[ Upstream commit 4f36cb36c9d14340bb200d2ad9117b03ce992cfe ]

The GFS2_RDF_UPTODATE flag in the rgrp is used to determine when
a rgrp buffer is valid. It's cleared when the glock is invalidated,
signifying that the buffer data is now invalid. But before this
patch, function update_rgrp_lvb was setting the flag when it
determined it had a valid lvb. But that's an invalid assumption:
just because you have a valid lvb doesn't mean you have valid
buffers. After all, another node may have made the lvb valid,
and this node just fetched it from the glock via dlm.

Consider this scenario:
1. The file system is mounted with RGRPLVB option.
2. In gfs2_inplace_reserve it locks the rgrp glock EX, but thanks
   to GL_SKIP, it skips the gfs2_rgrp_bh_get.
3. Since loops == 0 and the allocation target (ap->target) is
   bigger than the largest known chunk of blocks in the rgrp
   (rs->rs_rbm.rgd->rd_extfail_pt) it skips that rgrp and bypasses
   the call to gfs2_rgrp_bh_get there as well.
4. update_rgrp_lvb sees the lvb MAGIC number is valid, so bypasses
   gfs2_rgrp_bh_get, but it still sets sets GFS2_RDF_UPTODATE due
   to this invalid assumption.
5. The next time update_rgrp_lvb is called, it sees the bit is set
   and just returns 0, assuming both the lvb and rgrp are both
   uptodate. But since this is a smaller allocation, or space has
   been freed by another node, thus adjusting the lvb values,
   it decides to use the rgrp for allocations, with invalid rd_free
   due to the fact it was never updated.

This patch changes update_rgrp_lvb so it doesn't set the UPTODATE
flag anymore. That way, it has no choice but to fetch the latest
values.
Signed-off-by: NBob Peterson <rpeterso@redhat.com>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit fd2a981777d911b2e94cdec50779c85c58a0dec9 ]

Once reset is issued, the driver loses all control queue interfaces.
Exercising control queue operations during reset is incorrect and
may result in long timeouts.

This patch introduces a new field 'reset_ongoing' in the hw structure.
This is set to 1 by the core driver when it receives a reset interrupt.
ice_sq_send_cmd checks reset_ongoing before actually issuing the control
queue operation. If a reset is in progress, it returns a soft error code
(ICE_ERR_RESET_PENDING) to the caller. The caller may or may not have to
take any action based on this return. Once the driver knows that the
reset is done, it has to set reset_ongoing back to 0. This will allow
control queue operations to be posted to the hardware again.

This "bail out" logic was specifically added to ice_sq_send_cmd (which
is pretty low level function) so that we have one solution in one place
that applies to all types of control queues.
Signed-off-by: NAnirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: NTony Brelinski <tonyx.brelinski@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit ff2b13213a6a0baca105bc3bc724225f0adde1f8 ]

Align Request Resource Ownership AQ command (0x0008) to the latest
specification. This includes:

- Correcting the resource IDs for the Global Cfg and Change locks.
- new enum ICE_CHANGE_LOCK_RES_ID
- new enum ICE_GLOBAL_CFG_LOCK_RES_ID
- Altering the flow for Global Config Lock to allow only the first PF to
  download the package.
Signed-off-by: NDan Nowlin <dan.nowlin@intel.com>
Signed-off-by: NAnirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: NTony Brelinski <tonyx.brelinski@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit 34d5629d2ca89d847b7040762b87964c696c14da ]

Tri-band devices (1x 2.4GHz + 2x 5GHz) often incorporate special filters in
the RX and TX path. These filtered channel can in theory still be used by
the hardware but the signal strength is reduced so much that it makes no
sense.

There is already a DT property to limit the available channels but ath10k
has to manually call this functionality to limit the currrently set wiphy
channels further.
Signed-off-by: NSven Eckelmann <sven.eckelmann@openmesh.com>
Signed-off-by: NKalle Valo <kvalo@codeaurora.org>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit 4405b632e3da839defec966e4b0be44d0c5e3102 ]

Check rx_buff_mgmt is allocated before accessing its internal fields.
Signed-off-by: NDedy Lansky <dlansky@codeaurora.org>
Signed-off-by: NMaya Erez <merez@codeaurora.org>
Signed-off-by: NKalle Valo <kvalo@codeaurora.org>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit df2b53884a5a454bf441ca78e5b57307262c73f4 ]

In enhanced DMA ring 0 is used for RX ring, hence TX ring 0
is an unused element in ring_tx and ring2cid_tid arrays.
Initialize ring2cid_tid CID to WIL6210_MAX_CID to prevent a false
match of CID 0.
Go over the ring_tx and ring2cid_tid from wil_get_min_tx_ring_id
and on to prevent access to index 0 in eDMA.
Signed-off-by: NMaya Erez <merez@codeaurora.org>
Signed-off-by: NKalle Valo <kvalo@codeaurora.org>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit 596bdbcce90fa93f43ebcb99cefea34bd2e27707 ]

use_rx_hw_reordering is already set to true in wil_set_capabilities
for Talyn-MB devices. Remove its setting from wil_priv_init to
prevent its activation for older chips.
Similarly, move the setting of use_compressed_rx_status to
wil_set_capabilities.
Signed-off-by: NMaya Erez <merez@codeaurora.org>
Signed-off-by: NKalle Valo <kvalo@codeaurora.org>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit 9a65064abdf82934e0ed4744125f9f466f421f57 ]

Delivering a looped-back multicast packet to network stack can cause
higher layer protocols to fail like for example IPv6 DAD.
In STA mode, upon receiving Rx multicast packet, check if the source
MAC address is equal to our own MAC address and if so drop the packet.
Signed-off-by: NDedy Lansky <dlansky@codeaurora.org>
Signed-off-by: NMaya Erez <merez@codeaurora.org>
Signed-off-by: NKalle Valo <kvalo@codeaurora.org>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit d9c52fd17cb483bd8a470398afcb79f86c1b77c8 ]

Tx99 is typically configured via a monitor mode interface, which does
not get added to the driver as a vif. Since the code currently expects
a configured virtual interface for tx99, enabling tx99 via debugfs fails.
Since the vif is not needed anyway, remove all checks for it.
Signed-off-by: NFelix Fietkau <nbd@nbd.name>
[kvalo@codeaurora.org: s/CPTCFG/CONFIG/]
Signed-off-by: NKalle Valo <kvalo@codeaurora.org>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit 58da3b42307061b71f2dcce2bd1185d578a3aa53 ]

WCN3990 has the MAC_PCU_ADDR1 configured properly
and hence it will not send spurious ack frames
during boot up.

Hence the reset_rx_filter workaround is not needed
for WCN3990. Add a hw_param to indicate if hardware rx
filter reset is needed and skip the reset_rx_filter for
WCN3990.

Tested HW: WCN3990
Tested FW: WLAN.HL.2.0-01188-QCAHLSWMTPLZ-1
Signed-off-by: NRakesh Pillai <pillair@codeaurora.org>
Signed-off-by: NKalle Valo <kvalo@codeaurora.org>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit b8e131542b47b81236ecf6768c923128e1f5db6e ]

snd_seq_system_client_init() doesn't check the errors returned from
its port creations.  Let's do it properly and handle the error paths.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit 7417844b63d4b0dc8ab23f88259bf95de7d09b57 ]

When REGULATORY_COUNTRY_IE_IGNORE is set,  __reg_process_hint_country_ie()
ignores the country code change request from __cfg80211_connect_result()
via regulatory_hint_country_ie().

After Disconnect, similar to above, country code should not be reset to
world when country IE ignore is set. But this is violated and restore of
regulatory settings is invoked by cfg80211_disconnect_work via
regulatory_hint_disconnect().

To address this, avoid regulatory restore from regulatory_hint_disconnect()
when COUNTRY_IE_IGNORE is set.

Note: Currently, restore_regulatory_settings() takes care of clearing
beacon hints. But in the proposed change, regulatory restore is avoided.
Therefore, explicitly clear beacon hints when DISABLE_BEACON_HINTS
is not set.
Signed-off-by: NRajeev Kumar Sirasanagandla <rsirasan@codeaurora.org>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit 962341b54b99965ebec5f70c8d39f1c382eea833 ]

When we have first case to fall through it's not enough to put
single comment there to satisfy compiler. Instead of doing that,
return fall back value directly from default case.

This to avoid following warnings:

drivers/extcon/extcon-intel-cht-wc.c: In function ‘cht_wc_extcon_get_charger’:
include/linux/device.h:1420:2: warning: this statement may fall through [-Wimplicit-fallthrough=]
  _dev_warn(dev, dev_fmt(fmt), ##__VA_ARGS__)
  ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/extcon/extcon-intel-cht-wc.c:148:3: note: in expansion of macro ‘dev_warn’
   dev_warn(ext->dev,
   ^~~~~~~~
drivers/extcon/extcon-intel-cht-wc.c:152:2: note: here
  case CHT_WC_USBSRC_TYPE_SDP:
  ^~~~
Signed-off-by: NAndy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: NChanwoo Choi <cw00.choi@samsung.com>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit 1e28dbbeced6152b9ea7c417ff8cef3f7dcf0f19 ]

da8xx_rproc_mem size is of type size_t, so use %zx to format the debug
print of it to avoid a compile warning.
Acked-by: NSuman Anna <s-anna@ti.com>
Reviewed-by: NBartosz Golaszewski <bgolaszewski@baylibre.com>
Signed-off-by: NBjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit c856618d20662695fcdb47bf4d560dc457662aec ]

The ID for RV8803 must be rv_8803
Signed-off-by: NAlexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit 777d8ae56da18fb6440acd941edb3597c1b02bf0 ]

devm_kcalloc() returns NULL, it never returns error pointers.  In the
current code we would return PTR_ERR(NULL) which is success, instead of
returning the -ENOMEM error code.

Fixes: a0a1a1ba ("rtc: sysfs: facilitate attribute add to rtc device")
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NAlexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit 10af10db8c76fa5b9bf1f52a895c1cb2c0ac24da ]

Fix a typo. No functional change made by this patch.
Signed-off-by: NJay Foster <jayfoster@ieee.org>
Signed-off-by: NNicolas Ferre <nicolas.ferre@microchip.com>
Signed-off-by: NAlexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit 441f61e3aa9e7386731ce1405044d484bd81f911 ]

Update the SATA device nodes on R-Car H1, H2, and M2-W to use a 2 MiB
I/O space, as specified in Rev.1.0 of the R-Car H1 and R-Car Gen2
hardware user manuals.

See also commit e9f0089b ("arm64: dts: r8a7795: Correct SATA
device size to 2MiB").
Signed-off-by: NGeert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: NSimon Horman <horms+renesas@verge.net.au>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit 33e26418193f58d1895f2f968e1953b1caf8deb7 ]

get_seconds() and do_gettimeofday() are only used by a few modules now any
more (waiting for the respective patches to get accepted), and they are
among the last holdouts of code that is not y2038 safe in the core kernel.

Move the implementation into the timekeeping32.h header to clean up
the core kernel and isolate the old interfaces further.
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit 6ff7705da8806de45ca1490194f0b4eb07725804 ]

On p2180 sdmmc4 is powered from a fixed 1.8 V regulator.
Signed-off-by: NAapo Vienamo <avienamo@nvidia.com>
Reviewed-by: NMikko Perttunen <mperttunen@nvidia.com>
Signed-off-by: NThierry Reding <treding@nvidia.com>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit 13136a47a061c01c91df78b37f7708dd5ce7035f ]

Implement support for the PMC_IMPL_E_33V_PWR register which replaces
PMC_PWR_DET register interface of the SoC generations preceding
Tegra186. Also add the voltage bit offsets to the tegra186_io_pads[]
table and the AO_HV pad.
Signed-off-by: NAapo Vienamo <avienamo@nvidia.com>
Acked-by: NJon Hunter <jonathanh@nvidia.com>
Signed-off-by: NThierry Reding <treding@nvidia.com>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit 6f128fa41f310e1f39ebcea9621d2905549ecf52 ]

The "frames" variable is unsigned so the error handling doesn't work
properly.
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit 480f58cdbe392d4387a2193b6131a277e0111dd0 ]

According to the NanoPi-A64 schematics, DCDC1 is connected to a voltage
rail named "VDD_SYS_3.3V". All users seem to expect 3.3V here: the
Ethernet PHY, the uSD card slot, the camera interface and the GPIO pins
on the headers.
Fix up the voltage on the regulator to lift it up to 3.3V.
Signed-off-by: NAndre Przywara <andre.przywara@arm.com>
Acked-by: NMaxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: NChen-Yu Tsai <wens@csie.org>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit 93366b49a35f3a190052734b3f32c8fe2535b53f ]

The Olinuxino board uses DDR3L chips which are supposed to be driven
with 1.35V. The reset default of the AXP is properly set to 1.36V.

While technically the chips can also run at 1.5 volts, changing the
voltage on the fly while booting Linux is asking for trouble. Also
running at a lower voltage saves power.

So fix the DCDC5 value to match the actual board design.
Signed-off-by: NAndre Przywara <andre.przywara@arm.com>
Tested-by: NMartin Lucina <martin@lucina.net>
Acked-by: NMaxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: NChen-Yu Tsai <wens@csie.org>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit 09b964afca14d0594b2b2f265df3d987e2f43867 ]

The Orange Pi Win has a microSD card slot which is connected via all
four SD data lines. As the DT was not mentioning this fact, we got the
default single bit transfers, losing out on performance.
Also, as microSD does not have a write protect switch, we disable this
feature in the DT node.
Signed-off-by: NSamuel Holland <samuel@sholland.org>
Signed-off-by: NAndre Przywara <andre.przywara@arm.com>
Acked-by: NMaxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: NChen-Yu Tsai <wens@csie.org>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit e1c815f4b24a305e5bc9eceb541674bf4d02b709 ]

In case of error, we can dereference uninitialized 'adev'

drivers/soundwire/intel_init.c:154 sdw_intel_acpi_cb()
error: uninitialized symbol 'adev'.

Fix that by not using adev for warn print and make it pr_err.
Reported-by: NDan Carpenter <dan.carpenter@oracle.com>
Acked-by: NPierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: NVinod Koul <vkoul@kernel.org>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit a306a0e4a5326269b6c78d136407f08433ab5ece ]

Deferred messages are async messages used to synchronize
transitions mostly while doing a bank switch on multi links.
On successful transitions these messages are marked complete
and thereby confirming that all the buses performed bank switch
successfully.

So, initialize the completion structure for the same.
Signed-off-by: NSanyog Kale <sanyog.r.kale@intel.com>
Signed-off-by: NShreyas NC <shreyas.nc@intel.com>
Acked-by: NPierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: NVinod Koul <vkoul@kernel.org>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit 58c0f79887d5e425fe6a9fd542778e50df69e9c6 ]

Currently the register offset of the PWM bus gate in Allwinner H6 clock
driver is wrong.

Fix this issue.

Fixes: 542353ea ("clk: sunxi-ng: add support for the Allwinner H6 CCU")
Signed-off-by: NRongyi Chen <chenyi@tt-cool.com>
[Icenowy: refactor commit message]
Signed-off-by: NIcenowy Zheng <icenowy@aosc.io>
Signed-off-by: NChen-Yu Tsai <wens@csie.org>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit 0833627fc3f757a0dca11e2a9c46c96335a900ee ]

Do not try to write negative values and make sure that the write goes well.
Signed-off-by: NMarcus Folkesson <marcus.folkesson@gmail.com>
Signed-off-by: NJonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit bd1d395070cca4f42a93e520b0597274789274a4 ]

When continuously running wifi up/down sequence, the napi poll
can be scheduled after the CE buffers being freed by ath10k_pci_flush

Steps:
  In a certain condition, during wifi down below scenario might occur.

ath10k_stop->ath10k_hif_stop->napi_schedule->ath10k_pci_flush->napi_poll(napi_synchronize).

In the above scenario, CE buffer entries will be freed up and become NULL in
ath10k_pci_flush. And the napi_poll has been invoked after the flush process
and it will try to get the skb from the CE buffer entry and perform some action on that.
Since the CE buffer already cleaned by pci flush this action will create NULL
pointer dereference and trigger below kernel panic.

Unable to handle kernel NULL pointer dereference at virtual address 0000005c
PC is at ath10k_pci_htt_rx_cb+0x64/0x3ec [ath10k_pci]
ath10k_pci_htt_rx_cb [ath10k_pci]
ath10k_ce_per_engine_service+0x74/0xc4 [ath10k_pci]
ath10k_ce_per_engine_service [ath10k_pci]
ath10k_ce_per_engine_service_any+0x74/0x80 [ath10k_pci]
ath10k_ce_per_engine_service_any [ath10k_pci]
ath10k_pci_napi_poll+0x48/0xec [ath10k_pci]
ath10k_pci_napi_poll [ath10k_pci]
net_rx_action+0xac/0x160
net_rx_action
__do_softirq+0xdc/0x208
__do_softirq
irq_exit+0x84/0xe0
irq_exit
__handle_domain_irq+0x80/0xa0
__handle_domain_irq
gic_handle_irq+0x38/0x5c
gic_handle_irq
__irq_usr+0x44/0x60

Tested on QCA4019 and firmware version 10.4.3.2.1.1-00010
Signed-off-by: NTamizh chelvam <tamizhr@codeaurora.org>
Signed-off-by: NKalle Valo <kvalo@codeaurora.org>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit db878f76b9ff7487da9bb0f686153f81829f1230 ]

DT nodes may have a 'status' property which, if set to anything other
than 'ok' or 'okay', indicates to the OS that the DT node should be
treated as if it was not present. So add that missing logic to the
OP-TEE driver.
Signed-off-by: NArd Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: NJens Wiklander <jens.wiklander@linaro.org>
Signed-off-by: NSasha Levin <sashal@kernel.org>

[ Upstream commit b1ec0802503820ccbc894aadfd2a44da20232f5e ]

After finding a reasonable gain, the function converts the configured
gain to a gain configuration option selector enum max9611_csa_gain.
Make the conversion clearly visible by using an explicit cast. This
also avoids a warning seen with clang:
  drivers/iio/adc/max9611.c:292:16: warning: implicit conversion from
      enumeration type 'enum max9611_conf_ids' to different enumeration
      type 'enum max9611_csa_gain' [-Wenum-conversion]
                        *csa_gain = gain_selectors[i];
                                  ~ ^~~~~~~~~~~~~~~~~
Signed-off-by: NStefan Agner <stefan@agner.ch>
Signed-off-by: NJonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: NSasha Levin <sashal@kernel.org>

commit fed23c5829ecab4ddc712d7b0046e59610ca3ba4 upstream.

The quirks2 are parsed and set (e.g. from DT) before the quirk for broken
HS200 is set in the driver.
The driver needs to enable just this flag, not rewrite the whole quirk set.

Fixes: 7871aa60ae00 ("mmc: sdhci-of-at91: add quirk for broken HS200")
Signed-off-by: NEugen Hristev <eugen.hristev@microchip.com>
Acked-by: NAdrian Hunter <adrian.hunter@intel.com>
Cc: stable@vger.kernel.org
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

commit 0362f326d86c645b5e96b7dbc3ee515986ed019d upstream.

An exiting task might belong to an offline cgroup.  In this case an
attempt to grab a cgroup reference from the task can end up with an
infinite loop in hugetlb_cgroup_charge_cgroup(), because neither the
cgroup will become online, neither the task will be migrated to a live
cgroup.

Fix this by switching over to css_tryget().  As css_tryget_online()
can't guarantee that the cgroup won't go offline, in most cases the
check doesn't make sense.  In this particular case users of
hugetlb_cgroup_charge_cgroup() are not affected by this change.

A similar problem is described by commit 18fa84a2db0e ("cgroup: Use
css_tryget() instead of css_tryget_online() in task_get_css()").

Link: http://lkml.kernel.org/r/20191106225131.3543616-2-guro@fb.comSigned-off-by: NRoman Gushchin <guro@fb.com>
Acked-by: NJohannes Weiner <hannes@cmpxchg.org>
Acked-by: NTejun Heo <tj@kernel.org>
Reviewed-by: NShakeel Butt <shakeelb@google.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

commit 00d484f354d85845991b40141d40ba9e5eb60faf upstream.

We've encountered a rcu stall in get_mem_cgroup_from_mm():

  rcu: INFO: rcu_sched self-detected stall on CPU
  rcu: 33-....: (21000 ticks this GP) idle=6c6/1/0x4000000000000002 softirq=35441/35441 fqs=5017
  (t=21031 jiffies g=324821 q=95837) NMI backtrace for cpu 33
  <...>
  RIP: 0010:get_mem_cgroup_from_mm+0x2f/0x90
  <...>
   __memcg_kmem_charge+0x55/0x140
   __alloc_pages_nodemask+0x267/0x320
   pipe_write+0x1ad/0x400
   new_sync_write+0x127/0x1c0
   __kernel_write+0x4f/0xf0
   dump_emit+0x91/0xc0
   writenote+0xa0/0xc0
   elf_core_dump+0x11af/0x1430
   do_coredump+0xc65/0xee0
   get_signal+0x132/0x7c0
   do_signal+0x36/0x640
   exit_to_usermode_loop+0x61/0xd0
   do_syscall_64+0xd4/0x100
   entry_SYSCALL_64_after_hwframe+0x44/0xa9

The problem is caused by an exiting task which is associated with an
offline memcg.  We're iterating over and over in the do {} while
(!css_tryget_online()) loop, but obviously the memcg won't become online
and the exiting task won't be migrated to a live memcg.

Let's fix it by switching from css_tryget_online() to css_tryget().

As css_tryget_online() cannot guarantee that the memcg won't go offline,
the check is usually useless, except some rare cases when for example it
determines if something should be presented to a user.

A similar problem is described by commit 18fa84a2db0e ("cgroup: Use
css_tryget() instead of css_tryget_online() in task_get_css()").

Johannes:

: The bug aside, it doesn't matter whether the cgroup is online for the
: callers.  It used to matter when offlining needed to evacuate all charges
: from the memcg, and so needed to prevent new ones from showing up, but we
: don't care now.

Link: http://lkml.kernel.org/r/20191106225131.3543616-1-guro@fb.comSigned-off-by: NRoman Gushchin <guro@fb.com>
Acked-by: NJohannes Weiner <hannes@cmpxchg.org>
Acked-by: NTejun Heo <tj@kernel.org>
Reviewed-by: NShakeel Butt <shakeeb@google.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Michal Koutn <mkoutny@suse.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

commit a85dfc305a21acfc48fa28a0fa0a0cb6ad496120 upstream.

Commit d883544515aa ("mm: mempolicy: make the behavior consistent when
MPOL_MF_MOVE* and MPOL_MF_STRICT were specified") fixed the return value
of mbind() for a couple of corner cases.  But, it altered the errno for
some other cases, for example, mbind() should return -EFAULT when part
or all of the memory range specified by nodemask and maxnode points
outside your accessible address space, or there was an unmapped hole in
the specified memory range specified by addr and len.

Fix this by preserving the errno returned by queue_pages_range().  And,
the pagelist may be not empty even though queue_pages_range() returns
error, put the pages back to LRU since mbind_range() is not called to
really apply the policy so those pages should not be migrated, this is
also the old behavior before the problematic commit.

Link: http://lkml.kernel.org/r/1572454731-3925-1-git-send-email-yang.shi@linux.alibaba.com
Fixes: d883544515aa ("mm: mempolicy: make the behavior consistent when MPOL_MF_MOVE* and MPOL_MF_STRICT were specified")
Signed-off-by: NYang Shi <yang.shi@linux.alibaba.com>
Reported-by: NLi Xinhai <lixinhai.lxh@gmail.com>
Reviewed-by: NLi Xinhai <lixinhai.lxh@gmail.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: <stable@vger.kernel.org>	[4.19 and 5.2+]
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

commit 4e7120d79edb31e4ee68e6f8421448e4603be1e9 upstream.

For both PASID-based-Device-TLB Invalidate Descriptor and
Device-TLB Invalidate Descriptor, the Physical Function Source-ID
value is split according to this layout:

PFSID[3:0] is set at offset 12 and PFSID[15:4] is put at offset 52.
Fix the part laid out at offset 52.

Fixes: 0f725561 ("iommu/vt-d: Add definitions for PFSID")
Signed-off-by: NEric Auger <eric.auger@redhat.com>
Acked-by: NJacob Pan <jacob.jun.pan@linux.intel.com>
Cc: stable@vger.kernel.org # v4.19+
Acked-by: NLu Baolu <baolu.lu@linux.intel.com>
Signed-off-by: NJoerg Roedel <jroedel@suse.de>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

commit 40a1dcee2d1846a24619fe9ca45c661ca0db7dda upstream.

When PHY is not powered, the probe function fail and some resource are
still unallocated.
Furthermore some BUG happens:
dwmac-sun8i 5020000.ethernet: EMAC reset timeout
------------[ cut here ]------------
kernel BUG at /linux-next/net/core/dev.c:9844!

So let's use the right function (stmmac_pltfr_remove) in the error path.

Fixes: 9f93ac8d ("net-next: stmmac: Add dwmac-sun8i")
Cc: <stable@vger.kernel.org> # v4.15+
Signed-off-by: NCorentin Labbe <clabbe@baylibre.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

commit 762c69685ff7ad5ad7fee0656671e20a0c9c864d upstream.

We need to get the underlying dentry of parent; sure, absent the races
it is the parent of underlying dentry, but there's nothing to prevent
losing a timeslice to preemtion in the middle of evaluation of
lower_dentry->d_parent->d_inode, having another process move lower_dentry
around and have its (ex)parent not pinned anymore and freed on memory
pressure.  Then we regain CPU and try to fetch ->d_inode from memory
that is freed by that point.

dentry->d_parent *is* stable here - it's an argument of ->lookup() and
we are guaranteed that it won't be moved anywhere until we feed it
to d_add/d_splice_alias.  So we safely go that way to get to its
underlying dentry.

Cc: stable@vger.kernel.org # since 2009 or so
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

commit e72b9dd6a5f17d0fb51f16f8685f3004361e83d0 upstream.

lower_dentry can't go from positive to negative (we have it pinned),
but it *can* go from negative to positive.  So fetching ->d_inode
into a local variable, doing a blocking allocation, checking that
now ->d_inode is non-NULL and feeding the value we'd fetched
earlier to a function that won't accept NULL is not a good idea.

Cc: stable@vger.kernel.org
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>