linux-stable.git/drivers/soc, branch v4.9.166

soc: fsl: qbman: avoid race in clearing QMan interrupt

2019-03-13T21:04:58+00:00

[ Upstream commit 89857a8a5c89a406b967ab2be7bd2ccdbe75e73d ]

By clearing all interrupt sources, not only those that
already occurred, the existing code may acknowledge by
mistake interrupts that occurred after the code checks
for them.

Signed-off-by: Madalin Bucur 
Signed-off-by: Roy Pledge 
Signed-off-by: Li Yang 
Signed-off-by: Sasha Levin

soc: bcm: brcmstb: Don't leak device tree node reference

2019-02-12T18:44:53+00:00

[ Upstream commit 1861a7f07e02292830a1ca256328d370deefea30 ]

of_find_node_by_path() acquires a reference to the node returned by it
and that reference needs to be dropped by its caller. soc_is_brcmstb()
doesn't do that, so fix it.

[treding: slightly rewrite to avoid inline comparison]

Fixes: d52fad262041 ("soc: add stubs for brcmstb SoC's")
Signed-off-by: Yangtao Li 
Signed-off-by: Florian Fainelli 
Signed-off-by: Sasha Levin

soc/tegra: Don't leak device tree node reference

2019-02-12T18:44:52+00:00

[ Upstream commit 9eb40fa2cd2d1f6829e7b49bb22692f754b9cfe0 ]

of_find_node_by_path() acquires a reference to the node returned by it
and that reference needs to be dropped by its caller. soc_is_tegra()
doesn't do that, so fix it.

Signed-off-by: Yangtao Li 
Acked-by: Jon Hunter 
[treding: slightly rewrite to avoid inline comparison]
Signed-off-by: Thierry Reding 
Signed-off-by: Sasha Levin

soc/tegra: pmc: Fix child-node lookup

2018-11-13T19:17:04+00:00

commit 1dc6bd5e39a29453bdcc17348dd2a89f1aa4004e upstream.

Fix child-node lookup during probe, which ended up searching the whole
device tree depth-first starting at the parent rather than just matching
on its children.

To make things worse, the parent pmc node could end up being prematurely
freed as of_find_node_by_name() drops a reference to its first argument.

Fixes: 3568df3d31d6 ("soc: tegra: Add thermal reset (thermtrip) support to PMC")
Cc: stable      # 4.0
Cc: Mikko Perttunen 
Signed-off-by: Johan Hovold 
Reviewed-by: Mikko Perttunen 
Signed-off-by: Thierry Reding 
Signed-off-by: Greg Kroah-Hartman

soc: fsl: qe: Fix copy/paste bug in ucc_get_tdm_sync_shift()

2018-11-10T15:42:42+00:00

[ Upstream commit 96fc74333f84cfdf8d434c6c07254e215e2aad00 ]

There is a copy and paste bug so we accidentally use the RX_ shift when
we're in TX_ mode.

Fixes: bb8b2062aff3 ("fsl/qe: setup clock source for TDM mode")
Signed-off-by: Dan Carpenter 
Signed-off-by: Zhao Qiang 
Signed-off-by: Li Yang 
(cherry picked from commit 3cb31b634052ed458922e0c8e2b4b093d7fb60b9)
Signed-off-by: Olof Johansson 
Signed-off-by: Sasha Levin

soc: fsl: qbman: qman: avoid allocating from non existing gen_pool

2018-11-10T15:42:42+00:00

[ Upstream commit 64e9e22e68512da8df3c9a7430f07621e48db3c2 ]

If the qman driver didn't probe, calling qman_alloc_fqid_range,
qman_alloc_pool_range or qman_alloc_cgrid_range (as done in dpaa_eth) will
pass a NULL pointer to gen_pool_alloc, leading to a NULL pointer
dereference.

Signed-off-by: Alexandre Belloni 
Reviewed-by: Roy Pledge 
Signed-off-by: Li Yang 
(cherry picked from commit f72487a2788aa70c3aee1d0ebd5470de9bac953a)
Signed-off-by: Olof Johansson 
Signed-off-by: Sasha Levin

soc: qcom: wcnss_ctrl: Fix increment in NV upload

2018-05-30T05:50:47+00:00

[ Upstream commit 90c29ed7627b6b4aeb603ee197650173c8434512 ]

hdr.len includes both the size of the header and the fragment, so using
this when stepping through the firmware causes us to skip 16 bytes every
chunk of 3072 bytes; causing only the first fragment to actually be
valid data.

Instead use fragment size steps through the firmware blob.

Fixes: ea7a1f275cf0 ("soc: qcom: Introduce WCNSS_CTRL SMD client")
Reported-by: Will Newton 
Signed-off-by: Bjorn Andersson 
Signed-off-by: Andy Gross 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

soc/fsl/qbman: fix issue in qman_delete_cgr_safe()

2018-03-31T16:11:35+00:00

[ Upstream commit 96f413f47677366e0ae03797409bfcc4151dbf9e ]

The wait_for_completion() call in qman_delete_cgr_safe()
was triggering a scheduling while atomic bug, replacing the
kthread with a smp_call_function_single() call to fix it.

Signed-off-by: Madalin Bucur 
Signed-off-by: Roy Pledge 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

soc/fsl/qe: round brg_freq to 1kHz granularity

2018-03-24T10:00:19+00:00

[ Upstream commit 2ccf80b7566cc035d903dd0ac5d7ebd25c2c1060 ]

Because of integer computation rounding in u-boot (that sets the QE
brg-frequency DTS prop), the clk value is 99999999 Hz even though it is
100 MHz.

When setting brg clks that are exact divisors of 100 MHz, this small
differnce plays a role and can result in lower clks to be output (for
instance 20 MHz - divide by 5 - results in 16.666 MHz - divide by 6).

This patch fixes that by "forcing" the brg_clk to the nearest kHz when
the difference is below 2 integer rounding errors (i.e. 4).

Signed-off-by: Valentin Longchamp 
Signed-off-by: Scott Wood 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

soc: mediatek: pwrap: fix compiler errors

2017-12-20T09:07:32+00:00

[ Upstream commit fb2c1934f30577756e55e24e8870b45c78da3bc2 ]

When compiling using sparse, we got the following error:
drivers/soc/mediatek/mtk-pmic-wrap.c:686:25: error: dubious one-bit signed bitfield

Changing the data type to unsigned fixes this.

Signed-off-by: Matthias Brugger 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman