- 24 Mar, 2018 40 commits
-
-
Maor Gottlieb authored
[ Upstream commit 13776612 ] Anonymous VMA (->vm_ops == NULL) cannot be shared, otherwise it would lead to SIGBUS. Remove the shared flags from the vma after we change it to be anonymous. This is easily reproduced by doing modprobe -r while running a user-space application such as raw_ethernet_bw. Fixes: 7c2344c3 ('IB/mlx5: Implements disassociate_ucontext API') Signed-off-by:
Maor Gottlieb <maorg@mellanox.com> Signed-off-by:
Leon Romanovsky <leon@kernel.org> Signed-off-by:
Doug Ledford <dledford@redhat.com> Signed-off-by:
Sasha Levin <alexander.levin@microsoft.com> Signed-off-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-
Maor Gottlieb authored
[ Upstream commit ecc7d83b ] When the driver disassociate user context, it changes the vma to anonymous by setting the vm_ops to null and zap the vma ptes. In order to avoid race in the kernel, we need to take write lock before we change the vma entries. Fixes: 7c2344c3 ('IB/mlx5: Implements disassociate_ucontext API') Signed-off-by:
-
Maor Gottlieb authored
[ Upstream commit ca37a664 ] Anonymous VMA (->vm_ops == NULL) cannot be shared, otherwise it would lead to SIGBUS. Remove the shared flags from the vma after we change it to be anonymous. This is easily reproduced by doing modprobe -r while running a user-space application such as raw_ethernet_bw. Fixes: ae184dde ('IB/mlx4_ib: Disassociate support') Signed-off-by:
-
Maor Gottlieb authored
[ Upstream commit 22c3653d ] When the driver disassociate user context, it changes the vma to anonymous by setting the vm_ops to null and zap the vma ptes. In order to avoid race in the kernel, we need to take write lock before we change the vma entries. Fixes: ae184dde ('IB/mlx4_ib: Disassociate support') Signed-off-by:
-
Dan Carpenter authored
[ Upstream commit 30260501 ] If skb_pad() fails then it frees skb and we don't need to free it again at the end of the function. Fixes: dc7bf5d7 ("HSI: Introduce driver for SSI Protocol") Signed-off-by:
Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by:
Sebastian Reichel <sre@kernel.org> Signed-off-by:
-
Feras Daoud authored
[ Upstream commit 9a9b8112 ] Update the broadcast address in the priv->broadcast object when the Pkey value changes in index 0, otherwise the multicast GID value will keep the previous value of the PKey, and will not be updated. This leads to interface state down because the interface will keep the old PKey value. For example, in SR-IOV environment, if the PF changes the value of PKey index 0 for one of the VFs, then the VF receives PKey change event that triggers heavy flush. This flush calls update_parent_pkey that update the broadcast object and its relevant members. If in this case the multicast GID will not be updated, the interface state will be down. Fixes: c2904141 ("IPoIB: Fix pkey change flow for virtualization environments") Signed-off-by:
Feras Daoud <ferasda@mellanox.com> Signed-off-by:
Erez Shitrit <erezsh@mellanox.com> Reviewed-by:
Alex Vesker <valex@mellanox.com> Signed-off-by:
-
Feras Daoud authored
[ Upstream commit 3e31a490 ] Before calling ipoib_stop, rtnl_lock should be taken, then the flow clears the IPOIB_FLAG_ADMIN_UP and IPOIB_FLAG_OPER_UP flags, and waits for mcast completion if IPOIB_MCAST_FLAG_BUSY is set. On the other hand, the flow of multicast join task initializes a mcast completion, sets the IPOIB_MCAST_FLAG_BUSY and calls ipoib_mcast_join. If IPOIB_FLAG_OPER_UP flag is not set, this call returns EINVAL without setting the mcast completion and leads to a deadlock. ipoib_stop | | | clear_bit(IPOIB_FLAG_ADMIN_UP) | | | Context Switch | | ipoib_mcast_join_task | | | spin_lock_irq(lock) | | | init_completion(mcast) | | | set_bit(IPOIB_MCAST_FLAG_BUSY) | | | Context Switch | | clear_bit(IPOIB_FLAG_OPER_UP) | | | spin_lock_irqsave(lock) | | | Context Switch | | ipoib_mcast_join | return (-EINVAL) | | | spin_unlock_irq(lock) | | | Context Switch | | ipoib_mcast_dev_flush | wait_for_completion(mcast) | ipoib_stop will wait for mcast completion for ever, and will not release the rtnl_lock. As a result panic occurs with the following trace: [13441.639268] Call Trace: [13441.640150] [<ffffffff8168b579>] schedule+0x29/0x70 [13441.641038] [<ffffffff81688fc9>] schedule_timeout+0x239/0x2d0 [13441.641914] [<ffffffff810bc017>] ? complete+0x47/0x50 [13441.642765] [<ffffffff810a690d>] ? flush_workqueue_prep_pwqs+0x16d/0x200 [13441.643580] [<ffffffff8168b956>] wait_for_completion+0x116/0x170 [13441.644434] [<ffffffff810c4ec0>] ? wake_up_state+0x20/0x20 [13441.645293] [<ffffffffa05af170>] ipoib_mcast_dev_flush+0x150/0x190 [ib_ipoib] [13441.646159] [<ffffffffa05ac967>] ipoib_ib_dev_down+0x37/0x60 [ib_ipoib] [13441.647013] [<ffffffffa05a4805>] ipoib_stop+0x75/0x150 [ib_ipoib] Fixes: 08bc3276 ("IB/ipoib: fix for rare multicast join race condition") Signed-off-by:
-
Mikhail Paulyshka authored
[ Upstream commit fc7438b1 ] Headset microphone does not work out of the box on ASUS Nx51 laptops. This patch fixes it. Patch tested on Asus N551 laptop. Asus N751 part is not tested, but according to [1] this laptop uses the same audiosystem. 1. https://bugzilla.kernel.org/show_bug.cgi?id=117781 Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=195437Signed-off-by:
Mikhail Paulyshka <me@mixaill.tk> Signed-off-by:
Takashi Iwai <tiwai@suse.de> Signed-off-by:
-
Bernd Faust authored
[ Upstream commit 5313eecc ] After an upgrade to Linux kernel v4.x the hardware timestamps of the 82579 Gigabit Ethernet Controller are different than expected. The values that are being read are almost four times as big as before the kernel upgrade. The difference is that after the upgrade the driver sets the clock frequency to 25MHz, where before the upgrade it was set to 96MHz. Intel confirmed that the correct frequency for this network adapter is 96MHz. Signed-off-by:
Bernd Faust <berndfaust@gmail.com> Acked-by:
Sasha Neftin <sasha.neftin@intel.com> Acked-by:
Jacob Keller <jacob.e.keller@intel.com> Tested-by:
Aaron Brown <aaron.f.brown@intel.com> Signed-off-by:
Jeff Kirsher <jeffrey.t.kirsher@intel.com> Signed-off-by:
-
Eric Dumazet authored
[ Upstream commit 0f9fa831 ] When using TCP FastOpen for an active session, we send one wakeup event from tcp_finish_connect(), right before the data eventually contained in the received SYNACK is queued to sk->sk_receive_queue. This means that depending on machine load or luck, poll() users might receive POLLOUT events instead of POLLIN|POLLOUT To fix this, we need to move the call to sk->sk_state_change() after the (optional) call to tcp_rcv_fastopen_synack() Signed-off-by:
Eric Dumazet <edumazet@google.com> Acked-by:
Yuchung Cheng <ycheng@google.com> Signed-off-by:
David S. Miller <davem@davemloft.net> Signed-off-by:
-
Benjamin Coddington authored
[ Upstream commit 43b7d964 ] Commit a7d42ddb ("nfs: add mirroring support to pgio layer") moved pg_cleanup out of the path when there was non-sequental I/O that needed to be flushed. The result is that for layouts that have more than one layout segment per file, the pg_lseg is not cleared, so we can end up hitting the WARN_ON_ONCE(req_start >= seg_end) in pnfs_generic_pg_test since the pg_lseg will be pointing to that previously-flushed layout segment. Signed-off-by:
Benjamin Coddington <bcodding@redhat.com> Fixes: a7d42ddb ("nfs: add mirroring support to pgio layer") Signed-off-by:
Trond Myklebust <trond.myklebust@primarydata.com> Signed-off-by:
-
Guoqing Jiang authored
[ Upstream commit cf25ae78 ] Since nr_queued is changed, we need to call wake_up here if the array is already frozen and waiting for condition "nr_pending == nr_queued + extra" to be true. And commit 824e47da ("RAID1: avoid unnecessary spin locks in I/O barrier code") which has already added the wake_up for raid1. Signed-off-by:
Guoqing Jiang <gqjiang@suse.com> Reviewed-by:
NeilBrown <neilb@suse.com> Signed-off-by:
Shaohua Li <shli@fb.com> Signed-off-by:
-
Suman Anna authored
[ Upstream commit abaa7e5b ] Move the registration of the OMAP IOMMU platform driver before setting the IOMMU callbacks on the platform bus. This causes the IOMMU devices to be probed first before the .add_device() callback is invoked for all registered devices, and allows the iommu_group support to be added to the OMAP IOMMU driver. While at this, also check for the return status from bus_set_iommu. Signed-off-by:
Suman Anna <s-anna@ti.com> Signed-off-by:
Joerg Roedel <jroedel@suse.de> Signed-off-by:
-
Paul Burton authored
[ Upstream commit f8dcd9e8 ] Since commit 2af70a96 ("irqchip/mips-gic: Add a IPI hierarchy domain") introduced the GIC IPI IRQ domain we have tracked both reservation of interrupts & their use with a single bitmap - ipi_resrv. If an interrupt is reserved for use as an IPI but not actually in use then the appropriate bit is set in ipi_resrv. If an interrupt is either not reserved for use as an IPI or has been allocated as one then the appropriate bit is clear in ipi_resrv. Unfortunately this means that checking whether a bit is set in ipi_resrv to prevent IPI interrupts being allocated for use with a device is broken, because if the interrupt has been allocated as an IPI first then its bit will be clear. Fix this by separating the tracking of IPI reservation & usage, introducing a separate ipi_available bitmap for the latter. This means that ipi_resrv will now always have bits set corresponding to all interrupts reserved for use as IPIs, whether or not they have been allocated yet, and therefore that checking it when allocating device interrupts works as expected. Fixes: 2af70a96 ("irqchip/mips-gic: Add a IPI hierarchy domain") Signed-off-by:
Paul Burton <paul.burton@imgtec.com> Signed-off-by:
Matt Redfearn <matt.redfearn@imgtec.com> Cc: linux-mips@linux-mips.org Cc: Jason Cooper <jason@lakedaemon.net> Cc: Marc Zyngier <marc.zyngier@arm.com> Cc: Ralf Baechle <ralf@linux-mips.org> Link: http://lkml.kernel.org/r/1492679256-14513-2-git-send-email-matt.redfearn@imgtec.comSigned-off-by:
Thomas Gleixner <tglx@linutronix.de> Signed-off-by:
-
Abel Vesa authored
[ Upstream commit 6f05d076 ] The support for dynamic ftrace with CONFIG_DEBUG_RODATA involves overriding the weak arch_ftrace_update_code() with a variant which makes the kernel text writable around the patching. This override was however added under the CONFIG_OLD_MCOUNT ifdef, and CONFIG_OLD_MCOUNT is only enabled if frame pointers are enabled. This leads to non-functional dynamic ftrace (ftrace triggers a WARN_ON()) when CONFIG_DEBUG_RODATA is enabled and CONFIG_FRAME_POINTER is not. Move the override out of that ifdef and into the CONFIG_DYNAMIC_FTRACE ifdef where it belongs. Fixes: 80d6b0c2 ("ARM: mm: allow text and rodata sections to be read-only") Suggested-by:
Nicolai Stange <nicstange@gmail.com> Suggested-by:
Rabin Vincent <rabin@rab.in> Signed-off-by:
Abel Vesa <abelvesa@gmail.com> Acked-by:
Russell King <rmk+kernel@armlinux.org.uk> Signed-off-by:
-
Tiantian Feng authored
[ Upstream commit fba4f472 ] A CPU in VMX root mode will ignore INIT signals and will fail to bring up the APs after reboot. Therefore, on a panic we disable VMX on all CPUs before rebooting or triggering kdump. Do this when halting the machine as well, in case a firmware-level reboot does not perform a cold reset for all processors. Without doing this, rebooting the host may hang. Signed-off-by:
Tiantian Feng <fengtiantian@huawei.com> Signed-off-by:
Xishi Qiu <qiuxishi@huawei.com> [ Rewritten commit message. ] Signed-off-by:
Paolo Bonzini <pbonzini@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: kvm@vger.kernel.org Link: http://lkml.kernel.org/r/20170419161839.30550-1-pbonzini@redhat.comSigned-off-by:
Ingo Molnar <mingo@kernel.org> Signed-off-by:
-
Brian Norris authored
[ Upstream commit fb9e67be ] 'chan_stats' is (re)allocated in _mwifiex_fw_dpc() -> mwifiex_init_channel_scan_gap(), which is called whenever the device is initialized -- at probe or at reset. But we only free it in we completely unregister the adapter, meaning we leak a copy of it during every reset. Let's free it in the shutdown / removal paths instead (and in the error-handling path), to avoid the leak. Ideally, we can eventually unify much of mwifiex_shutdown_sw() and mwifiex_remove_card() (way too much copy-and-paste) to reduce the burden on bugfixes like this. But that's work for tomorrow. Signed-off-by:
Brian Norris <briannorris@chromium.org> Signed-off-by:
Kalle Valo <kvalo@codeaurora.org> Signed-off-by:
-
Alexey Kardashevskiy authored
[ Upstream commit bd9166ff ] At the moment kvmppc_mmu_map_page() returns -1 if mmu_hash_ops.hpte_insert() fails for any reason so the page fault handler resumes the guest and it faults on the same address again. This adds distinction to kvmppc_mmu_map_page() to return -EIO if mmu_hash_ops.hpte_insert() failed for a reason other than full pteg. At the moment only pSeries_lpar_hpte_insert() returns -2 if plpar_pte_enter() failed with a code other than H_PTEG_FULL. Other mmu_hash_ops.hpte_insert() instances can only fail with -1 "full pteg". With this change, if PR KVM fails to update HPT, it can signal the userspace about this instead of returning to guest and having the very same page fault over and over again. Signed-off-by:
Alexey Kardashevskiy <aik@ozlabs.ru> Reviewed-by:
David Gibson <david@gibson.dropbear.id.au> Signed-off-by:
Paul Mackerras <paulus@ozlabs.org> Signed-off-by:
-
David Gibson authored
[ Upstream commit 25d1d50e ] Passed through SCSI targets may have transfer limits which come from the host SCSI controller or something on the host side other than the target itself. To make this work properly, the hypervisor can adjust the target's VPD information to advertise these limits. But for that to work, the guest has to look at the VPD pages, which we won't do by default if it is an SPC-2 device, even if it does actually support it. This adds a workaround to address this, forcing devices attached to a virtio-scsi controller to always check the VPD pages. This is modelled on a similar workaround for the storvsc (Hyper-V) SCSI controller, although that exists for slightly different reasons. A specific case which causes this is a volume from IBM's IPR RAID controller (which presents as an SPC-2 device, although it does support VPD) passed through with qemu's 'scsi-block' device. [mkp: fixed typo] Signed-off-by:
Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by:
-
Liad Kaufman authored
[ Upstream commit f4d10479 ] Memory offsets and lengths for A000 HW is different than currently specified. Fixes: e34d975e ("iwlwifi: Add a000 HW family support") Signed-off-by:
Liad Kaufman <liad.kaufman@intel.com> Signed-off-by:
Luca Coelho <luciano.coelho@intel.com> Signed-off-by:
-
Emmanuel Grumbach authored
[ Upstream commit 2220fb29 ] The notification infrastructure (iwl_notification_wait_* functions) allows to wait until a list of notifications will come up from the firmware and to run a special handler (notif_wait handler) when those are received. The operation mode notifies the notification infrastructure about any Rx being received by the mean of iwl_notification_wait_notify() which will do two things: 1) call the notif_wait handler 2) wakeup the thread that was waiting for the notification Typically, only after those two steps happened, the operation mode will run its own handler for the notification that was received from the firmware. This means that the thread that was waiting for that notification can be running before the operation mode's handler was called. When the operation mode's handler is ASYNC, things get even worse since the thread that was waiting for the notification isn't even guaranteed that the ASYNC callback was added to async_handlers_list before it starts to run. This means that even calling iwl_mvm_wait_for_async_handlers() can't guarantee that absolutely everything related to that notification has run. The following can happen: Thread sending the command Operation mode's Rx path -------------------------- ------------------------ iwl_init_notification_wait() iwl_mvm_send_cmd() iwl_mvm_rx_common() iwl_notification_wait_notify() iwl_mvm_wait_for_async_handlers() // Possibly free some data // structure list_add_tail(async_handlers_list); schedule_work(async_handlers_wk); // Access the freed structure Split the 'run notif_wait's handler' and the 'wake up the thread' parts to fix this. This allows the operation mode to do the following: Thread sending the command Operation mode's Rx path -------------------------- ------------------------ iwl_init_notification_wait() iwl_mvm_send_cmd() iwl_mvm_rx_common() iwl_notification_wait() // Will run the notif_wait's handler list_add_tail(async_handlers_list); schedule_work(async_handlers_wk); iwl_notification_notify() iwl_mvm_wait_for_async_handlers() This way, the waiter is guaranteed that all the handlers have been run (if SYNC), or at least enqueued (if ASYNC) by the time it wakes up. Signed-off-by:
Emmanuel Grumbach <emmanuel.grumbach@intel.com> Signed-off-by:
-
Bharat Kumar Reddy Gooty authored
[ Upstream commit 8973aa4a ] Corrected the bits for power and iso. Signed-off-by:
Bharat Kumar Reddy Gooty <bharat.gooty@broadcom.com> Signed-off-by:
Jon Mason <jon.mason@broadcom.com> Fixes: f7225a83 ("clk: ns2: add clock support for Broadcom Northstar 2 SoC") Signed-off-by:
Stephen Boyd <sboyd@codeaurora.org> Signed-off-by:
-
Mohammed Shafi Shajakhan authored
[ Upstream commit c0c345d4 ] As per latest regulatory update for India, channel 52, 56, 60, 64 is no longer restricted to DFS. Enabling DFS/no infra flags in driver results in applying all DFS related restrictions (like doing CAC etc before this channel moves to 'available state') for these channels even though the country code is programmed as 'India' in he hardware, fix this by relaxing the frequency range while applying RADAR flags only if the country code is programmed to India. If the frequency range needs to modified based on different country code, ath_is_radar_freq can be extended/modified dynamically. Signed-off-by:
Mohammed Shafi Shajakhan <mohammed@qti.qualcomm.com> Signed-off-by:
Kalle Valo <kvalo@qca.qualcomm.com> Signed-off-by:
-
Nicholas Piggin authored
[ Upstream commit ca80d5d0 ] Power9 DD1 does not implement SAO. Although it's not widely used, its presence or absence is visible to user space via arch_validate_prot() so it's moderately important that we get the value right. Fixes: 7dccfbc3 ("powerpc/book3s: Add a cpu table entry for different POWER9 revs") Signed-off-by:
Nicholas Piggin <npiggin@gmail.com> Signed-off-by:
Michael Ellerman <mpe@ellerman.id.au> Signed-off-by:
-
Marek Vasut authored
[ Upstream commit 400c18e3 ] The dw_mmio driver disables the block clock before unregistering the host. The code unregistering the host may access the SPI block registers. If register access happens with block clock disabled, this may lead to a bus hang. Disable the clock after unregistering the host to prevent such situation. This bug was observed on Altera Cyclone V SoC. Signed-off-by:
Marek Vasut <marex@denx.de> Cc: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Mark Brown <broonie@kernel.org> Signed-off-by:
Mark Brown <broonie@kernel.org> Signed-off-by:
-
Dan Williams authored
[ Upstream commit 8b06b884 ] Keep the nfit_test instances alive until after nfit_test_teardown(), as we may be doing resource lookups until the final un-registrations have completed. This fixes crashes of the form. BUG: unable to handle kernel NULL pointer dereference at 0000000000000038 IP: __release_resource+0x12/0x90 Call Trace: remove_resource+0x23/0x40 __wrap_remove_resource+0x29/0x30 [nfit_test_iomap] acpi_nfit_remove_resource+0xe/0x10 [nfit] devm_action_release+0xf/0x20 release_nodes+0x16d/0x2b0 devres_release_all+0x3c/0x60 device_release+0x21/0x90 kobject_release+0x6a/0x170 kobject_put+0x2f/0x60 put_device+0x17/0x20 platform_device_unregister+0x20/0x30 nfit_test_exit+0x36/0x960 [nfit_test] Reported-by:
Linda Knippers <linda.knippers@hpe.com> Signed-off-by:
Dan Williams <dan.j.williams@intel.com> Signed-off-by:
-
Pierre-Louis Bossart authored
[ Upstream commit beb5989a ] There are multiple skews of the same Lenovo audio hardware based on the Realtek RT5670 codec. Manufacturer: LENOVO Product Name: 20C1CTO1WW Version: ThinkPad 10 Manufacturer: LENOVO Product Name: 20C3001VHH Version: ThinkPad 10 Manufacturer: LENOVO Product Name: 20C10024GE Version: ThinkPad Tablet B Manufacturer: LENOVO Product Name: 20359 Version: Lenovo Miix 2 10 For all these devices, the same quirk is used to force the machine driver to be based on RT5670 instead of RT5640 as indicated by the BIOS. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=96691Tested-by:
Nicole Faerber <nicole.faerber@dpin.de> Tested-by:
Viacheslav Ostroukh <v.dev@ostroukh.me> Signed-off-by:
Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Signed-off-by:
-
Adam Borowski authored
[ Upstream commit 14506127 ] If your filesystem has, eg, data:raid0 metadata:raid1, and you run "btrfs balance -dconvert=raid1", the meta.target field will be uninitialized. That's otherwise ok, as it's unused except for this warning. Thus, let's use the existing set of raid levels for the comparison. As a side effect, non-convert balances will now nag about data>metadata. Signed-off-by:
Adam Borowski <kilobyte@angband.pl> Reviewed-by:
Liu Bo <bo.li.liu@oracle.com> Signed-off-by:
David Sterba <dsterba@suse.com> Signed-off-by:
-
Jasmin J authored
[ Upstream commit e7080d44 ] It started with a sporadic message in syslog: "CAM tried to send a buffer larger than the ecount size" This message is not the fault itself, but a consecutive fault, after a read error from the CAM. This happens only on several CAMs, several hardware, and of course sporadic. It is a consecutive fault, if the last read from the CAM did fail. I guess this will not happen on all CAMs, but at least it did on mine. There was a write error to the CAM and during the re-initialization procedure, the CAM finished the last read, although it got a RS. The write error to the CAM happened because a race condition between HC write, checking DA and FR. This patch added an additional check for DA(RE), just after checking FR. It is important to read the CAMs status register again, to give the CAM the necessary time for a proper reaction to HC. Please note the description within the source code (patch below). [mchehab@s-opensource.com: make checkpatch happy] Signed-off-by:
Jasmin jessich <jasmin@anw.at> Tested-by:
Ralph Metzler <rjkm@metzlerbros.de> Signed-off-by:
Mauro Carvalho Chehab <mchehab@s-opensource.com> Signed-off-by:
-
David Ahern authored
[ Upstream commit 4a6e3c5d ] ndisc_notify is the ipv6 equivalent to arp_notify. When arp_notify is set to 1, gratuitous arp requests are sent when the device is brought up. The same is expected when ndisc_notify is set to 1 (per ndisc_notify in Documentation/networking/ip-sysctl.txt). The NA is not sent on NETDEV_UP event; add it. Fixes: 5cb04436 ("ipv6: add knob to send unsolicited ND on link-layer address change") Signed-off-by:
David Ahern <dsa@cumulusnetworks.com> Acked-by:
Hannes Frederic Sowa <hannes@stressinduktion.org> Signed-off-by:
-
Edgar Cherkasov authored
[ Upstream commit e058e7a4 ] Description of the problem: - i2c-scmi driver contains only two identifiers "SMBUS01" and "SMBUSIBM"; - the fist HID (SMBUS01) is clearly defined in "SMBus Control Method Interface Specification, version 1.0": "Each device must specify 'SMBUS01' as its _HID and use a unique _UID value"; - unfortunately, BIOS vendors (like AMI) seem to ignore this requirement and implement "SMB0001" HID instead of "SMBUS01"; - I speculate that they do this because only "SMB0001" is hard coded in Windows SMBus driver produced by Microsoft. This leads to following situation: - SMBus works out of box in Windows but not in Linux; - board vendors are forced to add correct "SMBUS01" HID to BIOS to make SMBus work in Linux. Moreover the same board vendors complain that tools (3-rd party ASL compiler) do not like the "SMBUS01" identifier and produce errors. So they need to constantly patch the compiler for each new version of BIOS. As it is very unlikely that BIOS vendors implement a correct HID in future, I would propose to consider whether it is possible to work around the problem by adding MS HID to the Linux i2c-scmi driver. v2: move the definition of the new HID to the driver itself. Signed-off-by:
Edgar Cherkasov <echerkasov@dev.rtsoft.ru> Signed-off-by:
Michael Brunner <Michael.Brunner@kontron.com> Acked-by:
Viktor Krasnov <vkrasnov@dev.rtsoft.ru> Reviewed-by:
Jean Delvare <jdelvare@suse.de> Reviewed-by:
Mika Westerberg <mika.westerberg@linux.intel.com> Signed-off-by:
Wolfram Sang <wsa@the-dreams.de> Signed-off-by:
-
Hans de Goede authored
[ Upstream commit 382bd4de ] When requesting a shared irq with IRQF_TRIGGER_NONE then the irqaction flags get filled with the trigger type from the irq_data: if (!(new->flags & IRQF_TRIGGER_MASK)) new->flags |= irqd_get_trigger_type(&desc->irq_data); On the first setup_irq() the trigger type in irq_data is NONE when the above code executes, then the irq is started up for the first time and then the actual trigger type gets established, but that's too late to fix up new->flags. When then a second user of the irq requests the irq with IRQF_TRIGGER_NONE its irqaction's triggertype gets set to the actual trigger type and the following check fails: if (!((old->flags ^ new->flags) & IRQF_TRIGGER_MASK)) Resulting in the request_irq failing with -EBUSY even though both users requested the irq with IRQF_SHARED | IRQF_TRIGGER_NONE Fix this by comparing the new irqaction's trigger type to the trigger type stored in the irq_data which correctly reflects the actual trigger type being used for the irq. Suggested-by:
Hans de Goede <hdegoede@redhat.com> Acked-by:
Marc Zyngier <marc.zyngier@arm.com> Link: http://lkml.kernel.org/r/20170415100831.17073-1-hdegoede@redhat.comSigned-off-by:
-
Thomas Gleixner authored
[ Upstream commit 205dcc1e ] The target() callback must run on the affected cpu. This is achieved by temporarily setting the affinity of the calling thread to the requested CPU and reset it to the original affinity afterwards. That's racy vs. concurrent affinity settings for that thread resulting in code executing on the wrong CPU. Replace it by work_on_cpu(). All call pathes which invoke the callbacks are already protected against CPU hotplug. Signed-off-by:
Viresh Kumar <viresh.kumar@linaro.org> Cc: Fenghua Yu <fenghua.yu@intel.com> Cc: Tony Luck <tony.luck@intel.com> Cc: Herbert Xu <herbert@gondor.apana.org.au> Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Sebastian Siewior <bigeasy@linutronix.de> Cc: linux-pm@vger.kernel.org Cc: Lai Jiangshan <jiangshanlai@gmail.com> Cc: Michael Ellerman <mpe@ellerman.id.au> Cc: Tejun Heo <tj@kernel.org> Cc: "David S. Miller" <davem@davemloft.net> Cc: Len Brown <lenb@kernel.org> Link: http://lkml.kernel.org/r/20170412201042.958216363@linutronix.deSigned-off-by:
-
Thomas Gleixner authored
[ Upstream commit 8153f9ac ] acpi_processor_get_throttling() requires to invoke the getter function on the target CPU. This is achieved by temporarily setting the affinity of the calling user space thread to the requested CPU and reset it to the original affinity afterwards. That's racy vs. CPU hotplug and concurrent affinity settings for that thread resulting in code executing on the wrong CPU and overwriting the new affinity setting. acpi_processor_get_throttling() is invoked in two ways: 1) The CPU online callback, which is already running on the target CPU and obviously protected against hotplug and not affected by affinity settings. 2) The ACPI driver probe function, which is not protected against hotplug during modprobe. Switch it over to work_on_cpu() and protect the probe function against CPU hotplug. Signed-off-by:
-
Thomas Gleixner authored
[ Upstream commit a5cbdf69 ] When acpi_install_notify_handler() fails the cooling device stays registered and the sysfs files created via acpi_pss_perf_init() are leaked and the function returns success. Undo acpi_pss_perf_init() and return a proper error code. Signed-off-by:
-
Deepa Dinamani authored
[ Upstream commit d340266e ] struct timespec is not y2038 safe on 32 bit machines. The posix clocks apis use struct timespec directly and through struct itimerspec. Replace the posix clock interfaces to use struct timespec64 and struct itimerspec64 instead. Also fix up their implementations accordingly. Note that the clock_getres() interface has also been changed to use timespec64 even though this particular interface is not affected by the y2038 problem. This helps verification for internal kernel code for y2038 readiness by getting rid of time_t/ timeval/ timespec. Signed-off-by:
Deepa Dinamani <deepa.kernel@gmail.com> Cc: arnd@arndb.de Cc: y2038@lists.linaro.org Cc: netdev@vger.kernel.org Cc: Richard Cochran <richardcochran@gmail.com> Cc: john.stultz@linaro.org Link: http://lkml.kernel.org/r/1490555058-4603-3-git-send-email-deepa.kernel@gmail.comSigned-off-by:
-
Dmitry Torokhov authored
[ Upstream commit 95123fc4 ] The name field in structure i2c_device_id is 20 characters, and we expect it to be NULL-terminated, however we are trying to stuff it with 21 bytes and thus NULL-terminator is lost. This causes issues when one creates device with name "MICROCHIP_AR1021_I2C" as i2c core cuts off the last "C", and automatic module loading by alias does not work as result. The -I2C suffix in the device name is superfluous, we know what bus we are dealing with, so let's drop it. Also, no other driver uses capitals, and the manufacturer name is normally not included, except in very rare cases of incompatible name collisions. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=116211 Fixes: dd4cae8b ("Input: Add Microchip AR1021 i2c touchscreen") Reviewed-By:
Christian Gmeiner <christian.gmeiner@gmail.com> Tested-by:
Martin Kepplinger <martin.kepplinger@ginzinger.com> Signed-off-by:
Dmitry Torokhov <dmitry.torokhov@gmail.com> Signed-off-by:
-
Hans de Goede authored
[ Upstream commit a1e23a42 ] On some systems (e.g. Intel Bay Trail systems) the legacy PIC is not used, in this case virq 8 will be a random irq, rather then hw_irq 8 from the PIC. Requesting virq 8 in this case will not help us to get alarm irqs and may cause problems for other drivers which actually do need virq 8, for example on an Asus Transformer T100TA this leads to: [ 28.745155] genirq: Flags mismatch irq 8. 00000088 (mmc0) vs. 00000080 (rtc0) <snip oops> [ 28.753700] mmc0: Failed to request IRQ 8: -16 [ 28.975934] sdhci-acpi: probe of 80860F14:01 failed with error -16 This commit fixes this by making the rtc-cmos driver continue without using an irq rather then claiming irq 8 when no irq is specified in the pnp-info and there are no legacy-irqs. Signed-off-by:
Alexandre Belloni <alexandre.belloni@free-electrons.com> Signed-off-by:
-
Hans de Goede authored
[ Upstream commit 7ee06cb2 ] The classic PC rtc-coms driver has a workaround for broken ACPI device nodes for it which lack an irq resource. This workaround used to unconditionally hardcode the irq to 8 in these cases. This was causing irq conflict problems on systems without a legacy-pic so a recent patch added an if (nr_legacy_irqs()) guard to the workaround to avoid this irq conflict. nr_legacy_irqs() uses the legacy_pic symbol under the hood causing an undefined symbol error if the rtc-cmos code is build as a module. This commit exports the legacy_pic symbol to fix this. Cc: rtc-linux@googlegroups.com Cc: alexandre.belloni@free-electrons.com Signed-off-by:
-
Liam Breck authored
[ Upstream commit d63d07c6 ] If the charger is unplugged before the battery is full we may see an over/under voltage fault. Ignore this rather then emitting a message or uevent. This fixes messages like these getting logged on charger unplug + replug: bq24190-charger 15-006b: Fault: boost 0, charge 1, battery 0, ntc 0 bq24190-charger 15-006b: Fault: boost 0, charge 0, battery 0, ntc 0 Cc: Tony Lindgren <tony@atomide.com> Cc: Hans de Goede <hdegoede@redhat.com> Signed-off-by:
Liam Breck <kernel@networkimprov.net> Acked-by:
Tony Lindgren <tony@atomide.com> Reviewed-by:
-
