If GuC firmware performs an engine reset while that engine had a
preemption pending, it will set the terminated attribute bit on our
preemption stage descriptor. GuC firmware retains all pending work
items for a high-priority GuC client, unlike the normal-priority GuC
client where work items are dropped. It wants to make sure the preempt-
to-idle work doesn't run when scheduling resumes, and uses this bit to
inform its scheduler and presumably us as well. Our job is to clear it
for the next preemption after reset, otherwise that and future
preemptions will never complete. We'll just clear it every time.
Signed-off-by: Jeff McGee <jeff.mcgee@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171101221630.25086-1-jeff.mcgee@intel.com
Reviewed-by: Michel Thierry <michel.thierry@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
On gcc-4.6, we get a harmless link-time warning:
WARNING: vmlinux.o(.text.unlikely+0x196a0): Section mismatch in reference from the function mtk_infrasys_init_early() to the function .init.text:mtk_clk_register_cpumuxes()
The function mtk_infrasys_init_early() references
the function __init mtk_clk_register_cpumuxes().
This is often because mtk_infrasys_init_early lacks a __init
annotation or the annotation of mtk_clk_register_cpumuxes is wrong.
Newer compilers inline this function so they don't warn, but
marking it __init is the right solution for all versions.
Fixes: e986211827 ("clk: mediatek: Add MT2701 clock support")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
LiquidIO firmware supports a vswitch that needs to know the names of the
VF representors in the host to maintain compatibility for direct
programming using external Openflow agents. So, for each VF representor,
send its name to the firmware when it gets registered and when its name
changes.
Signed-off-by: Vijaya Mohan Guvva <vijaya.guvva@cavium.com>
Signed-off-by: Raghu Vatsavayi <raghu.vatsavayi@cavium.com>
Signed-off-by: Felix Manlunas <felix.manlunas@cavium.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add MT2712 clock support, include topckgen, apmixedsys,
infracfg, pericfg, mcucfg and subsystem clocks.
Signed-off-by: Weiyi Lu <weiyi.lu@mediatek.com>
[sboyd@codeaurora.org: Static on top_clk_data]
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Hesoteric board configurations where port 0 is not available would still
make SYSTEMPORT inspect the switch port 0, queue 0, which, not being
enabled, would cause transmit timeouts over time. Just ignore those
unconfigured rings instead.
Fixes: 84ff33eeb23d ("net: systemport: Establish DSA network device queue mapping")
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch supports BPF_NEG under both BPF_ALU64 and BPF_ALU. LLVM recently
starts to generate it.
NOTE: BPF_NEG takes single operand which is an register and serve as both
input and output.
Signed-off-by: Jiong Wang <jiong.wang@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The current ALU_OP_NEG is Op encoding 0x4 for NPF ALU instruction. It is
actually performing "~B" operation which is bitwise NOT.
The using naming ALU_OP_NEG is misleading as NEG is -B which is not the
same as ~B.
Signed-off-by: Jiong Wang <jiong.wang@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
IMX7d does not have an M0 Core and this particular
clock doesn't seem connected to anything else.
Remove this entry from the CCM driver.
Signed-off-by: Adriana Reus <adriana.reus@nxp.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Allowing the lcdif_pre_sel to propagate rate changes to its parent PLL
allows more fine grained control over the LCDIF pixel clock rate.
For example, the Innovision AT043TN24 LCD panel described in the
imx6ul-14x14-evk device tree requires a 9 MHz pixel clock.
Before this patch, the lcdif_pre_sel clock rate is fixed, and just
setting the lcdif_pred and lcdif_podf dividers only allows to get as
close as about 8.44 MHz:
pll3 1 1 480000000 0 0
pll3_bypass 1 1 480000000 0 0
pll3_usb_otg 1 1 480000000 0 0
pll3_pfd1_540m 1 1 540000000 0 0
lcdif_pre_sel 1 1 540000000 0 0
lcdif_pred 1 1 67500000 0 0
lcdif_podf 1 1 8437500 0 0
lcdif_pix 1 1 8437500 0 0
Once lcdif_pre_sel is allowed to propagate rate requests to its parent,
the actual pixel clock matches the requested value:
pll3 1 1 480000000 0 0
pll3_bypass 1 1 480000000 0 0
pll3_usb_otg 1 1 480000000 0 0
pll3_pfd1_540m 1 1 288000000 0 0
lcdif_pre_sel 1 1 288000000 0 0
lcdif_pred 1 1 36000000 0 0
lcdif_podf 1 1 9000000 0 0
lcdif_pix 1 1 9000000 0 0
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Reviewed-by: Fabio Estevam <fabio.estevam@nxp.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
On i.MX6 SoCs without VPU (in my case MCIMX6D4AVT10AC), the hdmi driver
fails to probe:
[ 2.540030] dwhdmi-imx 120000.hdmi: Unsupported HDMI controller
(0000:00:00)
[ 2.548199] imx-drm display-subsystem: failed to bind 120000.hdmi
(ops dw_hdmi_imx_ops): -19
[ 2.557403] imx-drm display-subsystem: master bind failed: -19
That's because hdmi_isfr's parent, video_27m, is not correctly ungated.
As explained in commit 5ccc248cc5 ("ARM: imx6q: clk: Add support for
mipi_core_cfg clock as a shared clock gate"), video_27m is gated by
CCM_CCGR3[CG8].
On i.MX6 SoCs with VPU, the hdmi is working thanks to the
CCM_CMEOR[mod_en_ov_vpu] bit which makes the video_27m ungated whatever
is in CCM_CCGR3[CG8]. The issue can be reproduced by setting
CCMEOR[mod_en_ov_vpu] to 0.
Make the HDMI work in every case by setting hdmi_isfr's parent to
mipi_core_cfg.
Signed-off-by: Sébastien Szymanski <sebastien.szymanski@armadeus.com>
Reviewed-by: Fabio Estevam <fabio.estevam@nxp.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
FRA_L3MDEV is defined as U8, but is being added as a U32 attribute. On
big endian architecture, this results in the l3mdev entry not being
added to the FIB rules.
Fixes: 1aa6c4f6b8 ("net: vrf: Add l3mdev rules on first device create")
Signed-off-by: Jeff Barnhill <0xeffeff@gmail.com>
Acked-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This restores the original behaviour before the block callbacks were
introduced. Allow the drivers to do binding of block always, no matter
if the NETIF_F_HW_TC feature is on or off. Move the check to the block
callback which is called for rule insertion.
Reported-by: Alexander Duyck <alexander.duyck@gmail.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The assignment to kinfo is redundant as this is a duplicate of
the initialiation of kinfo a few lines earlier, so it can be
removed. The assignment to v_tc_info is never read, so this
variable is redundant and can be removed completely. Cleans
up two clang warnings:
drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_tm.c:433:34:
warning: Value stored to 'kinfo' during its initialization is never read
drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_tm.c:775:3:
warning: Value stored to 'v_tc_info' is never read
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The zero value assigned to inst_processed at the end of each
iteration of the do-while loop is overwritten on the next iteration
and hence it is a redundant assignment and can be removed. Cleans
up clang warning:
drivers/net/ethernet/cavium/liquidio/request_manager.c:480:3:
warning: Value stored to 'inst_processed' is never read
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The pointer np is initialized and then re-assigned the same value
a few lines later. Remove the redundant duplicated assignment. Cleans
up clang warning:
drivers/net/ethernet/dlink/dl2k.c:314:25: warning: Value stored to
'np' during its initialization is never read
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The RPM clocks were missing for MSM8660/APQ8060. For this to be
completed we need to add a special fixed rate RPM clock that is used
for the PLL4 on these SoCs. The rest of the clocks are pretty
similar to the other supported platforms.
The "active" clock pattern is mirrored in all the clocks. I guess
that the PLL4 that clocks the LPASS is actually never used as
"active only" since the low-power audio subsystem should be left
on when the CPU goes to idle, so that it can be used as a stand-alone
MP3 player type of device.
The PLL4 seems to be enabled only on behalf of the booting LPASS
Hexagon - which will cast its own vote once its booted - and as
such we only configure the active state (meaning both states will
have same configuration). The result is that PLL4 will be on
from prepare() to unprepare() regardless of what the application
CPU does.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
stat set to zero and the value is never read, instead stat is
set again in the do-loop. Hence the setting to zero is redundant
and can be removed. Cleans up clang warning:
drivers/net/wan/wanxl.c:737:2: warning: Value stored to 'stat'
is never read
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
By default, it is assumed that the UTMI clock is generated from a 12 MHz
reference clock (MAINCK). If it's not the case, the FREQ field of the
SFR_UTMICKTRIM has to be updated to generate the UTMI clock in the
proper way.
The UTMI clock has a fixed rate of 480 MHz. In fact, there is no
multiplier we can configure. The multiplier is managed internally,
depending on the reference clock frequency, to achieve the target of
480 MHz.
Signed-off-by: Ludovic Desroches <ludovic.desroches@microchip.com>
Acked-by: Ingo van Lil <inguin@gmx.de>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Now that we have devm APIs for the reset controller and of clk hw
provider APIs we can remove the custom code here.
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Sometimes we only have one of_clk_del_provider() call in driver
error and remove paths, because we're missing a
devm_of_clk_add_hw_provider() API. Introduce the API so we can
convert drivers to use this and potentially reduce the amount of
code needed to remove providers in drivers.
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Make these const as they are only stored in the init field of a clk_hw
structure, which is const.
Signed-off-by: Bhumika Goyal <bhumirks@gmail.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Make these const as they are only stored in the const field of a
clk_init_data structure.
Signed-off-by: Bhumika Goyal <bhumirks@gmail.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Make these const as they are only stored in the const field of a
clk_init_data structure.
Signed-off-by: Bhumika Goyal <bhumirks@gmail.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Make these const as they are only stored in the const field of a
clk_init_data structure.
Signed-off-by: Bhumika Goyal <bhumirks@gmail.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Make these const as they are only stored in the const field of a
clk_init_data structure.
Signed-off-by: Bhumika Goyal <bhumirks@gmail.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Make these const as they are only stored in the const field of a
clk_init_data structure.
Signed-off-by: Bhumika Goyal <bhumirks@gmail.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Smooth Cong Wang's bug fix into 'net-next'. Basically put
the bulk of the tcf_block_put() logic from 'net' into
tcf_block_put_ext(), but after the offload unbind.
Signed-off-by: David S. Miller <davem@davemloft.net>
Make these const as they are either stored in the masks 'const' field
of a clk_aux structure or passed to the function clk_register_aux
having the argument as const.
Signed-off-by: Bhumika Goyal <bhumirks@gmail.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Make the masks field of clk_aux structure const as it do not modify the
fields of the aux_clk_masks structure it points to.
Make the struct aux_clk_masks *aux argument of the function
clk_register_aux as const as the argument is only stored in the masks
field of a clk_aux structure which is now made const.
Signed-off-by: Bhumika Goyal <bhumirks@gmail.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
This patch fix the following build warning:
drivers/clk/sunxi/clk-factors.c:279:14: warning: variable 'name' set but not used [-Wunused-but-set-variable]
Fixes: 4cbeaebb8a ("clk: sunxi: factors: Add unregister function")
Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: Corentin Labbe <clabbe.montjoie@gmail.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Add support for displaying and modifying rx and tx ring sizes using
ethtool.
Also, increasing version to 2.3.0.45
Signed-off-by: Parvi Kaustubhi <pkaustub@cisco.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since we are allowing rx ring size modification, reset fetch index
everytime. Otherwise it could have a stale value that can lead to a null
pointer dereference.
Signed-off-by: Govindarajulu Varadarajan <gvaradar@cisco.com>
Signed-off-by: Parvi Kaustubhi <pkaustub@cisco.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Just two small patches for stable to fix the driver failing to load on polaris
cards with harvested VCE or UVD blocks.
* 'drm-fixes-4.14' of git://people.freedesktop.org/~agd5f/linux:
drm/amdgpu: allow harvesting check for Polaris VCE
drm/amdgpu: return -ENOENT from uvd 6.0 early init for harvesting
Only MD_SB_CHANGE_PENDING should be used to wait for transition from
clean to dirty. Checking also MD_SB_CHANGE_CLEAN is unnecessary and can
race with e.g. md_do_sync(). This sporadically causes a hang when
changing consistency policy during resync:
INFO: task mdadm:6183 blocked for more than 30 seconds.
Not tainted 4.14.0-rc3+ #391
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
mdadm D12752 6183 6022 0x00000000
Call Trace:
__schedule+0x93f/0x990
schedule+0x6b/0x90
md_allow_write+0x100/0x130 [md_mod]
? do_wait_intr_irq+0x90/0x90
resize_stripes+0x3a/0x5b0 [raid456]
? kernfs_fop_write+0xbe/0x180
raid5_change_consistency_policy+0xa6/0x200 [raid456]
consistency_policy_store+0x2e/0x70 [md_mod]
md_attr_store+0x90/0xc0 [md_mod]
sysfs_kf_write+0x42/0x50
kernfs_fop_write+0x119/0x180
__vfs_write+0x28/0x110
? rcu_sync_lockdep_assert+0x12/0x60
? __sb_start_write+0x15a/0x1c0
? vfs_write+0xa3/0x1a0
vfs_write+0xb4/0x1a0
SyS_write+0x49/0xa0
entry_SYSCALL_64_fastpath+0x18/0xad
Fixes: 2214c260c7 ("md: don't return -EAGAIN in md_allow_write for external metadata arrays")
Cc: <stable@vger.kernel.org>
Signed-off-by: Artur Paszkiewicz <artur.paszkiewicz@intel.com>
Signed-off-by: Shaohua Li <shli@fb.com>
The pointer q is assigned but never read; it is redundant and can
be removed. Cleans up clang warning:
drivers/md/md-multipath.c:260:4: warning: Value stored to 'q' is
never read
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Shaohua Li <shli@fb.com>
There are some lines could be removed due to recent
change for raid1 such as commit 3956df15d634 ("md:
move suspend_hi/lo handling into core md code").
Also, seems some comments are put to wrong place,
move them before wait_barrier.
Signed-off-by: Guoqing Jiang <gqjiang@suse.com>
Signed-off-by: Shaohua Li <shli@fb.com>
Suspending the entire device for resync could take
too long. Resync in small chunks.
cluster's resync window is maintained in r10conf as
cluster_sync_low and cluster_sync_high, and processed
in raid10's sync_request(). If the current resync is
outside the cluster resync window:
1. Set the cluster_sync_low to curr_resync_completed.
2. Set cluster_sync_high to cluster_sync_low + stripe
size.
3. Send a message to all nodes so they may add it in
their suspension list.
Note:
We only support "near" raid10 so far, resync a far or
offset raid10 array could have trouble. So raid10_run
checks the layout of clustered raid10, it will refuse
to run if the layout is not correct.
With the "near" layout we process one stripe at a time
progressing monotonically through the address space.
So we can have a sliding window of whole-stripes which
moves through the array suspending IO on other nodes,
and both resync which uses array addresses and recovery
which uses device addresses can stay within this window.
Signed-off-by: Guoqing Jiang <gqjiang@suse.com>
Signed-off-by: Shaohua Li <shli@fb.com>
If there is a resync going on, all nodes must suspend
writes to the range. This is recorded in suspend_info
and suspend_list.
If there is an I/O within the ranges of any of the
suspend_info, area_resyncing will return 1.
Signed-off-by: Guoqing Jiang <gqjiang@suse.com>
Signed-off-by: Shaohua Li <shli@fb.com>
Just like clustered raid1, it is impossible for cluster raid10
to choose the best device for read balance when the area of
array is resyncing. Because we cannot trust the data to be the
same on all devices at that time, so we choose just the first
one to use, so set do_balance to 0.
Signed-off-by: Guoqing Jiang <gqjiang@suse.com>
Signed-off-by: Shaohua Li <shli@fb.com>
If freeze_array is attempted in the middle of close_sync/
wait_all_barriers, deadlock can occur.
freeze_array will wait for nr_pending and nr_queued to line up.
wait_all_barriers increments nr_pending for each barrier bucket, one
at a time, but doesn't actually issue IO that could be counted in
nr_queued. So freeze_array is blocked until wait_all_barriers
completes and allow_all_barriers runs. At the same time, when
_wait_barrier sees array_frozen == 1, it stops and waits for
freeze_array to complete.
Prevent the deadlock by making close_sync call _wait_barrier and
_allow_barrier for one bucket at a time, instead of deferring the
_allow_barrier calls until after all _wait_barriers are complete.
Signed-off-by: Nate Dailey <nate.dailey@stratus.com>
Fix: fd76863e37fe(RAID1: a new I/O barrier implementation to remove resync window)
Reviewed-by: Coly Li <colyli@suse.de>
Cc: stable@vger.kernel.org (v4.11)
Signed-off-by: Shaohua Li <shli@fb.com>
Hi - I submit this patch for the next merge window:
Some times ago, I made a patch f9c79bc05a that blocks signals around the
schedule() calls in MD. The MD subsystem needs to do an uninterruptible
sleep that is not accounted in load average - so we block signals and use
interruptible sleep.
The kernel has a special TASK_IDLE state for this purpose, so we can use
it instead of blocking signals. This patch doesn't fix any bug, it just
makes the code simpler.
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Acked-by: NeilBrown <neilb@suse.com>
Signed-off-by: Shaohua Li <shli@fb.com>
