? set_kthread_struct+0x100/0x100 ret_from_fork+0x1f/0x30 Reported-by:
Dokyung Song <dokyungs@yonsei.ac.kr> Reported-by:
Jisoo Jang <jisoo.jang@yonsei.ac.kr> Reported-by:
Minsuk Kang <linuxlovemin@yonsei.ac.kr> Signed-off-by:
-
- 01 Nov, 2022 16 commits
-
-
Michael Walle authored
There are no SDIO module aliases included in the driver, therefore, module autoloading isn't working. Add the proper MODULE_DEVICE_TABLE(). Cc: stable@vger.kernel.org Signed-off-by:
Michael Walle <michael@walle.cc> Signed-off-by:
-
Kees Cook authored
Fix the atmel_private_handler to correctly sized (1 element) again. (I should have checked the data segment for differences.) This had no behavioral impact (no private callbacks), but it made a very large zero-filled array. Cc: Simon Kelley <simon@thekelleys.org.uk> Cc: Kalle Valo <kvalo@kernel.org> Cc: "David S. Miller" <davem@davemloft.net> Cc: Eric Dumazet <edumazet@google.com> Cc: Jakub Kicinski <kuba@kernel.org> Cc: Paolo Abeni <pabeni@redhat.com> Cc: linux-wireless@vger.kernel.org Cc: netdev@vger.kernel.org Fixes: 8af9d406 ("wifi: atmel: Avoid clashing function prototypes") Signed-off-by:
Kees Cook <keescook@chromium.org> Signed-off-by:
-
Jiapeng Chong authored
The function ipw_alive() is defined in the ipw2200.c file, but not called elsewhere, so delete this unused function. drivers/net/wireless/intel/ipw2x00/ipw2200.c:3007:19: warning: unused function 'ipw_alive'. Link: https://bugzilla.openanolis.cn/show_bug.cgi?id=2410Reported-by:
Abaci Robot <abaci@linux.alibaba.com> Signed-off-by:
Jiapeng Chong <jiapeng.chong@linux.alibaba.com> Signed-off-by:
-
Colin Ian King authored
Ther is a spelling mistake in a rtw89_debug message. Fix it. Signed-off-by:
-
Chin-Yen Lee authored
Pattern match is an option of WoWLAN to allow the device to be woken up from suspend mode when receiving packets matched user-designed patterns. The patterns are written into hardware via WoWLAN firmware in suspend flow if users have set up them. If packets matched designed pattern are received, WoWLAN firmware will send an interrupt and then wake up the device. Signed-off-by:
Chin-Yen Lee <timlee@realtek.com> Signed-off-by:
-
Chin-Yen Lee authored
WoWLAN is a feature which allows devices to be woken up from suspend state through WLAN events. When user enables WoWLAN feature and then let the device enter suspend state, WoWLAN firmware will be loaded by the driver and periodically monitors WiFi packets. Power consumption of WiFi chip will be reduced in this state. We now implement WoWLAN function in rtw8852ae and rtw8852ce chip. Currently supported WLAN events include receiving magic packet, rekey packet and deauth packet, and disconnecting from AP. Signed-off-by:
-
Chin-Yen Lee authored
In this patch we define some H2C, which will be called during suspend flow, to enable WoWLAN function provided by WoWLAN firmware. These H2C includes keep alive used to send null packet to AP periodically to avoid being disconnected by AP, disconnect detection used to configure how we check if AP is offline, wake up control used to decide which WiFi events could trigger resume flow, and global control used to enable WoWLAN function. Signed-off-by:
-
Chih-Kang Chang authored
When entering WoWLAN mode, we need to drop all transmit packets, including those in mac buffer, to avoid memory leakage, so implement the drop_tx function. Signed-off-by:
Chih-Kang Chang <gary.chang@realtek.com> Signed-off-by:
-
Chih-Kang Chang authored
PLE RX quota, which is the setting of RX buffer, is needed to be adjusted dynamically for WoWLAN mode, and restored when back to normal mode. The action is not needed for rtw8852c chip. Signed-off-by:
-
Chih-Kang Chang authored
For WoWLAN mode, we need to download WoWLAN firmware by calling fw_download(). Another, to disable/enable WiFi CPU is needed before calling fw_download. Since Firmware runs on WiFi CPU, it is intuitive to combine enable_cpu/disable_cpu functions into fw_download. Signed-off-by:
-
Chih-Kang Chang authored
For WoWLAN mode, we only collect and send RF parameters to Firmware without writing RF registers. So we add one function to practice it. Signed-off-by:
-
Zong-Zhe Yang authored
Add boundary check of mac_id when adding sta under AP/TDLS. And, return -ENOSPC if the acquired mac_id is invalid. Signed-off-by:
Zong-Zhe Yang <kevin_yang@realtek.com> Signed-off-by:
-
Eric Huang authored
In order to debug performance issue intuitively, add bandwidth information into debugfs entry phy_info. After applying this patch, it looks like: TX rate [0]: HE 2SS MCS-11 GI:0.8 BW:80 (hw_rate=0x19b) ==> agg_wait=1 (3500) RX rate [0]: HE 2SS MCS-9 GI:0.8 BW:80 (hw_rate=0x199) Signed-off-by:
Eric Huang <echuang@realtek.com> Signed-off-by:
-
Ping-Ke Shih authored
This table must be constant, so change it as expectation. Signed-off-by:
-
Zong-Zhe Yang authored
They are just default declarations and we won't modify them directly. Instead, we actually do moification on their memdup now. So, they should be declared with const. Signed-off-by:
-
Ping-Ke Shih authored
Since firmware size is limited, we create variant firmwares for variant application areas. To help driver to know firmware's capabilities, firmware dynamic header is introduced to have more information, such as firmware features and firmware compile flags. Since this driver rtw89 only uses single one specific firmware at runtime, this patch is just to ignore this dynamic header, not actually use the content. This patch can be backward compatible, and no this kind of firmware is added to linux-firmware yet, so I can prepare this in advance. Signed-off-by:
-
- 29 Oct, 2022 20 commits
-
-
git://git.kernel.org/pub/scm/linux/kernel/git/kvalo/ath.gitKalle Valo authored
ath.git patches for v6.2. Major changes: ath11k * support configuring channel dwell time during scan
-
Lukasz Czapnik authored
In the event of a Tx hang it can be useful to read a variety of hardware registers to capture some state about why the transmit queue got stuck. Extend the ETHTOOL_GREGS dump provided by the ice driver with several CSR registers that provide such relevant information regarding the hardware Tx state. This enables capturing relevant data to enable debugging such a Tx hang. Signed-off-by:
Lukasz Czapnik <lukasz.czapnik@intel.com> Signed-off-by:
Mateusz Palczewski <mateusz.palczewski@intel.com> Tested-by: Gurucharan <gurucharanx.g@intel.com> (A Contingent worker at Intel) Link: https://lore.kernel.org/r/20221027104239.1691549-1-jacob.e.keller@intel.comSigned-off-by:
Jakub Kicinski <kuba@kernel.org>
-
Jakub Kicinski authored
Russell King says: ==================== Clean up SFP register definitions This two-part patch series cleans up the SFP register definitions by 1. converting them from hex to decimal, as all the definitions in the documents use decimal, this makes it easier to cross-reference. 2. moving the bit definitions for each register along side their register address definition ==================== Link: https://lore.kernel.org/r/Y1qFvaDlLVM1fHdG@shell.armlinux.org.ukSigned-off-by:
-
Russell King (Oracle) authored
Just as we do for the A2h enum, arrange the A0h enum to have the field definitions next to their corresponding register index. Signed-off-by:
Russell King (Oracle) <rmk+kernel@armlinux.org.uk> Reviewed-by:
Andrew Lunn <andrew@lunn.ch> Signed-off-by:
-
Russell King (Oracle) authored
The register indexes in the standards are in decimal rather than hex, so lets specify them in decimal in the header file so we can easily cross-reference without converting between hex and decimal. Signed-off-by:
-
Jakub Kicinski authored
Russell King says: ==================== net: mtk_eth_soc: improve PCS implementation As a result of invesigations from Frank Wunderlich, we know a lot more about the Mediatek "SGMII" PCS block, and can implement the PCS support correctly. This series achieves that, and Frank has tested the final result and reports that it works for him. The series could do with further testing by others, but I suspect that is unlikely to happen until it is merged based on past performances with this driver. Briefly, the patches in order: 1. Add a new helper to get the link timer duration in nanoseconds 2. Add definitions for the newly discovered registers and updates to bit definitions, including bitmasks for the BMCR, BMSR and two advertisement registers. 3. Remove unnecessary/unused error handling (functions always returning zero.) 4. Adding the missing pcs_get_state() implementation. 5. Converting the code to use regmap_update_bits() rather than open-coding read-modify-write sequences. 6. Adding out-of-band speed and duplex forcing for all non-inband modes not just the 802.3z link modes the code currently does. 7. Moving the release of the PHY power down to the main pcs_config() function. 8. Moving the interface speed selection to the main pcs_config() function. 9. Adding advertisement programming. 10. Adding correct link timer programming using the new helper in the first patch. 11. Adding support for 802.3z negotiation. There is one remaining issue - when configuring the PCS for in-band, for some reason the AN restart bit is always set. This should not be necessary, but requires further investigation with the hardware to find out whether it is really necessary. I suspect this was a work around for a previous poor implementation. ==================== Link: https://lore.kernel.org/r/Y1qDMw+DJLAJHT40@shell.armlinux.org.ukSigned-off-by: -
Russell King (Oracle) authored
As a result of help from Frank Wunderlich to investigate and test, we now know how to program this PCS for in-band 802.3z negotiation. Add support for this by moving the contents of the two functions into the common mtk_pcs_config() function and adding the register settings for 802.3z negotiation. Signed-off-by:
-
Russell King (Oracle) authored
Program the link timer appropriately for the interface mode being used, using the newly introduced phylink helper that provides the nanosecond link timer interval. The intervals are 1.6ms for SGMII based protocols and 10ms for 802.3z based protocols. Signed-off-by:
-
Russell King (Oracle) authored
Program the advertisement into the mtk PCS block. Signed-off-by:
-
Russell King (Oracle) authored
Move the selection of the underlying interface speed to the pcs_config function, so we always program the interface speed. Signed-off-by:
-
Russell King (Oracle) authored
The PHY power up is common to both configuration paths, so move it into the parent function. We need to do this for all serdes modes. Signed-off-by:
-
Russell King (Oracle) authored
Add support for forcing the link speed and duplex setting in the pcs_link_up() method for out of band modes, which will be useful when we finish converting the pcs_config() method. Until then, we still have to force duplex for 802.3z modes to work correctly. Signed-off-by:
-
Russell King (Oracle) authored
mtk_sgmii does a lot of read-modify-write operations, for which there is a specific regmap function. Use this function instead of open-coding the operations. Signed-off-by:
-
Russell King (Oracle) authored
Add a pcs_get_state() implementation which uses the advertisements to compute the resulting link modes, and BMSR contents to determine negotiation and link status. Signed-off-by:
-
Russell King (Oracle) authored
The functions called by the pcs_config() method always return zero, so there is no point trying to handle an error from these functions. Make these functions void, eliminate the "err" variable and simply return zero from the pcs_config() function itself. Signed-off-by:
-
Russell King (Oracle) authored
As a result of help from Frank Wunderlich to investigate and test, we know a bit more about the PCS on the Mediatek platforms. Update the definitions from this investigation. This PCS appears similar, but not identical to the Lynx PCS. Although not included in this patch, but for future reference, the PHY ID registers at offset 4 read as 0x4d544950 'MTIP'. Signed-off-by:
-
Russell King (Oracle) authored
Add a helper to convert the PHY interface mode to the required link timer setting as stated by the appropriate standard. Inappropriate interface modes return an error. Signed-off-by:
-
Jakub Kicinski authored
Sebastian Andrzej Siewior says: ==================== net: Remove the obsolte u64_stats_fetch_*_irq() This is the removal of u64_stats_fetch_*_irq() users in networking. The prerequisites are part of v6.1-rc1. The spi and bpf bits are not part of the series and have been routed directly. ==================== Link: https://lore.kernel.org/r/20221026132215.696950-1-bigeasy@linutronix.deSigned-off-by:
-
Thomas Gleixner authored
Now that the 32bit UP oddity is gone and 32bit uses always a sequence count, there is no need for the fetch_irq() variants anymore. Convert to the regular interface. Signed-off-by:
Thomas Gleixner <tglx@linutronix.de> Signed-off-by:
Sebastian Andrzej Siewior <bigeasy@linutronix.de> Acked-by:
Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by:
-
Thomas Gleixner authored
Now that the 32bit UP oddity is gone and 32bit uses always a sequence count, there is no need for the fetch_irq() variants anymore. Convert to the regular interface. Signed-off-by:
-
