platform/kernel/linux-rpi.git
4 years agob43: Fix connection problem with WPA3
Larry Finger [Tue, 26 May 2020 15:59:08 +0000 (10:59 -0500)]
b43: Fix connection problem with WPA3

commit 75d057bda1fbca6ade21378aa45db712e5f7d962 upstream.

Since the driver was first introduced into the kernel, it has only
handled the ciphers associated with WEP, WPA, and WPA2. It fails with
WPA3 even though mac80211 can handle those additional ciphers in software,
b43 did not report that it could handle them. By setting MFP_CAPABLE using
ieee80211_set_hw(), the problem is fixed.

With this change, b43 will handle the ciphers it knows in hardware,
and let mac80211 handle the others in software. It is not necessary to
use the module parameter NOHWCRYPT to turn hardware encryption off.
Although this change essentially eliminates that module parameter,
I am choosing to keep it for cases where the hardware is broken,
and software encryption is required for all ciphers.

Reported-and-tested-by: Rui Salvaterra <rsalvaterra@gmail.com>
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Cc: Stable <stable@vger.kernel.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200526155909.5807-2-Larry.Finger@lwfinger.net
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agob43legacy: Fix case where channel status is corrupted
Larry Finger [Tue, 7 Apr 2020 19:00:43 +0000 (14:00 -0500)]
b43legacy: Fix case where channel status is corrupted

commit ec4d3e3a054578de34cd0b587ab8a1ac36f629d9 upstream.

This patch fixes commit 75388acd0cd8 ("add mac80211-based driver for
legacy BCM43xx devices")

In https://bugzilla.kernel.org/show_bug.cgi?id=207093, a defect in
b43legacy is reported. Upon testing, thus problem exists on PPC and
X86 platforms, is present in the oldest kernel tested (3.2), and
has been present in the driver since it was first added to the kernel.

The problem is a corrupted channel status received from the device.
Both the internal card in a PowerBook G4 and the PCMCIA version
(Broadcom BCM4306 with PCI ID 14e4:4320) have the problem. Only Rev, 2
(revision 4 of the 802.11 core) of the chip has been tested. No other
devices using b43legacy are available for testing.

Various sources of the problem were considered. Buffer overrun and
other sources of corruption within the driver were rejected because
the faulty channel status is always the same, not a random value.
It was concluded that the faulty data is coming from the device, probably
due to a firmware bug. As that source is not available, the driver
must take appropriate action to recover.

At present, the driver reports the error, and them continues to process
the bad packet. This is believed that to be a mistake, and the correct
action is to drop the correpted packet.

Fixes: 75388acd0cd8 ("add mac80211-based driver for legacy BCM43xx devices")
Cc: Stable <stable@vger.kernel.org>
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Reported-and-tested by: F. Erhard <erhard_f@mailbox.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200407190043.1686-1-Larry.Finger@lwfinger.net
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agoBluetooth: hci_bcm: fix freeing not-requested IRQ
Michał Mirosław [Thu, 2 Apr 2020 12:55:20 +0000 (14:55 +0200)]
Bluetooth: hci_bcm: fix freeing not-requested IRQ

commit 81bd5d0c62437c02caac6b3f942fcda874063cb0 upstream.

When BT module can't be initialized, but it has an IRQ, unloading
the driver WARNs when trying to free not-yet-requested IRQ. Fix it by
noting whether the IRQ was requested.

WARNING: CPU: 2 PID: 214 at kernel/irq/devres.c:144 devm_free_irq+0x49/0x4ca
[...]
WARNING: CPU: 2 PID: 214 at kernel/irq/manage.c:1746 __free_irq+0x8b/0x27c
Trying to free already-free IRQ 264
Modules linked in: hci_uart(-) btbcm bluetooth ecdh_generic ecc libaes
CPU: 2 PID: 214 Comm: rmmod Tainted: G        W         5.6.1mq-00044-ga5f9ea098318-dirty #928
[...]
[<b016aefb>] (devm_free_irq) from [<af8ba1ff>] (bcm_close+0x97/0x118 [hci_uart])
[<af8ba1ff>] (bcm_close [hci_uart]) from [<af8b736f>] (hci_uart_unregister_device+0x33/0x3c [hci_uart])
[<af8b736f>] (hci_uart_unregister_device [hci_uart]) from [<b035930b>] (serdev_drv_remove+0x13/0x20)
[<b035930b>] (serdev_drv_remove) from [<b037093b>] (device_release_driver_internal+0x97/0x118)
[<b037093b>] (device_release_driver_internal) from [<b0370a0b>] (driver_detach+0x2f/0x58)
[<b0370a0b>] (driver_detach) from [<b036f855>] (bus_remove_driver+0x41/0x94)
[<b036f855>] (bus_remove_driver) from [<af8ba8db>] (bcm_deinit+0x1b/0x740 [hci_uart])
[<af8ba8db>] (bcm_deinit [hci_uart]) from [<af8ba86f>] (hci_uart_exit+0x13/0x30 [hci_uart])
[<af8ba86f>] (hci_uart_exit [hci_uart]) from [<b01900bd>] (sys_delete_module+0x109/0x1d0)
[<b01900bd>] (sys_delete_module) from [<b0101001>] (ret_fast_syscall+0x1/0x5a)
[...]

Cc: stable@vger.kernel.org
Fixes: 6cc4396c8829 ("Bluetooth: hci_bcm: Add wake-up capability")
Signed-off-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agoserial: 8250: Avoid error message on reprobe
Lukas Wunner [Tue, 12 May 2020 12:40:01 +0000 (14:40 +0200)]
serial: 8250: Avoid error message on reprobe

commit e0a851fe6b9b619527bd928aa93caaddd003f70c upstream.

If the call to uart_add_one_port() in serial8250_register_8250_port()
fails, a half-initialized entry in the serial_8250ports[] array is left
behind.

A subsequent reprobe of the same serial port causes that entry to be
reused.  Because uart->port.dev is set, uart_remove_one_port() is called
for the half-initialized entry and bails out with an error message:

bcm2835-aux-uart 3f215040.serial: Removing wrong port: (null) != (ptrval)

The same happens on failure of mctrl_gpio_init() since commit
4a96895f74c9 ("tty/serial/8250: use mctrl_gpio helpers").

Fix by zeroing the uart->port.dev pointer in the probe error path.

The bug was introduced in v2.6.10 by historical commit befff6f5bf5f
("[SERIAL] Add new port registration/unregistration functions."):
https://git.kernel.org/tglx/history/c/befff6f5bf5f

The commit added an unconditional call to uart_remove_one_port() in
serial8250_register_port().  In v3.7, commit 835d844d1a28 ("8250_pnp:
do pnp probe before legacy probe") made that call conditional on
uart->port.dev which allows me to fix the issue by zeroing that pointer
in the error path.  Thus, the present commit will fix the problem as far
back as v3.7 whereas still older versions need to also cherry-pick
835d844d1a28.

Fixes: 835d844d1a28 ("8250_pnp: do pnp probe before legacy probe")
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Cc: stable@vger.kernel.org # v2.6.10
Cc: stable@vger.kernel.org # v2.6.10: 835d844d1a28: 8250_pnp: do pnp probe before legacy
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://lore.kernel.org/r/b4a072013ee1a1d13ee06b4325afb19bda57ca1b.1589285873.git.lukas@wunner.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agomedia: cedrus: Program output format during each run
Samuel Holland [Sat, 9 May 2020 20:06:42 +0000 (22:06 +0200)]
media: cedrus: Program output format during each run

commit a8876c22eab9a871834f85de83e98bbf7e6e264d upstream.

Previously, the output format was programmed as part of the ioctl()
handler. However, this has two problems:

  1) If there are multiple active streams with different output
     formats, the hardware will use whichever format was set last
     for both streams. Similarly, an ioctl() done in an inactive
     context will wrongly affect other active contexts.
  2) The registers are written while the device is not actively
     streaming. To enable runtime PM tied to the streaming state,
     all hardware access needs to be moved inside cedrus_device_run().

The call to cedrus_dst_format_set() is now placed just before the
codec-specific callback that programs the hardware.

Cc: <stable@vger.kernel.org>
Fixes: 50e761516f2b ("media: platform: Add Cedrus VPU decoder driver")
Suggested-by: Jernej Skrabec <jernej.skrabec@siol.net>
Suggested-by: Paul Kocialkowski <paul.kocialkowski@bootlin.com>
Signed-off-by: Samuel Holland <samuel@sholland.org>
Tested-by: Jernej Skrabec <jernej.skrabec@siol.net>
Reviewed-by: Jernej Skrabec <jernej.skrabec@siol.net>
Reviewed-by: Ezequiel Garcia <ezequiel@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agomedia: go7007: fix a miss of snd_card_free
Chuhong Yuan [Tue, 10 Dec 2019 03:15:48 +0000 (04:15 +0100)]
media: go7007: fix a miss of snd_card_free

commit 9453264ef58638ce8976121ac44c07a3ef375983 upstream.

go7007_snd_init() misses a snd_card_free() in an error path.
Add the missed call to fix it.

Signed-off-by: Chuhong Yuan <hslester96@gmail.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Cc: Salvatore Bonaccorso <carnil@debian.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agocarl9170: remove P2P_GO support
Christian Lamparter [Tue, 5 May 2020 07:42:09 +0000 (10:42 +0300)]
carl9170: remove P2P_GO support

commit b14fba7ebd04082f7767a11daea7f12f3593de22 upstream.

This patch follows up on a bug-report by Frank Schäfer that
discovered P2P GO wasn't working with wpa_supplicant.
This patch removes part of the broken P2P GO support but
keeps the vif switchover code in place.

Cc: <stable@vger.kernel.org>
Link: <https://lkml.kernel.org/r/3a9d86b6-744f-e670-8792-9167257edef8@googlemail.com>
Reported-by: Frank Schäfer <fschaefer.oss@googlemail.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200425092811.9494-1-chunkeey@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agoe1000e: Relax condition to trigger reset for ME workaround
Punit Agrawal [Fri, 15 May 2020 04:31:27 +0000 (13:31 +0900)]
e1000e: Relax condition to trigger reset for ME workaround

commit d601afcae2febc49665008e9a79e701248d56c50 upstream.

It's an error if the value of the RX/TX tail descriptor does not match
what was written. The error condition is true regardless the duration
of the interference from ME. But the driver only performs the reset if
E1000_ICH_FWSM_PCIM2PCI_COUNT (2000) iterations of 50us delay have
transpired. The extra condition can lead to inconsistency between the
state of hardware as expected by the driver.

Fix this by dropping the check for number of delay iterations.

While at it, also make __ew32_prepare() static as it's not used
anywhere else.

CC: stable <stable@vger.kernel.org>
Signed-off-by: Punit Agrawal <punit1.agrawal@toshiba.co.jp>
Reviewed-by: Alexander Duyck <alexander.h.duyck@linux.intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agoe1000e: Disable TSO for buffer overrun workaround
Kai-Heng Feng [Thu, 7 May 2020 14:21:07 +0000 (22:21 +0800)]
e1000e: Disable TSO for buffer overrun workaround

commit f29801030ac67bf98b7a65d3aea67b30769d4f7c upstream.

Commit b10effb92e27 ("e1000e: fix buffer overrun while the I219 is
processing DMA transactions") imposes roughly 30% performance penalty.

The commit log states that "Disabling TSO eliminates performance loss
for TCP traffic without a noticeable impact on CPU performance", so
let's disable TSO by default to regain the loss.

CC: stable <stable@vger.kernel.org>
Fixes: b10effb92e27 ("e1000e: fix buffer overrun while the I219 is processing DMA transactions")
BugLink: https://bugs.launchpad.net/bugs/1802691
Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agoPCI: Program MPS for RCiEP devices
Ashok Raj [Fri, 27 Mar 2020 21:16:15 +0000 (14:16 -0700)]
PCI: Program MPS for RCiEP devices

commit aa0ce96d72dd2e1b0dfd0fb868f82876e7790878 upstream.

Root Complex Integrated Endpoints (RCiEPs) do not have an upstream bridge,
so pci_configure_mps() previously ignored them, which may result in reduced
performance.

Instead, program the Max_Payload_Size of RCiEPs to the maximum supported
value (unless it is limited for the PCIE_BUS_PEER2PEER case).  This also
affects the subsequent programming of Max_Read_Request_Size because Linux
programs MRRS based on the MPS value.

Fixes: 9dae3a97297f ("PCI: Move MPS configuration check to pci_configure_device()")
Link: https://lore.kernel.org/r/1585343775-4019-1-git-send-email-ashok.raj@intel.com
Tested-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Ashok Raj <ashok.raj@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: stable@vger.kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agoima: Set again build_ima_appraise variable
Krzysztof Struczynski [Mon, 27 Apr 2020 10:28:59 +0000 (12:28 +0200)]
ima: Set again build_ima_appraise variable

[ Upstream commit b59fda449cf07f2db3be3a67142e6c000f5e8d79 ]

After adding the new add_rule() function in commit c52657d93b05
("ima: refactor ima_init_policy()"), all appraisal flags are added to the
temp_ima_appraise variable. Revert to the previous behavior instead of
removing build_ima_appraise, to benefit from the protection offered by
__ro_after_init.

The mentioned commit introduced a bug, as it makes all the flags
modifiable, while build_ima_appraise flags can be protected with
__ro_after_init.

Cc: stable@vger.kernel.org # 5.0.x
Fixes: c52657d93b05 ("ima: refactor ima_init_policy()")
Co-developed-by: Roberto Sassu <roberto.sassu@huawei.com>
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Signed-off-by: Krzysztof Struczynski <krzysztof.struczynski@huawei.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoima: Remove redundant policy rule set in add_rules()
Krzysztof Struczynski [Mon, 27 Apr 2020 10:28:58 +0000 (12:28 +0200)]
ima: Remove redundant policy rule set in add_rules()

[ Upstream commit 6ee28442a465ab4c4be45e3b15015af24b1ba906 ]

Function ima_appraise_flag() returns the flag to be set in
temp_ima_appraise depending on the hook identifier passed as an argument.
It is not necessary to set the flag again for the POLICY_CHECK hook.

Signed-off-by: Krzysztof Struczynski <krzysztof.struczynski@huawei.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agox86/amd_nb: Add AMD family 17h model 60h PCI IDs
Alexander Monakov [Sun, 10 May 2020 20:48:40 +0000 (20:48 +0000)]
x86/amd_nb: Add AMD family 17h model 60h PCI IDs

[ Upstream commit a4e91825d7e1252f7cba005f1451e5464b23c15d ]

Add PCI IDs for AMD Renoir (4000-series Ryzen CPUs). This is necessary
to enable support for temperature sensors via the k10temp module.

Signed-off-by: Alexander Monakov <amonakov@ispras.ru>
Signed-off-by: Borislav Petkov <bp@suse.de>
Acked-by: Yazen Ghannam <yazen.ghannam@amd.com>
Acked-by: Guenter Roeck <linux@roeck-us.net>
Link: https://lkml.kernel.org/r/20200510204842.2603-2-amonakov@ispras.ru
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoserial: 8250_pci: Move Pericom IDs to pci_ids.h
Kai-Heng Feng [Fri, 8 May 2020 06:53:40 +0000 (14:53 +0800)]
serial: 8250_pci: Move Pericom IDs to pci_ids.h

[ Upstream commit 62a7f3009a460001eb46984395280dd900bc4ef4 ]

Move the IDs to pci_ids.h so it can be used by next patch.

Link: https://lore.kernel.org/r/20200508065343.32751-1-kai.heng.feng@canonical.com
Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: stable@vger.kernel.org
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoPCI: Add Loongson vendor ID
Tiezhu Yang [Tue, 10 Mar 2020 12:50:07 +0000 (20:50 +0800)]
PCI: Add Loongson vendor ID

[ Upstream commit 9acb9fe18d863aacc99948963f8d5d447dc311be ]

Add the Loongson vendor ID to pci_ids.h to be used by the controller
driver in the future.

The Loongson vendor ID can be found at the following link:
https://git.kernel.org/pub/scm/utils/pciutils/pciutils.git/tree/pci.ids

Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agox86/amd_nb: Add Family 19h PCI IDs
Yazen Ghannam [Fri, 10 Jan 2020 01:56:49 +0000 (01:56 +0000)]
x86/amd_nb: Add Family 19h PCI IDs

[ Upstream commit b3f79ae45904ae987a7c06a9e8d6084d7b73e67f ]

Add the new PCI Device 18h IDs for AMD Family 19h systems. Note that
Family 19h systems will not have a new PCI root device ID.

Signed-off-by: Yazen Ghannam <yazen.ghannam@amd.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Link: https://lkml.kernel.org/r/20200110015651.14887-4-Yazen.Ghannam@amd.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoPCI: vmd: Add device id for VMD device 8086:9A0B
Jon Derrick [Tue, 12 Nov 2019 12:47:53 +0000 (05:47 -0700)]
PCI: vmd: Add device id for VMD device 8086:9A0B

[ Upstream commit ec11e5c213cc20cac5e8310728b06793448b9f6d ]

This patch adds support for this VMD device which supports the bus
restriction mode.

Signed-off-by: Jon Derrick <jonathan.derrick@intel.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoPCI: Add ACS quirk for Intel Root Complex Integrated Endpoints
Ashok Raj [Thu, 28 May 2020 20:57:42 +0000 (13:57 -0700)]
PCI: Add ACS quirk for Intel Root Complex Integrated Endpoints

[ Upstream commit 3247bd10a4502a3075ce8e1c3c7d31ef76f193ce ]

All Intel platforms guarantee that all root complex implementations must
send transactions up to IOMMU for address translations. Hence for Intel
RCiEP devices, we can assume some ACS-type isolation even without an ACS
capability.

From the Intel VT-d spec, r3.1, sec 3.16 ("Root-Complex Peer to Peer
Considerations"):

  When DMA remapping is enabled, peer-to-peer requests through the
  Root-Complex must be handled as follows:

  - The input address in the request is translated (through first-level,
    second-level or nested translation) to a host physical address (HPA).
    The address decoding for peer addresses must be done only on the
    translated HPA. Hardware implementations are free to further limit
    peer-to-peer accesses to specific host physical address regions (or
    to completely disallow peer-forwarding of translated requests).

  - Since address translation changes the contents (address field) of
    the PCI Express Transaction Layer Packet (TLP), for PCI Express
    peer-to-peer requests with ECRC, the Root-Complex hardware must use
    the new ECRC (re-computed with the translated address) if it
    decides to forward the TLP as a peer request.

  - Root-ports, and multi-function root-complex integrated endpoints, may
    support additional peer-to-peer control features by supporting PCI
    Express Access Control Services (ACS) capability. Refer to ACS
    capability in PCI Express specifications for details.

Since Linux didn't give special treatment to allow this exception, certain
RCiEP MFD devices were grouped in a single IOMMU group. This doesn't permit
a single device to be assigned to a guest for instance.

In one vendor system: Device 14.x were grouped in a single IOMMU group.

  /sys/kernel/iommu_groups/5/devices/0000:00:14.0
  /sys/kernel/iommu_groups/5/devices/0000:00:14.2
  /sys/kernel/iommu_groups/5/devices/0000:00:14.3

After this patch:

  /sys/kernel/iommu_groups/5/devices/0000:00:14.0
  /sys/kernel/iommu_groups/5/devices/0000:00:14.2
  /sys/kernel/iommu_groups/6/devices/0000:00:14.3 <<< new group

14.0 and 14.2 are integrated devices, but legacy end points, whereas 14.3
was a PCIe-compliant RCiEP.

  00:14.3 Network controller: Intel Corporation Device 9df0 (rev 30)
    Capabilities: [40] Express (v2) Root Complex Integrated Endpoint, MSI 00

This permits assigning this device to a guest VM.

[bhelgaas: drop "Fixes" tag since this doesn't fix a bug in that commit]
Link: https://lore.kernel.org/r/1590699462-7131-1-git-send-email-ashok.raj@intel.com
Tested-by: Darrel Goeddel <dgoeddel@forcepoint.com>
Signed-off-by: Ashok Raj <ashok.raj@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Alex Williamson <alex.williamson@redhat.com>
Cc: stable@vger.kernel.org
Cc: Lu Baolu <baolu.lu@linux.intel.com>
Cc: Mark Scott <mscott@forcepoint.com>,
Cc: Romil Sharma <rsharma@forcepoint.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoPCI: Avoid FLR for AMD Starship USB 3.0
Kevin Buettner [Sun, 24 May 2020 07:35:29 +0000 (00:35 -0700)]
PCI: Avoid FLR for AMD Starship USB 3.0

[ Upstream commit 5727043c73fdfe04597971b5f3f4850d879c1f4f ]

The AMD Starship USB 3.0 host controller advertises Function Level Reset
support, but it apparently doesn't work.  Add a quirk to prevent use of FLR
on this device.

Without this quirk, when attempting to assign (pass through) an AMD
Starship USB 3.0 host controller to a guest OS, the system becomes
increasingly unresponsive over the course of several minutes, eventually
requiring a hard reset.  Shortly after attempting to start the guest, I see
these messages:

  vfio-pci 0000:05:00.3: not ready 1023ms after FLR; waiting
  vfio-pci 0000:05:00.3: not ready 2047ms after FLR; waiting
  vfio-pci 0000:05:00.3: not ready 4095ms after FLR; waiting
  vfio-pci 0000:05:00.3: not ready 8191ms after FLR; waiting

And then eventually:

  vfio-pci 0000:05:00.3: not ready 65535ms after FLR; giving up
  INFO: NMI handler (perf_event_nmi_handler) took too long to run: 0.000 msecs
  perf: interrupt took too long (642744 > 2500), lowering kernel.perf_event_max_sample_rate to 1000
  INFO: NMI handler (perf_event_nmi_handler) took too long to run: 82.270 msecs
  INFO: NMI handler (perf_event_nmi_handler) took too long to run: 680.608 msecs
  INFO: NMI handler (perf_event_nmi_handler) took too long to run: 100.952 msecs
  ...
  watchdog: BUG: soft lockup - CPU#3 stuck for 22s! [qemu-system-x86:7487]

Tested on a Micro-Star International Co., Ltd. MS-7C59/Creator TRX40
motherboard with an AMD Ryzen Threadripper 3970X.

Link: https://lore.kernel.org/r/20200524003529.598434ff@f31-4.lan
Signed-off-by: Kevin Buettner <kevinb@redhat.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoPCI: Avoid FLR for AMD Matisse HD Audio & USB 3.0
Marcos Scriven [Wed, 20 May 2020 23:23:30 +0000 (18:23 -0500)]
PCI: Avoid FLR for AMD Matisse HD Audio & USB 3.0

[ Upstream commit 0d14f06cd6657ba3446a5eb780672da487b068e7 ]

The AMD Matisse HD Audio & USB 3.0 devices advertise Function Level Reset
support, but hang when an FLR is triggered.

To reproduce the problem, attach the device to a VM, then detach and try to
attach again.

Rename the existing quirk_intel_no_flr(), which was not Intel-specific, to
quirk_no_flr(), and apply it to prevent the use of FLR on these AMD
devices.

Link: https://lore.kernel.org/r/CAAri2DpkcuQZYbT6XsALhx2e6vRqPHwtbjHYeiH7MNp4zmt1RA@mail.gmail.com
Signed-off-by: Marcos Scriven <marcos@scriven.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoPCI: Avoid Pericom USB controller OHCI/EHCI PME# defect
Kai-Heng Feng [Fri, 8 May 2020 06:53:41 +0000 (14:53 +0800)]
PCI: Avoid Pericom USB controller OHCI/EHCI PME# defect

[ Upstream commit 68f5fc4ea9ddf9f77720d568144219c4e6452cde ]

Both Pericom OHCI and EHCI devices advertise PME# support from all power
states:

  06:00.0 USB controller [0c03]: Pericom Semiconductor PI7C9X442SL USB OHCI Controller [12d8:400e] (rev 01) (prog-if 10 [OHCI])
    Subsystem: Pericom Semiconductor PI7C9X442SL USB OHCI Controller [12d8:400e]
    Capabilities: [80] Power Management version 3
      Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=375mA PME(D0+,D1+,D2+,D3hot+,D3cold+)

  06:00.2 USB controller [0c03]: Pericom Semiconductor PI7C9X442SL USB EHCI Controller [12d8:400f] (rev 01) (prog-if 20 [EHCI])
    Subsystem: Pericom Semiconductor PI7C9X442SL USB EHCI Controller [12d8:400f]
    Capabilities: [80] Power Management version 3
      Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=375mA PME(D0+,D1+,D2+,D3hot+,D3cold+)

But testing shows that it's unreliable: there is a 20% chance PME# won't be
asserted when a USB device is plugged.

Remove PME support for both devices to make USB plugging work reliably.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=205981
Link: https://lore.kernel.org/r/20200508065343.32751-2-kai.heng.feng@canonical.com
Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: stable@vger.kernel.org
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoext4: fix race between ext4_sync_parent() and rename()
Eric Biggers [Wed, 6 May 2020 18:31:40 +0000 (11:31 -0700)]
ext4: fix race between ext4_sync_parent() and rename()

commit 08adf452e628b0e2ce9a01048cfbec52353703d7 upstream.

'igrab(d_inode(dentry->d_parent))' without holding dentry->d_lock is
broken because without d_lock, d_parent can be concurrently changed due
to a rename().  Then if the old directory is immediately deleted, old
d_parent->inode can be NULL.  That causes a NULL dereference in igrab().

To fix this, use dget_parent() to safely grab a reference to the parent
dentry, which pins the inode.  This also eliminates the need to use
d_find_any_alias() other than for the initial inode, as we no longer
throw away the dentry at each step.

This is an extremely hard race to hit, but it is possible.  Adding a
udelay() in between the reads of ->d_parent and its ->d_inode makes it
reproducible on a no-journal filesystem using the following program:

    #include <fcntl.h>
    #include <unistd.h>

    int main()
    {
        if (fork()) {
            for (;;) {
                mkdir("dir1", 0700);
                int fd = open("dir1/file", O_RDWR|O_CREAT|O_SYNC);
                write(fd, "X", 1);
                close(fd);
            }
        } else {
            mkdir("dir2", 0700);
            for (;;) {
                rename("dir1/file", "dir2/file");
                rmdir("dir1");
            }
        }
    }

Fixes: d59729f4e794 ("ext4: fix races in ext4_sync_parent()")
Cc: stable@vger.kernel.org
Signed-off-by: Eric Biggers <ebiggers@google.com>
Link: https://lore.kernel.org/r/20200506183140.541194-1-ebiggers@kernel.org
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agoext4: fix error pointer dereference
Jeffle Xu [Thu, 23 Apr 2020 07:46:44 +0000 (15:46 +0800)]
ext4: fix error pointer dereference

commit 8418897f1bf87da0cb6936489d57a4320c32c0af upstream.

Don't pass error pointers to brelse().

commit 7159a986b420 ("ext4: fix some error pointer dereferences") has fixed
some cases, fix the remaining one case.

Once ext4_xattr_block_find()->ext4_sb_bread() failed, error pointer is
stored in @bs->bh, which will be passed to brelse() in the cleanup
routine of ext4_xattr_set_handle(). This will then cause a NULL panic
crash in __brelse().

BUG: unable to handle kernel NULL pointer dereference at 000000000000005b
RIP: 0010:__brelse+0x1b/0x50
Call Trace:
 ext4_xattr_set_handle+0x163/0x5d0
 ext4_xattr_set+0x95/0x110
 __vfs_setxattr+0x6b/0x80
 __vfs_setxattr_noperm+0x68/0x1b0
 vfs_setxattr+0xa0/0xb0
 setxattr+0x12c/0x1a0
 path_setxattr+0x8d/0xc0
 __x64_sys_setxattr+0x27/0x30
 do_syscall_64+0x60/0x250
 entry_SYSCALL_64_after_hwframe+0x49/0xbe

In this case, @bs->bh stores '-EIO' actually.

Fixes: fb265c9cb49e ("ext4: add ext4_sb_bread() to disambiguate ENOMEM cases")
Signed-off-by: Jeffle Xu <jefflexu@linux.alibaba.com>
Reviewed-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Cc: stable@kernel.org # 2.6.19
Reviewed-by: Ritesh Harjani <riteshh@linux.ibm.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/1587628004-95123-1-git-send-email-jefflexu@linux.alibaba.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agoext4: fix EXT_MAX_EXTENT/INDEX to check for zeroed eh_max
Harshad Shirwadkar [Tue, 21 Apr 2020 02:39:59 +0000 (19:39 -0700)]
ext4: fix EXT_MAX_EXTENT/INDEX to check for zeroed eh_max

commit c36a71b4e35ab35340facdd6964a00956b9fef0a upstream.

If eh->eh_max is 0, EXT_MAX_EXTENT/INDEX would evaluate to unsigned
(-1) resulting in illegal memory accesses. Although there is no
consistent repro, we see that generic/019 sometimes crashes because of
this bug.

Ran gce-xfstests smoke and verified that there were no regressions.

Signed-off-by: Harshad Shirwadkar <harshadshirwadkar@gmail.com>
Link: https://lore.kernel.org/r/20200421023959.20879-2-harshadshirwadkar@gmail.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Cc: stable@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agoevm: Fix possible memory leak in evm_calc_hmac_or_hash()
Roberto Sassu [Tue, 14 Apr 2020 08:01:31 +0000 (10:01 +0200)]
evm: Fix possible memory leak in evm_calc_hmac_or_hash()

commit 0c4395fb2aa77341269ea619c5419ea48171883f upstream.

Don't immediately return if the signature is portable and security.ima is
not present. Just set error so that memory allocated is freed before
returning from evm_calc_hmac_or_hash().

Fixes: 50b977481fce9 ("EVM: Add support for portable signature format")
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Cc: stable@vger.kernel.org
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agoima: Remove __init annotation from ima_pcrread()
Roberto Sassu [Sun, 7 Jun 2020 21:00:29 +0000 (23:00 +0200)]
ima: Remove __init annotation from ima_pcrread()

commit 8b8c704d913b0fe490af370631a4200e26334ec0 upstream.

Commit 6cc7c266e5b4 ("ima: Call ima_calc_boot_aggregate() in
ima_eventdigest_init()") added a call to ima_calc_boot_aggregate() so that
the digest can be recalculated for the boot_aggregate measurement entry if
the 'd' template field has been requested. For the 'd' field, only SHA1 and
MD5 digests are accepted.

Given that ima_eventdigest_init() does not have the __init annotation, all
functions called should not have it. This patch removes __init from
ima_pcrread().

Cc: stable@vger.kernel.org
Fixes:  6cc7c266e5b4 ("ima: Call ima_calc_boot_aggregate() in ima_eventdigest_init()")
Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agoima: Call ima_calc_boot_aggregate() in ima_eventdigest_init()
Roberto Sassu [Wed, 3 Jun 2020 15:08:21 +0000 (17:08 +0200)]
ima: Call ima_calc_boot_aggregate() in ima_eventdigest_init()

commit 6cc7c266e5b47d3cd2b5bb7fd3aac4e6bb2dd1d2 upstream.

If the template field 'd' is chosen and the digest to be added to the
measurement entry was not calculated with SHA1 or MD5, it is
recalculated with SHA1, by using the passed file descriptor. However, this
cannot be done for boot_aggregate, because there is no file descriptor.

This patch adds a call to ima_calc_boot_aggregate() in
ima_eventdigest_init(), so that the digest can be recalculated also for the
boot_aggregate entry.

Cc: stable@vger.kernel.org # 3.13.x
Fixes: 3ce1217d6cd5d ("ima: define template fields library and new helpers")
Reported-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agoima: Directly assign the ima_default_policy pointer to ima_rules
Roberto Sassu [Wed, 3 Jun 2020 15:08:20 +0000 (17:08 +0200)]
ima: Directly assign the ima_default_policy pointer to ima_rules

commit 067a436b1b0aafa593344fddd711a755a58afb3b upstream.

This patch prevents the following oops:

[   10.771813] BUG: kernel NULL pointer dereference, address: 0000000000000
[...]
[   10.779790] RIP: 0010:ima_match_policy+0xf7/0xb80
[...]
[   10.798576] Call Trace:
[   10.798993]  ? ima_lsm_policy_change+0x2b0/0x2b0
[   10.799753]  ? inode_init_owner+0x1a0/0x1a0
[   10.800484]  ? _raw_spin_lock+0x7a/0xd0
[   10.801592]  ima_must_appraise.part.0+0xb6/0xf0
[   10.802313]  ? ima_fix_xattr.isra.0+0xd0/0xd0
[   10.803167]  ima_must_appraise+0x4f/0x70
[   10.804004]  ima_post_path_mknod+0x2e/0x80
[   10.804800]  do_mknodat+0x396/0x3c0

It occurs when there is a failure during IMA initialization, and
ima_init_policy() is not called. IMA hooks still call ima_match_policy()
but ima_rules is NULL. This patch prevents the crash by directly assigning
the ima_default_policy pointer to ima_rules when ima_rules is defined. This
wouldn't alter the existing behavior, as ima_rules is always set at the end
of ima_init_policy().

Cc: stable@vger.kernel.org # 3.7.x
Fixes: 07f6a79415d7d ("ima: add appraise action keywords and default rules")
Reported-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agoima: Evaluate error in init_ima()
Roberto Sassu [Wed, 25 Mar 2020 10:47:07 +0000 (11:47 +0100)]
ima: Evaluate error in init_ima()

commit e144d6b265415ddbdc54b3f17f4f95133effa5a8 upstream.

Evaluate error in init_ima() before register_blocking_lsm_notifier() and
return if not zero.

Cc: stable@vger.kernel.org # 5.3.x
Fixes: b16942455193 ("ima: use the lsm policy update notifier")
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Reviewed-by: James Morris <jamorris@linux.microsoft.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agoima: Switch to ima_hash_algo for boot aggregate
Roberto Sassu [Wed, 25 Mar 2020 10:47:06 +0000 (11:47 +0100)]
ima: Switch to ima_hash_algo for boot aggregate

commit 6f1a1d103b48b1533a9c804e7a069e2c8e937ce7 upstream.

boot_aggregate is the first entry of IMA measurement list. Its purpose is
to link pre-boot measurements to IMA measurements. As IMA was designed to
work with a TPM 1.2, the SHA1 PCR bank was always selected even if a
TPM 2.0 with support for stronger hash algorithms is available.

This patch first tries to find a PCR bank with the IMA default hash
algorithm. If it does not find it, it selects the SHA256 PCR bank for
TPM 2.0 and SHA1 for TPM 1.2. Ultimately, it selects SHA1 also for TPM 2.0
if the SHA256 PCR bank is not found.

If none of the PCR banks above can be found, boot_aggregate file digest is
filled with zeros, as for TPM bypass, making it impossible to perform a
remote attestation of the system.

Cc: stable@vger.kernel.org # 5.1.x
Fixes: 879b589210a9 ("tpm: retrieve digest size of unknown algorithms with PCR read")
Reported-by: Jerry Snitselaar <jsnitsel@redhat.com>
Suggested-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agoima: Fix ima digest hash table key calculation
Krzysztof Struczynski [Tue, 28 Apr 2020 07:30:10 +0000 (09:30 +0200)]
ima: Fix ima digest hash table key calculation

commit 1129d31b55d509f15e72dc68e4b5c3a4d7b4da8d upstream.

Function hash_long() accepts unsigned long, while currently only one byte
is passed from ima_hash_key(), which calculates a key for ima_htable.

Given that hashing the digest does not give clear benefits compared to
using the digest itself, remove hash_long() and return the modulus
calculated on the first two bytes of the digest with the number of slots.
Also reduce the depth of the hash table by doubling the number of slots.

Cc: stable@vger.kernel.org
Fixes: 3323eec921ef ("integrity: IMA as an integrity service provider")
Co-developed-by: Roberto Sassu <roberto.sassu@huawei.com>
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Signed-off-by: Krzysztof Struczynski <krzysztof.struczynski@huawei.com>
Acked-by: David.Laight@aculab.com (big endian system concerns)
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agomm: call cond_resched() from deferred_init_memmap()
Pavel Tatashin [Wed, 3 Jun 2020 22:59:27 +0000 (15:59 -0700)]
mm: call cond_resched() from deferred_init_memmap()

commit da97f2d56bbd880b4138916a7ef96f9881a551b2 upstream.

Now that deferred pages are initialized with interrupts enabled we can
replace touch_nmi_watchdog() with cond_resched(), as it was before
3a2d7fa8a3d5.

For now, we cannot do the same in deferred_grow_zone() as it is still
initializes pages with interrupts disabled.

This change fixes RCU problem described in
https://lkml.kernel.org/r/20200401104156.11564-2-david@redhat.com

[   60.474005] rcu: INFO: rcu_sched detected stalls on CPUs/tasks:
[   60.475000] rcu:  1-...0: (0 ticks this GP) idle=02a/1/0x4000000000000000 softirq=1/1 fqs=15000
[   60.475000] rcu:  (detected by 0, t=60002 jiffies, g=-1199, q=1)
[   60.475000] Sending NMI from CPU 0 to CPUs 1:
[    1.760091] NMI backtrace for cpu 1
[    1.760091] CPU: 1 PID: 20 Comm: pgdatinit0 Not tainted 4.18.0-147.9.1.el8_1.x86_64 #1
[    1.760091] Hardware name: Red Hat KVM, BIOS 1.13.0-1.module+el8.2.0+5520+4e5817f3 04/01/2014
[    1.760091] RIP: 0010:__init_single_page.isra.65+0x10/0x4f
[    1.760091] Code: 48 83 cf 63 48 89 f8 0f 1f 40 00 48 89 c6 48 89 d7 e8 6b 18 80 ff 66 90 5b c3 31 c0 b9 10 00 00 00 49 89 f8 48 c1 e6 33 f3 ab <b8> 07 00 00 00 48 c1 e2 36 41 c7 40 34 01 00 00 00 48 c1 e0 33 41
[    1.760091] RSP: 0000:ffffba783123be40 EFLAGS: 00000006
[    1.760091] RAX: 0000000000000000 RBX: fffffad34405e300 RCX: 0000000000000000
[    1.760091] RDX: 0000000000000000 RSI: 0010000000000000 RDI: fffffad34405e340
[    1.760091] RBP: 0000000033f3177e R08: fffffad34405e300 R09: 0000000000000002
[    1.760091] R10: 000000000000002b R11: ffff98afb691a500 R12: 0000000000000002
[    1.760091] R13: 0000000000000000 R14: 000000003f03ea00 R15: 000000003e10178c
[    1.760091] FS:  0000000000000000(0000) GS:ffff9c9ebeb00000(0000) knlGS:0000000000000000
[    1.760091] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[    1.760091] CR2: 00000000ffffffff CR3: 000000a1cf20a001 CR4: 00000000003606e0
[    1.760091] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[    1.760091] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[    1.760091] Call Trace:
[    1.760091]  deferred_init_pages+0x8f/0xbf
[    1.760091]  deferred_init_memmap+0x184/0x29d
[    1.760091]  ? deferred_free_pages.isra.97+0xba/0xba
[    1.760091]  kthread+0x112/0x130
[    1.760091]  ? kthread_flush_work_fn+0x10/0x10
[    1.760091]  ret_from_fork+0x35/0x40
[   89.123011] node 0 initialised, 1055935372 pages in 88650ms

Fixes: 3a2d7fa8a3d5 ("mm: disable interrupts while initializing deferred pages")
Reported-by: Yiqian Wei <yiwei@redhat.com>
Signed-off-by: Pavel Tatashin <pasha.tatashin@soleen.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Tested-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Daniel Jordan <daniel.m.jordan@oracle.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: James Morris <jmorris@namei.org>
Cc: Kirill Tkhai <ktkhai@virtuozzo.com>
Cc: Sasha Levin <sashal@kernel.org>
Cc: Shile Zhang <shile.zhang@linux.alibaba.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: <stable@vger.kernel.org> [4.17+]
Link: http://lkml.kernel.org/r/20200403140952.17177-4-pasha.tatashin@soleen.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agomm/pagealloc.c: call touch_nmi_watchdog() on max order boundaries in deferred init
Daniel Jordan [Wed, 3 Jun 2020 22:59:20 +0000 (15:59 -0700)]
mm/pagealloc.c: call touch_nmi_watchdog() on max order boundaries in deferred init

commit 117003c32771df617acf66e140fbdbdeb0ac71f5 upstream.

Patch series "initialize deferred pages with interrupts enabled", v4.

Keep interrupts enabled during deferred page initialization in order to
make code more modular and allow jiffies to update.

Original approach, and discussion can be found here:
 http://lkml.kernel.org/r/20200311123848.118638-1-shile.zhang@linux.alibaba.com

This patch (of 3):

deferred_init_memmap() disables interrupts the entire time, so it calls
touch_nmi_watchdog() periodically to avoid soft lockup splats.  Soon it
will run with interrupts enabled, at which point cond_resched() should be
used instead.

deferred_grow_zone() makes the same watchdog calls through code shared
with deferred init but will continue to run with interrupts disabled, so
it can't call cond_resched().

Pull the watchdog calls up to these two places to allow the first to be
changed later, independently of the second.  The frequency reduces from
twice per pageblock (init and free) to once per max order block.

Fixes: 3a2d7fa8a3d5 ("mm: disable interrupts while initializing deferred pages")
Signed-off-by: Daniel Jordan <daniel.m.jordan@oracle.com>
Signed-off-by: Pavel Tatashin <pasha.tatashin@soleen.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: David Hildenbrand <david@redhat.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Shile Zhang <shile.zhang@linux.alibaba.com>
Cc: Kirill Tkhai <ktkhai@virtuozzo.com>
Cc: James Morris <jmorris@namei.org>
Cc: Sasha Levin <sashal@kernel.org>
Cc: Yiqian Wei <yiwei@redhat.com>
Cc: <stable@vger.kernel.org> [4.17+]
Link: http://lkml.kernel.org/r/20200403140952.17177-2-pasha.tatashin@soleen.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agomm: initialize deferred pages with interrupts enabled
Pavel Tatashin [Wed, 3 Jun 2020 22:59:24 +0000 (15:59 -0700)]
mm: initialize deferred pages with interrupts enabled

commit 3d060856adfc59afb9d029c233141334cfaba418 upstream.

Initializing struct pages is a long task and keeping interrupts disabled
for the duration of this operation introduces a number of problems.

1. jiffies are not updated for long period of time, and thus incorrect time
   is reported. See proposed solution and discussion here:
   lkml/20200311123848.118638-1-shile.zhang@linux.alibaba.com
2. It prevents farther improving deferred page initialization by allowing
   intra-node multi-threading.

We are keeping interrupts disabled to solve a rather theoretical problem
that was never observed in real world (See 3a2d7fa8a3d5).

Let's keep interrupts enabled. In case we ever encounter a scenario where
an interrupt thread wants to allocate large amount of memory this early in
boot we can deal with that by growing zone (see deferred_grow_zone()) by
the needed amount before starting deferred_init_memmap() threads.

Before:
[    1.232459] node 0 initialised, 12058412 pages in 1ms

After:
[    1.632580] node 0 initialised, 12051227 pages in 436ms

Fixes: 3a2d7fa8a3d5 ("mm: disable interrupts while initializing deferred pages")
Reported-by: Shile Zhang <shile.zhang@linux.alibaba.com>
Signed-off-by: Pavel Tatashin <pasha.tatashin@soleen.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Daniel Jordan <daniel.m.jordan@oracle.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: James Morris <jmorris@namei.org>
Cc: Kirill Tkhai <ktkhai@virtuozzo.com>
Cc: Sasha Levin <sashal@kernel.org>
Cc: Yiqian Wei <yiwei@redhat.com>
Cc: <stable@vger.kernel.org> [4.17+]
Link: http://lkml.kernel.org/r/20200403140952.17177-3-pasha.tatashin@soleen.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agomm: thp: make the THP mapcount atomic against __split_huge_pmd_locked()
Andrea Arcangeli [Wed, 27 May 2020 23:06:24 +0000 (19:06 -0400)]
mm: thp: make the THP mapcount atomic against __split_huge_pmd_locked()

commit c444eb564fb16645c172d550359cb3d75fe8a040 upstream.

Write protect anon page faults require an accurate mapcount to decide
if to break the COW or not. This is implemented in the THP path with
reuse_swap_page() ->
page_trans_huge_map_swapcount()/page_trans_huge_mapcount().

If the COW triggers while the other processes sharing the page are
under a huge pmd split, to do an accurate reading, we must ensure the
mapcount isn't computed while it's being transferred from the head
page to the tail pages.

reuse_swap_cache() already runs serialized by the page lock, so it's
enough to add the page lock around __split_huge_pmd_locked too, in
order to add the missing serialization.

Note: the commit in "Fixes" is just to facilitate the backporting,
because the code before such commit didn't try to do an accurate THP
mapcount calculation and it instead used the page_count() to decide if
to COW or not. Both the page_count and the pin_count are THP-wide
refcounts, so they're inaccurate if used in
reuse_swap_page(). Reverting such commit (besides the unrelated fix to
the local anon_vma assignment) would have also opened the window for
memory corruption side effects to certain workloads as documented in
such commit header.

Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
Suggested-by: Jann Horn <jannh@google.com>
Reported-by: Jann Horn <jannh@google.com>
Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Fixes: 6d0a07edd17c ("mm: thp: calculate the mapcount correctly for THP pages during WP faults")
Cc: stable@vger.kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agopowerpc/mm: Fix conditions to perform MMU specific management by blocks on PPC32.
Christophe Leroy [Tue, 19 May 2020 05:48:59 +0000 (05:48 +0000)]
powerpc/mm: Fix conditions to perform MMU specific management by blocks on PPC32.

commit 4e3319c23a66dabfd6c35f4d2633d64d99b68096 upstream.

Setting init mem to NX shall depend on sinittext being mapped by
block, not on stext being mapped by block.

Setting text and rodata to RO shall depend on stext being mapped by
block, not on sinittext being mapped by block.

Fixes: 63b2bc619565 ("powerpc/mm/32s: Use BATs for STRICT_KERNEL_RWX")
Cc: stable@vger.kernel.org
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/7d565fb8f51b18a3d98445a830b2f6548cb2da2a.1589866984.git.christophe.leroy@csgroup.eu
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agobtrfs: fix space_info bytes_may_use underflow during space cache writeout
Filipe Manana [Wed, 27 May 2020 10:16:19 +0000 (11:16 +0100)]
btrfs: fix space_info bytes_may_use underflow during space cache writeout

commit 2166e5edce9ac1edf3b113d6091ef72fcac2d6c4 upstream.

We always preallocate a data extent for writing a free space cache, which
causes writeback to always try the nocow path first, since the free space
inode has the prealloc bit set in its flags.

However if the block group that contains the data extent for the space
cache has been turned to RO mode due to a running scrub or balance for
example, we have to fallback to the cow path. In that case once a new data
extent is allocated we end up calling btrfs_add_reserved_bytes(), which
decrements the counter named bytes_may_use from the data space_info object
with the expection that this counter was previously incremented with the
same amount (the size of the data extent).

However when we started writeout of the space cache at cache_save_setup(),
we incremented the value of the bytes_may_use counter through a call to
btrfs_check_data_free_space() and then decremented it through a call to
btrfs_prealloc_file_range_trans() immediately after. So when starting the
writeback if we fallback to cow mode we have to increment the counter
bytes_may_use of the data space_info again to compensate for the extent
allocation done by the cow path.

When this issue happens we are incorrectly decrementing the bytes_may_use
counter and when its current value is smaller then the amount we try to
subtract we end up with the following warning:

 ------------[ cut here ]------------
 WARNING: CPU: 3 PID: 657 at fs/btrfs/space-info.h:115 btrfs_add_reserved_bytes+0x3d6/0x4e0 [btrfs]
 Modules linked in: btrfs blake2b_generic xor raid6_pq libcrc32c (...)
 CPU: 3 PID: 657 Comm: kworker/u8:7 Tainted: G        W         5.6.0-rc7-btrfs-next-58 #5
 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.0-59-gc9ba5276e321-prebuilt.qemu.org 04/01/2014
 Workqueue: writeback wb_workfn (flush-btrfs-1591)
 RIP: 0010:btrfs_add_reserved_bytes+0x3d6/0x4e0 [btrfs]
 Code: ff ff 48 (...)
 RSP: 0000:ffffa41608f13660 EFLAGS: 00010287
 RAX: 0000000000001000 RBX: ffff9615b93ae400 RCX: 0000000000000000
 RDX: 0000000000000002 RSI: 0000000000000000 RDI: ffff9615b96ab410
 RBP: fffffffffffee000 R08: 0000000000000001 R09: 0000000000000000
 R10: ffff961585e62a40 R11: 0000000000000000 R12: ffff9615b96ab400
 R13: ffff9615a1a2a000 R14: 0000000000012000 R15: ffff9615b93ae400
 FS:  0000000000000000(0000) GS:ffff9615bb200000(0000) knlGS:0000000000000000
 CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
 CR2: 000055cbbc2ae178 CR3: 0000000115794006 CR4: 00000000003606e0
 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
 Call Trace:
  find_free_extent+0x4a0/0x16c0 [btrfs]
  btrfs_reserve_extent+0x91/0x180 [btrfs]
  cow_file_range+0x12d/0x490 [btrfs]
  btrfs_run_delalloc_range+0x9f/0x6d0 [btrfs]
  ? find_lock_delalloc_range+0x221/0x250 [btrfs]
  writepage_delalloc+0xe8/0x150 [btrfs]
  __extent_writepage+0xe8/0x4c0 [btrfs]
  extent_write_cache_pages+0x237/0x530 [btrfs]
  extent_writepages+0x44/0xa0 [btrfs]
  do_writepages+0x23/0x80
  __writeback_single_inode+0x59/0x700
  writeback_sb_inodes+0x267/0x5f0
  __writeback_inodes_wb+0x87/0xe0
  wb_writeback+0x382/0x590
  ? wb_workfn+0x4a2/0x6c0
  wb_workfn+0x4a2/0x6c0
  process_one_work+0x26d/0x6a0
  worker_thread+0x4f/0x3e0
  ? process_one_work+0x6a0/0x6a0
  kthread+0x103/0x140
  ? kthread_create_worker_on_cpu+0x70/0x70
  ret_from_fork+0x3a/0x50
 irq event stamp: 0
 hardirqs last  enabled at (0): [<0000000000000000>] 0x0
 hardirqs last disabled at (0): [<ffffffffb2abdedf>] copy_process+0x74f/0x2020
 softirqs last  enabled at (0): [<ffffffffb2abdedf>] copy_process+0x74f/0x2020
 softirqs last disabled at (0): [<0000000000000000>] 0x0
 ---[ end trace bd7c03622e0b0a52 ]---
 ------------[ cut here ]------------

So fix this by incrementing the bytes_may_use counter of the data
space_info when we fallback to the cow path. If the cow path is successful
the counter is decremented after extent allocation (by
btrfs_add_reserved_bytes()), if it fails it ends up being decremented as
well when clearing the delalloc range (extent_clear_unlock_delalloc()).

This could be triggered sporadically by the test case btrfs/061 from
fstests.

Fixes: 82d5902d9c681b ("Btrfs: Support reading/writing on disk free ino cache")
CC: stable@vger.kernel.org # 4.4+
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agobtrfs: fix space_info bytes_may_use underflow after nocow buffered write
Filipe Manana [Wed, 27 May 2020 10:16:07 +0000 (11:16 +0100)]
btrfs: fix space_info bytes_may_use underflow after nocow buffered write

commit 467dc47ea99c56e966e99d09dae54869850abeeb upstream.

When doing a buffered write we always try to reserve data space for it,
even when the file has the NOCOW bit set or the write falls into a file
range covered by a prealloc extent. This is done both because it is
expensive to check if we can do a nocow write (checking if an extent is
shared through reflinks or if there's a hole in the range for example),
and because when writeback starts we might actually need to fallback to
COW mode (for example the block group containing the target extents was
turned into RO mode due to a scrub or balance).

When we are unable to reserve data space we check if we can do a nocow
write, and if we can, we proceed with dirtying the pages and setting up
the range for delalloc. In this case the bytes_may_use counter of the
data space_info object is not incremented, unlike in the case where we
are able to reserve data space (done through btrfs_check_data_free_space()
which calls btrfs_alloc_data_chunk_ondemand()).

Later when running delalloc we attempt to start writeback in nocow mode
but we might revert back to cow mode, for example because in the meanwhile
a block group was turned into RO mode by a scrub or relocation. The cow
path after successfully allocating an extent ends up calling
btrfs_add_reserved_bytes(), which expects the bytes_may_use counter of
the data space_info object to have been incremented before - but we did
not do it when the buffered write started, since there was not enough
available data space. So btrfs_add_reserved_bytes() ends up decrementing
the bytes_may_use counter anyway, and when the counter's current value
is smaller then the size of the allocated extent we get a stack trace
like the following:

 ------------[ cut here ]------------
 WARNING: CPU: 0 PID: 20138 at fs/btrfs/space-info.h:115 btrfs_add_reserved_bytes+0x3d6/0x4e0 [btrfs]
 Modules linked in: btrfs blake2b_generic xor raid6_pq libcrc32c (...)
 CPU: 0 PID: 20138 Comm: kworker/u8:15 Not tainted 5.6.0-rc7-btrfs-next-58 #5
 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.0-59-gc9ba5276e321-prebuilt.qemu.org 04/01/2014
 Workqueue: writeback wb_workfn (flush-btrfs-1754)
 RIP: 0010:btrfs_add_reserved_bytes+0x3d6/0x4e0 [btrfs]
 Code: ff ff 48 (...)
 RSP: 0018:ffffbda18a4b3568 EFLAGS: 00010287
 RAX: 0000000000000000 RBX: ffff9ca076f5d800 RCX: 0000000000000000
 RDX: 0000000000000002 RSI: 0000000000000000 RDI: ffff9ca068470410
 RBP: fffffffffffff000 R08: 0000000000000001 R09: 0000000000000000
 R10: ffff9ca079d58040 R11: 0000000000000000 R12: ffff9ca068470400
 R13: ffff9ca0408b2000 R14: 0000000000001000 R15: ffff9ca076f5d800
 FS:  0000000000000000(0000) GS:ffff9ca07a600000(0000) knlGS:0000000000000000
 CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
 CR2: 00005605dbfe7048 CR3: 0000000138570006 CR4: 00000000003606f0
 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
 Call Trace:
  find_free_extent+0x4a0/0x16c0 [btrfs]
  btrfs_reserve_extent+0x91/0x180 [btrfs]
  cow_file_range+0x12d/0x490 [btrfs]
  run_delalloc_nocow+0x341/0xa40 [btrfs]
  btrfs_run_delalloc_range+0x1ea/0x6d0 [btrfs]
  ? find_lock_delalloc_range+0x221/0x250 [btrfs]
  writepage_delalloc+0xe8/0x150 [btrfs]
  __extent_writepage+0xe8/0x4c0 [btrfs]
  extent_write_cache_pages+0x237/0x530 [btrfs]
  ? btrfs_wq_submit_bio+0x9f/0xc0 [btrfs]
  extent_writepages+0x44/0xa0 [btrfs]
  do_writepages+0x23/0x80
  __writeback_single_inode+0x59/0x700
  writeback_sb_inodes+0x267/0x5f0
  __writeback_inodes_wb+0x87/0xe0
  wb_writeback+0x382/0x590
  ? wb_workfn+0x4a2/0x6c0
  wb_workfn+0x4a2/0x6c0
  process_one_work+0x26d/0x6a0
  worker_thread+0x4f/0x3e0
  ? process_one_work+0x6a0/0x6a0
  kthread+0x103/0x140
  ? kthread_create_worker_on_cpu+0x70/0x70
  ret_from_fork+0x3a/0x50
 irq event stamp: 0
 hardirqs last  enabled at (0): [<0000000000000000>] 0x0
 hardirqs last disabled at (0): [<ffffffff94ebdedf>] copy_process+0x74f/0x2020
 softirqs last  enabled at (0): [<ffffffff94ebdedf>] copy_process+0x74f/0x2020
 softirqs last disabled at (0): [<0000000000000000>] 0x0
 ---[ end trace f9f6ef8ec4cd8ec9 ]---

So to fix this, when falling back into cow mode check if space was not
reserved, by testing for the bit EXTENT_NORESERVE in the respective file
range, and if not, increment the bytes_may_use counter for the data
space_info object. Also clear the EXTENT_NORESERVE bit from the range, so
that if the cow path fails it decrements the bytes_may_use counter when
clearing the delalloc range (through the btrfs_clear_delalloc_extent()
callback).

Fixes: 7ee9e4405f264e ("Btrfs: check if we can nocow if we don't have data space")
CC: stable@vger.kernel.org # 4.4+
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agobtrfs: fix wrong file range cleanup after an error filling dealloc range
Filipe Manana [Wed, 27 May 2020 10:15:53 +0000 (11:15 +0100)]
btrfs: fix wrong file range cleanup after an error filling dealloc range

commit e2c8e92d1140754073ad3799eb6620c76bab2078 upstream.

If an error happens while running dellaloc in COW mode for a range, we can
end up calling extent_clear_unlock_delalloc() for a range that goes beyond
our range's end offset by 1 byte, which affects 1 extra page. This results
in clearing bits and doing page operations (such as a page unlock) outside
our target range.

Fix that by calling extent_clear_unlock_delalloc() with an inclusive end
offset, instead of an exclusive end offset, at cow_file_range().

Fixes: a315e68f6e8b30 ("Btrfs: fix invalid attempt to free reserved space on failure to cow range")
CC: stable@vger.kernel.org # 4.14+
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agobtrfs: fix error handling when submitting direct I/O bio
Omar Sandoval [Thu, 16 Apr 2020 21:46:12 +0000 (14:46 -0700)]
btrfs: fix error handling when submitting direct I/O bio

commit 6d3113a193e3385c72240096fe397618ecab6e43 upstream.

In btrfs_submit_direct_hook(), if a direct I/O write doesn't span a RAID
stripe or chunk, we submit orig_bio without cloning it. In this case, we
don't increment pending_bios. Then, if btrfs_submit_dio_bio() fails, we
decrement pending_bios to -1, and we never complete orig_bio. Fix it by
initializing pending_bios to 1 instead of incrementing later.

Fixing this exposes another bug: we put orig_bio prematurely and then
put it again from end_io. Fix it by not putting orig_bio.

After this change, pending_bios is really more of a reference count, but
I'll leave that cleanup separate to keep the fix small.

Fixes: e65e15355429 ("btrfs: fix panic caused by direct IO")
CC: stable@vger.kernel.org # 4.4+
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agobtrfs: force chunk allocation if our global rsv is larger than metadata
Josef Bacik [Fri, 13 Mar 2020 19:28:48 +0000 (15:28 -0400)]
btrfs: force chunk allocation if our global rsv is larger than metadata

commit 9c343784c4328781129bcf9e671645f69fe4b38a upstream.

Nikolay noticed a bunch of test failures with my global rsv steal
patches.  At first he thought they were introduced by them, but they've
been failing for a while with 64k nodes.

The problem is with 64k nodes we have a global reserve that calculates
out to 13MiB on a freshly made file system, which only has 8MiB of
metadata space.  Because of changes I previously made we no longer
account for the global reserve in the overcommit logic, which means we
correctly allow overcommit to happen even though we are already
overcommitted.

However in some corner cases, for example btrfs/170, we will allocate
the entire file system up with data chunks before we have enough space
pressure to allocate a metadata chunk.  Then once the fs is full we
ENOSPC out because we cannot overcommit and the global reserve is taking
up all of the available space.

The most ideal way to deal with this is to change our space reservation
stuff to take into account the height of the tree's that we're
modifying, so that our global reserve calculation does not end up so
obscenely large.

However that is a huge undertaking.  Instead fix this by forcing a chunk
allocation if the global reserve is larger than the total metadata
space.  This gives us essentially the same behavior that happened
before, we get a chunk allocated and these tests can pass.

This is meant to be a stop-gap measure until we can tackle the "tree
height only" project.

Fixes: 0096420adb03 ("btrfs: do not account global reserve in can_overcommit")
CC: stable@vger.kernel.org # 5.4+
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Tested-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agobtrfs: send: emit file capabilities after chown
Marcos Paulo de Souza [Mon, 11 May 2020 02:15:07 +0000 (23:15 -0300)]
btrfs: send: emit file capabilities after chown

commit 89efda52e6b6930f80f5adda9c3c9edfb1397191 upstream.

Whenever a chown is executed, all capabilities of the file being touched
are lost.  When doing incremental send with a file with capabilities,
there is a situation where the capability can be lost on the receiving
side. The sequence of actions bellow shows the problem:

  $ mount /dev/sda fs1
  $ mount /dev/sdb fs2

  $ touch fs1/foo.bar
  $ setcap cap_sys_nice+ep fs1/foo.bar
  $ btrfs subvolume snapshot -r fs1 fs1/snap_init
  $ btrfs send fs1/snap_init | btrfs receive fs2

  $ chgrp adm fs1/foo.bar
  $ setcap cap_sys_nice+ep fs1/foo.bar

  $ btrfs subvolume snapshot -r fs1 fs1/snap_complete
  $ btrfs subvolume snapshot -r fs1 fs1/snap_incremental

  $ btrfs send fs1/snap_complete | btrfs receive fs2
  $ btrfs send -p fs1/snap_init fs1/snap_incremental | btrfs receive fs2

At this point, only a chown was emitted by "btrfs send" since only the
group was changed. This makes the cap_sys_nice capability to be dropped
from fs2/snap_incremental/foo.bar

To fix that, only emit capabilities after chown is emitted. The current
code first checks for xattrs that are new/changed, emits them, and later
emit the chown. Now, __process_new_xattr skips capabilities, letting
only finish_inode_if_needed to emit them, if they exist, for the inode
being processed.

This behavior was being worked around in "btrfs receive" side by caching
the capability and only applying it after chown. Now, xattrs are only
emmited _after_ chown, making that workaround not needed anymore.

Link: https://github.com/kdave/btrfs-progs/issues/202
CC: stable@vger.kernel.org # 4.4+
Suggested-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Marcos Paulo de Souza <mpdesouza@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agobtrfs: include non-missing as a qualifier for the latest_bdev
Anand Jain [Mon, 4 May 2020 18:58:25 +0000 (02:58 +0800)]
btrfs: include non-missing as a qualifier for the latest_bdev

commit 998a0671961f66e9fad4990ed75f80ba3088c2f1 upstream.

btrfs_free_extra_devids() updates fs_devices::latest_bdev to point to
the bdev with greatest device::generation number.  For a typical-missing
device the generation number is zero so fs_devices::latest_bdev will
never point to it.

But if the missing device is due to alienation [1], then
device::generation is not zero and if it is greater or equal to the rest
of device  generations in the list, then fs_devices::latest_bdev ends up
pointing to the missing device and reports the error like [2].

[1] We maintain devices of a fsid (as in fs_device::fsid) in the
fs_devices::devices list, a device is considered as an alien device
if its fsid does not match with the fs_device::fsid

Consider a working filesystem with raid1:

  $ mkfs.btrfs -f -d raid1 -m raid1 /dev/sda /dev/sdb
  $ mount /dev/sda /mnt-raid1
  $ umount /mnt-raid1

While mnt-raid1 was unmounted the user force-adds one of its devices to
another btrfs filesystem:

  $ mkfs.btrfs -f /dev/sdc
  $ mount /dev/sdc /mnt-single
  $ btrfs dev add -f /dev/sda /mnt-single

Now the original mnt-raid1 fails to mount in degraded mode, because
fs_devices::latest_bdev is pointing to the alien device.

  $ mount -o degraded /dev/sdb /mnt-raid1

[2]
mount: wrong fs type, bad option, bad superblock on /dev/sdb,
       missing codepage or helper program, or other error

       In some cases useful info is found in syslog - try
       dmesg | tail or so.

  kernel: BTRFS warning (device sdb): devid 1 uuid 072a0192-675b-4d5a-8640-a5cf2b2c704d is missing
  kernel: BTRFS error (device sdb): failed to read devices
  kernel: BTRFS error (device sdb): open_ctree failed

Fix the root cause by checking if the device is not missing before it
can be considered for the fs_devices::latest_bdev.

CC: stable@vger.kernel.org # 4.19+
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Anand Jain <anand.jain@oracle.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agobtrfs: free alien device after device add
Anand Jain [Mon, 4 May 2020 18:58:26 +0000 (02:58 +0800)]
btrfs: free alien device after device add

commit 7f551d969037cc128eca60688d9c5a300d84e665 upstream.

When an old device has new fsid through 'btrfs device add -f <dev>' our
fs_devices list has an alien device in one of the fs_devices lists.

By having an alien device in fs_devices, we have two issues so far

1. missing device does not not show as missing in the userland

2. degraded mount will fail

Both issues are caused by the fact that there's an alien device in the
fs_devices list. (Alien means that it does not belong to the filesystem,
identified by fsid, or does not contain btrfs filesystem at all, eg. due
to overwrite).

A device can be scanned/added through the control device ioctls
SCAN_DEV, DEVICES_READY or by ADD_DEV.

And device coming through the control device is checked against the all
other devices in the lists, but this was not the case for ADD_DEV.

This patch fixes both issues above by removing the alien device.

CC: stable@vger.kernel.org # 5.4+
Signed-off-by: Anand Jain <anand.jain@oracle.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4 years agostring.h: fix incompatibility between FORTIFY_SOURCE and KASAN
Daniel Axtens [Wed, 3 Jun 2020 22:56:46 +0000 (15:56 -0700)]
string.h: fix incompatibility between FORTIFY_SOURCE and KASAN

[ Upstream commit 47227d27e2fcb01a9e8f5958d8997cf47a820afc ]

The memcmp KASAN self-test fails on a kernel with both KASAN and
FORTIFY_SOURCE.

When FORTIFY_SOURCE is on, a number of functions are replaced with
fortified versions, which attempt to check the sizes of the operands.
However, these functions often directly invoke __builtin_foo() once they
have performed the fortify check.  Using __builtins may bypass KASAN
checks if the compiler decides to inline it's own implementation as
sequence of instructions, rather than emit a function call that goes out
to a KASAN-instrumented implementation.

Why is only memcmp affected?
============================

Of the string and string-like functions that kasan_test tests, only memcmp
is replaced by an inline sequence of instructions in my testing on x86
with gcc version 9.2.1 20191008 (Ubuntu 9.2.1-9ubuntu2).

I believe this is due to compiler heuristics.  For example, if I annotate
kmalloc calls with the alloc_size annotation (and disable some fortify
compile-time checking!), the compiler will replace every memset except the
one in kmalloc_uaf_memset with inline instructions.  (I have some WIP
patches to add this annotation.)

Does this affect other functions in string.h?
=============================================

Yes. Anything that uses __builtin_* rather than __real_* could be
affected. This looks like:

 - strncpy
 - strcat
 - strlen
 - strlcpy maybe, under some circumstances?
 - strncat under some circumstances
 - memset
 - memcpy
 - memmove
 - memcmp (as noted)
 - memchr
 - strcpy

Whether a function call is emitted always depends on the compiler.  Most
bugs should get caught by FORTIFY_SOURCE, but the missed memcmp test shows
that this is not always the case.

Isn't FORTIFY_SOURCE disabled with KASAN?
========================================-

The string headers on all arches supporting KASAN disable fortify with
kasan, but only when address sanitisation is _also_ disabled.  For example
from x86:

 #if defined(CONFIG_KASAN) && !defined(__SANITIZE_ADDRESS__)
 /*
  * For files that are not instrumented (e.g. mm/slub.c) we
  * should use not instrumented version of mem* functions.
  */
 #define memcpy(dst, src, len) __memcpy(dst, src, len)
 #define memmove(dst, src, len) __memmove(dst, src, len)
 #define memset(s, c, n) __memset(s, c, n)

 #ifndef __NO_FORTIFY
 #define __NO_FORTIFY /* FORTIFY_SOURCE uses __builtin_memcpy, etc. */
 #endif

 #endif

This comes from commit 6974f0c4555e ("include/linux/string.h: add the
option of fortified string.h functions"), and doesn't work when KASAN is
enabled and the file is supposed to be sanitised - as with test_kasan.c

I'm pretty sure this is not wrong, but not as expansive it should be:

 * we shouldn't use __builtin_memcpy etc in files where we don't have
   instrumentation - it could devolve into a function call to memcpy,
   which will be instrumented. Rather, we should use __memcpy which
   by convention is not instrumented.

 * we also shouldn't be using __builtin_memcpy when we have a KASAN
   instrumented file, because it could be replaced with inline asm
   that will not be instrumented.

What is correct behaviour?
==========================

Firstly, there is some overlap between fortification and KASAN: both
provide some level of _runtime_ checking. Only fortify provides
compile-time checking.

KASAN and fortify can pick up different things at runtime:

 - Some fortify functions, notably the string functions, could easily be
   modified to consider sub-object sizes (e.g. members within a struct),
   and I have some WIP patches to do this. KASAN cannot detect these
   because it cannot insert poision between members of a struct.

 - KASAN can detect many over-reads/over-writes when the sizes of both
   operands are unknown, which fortify cannot.

So there are a couple of options:

 1) Flip the test: disable fortify in santised files and enable it in
    unsanitised files. This at least stops us missing KASAN checking, but
    we lose the fortify checking.

 2) Make the fortify code always call out to real versions. Do this only
    for KASAN, for fear of losing the inlining opportunities we get from
    __builtin_*.

(We can't use kasan_check_{read,write}: because the fortify functions are
_extern inline_, you can't include _static_ inline functions without a
compiler warning. kasan_check_{read,write} are static inline so we can't
use them even when they would otherwise be suitable.)

Take approach 2 and call out to real versions when KASAN is enabled.

Use __underlying_foo to distinguish from __real_foo: __real_foo always
refers to the kernel's implementation of foo, __underlying_foo could be
either the kernel implementation or the __builtin_foo implementation.

This is sometimes enough to make the memcmp test succeed with
FORTIFY_SOURCE enabled. It is at least enough to get the function call
into the module. One more fix is needed to make it reliable: see the next
patch.

Fixes: 6974f0c4555e ("include/linux/string.h: add the option of fortified string.h functions")
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Tested-by: David Gow <davidgow@google.com>
Reviewed-by: Dmitry Vyukov <dvyukov@google.com>
Cc: Daniel Micay <danielmicay@gmail.com>
Cc: Andrey Ryabinin <aryabinin@virtuozzo.com>
Cc: Alexander Potapenko <glider@google.com>
Link: http://lkml.kernel.org/r/20200423154503.5103-3-dja@axtens.net
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agokasan: stop tests being eliminated as dead code with FORTIFY_SOURCE
Daniel Axtens [Wed, 3 Jun 2020 22:56:43 +0000 (15:56 -0700)]
kasan: stop tests being eliminated as dead code with FORTIFY_SOURCE

[ Upstream commit adb72ae1915db28f934e9e02c18bfcea2f3ed3b7 ]

Patch series "Fix some incompatibilites between KASAN and FORTIFY_SOURCE", v4.

3 KASAN self-tests fail on a kernel with both KASAN and FORTIFY_SOURCE:
memchr, memcmp and strlen.

When FORTIFY_SOURCE is on, a number of functions are replaced with
fortified versions, which attempt to check the sizes of the operands.
However, these functions often directly invoke __builtin_foo() once they
have performed the fortify check.  The compiler can detect that the
results of these functions are not used, and knows that they have no other
side effects, and so can eliminate them as dead code.

Why are only memchr, memcmp and strlen affected?
================================================

Of string and string-like functions, kasan_test tests:

 * strchr  ->  not affected, no fortified version
 * strrchr ->  likewise
 * strcmp  ->  likewise
 * strncmp ->  likewise

 * strnlen ->  not affected, the fortify source implementation calls the
               underlying strnlen implementation which is instrumented, not
               a builtin

 * strlen  ->  affected, the fortify souce implementation calls a __builtin
               version which the compiler can determine is dead.

 * memchr  ->  likewise
 * memcmp  ->  likewise

 * memset ->   not affected, the compiler knows that memset writes to its
       first argument and therefore is not dead.

Why does this not affect the functions normally?
================================================

In string.h, these functions are not marked as __pure, so the compiler
cannot know that they do not have side effects.  If relevant functions are
marked as __pure in string.h, we see the following warnings and the
functions are elided:

lib/test_kasan.c: In function `kasan_memchr':
lib/test_kasan.c:606:2: warning: statement with no effect [-Wunused-value]
  memchr(ptr, '1', size + 1);
  ^~~~~~~~~~~~~~~~~~~~~~~~~~
lib/test_kasan.c: In function `kasan_memcmp':
lib/test_kasan.c:622:2: warning: statement with no effect [-Wunused-value]
  memcmp(ptr, arr, size+1);
  ^~~~~~~~~~~~~~~~~~~~~~~~
lib/test_kasan.c: In function `kasan_strings':
lib/test_kasan.c:645:2: warning: statement with no effect [-Wunused-value]
  strchr(ptr, '1');
  ^~~~~~~~~~~~~~~~
...

This annotation would make sense to add and could be added at any point,
so the behaviour of test_kasan.c should change.

The fix
=======

Make all the functions that are pure write their results to a global,
which makes them live.  The strlen and memchr tests now pass.

The memcmp test still fails to trigger, which is addressed in the next
patch.

[dja@axtens.net: drop patch 3]
Link: http://lkml.kernel.org/r/20200424145521.8203-2-dja@axtens.net
Fixes: 0c96350a2d2f ("lib/test_kasan.c: add tests for several string/memory API functions")
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Tested-by: David Gow <davidgow@google.com>
Reviewed-by: Dmitry Vyukov <dvyukov@google.com>
Cc: Daniel Micay <danielmicay@gmail.com>
Cc: Andrey Ryabinin <aryabinin@virtuozzo.com>
Cc: Alexander Potapenko <glider@google.com>
Link: http://lkml.kernel.org/r/20200423154503.5103-1-dja@axtens.net
Link: http://lkml.kernel.org/r/20200423154503.5103-2-dja@axtens.net
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoselftests/bpf, flow_dissector: Close TAP device FD after the test
Jakub Sitnicki [Sun, 31 May 2020 08:28:44 +0000 (10:28 +0200)]
selftests/bpf, flow_dissector: Close TAP device FD after the test

[ Upstream commit b8215dce7dfd817ca38807f55165bf502146cd68 ]

test_flow_dissector leaves a TAP device after it's finished, potentially
interfering with other tests that will run after it. Fix it by closing the
TAP descriptor on cleanup.

Fixes: 0905beec9f52 ("selftests/bpf: run flow dissector tests in skb-less mode")
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-11-jakub@cloudflare.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agobpf: Fix running sk_skb program types with ktls
John Fastabend [Fri, 29 May 2020 23:06:59 +0000 (16:06 -0700)]
bpf: Fix running sk_skb program types with ktls

[ Upstream commit e91de6afa81c10e9f855c5695eb9a53168d96b73 ]

KTLS uses a stream parser to collect TLS messages and send them to
the upper layer tls receive handler. This ensures the tls receiver
has a full TLS header to parse when it is run. However, when a
socket has BPF_SK_SKB_STREAM_VERDICT program attached before KTLS
is enabled we end up with two stream parsers running on the same
socket.

The result is both try to run on the same socket. First the KTLS
stream parser runs and calls read_sock() which will tcp_read_sock
which in turn calls tcp_rcv_skb(). This dequeues the skb from the
sk_receive_queue. When this is done KTLS code then data_ready()
callback which because we stacked KTLS on top of the bpf stream
verdict program has been replaced with sk_psock_start_strp(). This
will in turn kick the stream parser again and eventually do the
same thing KTLS did above calling into tcp_rcv_skb() and dequeuing
a skb from the sk_receive_queue.

At this point the data stream is broke. Part of the stream was
handled by the KTLS side some other bytes may have been handled
by the BPF side. Generally this results in either missing data
or more likely a "Bad Message" complaint from the kTLS receive
handler as the BPF program steals some bytes meant to be in a
TLS header and/or the TLS header length is no longer correct.

We've already broke the idealized model where we can stack ULPs
in any order with generic callbacks on the TX side to handle this.
So in this patch we do the same thing but for RX side. We add
a sk_psock_strp_enabled() helper so TLS can learn a BPF verdict
program is running and add a tls_sw_has_ctx_rx() helper so BPF
side can learn there is a TLS ULP on the socket.

Then on BPF side we omit calling our stream parser to avoid
breaking the data stream for the KTLS receiver. Then on the
KTLS side we call BPF_SK_SKB_STREAM_VERDICT once the KTLS
receiver is done with the packet but before it posts the
msg to userspace. This gives us symmetry between the TX and
RX halfs and IMO makes it usable again. On the TX side we
process packets in this order BPF -> TLS -> TCP and on
the receive side in the reverse order TCP -> TLS -> BPF.

Discovered while testing OpenSSL 3.0 Alpha2.0 release.

Fixes: d829e9c4112b5 ("tls: convert to generic sk_msg interface")
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/159079361946.5745.605854335665044485.stgit@john-Precision-5820-Tower
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agobpf: Refactor sockmap redirect code so its easy to reuse
John Fastabend [Fri, 29 May 2020 23:06:41 +0000 (16:06 -0700)]
bpf: Refactor sockmap redirect code so its easy to reuse

[ Upstream commit ca2f5f21dbbd5e3a00cd3e97f728aa2ca0b2e011 ]

We will need this block of code called from tls context shortly
lets refactor the redirect logic so its easy to use. This also
cleans up the switch stmt so we have fewer fallthrough cases.

No logic changes are intended.

Fixes: d829e9c4112b5 ("tls: convert to generic sk_msg interface")
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/159079360110.5745.7024009076049029819.stgit@john-Precision-5820-Tower
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agobpf: Fix map permissions check
Anton Protopopov [Wed, 27 May 2020 18:56:59 +0000 (18:56 +0000)]
bpf: Fix map permissions check

[ Upstream commit 1ea0f9120c8ce105ca181b070561df5cbd6bc049 ]

The map_lookup_and_delete_elem() function should check for both FMODE_CAN_WRITE
and FMODE_CAN_READ permissions because it returns a map element to user space.

Fixes: bd513cd08f10 ("bpf: add MAP_LOOKUP_AND_DELETE_ELEM syscall")
Signed-off-by: Anton Protopopov <a.s.protopopov@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20200527185700.14658-5-a.s.protopopov@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agolibbpf: Fix perf_buffer__free() API for sparse allocs
Eelco Chaudron [Wed, 27 May 2020 08:42:00 +0000 (10:42 +0200)]
libbpf: Fix perf_buffer__free() API for sparse allocs

[ Upstream commit 601b05ca6edb0422bf6ce313fbfd55ec7bbbc0fd ]

In case the cpu_bufs are sparsely allocated they are not all
free'ed. These changes will fix this.

Fixes: fb84b8224655 ("libbpf: add perf buffer API")
Signed-off-by: Eelco Chaudron <echaudro@redhat.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Link: https://lore.kernel.org/bpf/159056888305.330763.9684536967379110349.stgit@ebuild
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoplatform/x86: asus_wmi: Reserve more space for struct bias_args
Chris Chiu [Fri, 22 May 2020 07:44:24 +0000 (15:44 +0800)]
platform/x86: asus_wmi: Reserve more space for struct bias_args

[ Upstream commit 7b91f1565fbfbe5a162d91f8a1f6c5580c2fc1d0 ]

On the ASUS laptop UX325JA/UX425JA, most of the media keys are not
working due to the ASUS WMI driver fails to be loaded. The ACPI error
as follows leads to the failure of asus_wmi_evaluate_method.
  ACPI BIOS Error (bug): AE_AML_BUFFER_LIMIT, Field [IIA3] at bit offset/length 96/32 exceeds size of target Buffer (96 bits) (20200326/dsopcode-203)
  No Local Variables are initialized for Method [WMNB]
  ACPI Error: Aborting method \_SB.ATKD.WMNB due to previous error (AE_AML_BUFFER_LIMIT) (20200326/psparse-531)

The DSDT for the WMNB part shows that 5 DWORD required for local
variables and the 3rd variable IIA3 hit the buffer limit.

Method (WMNB, 3, Serialized)
{ ..
    CreateDWordField (Arg2, Zero, IIA0)
    CreateDWordField (Arg2, 0x04, IIA1)
    CreateDWordField (Arg2, 0x08, IIA2)
    CreateDWordField (Arg2, 0x0C, IIA3)
    CreateDWordField (Arg2, 0x10, IIA4)
    Local0 = (Arg1 & 0xFFFFFFFF)
    If ((Local0 == 0x54494E49))
  ..
}

The limitation is determined by the input acpi_buffer size passed
to the wmi_evaluate_method. Since the struct bios_args is the data
structure used as input buffer by default for all ASUS WMI calls,
the size needs to be expanded to fix the problem.

Signed-off-by: Chris Chiu <chiu@endlessm.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoplatform/x86: intel-vbtn: Only blacklist SW_TABLET_MODE on the 9 / "Laptop" chasis...
Hans de Goede [Fri, 15 May 2020 18:39:16 +0000 (20:39 +0200)]
platform/x86: intel-vbtn: Only blacklist SW_TABLET_MODE on the 9 / "Laptop" chasis-type

[ Upstream commit cfae58ed681c5fe0185db843013ecc71cd265ebf ]

The HP Stream x360 11-p000nd no longer report SW_TABLET_MODE state / events
with recent kernels. This model reports a chassis-type of 10 / "Notebook"
which is not on the recently introduced chassis-type whitelist

Commit de9647efeaa9 ("platform/x86: intel-vbtn: Only activate tablet mode
switch on 2-in-1's") added a chassis-type whitelist and only listed 31 /
"Convertible" as being capable of generating valid SW_TABLET_MOD events.

Commit 1fac39fd0316 ("platform/x86: intel-vbtn: Also handle tablet-mode
switch on "Detachable" and "Portable" chassis-types") extended the
whitelist with chassis-types 8 / "Portable" and 32 / "Detachable".

And now we need to exten the whitelist again with 10 / "Notebook"...

The issue original fixed by the whitelist is really a ACPI DSDT bug on
the Dell XPS 9360 where it has a VGBS which reports it is in tablet mode
even though it is not a 2-in-1 at all, but a regular laptop.

So since this is a workaround for a DSDT issue on that specific model,
instead of extending the whitelist over and over again, lets switch to
a blacklist and only blacklist the chassis-type of the model for which
the chassis-type check was added.

Note this also fixes the current version of the code no longer checking
if dmi_get_system_info(DMI_CHASSIS_TYPE) returns NULL.

Fixes: 1fac39fd0316 ("platform/x86: intel-vbtn: Also handle tablet-mode switch on "Detachable" and "Portable" chassis-types")
Cc: Mario Limonciello <mario.limonciello@dell.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Mario Limonciello <Mario.limonciello@dell.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoplatform/x86: intel-hid: Add a quirk to support HP Spectre X2 (2015)
Nickolai Kozachenko [Sat, 30 May 2020 17:07:20 +0000 (22:07 +0500)]
platform/x86: intel-hid: Add a quirk to support HP Spectre X2 (2015)

[ Upstream commit 8fe63eb757ac6e661a384cc760792080bdc738dc ]

HEBC method reports capabilities of 5 button array but HP Spectre X2 (2015)
does not have this control method (the same was for Wacom MobileStudio Pro).
Expand previous DMI quirk by Alex Hung to also enable 5 button array
for this system.

Signed-off-by: Nickolai Kozachenko <daemongloom@gmail.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoplatform/x86: hp-wmi: Convert simple_strtoul() to kstrtou32()
Andy Shevchenko [Fri, 15 May 2020 13:27:04 +0000 (16:27 +0300)]
platform/x86: hp-wmi: Convert simple_strtoul() to kstrtou32()

[ Upstream commit 5cdc45ed3948042f0d73c6fec5ee9b59e637d0d2 ]

First of all, unsigned long can overflow u32 value on 64-bit machine.
Second, simple_strtoul() doesn't check for overflow in the input.

Convert simple_strtoul() to kstrtou32() to eliminate above issues.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agocpuidle: Fix three reference count leaks
Qiushi Wu [Thu, 28 May 2020 18:20:46 +0000 (13:20 -0500)]
cpuidle: Fix three reference count leaks

[ Upstream commit c343bf1ba5efcbf2266a1fe3baefec9cc82f867f ]

kobject_init_and_add() takes reference even when it fails.
If this function returns an error, kobject_put() must be called to
properly clean up the memory associated with the object.

Previous commit "b8eb718348b8" fixed a similar problem.

Signed-off-by: Qiushi Wu <wu000273@umn.edu>
[ rjw: Subject ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agospi: dw: Return any value retrieved from the dma_transfer callback
Serge Semin [Fri, 29 May 2020 13:11:51 +0000 (16:11 +0300)]
spi: dw: Return any value retrieved from the dma_transfer callback

[ Upstream commit f0410bbf7d0fb80149e3b17d11d31f5b5197873e ]

DW APB SSI DMA-part of the driver may need to perform the requested
SPI-transfer synchronously. In that case the dma_transfer() callback
will return 0 as a marker of the SPI transfer being finished so the
SPI core doesn't need to wait and may proceed with the SPI message
trasnfers pumping procedure. This will be needed to fix the problem
when DMA transactions are finished, but there is still data left in
the SPI Tx/Rx FIFOs being sent/received. But for now make dma_transfer
to return 1 as the normal dw_spi_transfer_one() method.

Signed-off-by: Serge Semin <Sergey.Semin@baikalelectronics.ru>
Cc: Georgy Vlasov <Georgy.Vlasov@baikalelectronics.ru>
Cc: Ramil Zaripov <Ramil.Zaripov@baikalelectronics.ru>
Cc: Alexey Malahov <Alexey.Malahov@baikalelectronics.ru>
Cc: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Cc: Feng Tang <feng.tang@intel.com>
Cc: Rob Herring <robh+dt@kernel.org>
Cc: linux-mips@vger.kernel.org
Cc: devicetree@vger.kernel.org
Link: https://lore.kernel.org/r/20200529131205.31838-3-Sergey.Semin@baikalelectronics.ru
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agommc: sdhci-esdhc-imx: fix the mask for tuning start point
Haibo Chen [Tue, 26 May 2020 10:22:01 +0000 (18:22 +0800)]
mmc: sdhci-esdhc-imx: fix the mask for tuning start point

[ Upstream commit 1194be8c949b8190b2882ad8335a5d98aa50c735 ]

According the RM, the bit[6~0] of register ESDHC_TUNING_CTRL is
TUNING_START_TAP, bit[7] of this register is to disable the command
CRC check for standard tuning. So fix it here.

Fixes: d87fc9663688 ("mmc: sdhci-esdhc-imx: support setting tuning start point")
Signed-off-by: Haibo Chen <haibo.chen@nxp.com>
Link: https://lore.kernel.org/r/1590488522-9292-1-git-send-email-haibo.chen@nxp.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoiwlwifi: mvm: fix aux station leak
Sharon [Fri, 29 May 2020 06:39:29 +0000 (09:39 +0300)]
iwlwifi: mvm: fix aux station leak

[ Upstream commit f327236df2afc8c3c711e7e070f122c26974f4da ]

When mvm is initialized we alloc aux station with aux queue.
We later free the station memory when driver is stopped, but we
never free the queue's memory, which casues a leak.

Add a proper de-initialization of the station.

Signed-off-by: Sharon <sara.sharon@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Link: https://lore.kernel.org/r/iwlwifi.20200529092401.0121c5be55e9.Id7516fbb3482131d0c9dfb51ff20b226617ddb49@changeid
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoixgbe: fix signed-integer-overflow warning
Xie XiuQi [Tue, 5 May 2020 02:45:21 +0000 (10:45 +0800)]
ixgbe: fix signed-integer-overflow warning

[ Upstream commit 3b70683fc4d68f5d915d9dc7e5ba72c732c7315c ]

ubsan report this warning, fix it by adding a unsigned suffix.

UBSAN: signed-integer-overflow in
drivers/net/ethernet/intel/ixgbe/ixgbe_common.c:2246:26
65535 * 65537 cannot be represented in type 'int'
CPU: 21 PID: 7 Comm: kworker/u256:0 Not tainted 5.7.0-rc3-debug+ #39
Hardware name: Huawei TaiShan 2280 V2/BC82AMDC, BIOS 2280-V2 03/27/2020
Workqueue: ixgbe ixgbe_service_task [ixgbe]
Call trace:
 dump_backtrace+0x0/0x3f0
 show_stack+0x28/0x38
 dump_stack+0x154/0x1e4
 ubsan_epilogue+0x18/0x60
 handle_overflow+0xf8/0x148
 __ubsan_handle_mul_overflow+0x34/0x48
 ixgbe_fc_enable_generic+0x4d0/0x590 [ixgbe]
 ixgbe_service_task+0xc20/0x1f78 [ixgbe]
 process_one_work+0x8f0/0xf18
 worker_thread+0x430/0x6d0
 kthread+0x218/0x238
 ret_from_fork+0x10/0x18

Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Xie XiuQi <xiexiuqi@huawei.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoice: fix potential double free in probe unrolling
Jacob Keller [Sat, 16 May 2020 00:42:24 +0000 (17:42 -0700)]
ice: fix potential double free in probe unrolling

[ Upstream commit bc3a024101ca497bea4c69be4054c32a5c349f1d ]

If ice_init_interrupt_scheme fails, ice_probe will jump to clearing up
the interrupts. This can lead to some static analysis tools such as the
compiler sanitizers complaining about double free problems.

Since ice_init_interrupt_scheme already unrolls internally on failure,
there is no need to call ice_clear_interrupt_scheme when it fails. Add
a new unroll label and use that instead.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agommc: via-sdmmc: Respect the cmd->busy_timeout from the mmc core
Ulf Hansson [Tue, 14 Apr 2020 16:14:10 +0000 (18:14 +0200)]
mmc: via-sdmmc: Respect the cmd->busy_timeout from the mmc core

[ Upstream commit 966244ccd2919e28f25555a77f204cd1c109cad8 ]

Using a fixed 1s timeout for all commands (and data transfers) is a bit
problematic.

For some commands it means waiting longer than needed for the timer to
expire, which may not a big issue, but still. For other commands, like for
an erase (CMD38) that uses a R1B response, may require longer timeouts than
1s. In these cases, we may end up treating the command as it failed, while
it just needed some more time to complete successfully.

Fix the problem by respecting the cmd->busy_timeout, which is provided by
the mmc core.

Cc: Bruce Chang <brucechang@via.com.tw>
Cc: Harald Welte <HaraldWelte@viatech.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/20200414161413.3036-17-ulf.hansson@linaro.org
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agostaging: greybus: sdio: Respect the cmd->busy_timeout from the mmc core
Ulf Hansson [Tue, 14 Apr 2020 16:14:13 +0000 (18:14 +0200)]
staging: greybus: sdio: Respect the cmd->busy_timeout from the mmc core

[ Upstream commit a389087ee9f195fcf2f31cd771e9ec5f02c16650 ]

Using a fixed 1s timeout for all commands is a bit problematic.

For some commands it means waiting longer than needed for the timeout to
expire, which may not a big issue, but still. For other commands, like for
an erase (CMD38) that uses a R1B response, may require longer timeouts than
1s. In these cases, we may end up treating the command as it failed, while
it just needed some more time to complete successfully.

Fix the problem by respecting the cmd->busy_timeout, which is provided by
the mmc core.

Cc: Rui Miguel Silva <rmfrfs@gmail.com>
Cc: Johan Hovold <johan@kernel.org>
Cc: Alex Elder <elder@kernel.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: greybus-dev@lists.linaro.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Acked-by: Rui Miguel Silva <rmfrfs@gmail.com>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Link: https://lore.kernel.org/r/20200414161413.3036-20-ulf.hansson@linaro.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agommc: sdhci-msm: Set SDHCI_QUIRK_MULTIBLOCK_READ_ACMD12 quirk
Veerabhadrarao Badiganti [Mon, 20 Apr 2020 06:20:24 +0000 (11:50 +0530)]
mmc: sdhci-msm: Set SDHCI_QUIRK_MULTIBLOCK_READ_ACMD12 quirk

[ Upstream commit d863cb03fb2aac07f017b2a1d923cdbc35021280 ]

sdhci-msm can support auto cmd12.
So enable SDHCI_QUIRK_MULTIBLOCK_READ_ACMD12 quirk.

Signed-off-by: Veerabhadrarao Badiganti <vbadigan@codeaurora.org>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/1587363626-20413-3-git-send-email-vbadigan@codeaurora.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agobcache: fix refcount underflow in bcache_device_free()
Coly Li [Wed, 27 May 2020 04:01:53 +0000 (12:01 +0800)]
bcache: fix refcount underflow in bcache_device_free()

[ Upstream commit 86da9f736740eba602389908574dfbb0f517baa5 ]

The problematic code piece in bcache_device_free() is,

 785 static void bcache_device_free(struct bcache_device *d)
 786 {
 787     struct gendisk *disk = d->disk;
 [snipped]
 799     if (disk) {
 800             if (disk->flags & GENHD_FL_UP)
 801                     del_gendisk(disk);
 802
 803             if (disk->queue)
 804                     blk_cleanup_queue(disk->queue);
 805
 806             ida_simple_remove(&bcache_device_idx,
 807                               first_minor_to_idx(disk->first_minor));
 808             put_disk(disk);
 809         }
 [snipped]
 816 }

At line 808, put_disk(disk) may encounter kobject refcount of 'disk'
being underflow.

Here is how to reproduce the issue,
- Attche the backing device to a cache device and do random write to
  make the cache being dirty.
- Stop the bcache device while the cache device has dirty data of the
  backing device.
- Only register the backing device back, NOT register cache device.
- The bcache device node /dev/bcache0 won't show up, because backing
  device waits for the cache device shows up for the missing dirty
  data.
- Now echo 1 into /sys/fs/bcache/pendings_cleanup, to stop the pending
  backing device.
- After the pending backing device stopped, use 'dmesg' to check kernel
  message, a use-after-free warning from KASA reported the refcount of
  kobject linked to the 'disk' is underflow.

The dropping refcount at line 808 in the above code piece is added by
add_disk(d->disk) in bch_cached_dev_run(). But in the above condition
the cache device is not registered, bch_cached_dev_run() has no chance
to be called and the refcount is not added. The put_disk() for a non-
added refcount of gendisk kobject triggers a underflow warning.

This patch checks whether GENHD_FL_UP is set in disk->flags, if it is
not set then the bcache device was not added, don't call put_disk()
and the the underflow issue can be avoided.

Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoMIPS: Fix IRQ tracing when call handle_fpe() and handle_msa_fpe()
YuanJunQing [Wed, 27 May 2020 06:11:30 +0000 (14:11 +0800)]
MIPS: Fix IRQ tracing when call handle_fpe() and handle_msa_fpe()

[ Upstream commit 31e1b3efa802f97a17628dde280006c4cee4ce5e ]

Register "a1" is unsaved in this function,
 when CONFIG_TRACE_IRQFLAGS is enabled,
 the TRACE_IRQS_OFF macro will call trace_hardirqs_off(),
 and this may change register "a1".
 The changed register "a1" as argument will be send
 to do_fpe() and do_msa_fpe().

Signed-off-by: YuanJunQing <yuanjunqing66@163.com>
Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoPCI: Don't disable decoding when mmio_always_on is set
Jiaxun Yang [Tue, 26 May 2020 09:21:12 +0000 (17:21 +0800)]
PCI: Don't disable decoding when mmio_always_on is set

[ Upstream commit b6caa1d8c80cb71b6162cb1f1ec13aa655026c9f ]

Don't disable MEM/IO decoding when a device have both non_compliant_bars
and mmio_always_on.

That would allow us quirk devices with junk in BARs but can't disable
their decoding.

Signed-off-by: Jiaxun Yang <jiaxun.yang@flygoat.com>
Acked-by: Bjorn Helgaas <helgaas@kernel.org>
Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agomacvlan: Skip loopback packets in RX handler
Alexander Sverdlin [Tue, 26 May 2020 12:27:51 +0000 (14:27 +0200)]
macvlan: Skip loopback packets in RX handler

[ Upstream commit 81f3dc9349ce0bf7b8447f147f45e70f0a5b36a6 ]

Ignore loopback-originatig packets soon enough and don't try to process L2
header where it doesn't exist. The very similar br_handle_frame() in bridge
code performs exactly the same check.

This is an example of such ICMPv6 packet:

skb len=96 headroom=40 headlen=96 tailroom=56
mac=(40,0) net=(40,40) trans=80
shinfo(txflags=0 nr_frags=0 gso(size=0 type=0 segs=0))
csum(0xae2e9a2f ip_summed=1 complete_sw=0 valid=0 level=0)
hash(0xc97ebd88 sw=1 l4=1) proto=0x86dd pkttype=5 iif=24
dev name=etha01.212 feat=0x0x0000000040005000
skb headroom: 00000000: 00 7c 86 52 84 88 ff ff 00 00 00 00 00 00 08 00
skb headroom: 00000010: 45 00 00 9e 5d 5c 40 00 40 11 33 33 00 00 00 01
skb headroom: 00000020: 02 40 43 80 00 00 86 dd
skb linear:   00000000: 60 09 88 bd 00 38 3a ff fe 80 00 00 00 00 00 00
skb linear:   00000010: 00 40 43 ff fe 80 00 00 ff 02 00 00 00 00 00 00
skb linear:   00000020: 00 00 00 00 00 00 00 01 86 00 61 00 40 00 00 2d
skb linear:   00000030: 00 00 00 00 00 00 00 00 03 04 40 e0 00 00 01 2c
skb linear:   00000040: 00 00 00 78 00 00 00 00 fd 5f 42 68 23 87 a8 81
skb linear:   00000050: 00 00 00 00 00 00 00 00 01 01 02 40 43 80 00 00
skb tailroom: 00000000: ...
skb tailroom: 00000010: ...
skb tailroom: 00000020: ...
skb tailroom: 00000030: ...

Call Trace, how it happens exactly:
 ...
 macvlan_handle_frame+0x321/0x425 [macvlan]
 ? macvlan_forward_source+0x110/0x110 [macvlan]
 __netif_receive_skb_core+0x545/0xda0
 ? enqueue_task_fair+0xe5/0x8e0
 ? __netif_receive_skb_one_core+0x36/0x70
 __netif_receive_skb_one_core+0x36/0x70
 process_backlog+0x97/0x140
 net_rx_action+0x1eb/0x350
 ? __hrtimer_run_queues+0x136/0x2e0
 __do_softirq+0xe3/0x383
 do_softirq_own_stack+0x2a/0x40
 </IRQ>
 do_softirq.part.4+0x4e/0x50
 netif_rx_ni+0x60/0xd0
 dev_loopback_xmit+0x83/0xf0
 ip6_finish_output2+0x575/0x590 [ipv6]
 ? ip6_cork_release.isra.1+0x64/0x90 [ipv6]
 ? __ip6_make_skb+0x38d/0x680 [ipv6]
 ? ip6_output+0x6c/0x140 [ipv6]
 ip6_output+0x6c/0x140 [ipv6]
 ip6_send_skb+0x1e/0x60 [ipv6]
 rawv6_sendmsg+0xc4b/0xe10 [ipv6]
 ? proc_put_long+0xd0/0xd0
 ? rw_copy_check_uvector+0x4e/0x110
 ? sock_sendmsg+0x36/0x40
 sock_sendmsg+0x36/0x40
 ___sys_sendmsg+0x2b6/0x2d0
 ? proc_dointvec+0x23/0x30
 ? addrconf_sysctl_forward+0x8d/0x250 [ipv6]
 ? dev_forward_change+0x130/0x130 [ipv6]
 ? _raw_spin_unlock+0x12/0x30
 ? proc_sys_call_handler.isra.14+0x9f/0x110
 ? __call_rcu+0x213/0x510
 ? get_max_files+0x10/0x10
 ? trace_hardirqs_on+0x2c/0xe0
 ? __sys_sendmsg+0x63/0xa0
 __sys_sendmsg+0x63/0xa0
 do_syscall_64+0x6c/0x1e0
 entry_SYSCALL_64_after_hwframe+0x49/0xbe

Signed-off-by: Alexander Sverdlin <alexander.sverdlin@nokia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agobtrfs: qgroup: mark qgroup inconsistent if we're inherting snapshot to a new qgroup
Qu Wenruo [Thu, 2 Apr 2020 06:37:35 +0000 (14:37 +0800)]
btrfs: qgroup: mark qgroup inconsistent if we're inherting snapshot to a new qgroup

[ Upstream commit cbab8ade585a18c4334b085564d9d046e01a3f70 ]

[BUG]
For the following operation, qgroup is guaranteed to be screwed up due
to snapshot adding to a new qgroup:

  # mkfs.btrfs -f $dev
  # mount $dev $mnt
  # btrfs qgroup en $mnt
  # btrfs subv create $mnt/src
  # xfs_io -f -c "pwrite 0 1m" $mnt/src/file
  # sync
  # btrfs qgroup create 1/0 $mnt/src
  # btrfs subv snapshot -i 1/0 $mnt/src $mnt/snapshot
  # btrfs qgroup show -prce $mnt/src
  qgroupid         rfer         excl     max_rfer     max_excl parent  child
  --------         ----         ----     --------     -------- ------  -----
  0/5          16.00KiB     16.00KiB         none         none ---     ---
  0/257         1.02MiB     16.00KiB         none         none ---     ---
  0/258         1.02MiB     16.00KiB         none         none 1/0     ---
  1/0             0.00B        0.00B         none         none ---     0/258
        ^^^^^^^^^^^^^^^^^^^^

[CAUSE]
The problem is in btrfs_qgroup_inherit(), we don't have good enough
check to determine if the new relation would break the existing
accounting.

Unlike btrfs_add_qgroup_relation(), which has proper check to determine
if we can do quick update without a rescan, in btrfs_qgroup_inherit() we
can even assign a snapshot to multiple qgroups.

[FIX]
Fix it by manually marking qgroup inconsistent for snapshot inheritance.

For subvolume creation, since all its extents are exclusively owned, we
don't need to rescan.

In theory, we should call relation check like quick_update_accounting()
when doing qgroup inheritance and inform user about qgroup accounting
inconsistency.

But we don't have good mechanism to relay that back to the user in the
snapshot creation context, thus we can only silently mark the qgroup
inconsistent.

Anyway, user shouldn't use qgroup inheritance during snapshot creation,
and should add qgroup relationship after snapshot creation by 'btrfs
qgroup assign', which has a much better UI to inform user about qgroup
inconsistent and kick in rescan automatically.

Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agobtrfs: improve global reserve stealing logic
Josef Bacik [Fri, 13 Mar 2020 19:58:05 +0000 (15:58 -0400)]
btrfs: improve global reserve stealing logic

[ Upstream commit 7f9fe614407692f670601a634621138233ac00d7 ]

For unlink transactions and block group removal
btrfs_start_transaction_fallback_global_rsv will first try to start an
ordinary transaction and if it fails it will fall back to reserving the
required amount by stealing from the global reserve. This is problematic
because of all the same reasons we had with previous iterations of the
ENOSPC handling, thundering herd.  We get a bunch of failures all at
once, everybody tries to allocate from the global reserve, some win and
some lose, we get an ENSOPC.

Fix this behavior by introducing BTRFS_RESERVE_FLUSH_ALL_STEAL. It's
used to mark unlink reservation. To fix this we need to integrate this
logic into the normal ENOSPC infrastructure.  We still go through all of
the normal flushing work, and at the moment we begin to fail all the
tickets we try to satisfy any tickets that are allowed to steal by
stealing from the global reserve.  If this works we start the flushing
system over again just like we would with a normal ticket satisfaction.
This serializes our global reserve stealing, so we don't have the
thundering herd problem.

Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Tested-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agom68k: mac: Don't call via_flush_cache() on Mac IIfx
Finn Thain [Wed, 20 May 2020 04:32:02 +0000 (14:32 +1000)]
m68k: mac: Don't call via_flush_cache() on Mac IIfx

[ Upstream commit bcc44f6b74106b31f0b0408b70305a40360d63b7 ]

There is no VIA2 chip on the Mac IIfx, so don't call via_flush_cache().
This avoids a boot crash which appeared in v5.4.

printk: console [ttyS0] enabled
printk: bootconsole [debug0] disabled
printk: bootconsole [debug0] disabled
Calibrating delay loop... 9.61 BogoMIPS (lpj=48064)
pid_max: default: 32768 minimum: 301
Mount-cache hash table entries: 1024 (order: 0, 4096 bytes, linear)
Mountpoint-cache hash table entries: 1024 (order: 0, 4096 bytes, linear)
devtmpfs: initialized
random: get_random_u32 called from bucket_table_alloc.isra.27+0x68/0x194 with crng_init=0
clocksource: jiffies: mask: 0xffffffff max_cycles: 0xffffffff, max_idle_ns: 19112604462750000 ns
futex hash table entries: 256 (order: -1, 3072 bytes, linear)
NET: Registered protocol family 16
Data read fault at 0x00000000 in Super Data (pc=0x8a6a)
BAD KERNEL BUSERR
Oops: 00000000
Modules linked in:
PC: [<00008a6a>] via_flush_cache+0x12/0x2c
SR: 2700  SP: 01c1fe3c  a2: 01c24000
d0: 00001119    d1: 0000000c    d2: 00012000    d3: 0000000f
d4: 01c06840    d5: 00033b92    a0: 00000000    a1: 00000000
Process swapper (pid: 1, task=01c24000)
Frame format=B ssw=0755 isc=0200 isb=fff7 daddr=00000000 dobuf=01c1fed0
baddr=00008a6e dibuf=0000004e ver=f
Stack from 01c1fec4:
        01c1fed0 00007d7e 00010080 01c1fedc 0000792e 00000001 01c1fef4 00006b40
        01c80000 00040000 00000006 00000003 01c1ff1c 004a545e 004ff200 00040000
        00000000 00000003 01c06840 00033b92 004a5410 004b6c88 01c1ff84 000021e2
        00000073 00000003 01c06840 00033b92 0038507a 004bb094 004b6ca8 004b6c88
        004b6ca4 004b6c88 000021ae 00020002 00000000 01c0685d 00000000 01c1ffb4
        0049f938 00409c85 01c06840 0045bd40 00000073 00000002 00000002 00000000
Call Trace: [<00007d7e>] mac_cache_card_flush+0x12/0x1c
 [<00010080>] fix_dnrm+0x2/0x18
 [<0000792e>] cache_push+0x46/0x5a
 [<00006b40>] arch_dma_prep_coherent+0x60/0x6e
 [<00040000>] switched_to_dl+0x76/0xd0
 [<004a545e>] dma_atomic_pool_init+0x4e/0x188
 [<00040000>] switched_to_dl+0x76/0xd0
 [<00033b92>] parse_args+0x0/0x370
 [<004a5410>] dma_atomic_pool_init+0x0/0x188
 [<000021e2>] do_one_initcall+0x34/0x1be
 [<00033b92>] parse_args+0x0/0x370
 [<0038507a>] strcpy+0x0/0x1e
 [<000021ae>] do_one_initcall+0x0/0x1be
 [<00020002>] do_proc_dointvec_conv+0x54/0x74
 [<0049f938>] kernel_init_freeable+0x126/0x190
 [<0049f94c>] kernel_init_freeable+0x13a/0x190
 [<004a5410>] dma_atomic_pool_init+0x0/0x188
 [<00041798>] complete+0x0/0x3c
 [<000b9b0c>] kfree+0x0/0x20a
 [<0038df98>] schedule+0x0/0xd0
 [<0038d604>] kernel_init+0x0/0xda
 [<0038d610>] kernel_init+0xc/0xda
 [<0038d604>] kernel_init+0x0/0xda
 [<00002d38>] ret_from_kernel_thread+0xc/0x14
Code: 0000 2079 0048 10da 2279 0048 10c8 d3c8 <1011> 0200 fff7 1280 d1f9 0048 10c8 1010 0000 0008 1080 4e5e 4e75 4e56 0000 2039
Disabling lock debugging due to kernel taint
Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b

Thanks to Stan Johnson for capturing the console log and running git
bisect.

Git bisect said commit 8e3a68fb55e0 ("dma-mapping: make
dma_atomic_pool_init self-contained") is the first "bad" commit. I don't
know why. Perhaps mach_l2_flush first became reachable with that commit.

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Reported-and-tested-by: Stan Johnson <userm57@yahoo.com>
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Cc: Joshua Thompson <funaho@jurai.org>
Link: https://lore.kernel.org/r/b8bbeef197d6b3898e82ed0d231ad08f575a4b34.1589949122.git.fthain@telegraphics.com.au
Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoMIPS: tools: Fix resource leak in elf-entry.c
Kaige Li [Thu, 14 May 2020 12:59:41 +0000 (20:59 +0800)]
MIPS: tools: Fix resource leak in elf-entry.c

[ Upstream commit f33a0b941017b9cb5a4e975af198b855b2f2b455 ]

There is a file descriptor resource leak in elf-entry.c, fix this
by adding fclose() before return and die.

Signed-off-by: Kaige Li <likaige@loongson.cn>
Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agox86/mm: Stop printing BRK addresses
Arvind Sankar [Sat, 29 Feb 2020 23:11:20 +0000 (18:11 -0500)]
x86/mm: Stop printing BRK addresses

[ Upstream commit 67d631b7c05eff955ccff4139327f0f92a5117e5 ]

This currently leaks kernel physical addresses into userspace.

Signed-off-by: Arvind Sankar <nivedita@alum.mit.edu>
Signed-off-by: Borislav Petkov <bp@suse.de>
Acked-by: Kees Cook <keescook@chromium.org>
Acked-by: Dave Hansen <dave.hansen@intel.com>
Link: https://lkml.kernel.org/r/20200229231120.1147527-1-nivedita@alum.mit.edu
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoselftests/bpf: CONFIG_IPV6_SEG6_BPF required for test_seg6_loop.o
Alan Maguire [Fri, 22 May 2020 11:36:28 +0000 (12:36 +0100)]
selftests/bpf: CONFIG_IPV6_SEG6_BPF required for test_seg6_loop.o

[ Upstream commit 3c8e8cf4b18b3a7034fab4c4504fc4b54e4b6195 ]

test_seg6_loop.o uses the helper bpf_lwt_seg6_adjust_srh();
it will not be present if CONFIG_IPV6_SEG6_BPF is not specified.

Fixes: b061017f8b4d ("selftests/bpf: add realistic loop tests")
Signed-off-by: Alan Maguire <alan.maguire@oracle.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/1590147389-26482-2-git-send-email-alan.maguire@oracle.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agodrm/amdgpu: Sync with VM root BO when switching VM to CPU update mode
Felix Kuehling [Wed, 20 May 2020 01:02:45 +0000 (21:02 -0400)]
drm/amdgpu: Sync with VM root BO when switching VM to CPU update mode

[ Upstream commit 90ca78deb004abe75b5024968a199acb96bb70f9 ]

This fixes an intermittent bug where a root PD clear operation still in
progress could overwrite a PDE update done by the CPU, resulting in a
VM fault.

Fixes: 108b4d928c03 ("drm/amd/amdgpu: Update VM function pointer")
Reported-by: Jay Cornwall <Jay.Cornwall@amd.com>
Tested-by: Jay Cornwall <Jay.Cornwall@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agodrm/amd/powerpay: Disable gfxoff when setting manual mode on picasso and raven
chen gong [Thu, 21 May 2020 09:15:34 +0000 (17:15 +0800)]
drm/amd/powerpay: Disable gfxoff when setting manual mode on picasso and raven

[ Upstream commit cbd2d08c7463e78d625a69e9db27ad3004cbbd99 ]

[Problem description]
1. Boot up picasso platform, launches desktop, Don't do anything (APU enter into "gfxoff" state)
2. Remote login to platform using SSH, then type the command line:
sudo su -c "echo manual > /sys/class/drm/card0/device/power_dpm_force_performance_level"
sudo su -c "echo 2 > /sys/class/drm/card0/device/pp_dpm_sclk" (fix SCLK to 1400MHz)
3. Move the mouse around in Window
4. Phenomenon :  The screen frozen

Tester will switch sclk level during glmark2 run time.
APU will enter "gfxoff" state intermittently during glmark2 run time.
The system got hanged if fix GFXCLK to 1400MHz when APU is in "gfxoff"
state.

[Debug]
1. Fix SCLK to X MHz
1400: screen frozen, screen black, then OS will reboot.
1300: screen frozen.
1200: screen frozen, screen black.
1100: screen frozen, screen black, then OS will reboot.
1000: screen frozen, screen black.
900:  screen frozen, screen black, then OS will reboot.
800:  Situation Nomal, issue disappear.
700:  Situation Nomal, issue disappear.
2. SBIOS setting: AMD CBS --> SMU Debug Options -->SMU Debug --> "GFX DLDO Psm Margin Control":
50 : Situation Nomal, issue disappear.
45 : Situation Nomal, issue disappear.
40 : Situation Nomal, issue disappear.
35 : Situation Nomal, issue disappear.
30 : screen black.
25 : screen frozen, then blurred screen.
20 : screen frozen.
15 : screen black.
10 : screen frozen.
5  : screen frozen, then blurred screen.
3. Disable GFXOFF feature
Situation Nomal, issue disappear.

[Why]
Through a period of time debugging with Sys Eng team and SMU team, Sys
Eng team said this is voltage/frequency marginal issue not a F/W or H/W
bug. This experiment proves that default targetPsm [for f=1400MHz] is
not sufficient when GFXOFF is enabled on Picasso.

SMU team think it is an odd test conditions to force sclk="1400MHz" when
GPU is in "gfxoff" state,then wake up the GFX. SCLK should be in the
"lowest frequency" when gfxoff.

[How]
Disable gfxoff when setting manual mode.
Enable gfxoff when setting other mode(exiting manual mode) again.

By the way, from the user point of view, now that user switch to manual
mode and force SCLK Frequency, he don't want SCLK be controlled by
workload.It becomes meaningless to "switch to manual mode" if APU enter "gfxoff"
due to lack of workload at this point.

Tips: Same issue observed on Raven.

Signed-off-by: chen gong <curry.gong@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agocrypto: stm32/crc32 - fix multi-instance
Nicolas Toromanoff [Tue, 12 May 2020 14:11:11 +0000 (16:11 +0200)]
crypto: stm32/crc32 - fix multi-instance

[ Upstream commit 10b89c43a64eb0d236903b79a3bc9d8f6cbfd9c7 ]

Ensure CRC algorithm is registered only once in crypto framework when
there are several instances of CRC devices.

Update the CRC device list management to avoid that only the first CRC
instance is used.

Fixes: b51dbe90912a ("crypto: stm32 - Support for STM32 CRC32 crypto module")

Signed-off-by: Nicolas Toromanoff <nicolas.toromanoff@st.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agocrypto: stm32/crc32 - fix run-time self test issue.
Nicolas Toromanoff [Tue, 12 May 2020 14:11:10 +0000 (16:11 +0200)]
crypto: stm32/crc32 - fix run-time self test issue.

[ Upstream commit a8cc3128bf2c01c4d448fe17149e87132113b445 ]

Fix wrong crc32 initialisation value:
"alg: shash: stm32_crc32 test failed (wrong result) on test vector 0,
cfg="init+update+final aligned buffer"
cra_name="crc32c" expects an init value of 0XFFFFFFFF,
cra_name="crc32" expects an init value of 0.

Fixes: b51dbe90912a ("crypto: stm32 - Support for STM32 CRC32 crypto module")

Signed-off-by: Nicolas Toromanoff <nicolas.toromanoff@st.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agocrypto: stm32/crc32 - fix ext4 chksum BUG_ON()
Nicolas Toromanoff [Tue, 12 May 2020 14:11:09 +0000 (16:11 +0200)]
crypto: stm32/crc32 - fix ext4 chksum BUG_ON()

[ Upstream commit 49c2c082e00e0bc4f5cbb7c21c7f0f873b35ab09 ]

Allow use of crc_update without prior call to crc_init.
And change (and fix) driver to use CRC device even on unaligned buffers.

Fixes: b51dbe90912a ("crypto: stm32 - Support for STM32 CRC32 crypto module")

Signed-off-by: Nicolas Toromanoff <nicolas.toromanoff@st.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agomips: Add udelay lpj numbers adjustment
Serge Semin [Thu, 21 May 2020 14:07:22 +0000 (17:07 +0300)]
mips: Add udelay lpj numbers adjustment

[ Upstream commit ed26aacfb5f71eecb20a51c4467da440cb719d66 ]

Loops-per-jiffies is a special number which represents a number of
noop-loop cycles per CPU-scheduler quantum - jiffies. As you
understand aside from CPU-specific implementation it depends on
the CPU frequency. So when a platform has the CPU frequency fixed,
we have no problem and the current udelay interface will work
just fine. But as soon as CPU-freq driver is enabled and the cores
frequency changes, we'll end up with distorted udelay's. In order
to fix this we have to accordinly adjust the per-CPU udelay_val
(the same as the global loops_per_jiffy) number. This can be done
in the CPU-freq transition event handler. We subscribe to that event
in the MIPS arch time-inititalization method.

Co-developed-by: Alexey Malahov <Alexey.Malahov@baikalelectronics.ru>
Signed-off-by: Alexey Malahov <Alexey.Malahov@baikalelectronics.ru>
Signed-off-by: Serge Semin <Sergey.Semin@baikalelectronics.ru>
Reviewed-by: Jiaxun Yang <jiaxun.yang@flygoat.com>
Cc: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Cc: Paul Burton <paulburton@kernel.org>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Rob Herring <robh+dt@kernel.org>
Cc: devicetree@vger.kernel.org
Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agomips: MAAR: Use more precise address mask
Serge Semin [Thu, 21 May 2020 00:34:37 +0000 (03:34 +0300)]
mips: MAAR: Use more precise address mask

[ Upstream commit bbb5946eb545fab8ad8f46bce8a803e1c0c39d47 ]

Indeed according to the MIPS32 Privileged Resource Architecgture the MAAR
pair register address field either takes [12:31] bits for non-XPA systems
and [12:55] otherwise. In any case the current address mask is just
wrong for 64-bit and 32-bits XPA chips. So lets extend it to 59-bits
of physical address value. This shall cover the 64-bits architecture and
systems with XPA enabled, and won't cause any problem for non-XPA 32-bit
systems, since address values exceeding the architecture specific MAAR
mask will be just truncated with setting zeros in the unsupported upper
bits.

Co-developed-by: Alexey Malahov <Alexey.Malahov@baikalelectronics.ru>
Signed-off-by: Alexey Malahov <Alexey.Malahov@baikalelectronics.ru>
Signed-off-by: Serge Semin <Sergey.Semin@baikalelectronics.ru>
Cc: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Cc: Paul Burton <paulburton@kernel.org>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Rob Herring <robh+dt@kernel.org>
Cc: devicetree@vger.kernel.org
Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agosched: Defend cfs and rt bandwidth quota against overflow
Huaixin Chang [Sat, 25 Apr 2020 10:52:48 +0000 (18:52 +0800)]
sched: Defend cfs and rt bandwidth quota against overflow

[ Upstream commit d505b8af58912ae1e1a211fabc9995b19bd40828 ]

When users write some huge number into cpu.cfs_quota_us or
cpu.rt_runtime_us, overflow might happen during to_ratio() shifts of
schedulable checks.

to_ratio() could be altered to avoid unnecessary internal overflow, but
min_cfs_quota_period is less than 1 << BW_SHIFT, so a cutoff would still
be needed. Set a cap MAX_BW for cfs_quota_us and rt_runtime_us to
prevent overflow.

Signed-off-by: Huaixin Chang <changhuaixin@linux.alibaba.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Ben Segall <bsegall@google.com>
Link: https://lkml.kernel.org/r/20200425105248.60093-1-changhuaixin@linux.alibaba.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agox86/boot: Correct relocation destination on old linkers
Arvind Sankar [Fri, 7 Feb 2020 21:49:26 +0000 (16:49 -0500)]
x86/boot: Correct relocation destination on old linkers

[ Upstream commit 5214028dd89e49ba27007c3ee475279e584261f0 ]

For the 32-bit kernel, as described in

  6d92bc9d483a ("x86/build: Build compressed x86 kernels as PIE"),

pre-2.26 binutils generates R_386_32 relocations in PIE mode. Since the
startup code does not perform relocation, any reloc entry with R_386_32
will remain as 0 in the executing code.

Commit

  974f221c84b0 ("x86/boot: Move compressed kernel to the end of the
                 decompression buffer")

added a new symbol _end but did not mark it hidden, which doesn't give
the correct offset on older linkers. This causes the compressed kernel
to be copied beyond the end of the decompression buffer, rather than
flush against it. This region of memory may be reserved or already
allocated for other purposes by the bootloader.

Mark _end as hidden to fix. This changes the relocation from R_386_32 to
R_386_RELATIVE even on the pre-2.26 binutils.

For 64-bit, this is not strictly necessary, as the 64-bit kernel is only
built as PIE if the linker supports -z noreloc-overflow, which implies
binutils-2.27+, but for consistency, mark _end as hidden here too.

The below illustrates the before/after impact of the patch using
binutils-2.25 and gcc-4.6.4 (locally compiled from source) and QEMU.

  Disassembly before patch:
    48:   8b 86 60 02 00 00       mov    0x260(%esi),%eax
    4e:   2d 00 00 00 00          sub    $0x0,%eax
                          4f: R_386_32    _end
  Disassembly after patch:
    48:   8b 86 60 02 00 00       mov    0x260(%esi),%eax
    4e:   2d 00 f0 76 00          sub    $0x76f000,%eax
                          4f: R_386_RELATIVE      *ABS*

Dump from extract_kernel before patch:
early console in extract_kernel
input_data: 0x0207c098 <--- this is at output + init_size
input_len: 0x0074fef1
output: 0x01000000
output_len: 0x00fa63d0
kernel_total_size: 0x0107c000
needed_size: 0x0107c000

Dump from extract_kernel after patch:
early console in extract_kernel
input_data: 0x0190d098 <--- this is at output + init_size - _end
input_len: 0x0074fef1
output: 0x01000000
output_len: 0x00fa63d0
kernel_total_size: 0x0107c000
needed_size: 0x0107c000

Fixes: 974f221c84b0 ("x86/boot: Move compressed kernel to the end of the decompression buffer")
Signed-off-by: Arvind Sankar <nivedita@alum.mit.edu>
Signed-off-by: Borislav Petkov <bp@suse.de>
Link: https://lkml.kernel.org/r/20200207214926.3564079-1-nivedita@alum.mit.edu
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agokgdboc: Use a platform device to handle tty drivers showing up late
Douglas Anderson [Thu, 7 May 2020 20:08:41 +0000 (13:08 -0700)]
kgdboc: Use a platform device to handle tty drivers showing up late

[ Upstream commit 68e55f61c13842baf825958129698c5371db432c ]

If you build CONFIG_KGDB_SERIAL_CONSOLE into the kernel then you
should be able to have KGDB init itself at bootup by specifying the
"kgdboc=..." kernel command line parameter.  This has worked OK for me
for many years, but on a new device I switched to it stopped working.

The problem is that on this new device the serial driver gets its
probe deferred.  Now when kgdb initializes it can't find the tty
driver and when it gives up it never tries again.

We could try to find ways to move up the initialization of the serial
driver and such a thing might be worthwhile, but it's nice to be
robust against serial drivers that load late.  We could move kgdb to
init itself later but that penalizes our ability to debug early boot
code on systems where the driver inits early.  We could roll our own
system of detecting when new tty drivers get loaded and then use that
to figure out when kgdb can init, but that's ugly.

Instead, let's jump on the -EPROBE_DEFER bandwagon.  We'll create a
singleton instance of a "kgdboc" platform device.  If we can't find
our tty device when the singleton "kgdboc" probes we'll return
-EPROBE_DEFER which means that the system will call us back later to
try again when the tty device might be there.

We won't fully transition all of the kgdboc to a platform device
because early kgdb initialization (via the "ekgdboc" kernel command
line parameter) still runs before the platform device has been
created.  The kgdb platform device is merely used as a convenient way
to hook into the system's normal probe deferral mechanisms.

As part of this, we'll ever-so-slightly change how the "kgdboc=..."
kernel command line parameter works.  Previously if you booted up and
kgdb couldn't find the tty driver then later reading
'/sys/module/kgdboc/parameters/kgdboc' would return a blank string.
Now kgdb will keep track of the string that came as part of the
command line and give it back to you.  It's expected that this should
be an OK change.

Signed-off-by: Douglas Anderson <dianders@chromium.org>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reviewed-by: Daniel Thompson <daniel.thompson@linaro.org>
Link: https://lore.kernel.org/r/20200507130644.v4.3.I4a493cfb0f9f740ce8fd2ab58e62dc92d18fed30@changeid
[daniel.thompson@linaro.org: Make config_mutex static]
Signed-off-by: Daniel Thompson <daniel.thompson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agomwifiex: Fix memory corruption in dump_station
Pali Rohár [Fri, 15 May 2020 07:59:24 +0000 (09:59 +0200)]
mwifiex: Fix memory corruption in dump_station

[ Upstream commit 3aa42bae9c4d1641aeb36f1a8585cd1d506cf471 ]

The mwifiex_cfg80211_dump_station() uses static variable for iterating
over a linked list of all associated stations (when the driver is in UAP
role). This has a race condition if .dump_station is called in parallel
for multiple interfaces. This corruption can be triggered by registering
multiple SSIDs and calling, in parallel for multiple interfaces
    iw dev <iface> station dump

[16750.719775] Unable to handle kernel paging request at virtual address dead000000000110
...
[16750.899173] Call trace:
[16750.901696]  mwifiex_cfg80211_dump_station+0x94/0x100 [mwifiex]
[16750.907824]  nl80211_dump_station+0xbc/0x278 [cfg80211]
[16750.913160]  netlink_dump+0xe8/0x320
[16750.916827]  netlink_recvmsg+0x1b4/0x338
[16750.920861]  ____sys_recvmsg+0x7c/0x2b0
[16750.924801]  ___sys_recvmsg+0x70/0x98
[16750.928564]  __sys_recvmsg+0x58/0xa0
[16750.932238]  __arm64_sys_recvmsg+0x28/0x30
[16750.936453]  el0_svc_common.constprop.3+0x90/0x158
[16750.941378]  do_el0_svc+0x74/0x90
[16750.944784]  el0_sync_handler+0x12c/0x1a8
[16750.948903]  el0_sync+0x114/0x140
[16750.952312] Code: f9400003 f907f423 eb02007f 54fffd60 (b9401060)
[16750.958583] ---[ end trace c8ad181c2f4b8576 ]---

This patch drops the use of the static iterator, and instead every time
the function is called iterates to the idx-th position of the
linked-list.

It would be better to convert the code not to use linked list for
associated stations storage (since the chip has a limited number of
associated stations anyway - it could just be an array). Such a change
may be proposed in the future. In the meantime this patch can backported
into stable kernels in this simple form.

Fixes: 8baca1a34d4c ("mwifiex: dump station support in uap mode")
Signed-off-by: Pali Rohár <pali@kernel.org>
Acked-by: Ganapathi Bhat <ganapathi.bhat@nxp.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200515075924.13841-1-pali@kernel.org
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agortlwifi: Fix a double free in _rtl_usb_tx_urb_setup()
Dan Carpenter [Wed, 13 May 2020 09:39:51 +0000 (12:39 +0300)]
rtlwifi: Fix a double free in _rtl_usb_tx_urb_setup()

[ Upstream commit beb12813bc75d4a23de43b85ad1c7cb28d27631e ]

Seven years ago we tried to fix a leak but actually introduced a double
free instead.  It was an understandable mistake because the code was a
bit confusing and the free was done in the wrong place.  The "skb"
pointer is freed in both _rtl_usb_tx_urb_setup() and _rtl_usb_transmit().
The free belongs _rtl_usb_transmit() instead of _rtl_usb_tx_urb_setup()
and I've cleaned the code up a bit to hopefully make it more clear.

Fixes: 36ef0b473fbf ("rtlwifi: usb: add missing freeing of skbuff")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200513093951.GD347693@mwanda
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agonet/mlx5e: IPoIB, Drop multicast packets that this interface sent
Erez Shitrit [Mon, 4 May 2020 08:46:25 +0000 (11:46 +0300)]
net/mlx5e: IPoIB, Drop multicast packets that this interface sent

[ Upstream commit 8b46d424a743ddfef8056d5167f13ee7ebd1dcad ]

After enabled loopback packets for IPoIB, we need to drop these packets
that this HCA has replicated and came back to the same interface that
sent them.

Fixes: 4c6c615e3f30 ("net/mlx5e: IPoIB, Add PKEY child interface nic profile")
Signed-off-by: Erez Shitrit <erezsh@mellanox.com>
Reviewed-by: Alex Vesker <valex@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoveth: Adjust hard_start offset on redirect XDP frames
Jesper Dangaard Brouer [Thu, 14 May 2020 10:49:43 +0000 (12:49 +0200)]
veth: Adjust hard_start offset on redirect XDP frames

[ Upstream commit 5c8572251fabc5bb49fd623c064e95a9daf6a3e3 ]

When native XDP redirect into a veth device, the frame arrives in the
xdp_frame structure. It is then processed in veth_xdp_rcv_one(),
which can run a new XDP bpf_prog on the packet. Doing so requires
converting xdp_frame to xdp_buff, but the tricky part is that
xdp_frame memory area is located in the top (data_hard_start) memory
area that xdp_buff will point into.

The current code tried to protect the xdp_frame area, by assigning
xdp_buff.data_hard_start past this memory. This results in 32 bytes
less headroom to expand into via BPF-helper bpf_xdp_adjust_head().

This protect step is actually not needed, because BPF-helper
bpf_xdp_adjust_head() already reserve this area, and don't allow
BPF-prog to expand into it. Thus, it is safe to point data_hard_start
directly at xdp_frame memory area.

Fixes: 9fc8d518d9d5 ("veth: Handle xdp_frames in xdp napi ring")
Reported-by: Mao Wenan <maowenan@huawei.com>
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Toshiaki Makita <toshiaki.makita1@gmail.com>
Acked-by: Toke Høiland-Jørgensen <toke@redhat.com>
Link: https://lore.kernel.org/bpf/158945338331.97035.5923525383710752178.stgit@firesoul
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoiocost: don't let vrate run wild while there's no saturation signal
Tejun Heo [Tue, 15 Oct 2019 00:18:11 +0000 (17:18 -0700)]
iocost: don't let vrate run wild while there's no saturation signal

[ Upstream commit 81ca627a933063fa63a6d4c66425de822a2ab7f5 ]

When the QoS targets are met and nothing is being throttled, there's
no way to tell how saturated the underlying device is - it could be
almost entirely idle, at the cusp of saturation or anywhere inbetween.
Given that there's no information, it's best to keep vrate as-is in
this state.  Before 7cd806a9a953 ("iocost: improve nr_lagging
handling"), this was the case - if the device isn't missing QoS
targets and nothing is being throttled, busy_level was reset to zero.

While fixing nr_lagging handling, 7cd806a9a953 ("iocost: improve
nr_lagging handling") broke this.  Now, while the device is hitting
QoS targets and nothing is being throttled, vrate keeps getting
adjusted according to the existing busy_level.

This led to vrate keeping climing till it hits max when there's an IO
issuer with limited request concurrency if the vrate started low.
vrate starts getting adjusted upwards until the issuer can issue IOs
w/o being throttled.  From then on, QoS targets keeps getting met and
nothing on the system needs throttling and vrate keeps getting
increased due to the existing busy_level.

This patch makes the following changes to the busy_level logic.

* Reset busy_level if nr_shortages is zero to avoid the above
  scenario.

* Make non-zero nr_lagging block lowering nr_level but still clear
  positive busy_level if there's clear non-saturation signal - QoS
  targets are met and nr_shortages is non-zero.  nr_lagging's role is
  preventing adjusting vrate upwards while there are long-running
  commands and it shouldn't keep busy_level positive while there's
  clear non-saturation signal.

* Restructure code for clarity and add comments.

Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Andy Newell <newella@fb.com>
Fixes: 7cd806a9a953 ("iocost: improve nr_lagging handling")
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoraid5: remove gfp flags from scribble_alloc()
Coly Li [Thu, 9 Apr 2020 14:17:21 +0000 (22:17 +0800)]
raid5: remove gfp flags from scribble_alloc()

[ Upstream commit ba54d4d4d2844c234f1b4692bd8c9e0f833c8a54 ]

Using GFP_NOIO flag to call scribble_alloc() from resize_chunk() does
not have the expected behavior. kvmalloc_array() inside scribble_alloc()
which receives the GFP_NOIO flag will eventually call kmalloc_node() to
allocate physically continuous pages.

Now we have memalloc scope APIs in mddev_suspend()/mddev_resume() to
prevent memory reclaim I/Os during raid array suspend context, calling
to kvmalloc_array() with GFP_KERNEL flag may avoid deadlock of recursive
I/O as expected.

This patch removes the useless gfp flags from parameters list of
scribble_alloc(), and call kvmalloc_array() with GFP_KERNEL flag. The
incorrect GFP_NOIO flag does not exist anymore.

Fixes: b330e6a49dc3 ("md: convert to kvmalloc")
Suggested-by: Michal Hocko <mhocko@suse.com>
Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Song Liu <songliubraving@fb.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agomd: don't flush workqueue unconditionally in md_open
Guoqing Jiang [Sat, 4 Apr 2020 21:57:09 +0000 (23:57 +0200)]
md: don't flush workqueue unconditionally in md_open

[ Upstream commit f6766ff6afff70e2aaf39e1511e16d471de7c3ae ]

We need to check mddev->del_work before flush workqueu since the purpose
of flush is to ensure the previous md is disappeared. Otherwise the similar
deadlock appeared if LOCKDEP is enabled, it is due to md_open holds the
bdev->bd_mutex before flush workqueue.

kernel: [  154.522645] ======================================================
kernel: [  154.522647] WARNING: possible circular locking dependency detected
kernel: [  154.522650] 5.6.0-rc7-lp151.27-default #25 Tainted: G           O
kernel: [  154.522651] ------------------------------------------------------
kernel: [  154.522653] mdadm/2482 is trying to acquire lock:
kernel: [  154.522655] ffff888078529128 ((wq_completion)md_misc){+.+.}, at: flush_workqueue+0x84/0x4b0
kernel: [  154.522673]
kernel: [  154.522673] but task is already holding lock:
kernel: [  154.522675] ffff88804efa9338 (&bdev->bd_mutex){+.+.}, at: __blkdev_get+0x79/0x590
kernel: [  154.522691]
kernel: [  154.522691] which lock already depends on the new lock.
kernel: [  154.522691]
kernel: [  154.522694]
kernel: [  154.522694] the existing dependency chain (in reverse order) is:
kernel: [  154.522696]
kernel: [  154.522696] -> #4 (&bdev->bd_mutex){+.+.}:
kernel: [  154.522704]        __mutex_lock+0x87/0x950
kernel: [  154.522706]        __blkdev_get+0x79/0x590
kernel: [  154.522708]        blkdev_get+0x65/0x140
kernel: [  154.522709]        blkdev_get_by_dev+0x2f/0x40
kernel: [  154.522716]        lock_rdev+0x3d/0x90 [md_mod]
kernel: [  154.522719]        md_import_device+0xd6/0x1b0 [md_mod]
kernel: [  154.522723]        new_dev_store+0x15e/0x210 [md_mod]
kernel: [  154.522728]        md_attr_store+0x7a/0xc0 [md_mod]
kernel: [  154.522732]        kernfs_fop_write+0x117/0x1b0
kernel: [  154.522735]        vfs_write+0xad/0x1a0
kernel: [  154.522737]        ksys_write+0xa4/0xe0
kernel: [  154.522745]        do_syscall_64+0x64/0x2b0
kernel: [  154.522748]        entry_SYSCALL_64_after_hwframe+0x49/0xbe
kernel: [  154.522749]
kernel: [  154.522749] -> #3 (&mddev->reconfig_mutex){+.+.}:
kernel: [  154.522752]        __mutex_lock+0x87/0x950
kernel: [  154.522756]        new_dev_store+0xc9/0x210 [md_mod]
kernel: [  154.522759]        md_attr_store+0x7a/0xc0 [md_mod]
kernel: [  154.522761]        kernfs_fop_write+0x117/0x1b0
kernel: [  154.522763]        vfs_write+0xad/0x1a0
kernel: [  154.522765]        ksys_write+0xa4/0xe0
kernel: [  154.522767]        do_syscall_64+0x64/0x2b0
kernel: [  154.522769]        entry_SYSCALL_64_after_hwframe+0x49/0xbe
kernel: [  154.522770]
kernel: [  154.522770] -> #2 (kn->count#253){++++}:
kernel: [  154.522775]        __kernfs_remove+0x253/0x2c0
kernel: [  154.522778]        kernfs_remove+0x1f/0x30
kernel: [  154.522780]        kobject_del+0x28/0x60
kernel: [  154.522783]        mddev_delayed_delete+0x24/0x30 [md_mod]
kernel: [  154.522786]        process_one_work+0x2a7/0x5f0
kernel: [  154.522788]        worker_thread+0x2d/0x3d0
kernel: [  154.522793]        kthread+0x117/0x130
kernel: [  154.522795]        ret_from_fork+0x3a/0x50
kernel: [  154.522796]
kernel: [  154.522796] -> #1 ((work_completion)(&mddev->del_work)){+.+.}:
kernel: [  154.522800]        process_one_work+0x27e/0x5f0
kernel: [  154.522802]        worker_thread+0x2d/0x3d0
kernel: [  154.522804]        kthread+0x117/0x130
kernel: [  154.522806]        ret_from_fork+0x3a/0x50
kernel: [  154.522807]
kernel: [  154.522807] -> #0 ((wq_completion)md_misc){+.+.}:
kernel: [  154.522813]        __lock_acquire+0x1392/0x1690
kernel: [  154.522816]        lock_acquire+0xb4/0x1a0
kernel: [  154.522818]        flush_workqueue+0xab/0x4b0
kernel: [  154.522821]        md_open+0xb6/0xc0 [md_mod]
kernel: [  154.522823]        __blkdev_get+0xea/0x590
kernel: [  154.522825]        blkdev_get+0x65/0x140
kernel: [  154.522828]        do_dentry_open+0x1d1/0x380
kernel: [  154.522831]        path_openat+0x567/0xcc0
kernel: [  154.522834]        do_filp_open+0x9b/0x110
kernel: [  154.522836]        do_sys_openat2+0x201/0x2a0
kernel: [  154.522838]        do_sys_open+0x57/0x80
kernel: [  154.522840]        do_syscall_64+0x64/0x2b0
kernel: [  154.522842]        entry_SYSCALL_64_after_hwframe+0x49/0xbe
kernel: [  154.522844]
kernel: [  154.522844] other info that might help us debug this:
kernel: [  154.522844]
kernel: [  154.522846] Chain exists of:
kernel: [  154.522846]   (wq_completion)md_misc --> &mddev->reconfig_mutex --> &bdev->bd_mutex
kernel: [  154.522846]
kernel: [  154.522850]  Possible unsafe locking scenario:
kernel: [  154.522850]
kernel: [  154.522852]        CPU0                    CPU1
kernel: [  154.522853]        ----                    ----
kernel: [  154.522854]   lock(&bdev->bd_mutex);
kernel: [  154.522856]                                lock(&mddev->reconfig_mutex);
kernel: [  154.522858]                                lock(&bdev->bd_mutex);
kernel: [  154.522860]   lock((wq_completion)md_misc);
kernel: [  154.522861]
kernel: [  154.522861]  *** DEADLOCK ***
kernel: [  154.522861]
kernel: [  154.522864] 1 lock held by mdadm/2482:
kernel: [  154.522865]  #0: ffff88804efa9338 (&bdev->bd_mutex){+.+.}, at: __blkdev_get+0x79/0x590
kernel: [  154.522868]
kernel: [  154.522868] stack backtrace:
kernel: [  154.522873] CPU: 1 PID: 2482 Comm: mdadm Tainted: G           O      5.6.0-rc7-lp151.27-default #25
kernel: [  154.522875] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.10.2-1ubuntu1 04/01/2014
kernel: [  154.522878] Call Trace:
kernel: [  154.522881]  dump_stack+0x8f/0xcb
kernel: [  154.522884]  check_noncircular+0x194/0x1b0
kernel: [  154.522888]  ? __lock_acquire+0x1392/0x1690
kernel: [  154.522890]  __lock_acquire+0x1392/0x1690
kernel: [  154.522893]  lock_acquire+0xb4/0x1a0
kernel: [  154.522895]  ? flush_workqueue+0x84/0x4b0
kernel: [  154.522898]  flush_workqueue+0xab/0x4b0
kernel: [  154.522900]  ? flush_workqueue+0x84/0x4b0
kernel: [  154.522905]  ? md_open+0xb6/0xc0 [md_mod]
kernel: [  154.522908]  md_open+0xb6/0xc0 [md_mod]
kernel: [  154.522910]  __blkdev_get+0xea/0x590
kernel: [  154.522912]  ? bd_acquire+0xc0/0xc0
kernel: [  154.522914]  blkdev_get+0x65/0x140
kernel: [  154.522916]  ? bd_acquire+0xc0/0xc0
kernel: [  154.522918]  do_dentry_open+0x1d1/0x380
kernel: [  154.522921]  path_openat+0x567/0xcc0
kernel: [  154.522923]  ? __lock_acquire+0x380/0x1690
kernel: [  154.522926]  do_filp_open+0x9b/0x110
kernel: [  154.522929]  ? __alloc_fd+0xe5/0x1f0
kernel: [  154.522935]  ? kmem_cache_alloc+0x28c/0x630
kernel: [  154.522939]  ? do_sys_openat2+0x201/0x2a0
kernel: [  154.522941]  do_sys_openat2+0x201/0x2a0
kernel: [  154.522944]  do_sys_open+0x57/0x80
kernel: [  154.522946]  do_syscall_64+0x64/0x2b0
kernel: [  154.522948]  entry_SYSCALL_64_after_hwframe+0x49/0xbe
kernel: [  154.522951] RIP: 0033:0x7f98d279d9ae

And md_alloc also flushed the same workqueue, but the thing is different
here. Because all the paths call md_alloc don't hold bdev->bd_mutex, and
the flush is necessary to avoid race condition, so leave it as it is.

Signed-off-by: Guoqing Jiang <guoqing.jiang@cloud.ionos.com>
Signed-off-by: Song Liu <songliubraving@fb.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agomt76: avoid rx reorder buffer overflow
Ryder Lee [Fri, 24 Apr 2020 19:32:22 +0000 (03:32 +0800)]
mt76: avoid rx reorder buffer overflow

[ Upstream commit 7c4f744d6703757be959f521a7a441bf34745d99 ]

Enlarge slot to support 11ax 256 BA (256 MPDUs in an AMPDU)

Signed-off-by: Chih-Min Chen <chih-min.chen@mediatek.com>
Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agodrm/mcde: dsi: Fix return value check in mcde_dsi_bind()
Wei Yongjun [Thu, 30 Apr 2020 07:31:45 +0000 (07:31 +0000)]
drm/mcde: dsi: Fix return value check in mcde_dsi_bind()

[ Upstream commit 761e9f4f80a21a4b845097027030bef863001636 ]

The of_drm_find_bridge() function returns NULL on error, it doesn't return
error pointers so this check doesn't work.

Fixes: 5fc537bfd000 ("drm/mcde: Add new driver for ST-Ericsson MCDE")
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200430073145.52321-1-weiyongjun1@huawei.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agonet: qed*: Reduce RX and TX default ring count when running inside kdump kernel
Bhupesh Sharma [Mon, 11 May 2020 10:11:41 +0000 (15:41 +0530)]
net: qed*: Reduce RX and TX default ring count when running inside kdump kernel

[ Upstream commit 73e030977f7884dbe1be0018bab517e8d02760f8 ]

Normally kdump kernel(s) run under severe memory constraint with the
basic idea being to save the crashdump vmcore reliably when the primary
kernel panics/hangs.

Currently the qed* ethernet driver ends up consuming a lot of memory in
the kdump kernel, leading to kdump kernel panic when one tries to save
the vmcore via ssh/nfs (thus utilizing the services of the underlying
qed* network interfaces).

An example OOM message log seen in the kdump kernel can be seen here
[1], with crashkernel size reservation of 512M.

Using tools like memstrack (see [2]), we can track the modules taking up
the bulk of memory in the kdump kernel and organize the memory usage
output as per 'highest allocator first'. An example log for the OOM case
indicates that the qed* modules end up allocating approximately 216M
memory, which is a large part of the total crashkernel size:

 dracut-pre-pivot[676]: ======== Report format module_summary: ========
 dracut-pre-pivot[676]: Module qed using 149.6MB (2394 pages), peak allocation 149.6MB (2394 pages)
 dracut-pre-pivot[676]: Module qede using 65.3MB (1045 pages), peak allocation 65.3MB (1045 pages)

This patch reduces the default RX and TX ring count from 1024 to 64
when running inside kdump kernel, which leads to a significant memory
saving.

An example log with the patch applied shows the reduced memory
allocation in the kdump kernel:
 dracut-pre-pivot[674]: ======== Report format module_summary: ========
 dracut-pre-pivot[674]: Module qed using 141.8MB (2268 pages), peak allocation 141.8MB (2268 pages)
 <..snip..>
[dracut-pre-pivot[674]: Module qede using 4.8MB (76 pages), peak allocation 4.9MB (78 pages)

Tested crashdump vmcore save via ssh/nfs protocol using underlying qed*
network interface after applying this patch.

[1] OOM log:
------------

 kworker/0:6: page allocation failure: order:6,
 mode:0x60c0c0(GFP_KERNEL|__GFP_COMP|__GFP_ZERO), nodemask=(null)
 kworker/0:6 cpuset=/ mems_allowed=0
 CPU: 0 PID: 145 Comm: kworker/0:6 Not tainted 4.18.0-109.el8.aarch64 #1
 Hardware name: To be filled by O.E.M. Saber/Saber, BIOS 0ACKL025
 01/18/2019
 Workqueue: events work_for_cpu_fn
 Call trace:
  dump_backtrace+0x0/0x188
  show_stack+0x24/0x30
  dump_stack+0x90/0xb4
  warn_alloc+0xf4/0x178
  __alloc_pages_nodemask+0xcac/0xd58
  alloc_pages_current+0x8c/0xf8
  kmalloc_order_trace+0x38/0x108
  qed_iov_alloc+0x40/0x248 [qed]
  qed_resc_alloc+0x224/0x518 [qed]
  qed_slowpath_start+0x254/0x928 [qed]
   __qede_probe+0xf8/0x5e0 [qede]
  qede_probe+0x68/0xd8 [qede]
  local_pci_probe+0x44/0xa8
  work_for_cpu_fn+0x20/0x30
  process_one_work+0x1ac/0x3e8
  worker_thread+0x44/0x448
  kthread+0x130/0x138
  ret_from_fork+0x10/0x18
  Cannot start slowpath
  qede: probe of 0000:05:00.1 failed with error -12

[2]. Memstrack tool: https://github.com/ryncsn/memstrack

Cc: kexec@lists.infradead.org
Cc: linux-kernel@vger.kernel.org
Cc: Ariel Elior <aelior@marvell.com>
Cc: GR-everest-linux-l2@marvell.com
Cc: Manish Chopra <manishc@marvell.com>
Cc: David S. Miller <davem@davemloft.net>
Signed-off-by: Bhupesh Sharma <bhsharma@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agowcn36xx: Fix error handling path in 'wcn36xx_probe()'
Christophe JAILLET [Fri, 8 May 2020 02:56:03 +0000 (05:56 +0300)]
wcn36xx: Fix error handling path in 'wcn36xx_probe()'

[ Upstream commit a86308fc534edeceaf64670c691e17485436a4f4 ]

In case of error, 'qcom_wcnss_open_channel()' must be undone by a call to
'rpmsg_destroy_ept()', as already done in the remove function.

Fixes: 5052de8deff5 ("soc: qcom: smd: Transition client drivers from smd to rpmsg")
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200507043619.200051-1-christophe.jaillet@wanadoo.fr
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoath10k: Remove msdu from idr when management pkt send fails
Rakesh Pillai [Fri, 8 May 2020 02:55:18 +0000 (05:55 +0300)]
ath10k: Remove msdu from idr when management pkt send fails

[ Upstream commit c730c477176ad4af86d9aae4d360a7ad840b073a ]

Currently when the sending of any management pkt
via wmi command fails, the packet is being unmapped
freed in the error handling. But the idr entry added,
which is used to track these packet is not getting removed.

Hence, during unload, in wmi cleanup, all the entries
in IDR are removed and the corresponding buffer is
attempted to be freed. This can cause a situation where
one packet is attempted to be freed twice.

Fix this error by rmeoving the msdu from the idr
list when the sending of a management packet over
wmi fails.

Tested HW: WCN3990
Tested FW: WLAN.HL.3.1-01040-QCAHLSWMTPLZ-1

Fixes: 1807da49733e ("ath10k: wmi: add management tx by reference support over wmi")
Signed-off-by: Rakesh Pillai <pillair@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1588667015-25490-1-git-send-email-pillair@codeaurora.org
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agonvme-tcp: use bh_lock in data_ready
Sagi Grimberg [Thu, 30 Apr 2020 20:59:32 +0000 (13:59 -0700)]
nvme-tcp: use bh_lock in data_ready

[ Upstream commit 386e5e6e1aa90b479fcf0467935922df8524393d ]

data_ready may be invoked from send context or from
softirq, so need bh locking for that.

Fixes: 3f2304f8c6d6 ("nvme-tcp: add NVMe over TCP host driver")
Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agonvme-pci: align io queue count with allocted nvme_queue in nvme_probe
Weiping Zhang [Sat, 2 May 2020 07:29:41 +0000 (15:29 +0800)]
nvme-pci: align io queue count with allocted nvme_queue in nvme_probe

[ Upstream commit 2a5bcfdd41d68559567cec3c124a75e093506cc1 ]

Since commit 147b27e4bd08 ("nvme-pci: allocate device queues storage
space at probe"), nvme_alloc_queue does not alloc the nvme queues
itself anymore.

If the write/poll_queues module parameters are changed at runtime to
values larger than the number of allocated queues in nvme_probe,
nvme_alloc_queue will access unallocated memory.

Add a new nr_allocated_queues member to struct nvme_dev to record how
many queues were alloctated in nvme_probe to avoid using more than the
allocated queues after a reset following a change to the
write/poll_queues module parameters.

Also add nr_write_queues and nr_poll_queues members to allow refreshing
the number of write and poll queues based on a change to the module
parameters when resetting the controller.

Fixes: 147b27e4bd08 ("nvme-pci: allocate device queues storage space at probe")
Signed-off-by: Weiping Zhang <zhangweiping@didiglobal.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Reviewed-by: Max Gurtovoy <maxg@mellanox.com>
[hch: add nvme_max_io_queues, update the commit message]
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agonvme: refine the Qemu Identify CNS quirk
Christoph Hellwig [Sat, 4 Apr 2020 08:11:28 +0000 (10:11 +0200)]
nvme: refine the Qemu Identify CNS quirk

[ Upstream commit b9a5c3d4c34d8bd9fd75f7f28d18a57cb68da237 ]

Add a helper to check if we can use Identify CNS values > 1, and refine
the Qemu quirk to not apply to reported versions larger than 1.1, as the
Qemu implementation had been fixed by then.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoiwlwifi: avoid debug max amsdu config overwriting itself
Mordechay Goodstein [Fri, 24 Apr 2020 15:48:15 +0000 (18:48 +0300)]
iwlwifi: avoid debug max amsdu config overwriting itself

[ Upstream commit a65a5824298b06049dbaceb8a9bd19709dc9507c ]

If we set amsdu_len one after another the second one overwrites
the orig_amsdu_len so allow only moving from debug to non debug state.

Also the TLC update check was wrong: it was checking that also the orig
is smaller then the new updated size, which is not the case in debug
amsdu mode.

Signed-off-by: Mordechay Goodstein <mordechay.goodstein@intel.com>
Fixes: af2984e9e625 ("iwlwifi: mvm: add a debugfs entry to set a fixed size AMSDU for all TX packets")
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Link: https://lore.kernel.org/r/iwlwifi.20200424182644.e565446a4fce.I9729d8c520d8b8bb4de9a5cdc62e01eb85168aac@changeid
Signed-off-by: Sasha Levin <sashal@kernel.org>