| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
This patch unexports the low-level __get_user_pages() function.
Recent refactoring of the get_user_pages* functions allow flags to be
passed through get_user_pages() which eliminates the need for access to
this function from its one user, kvm.
We can see that the two calls to get_user_pages() which replace
__get_user_pages() in kvm_main.c are equivalent by examining their call
stacks:
get_user_page_nowait():
get_user_pages(start, 1, flags, page, NULL)
__get_user_pages_locked(current, current->mm, start, 1, page, NULL, NULL,
false, flags | FOLL_TOUCH)
__get_user_pages(current, current->mm, start, 1,
flags | FOLL_TOUCH | FOLL_GET, page, NULL, NULL)
check_user_page_hwpoison():
get_user_pages(addr, 1, flags, NULL, NULL)
__get_user_pages_locked(current, current->mm, addr, 1, NULL, NULL, NULL,
false, flags | FOLL_TOUCH)
__get_user_pages(current, current->mm, addr, 1, flags | FOLL_TOUCH, NULL,
NULL, NULL)
Signed-off-by: Lorenzo Stoakes <[email protected]>
Acked-by: Paolo Bonzini <[email protected]>
Acked-by: Michal Hocko <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
|
|
Now that Lorenzo cleaned things up and made the FOLL_FORCE users
explicit, it becomes obvious how some of them don't really need
FOLL_FORCE at all.
So remove FOLL_FORCE from the proc code that reads the command line and
arguments from user space.
The mem_rw() function actually does want FOLL_FORCE, because gdd (and
possibly many other debuggers) use it as a much more convenient version
of PTRACE_PEEKDATA, but we should consider making the FOLL_FORCE part
conditional on actually being a ptracer. This does not actually do
that, just moves adds a comment to that effect and moves the gup_flags
settings next to each other.
Signed-off-by: Linus Torvalds <[email protected]>
|
|
Report the exact number of bytes which have not been successfully
copied when an exception occurs, using the running remaining length.
Signed-off-by: David S. Miller <[email protected]>
|
|
The only times at which intel_pstate checks the policy set for
a given CPU is the initialization of that CPU and updates of its
policy settings from cpufreq when intel_pstate_set_policy() is
invoked.
That is insufficient, however, because intel_pstate uses the same
P-state selection function for all CPUs regardless of the policy
setting for each of them and the P-state limits are shared between
them. Thus if the policy is set to "performance" for a particular
CPU, it may not behave as expected if the cpufreq settings are
changed subsequently for another CPU.
That can be easily demonstrated by writing "performance" to
scaling_governor for all CPUs and then switching it to "powersave"
for one of them in which case all of the CPUs will behave as though
their scaling_governor were all "powersave" (even though the policy
still appears to be "performance" for the remaining CPUs).
Fix this problem by modifying intel_pstate_adjust_busy_pstate() to
always set the P-state to the maximum allowed by the current limits
for all CPUs whose policy is set to "performance".
Note that it still is recommended to always change the policy setting
in the same way for all CPUs even with this fix applied to avoid
confusion.
Signed-off-by: Rafael J. Wysocki <[email protected]>
|
|
Commit 0eadf37afc250 ("nbd: allow block mq to deal with timeouts")
changed normal usage of nbd->sock_lock to use spin_lock/spin_unlock
rather than the *_irq variants, but it missed this unlock in an
error path.
Found by Coverity, CID 1373871.
Signed-off-by: John W. Linville <[email protected]>
Cc: Josef Bacik <[email protected]>
Cc: Jens Axboe <[email protected]>
Cc: Markus Pargmann <[email protected]>
Fixes: 0eadf37afc250 ("nbd: allow block mq to deal with timeouts")
Signed-off-by: Jens Axboe <[email protected]>
|
|
Instead use d_fsdata which is the same size. Hoping to get rid of d_time,
which is used by very few filesystems by this time.
Signed-off-by: Miklos Szeredi <[email protected]>
Reviewed-by: Martin Brandenburg <[email protected]>
Signed-off-by: Mike Marshall <[email protected]>
|
|
Report the exact number of bytes which have not been successfully
copied when an exception occurs, using the running remaining length.
Signed-off-by: David S. Miller <[email protected]>
|
|
Report the exact number of bytes which have not been successfully
copied when an exception occurs, using the running remaining length.
Signed-off-by: David S. Miller <[email protected]>
|
|
Report the exact number of bytes which have not been successfully
copied when an exception occurs, using the running remaining length.
Signed-off-by: David S. Miller <[email protected]>
|
|
The fixup helper function mechanism for handling user copy fault
handling is not %100 accurrate, and can never be made so.
We are going to transition the code to return the running return
return length, which is always kept track in one or more registers
of each of these routines.
In order to convert them one by one, we have to allow the existing
behavior to continue functioning.
Therefore make all the copy code that wants the fixup helper to be
used return negative one.
After all of the user copy routines have been converted, this logic
and the fixup helpers themselves can be removed completely.
Signed-off-by: David S. Miller <[email protected]>
|
|
Replace wrong use of file->f_path.dentry->d_inode with file_inode(file).
In case orangefs ever finds itself as an overelayfs layer, it would want
to get its own inode and not overlayfs's inode.
DISCLAIMER: I did not test this patch because I do not know how to setup
an orangefs mount
Signed-off-by: Amir Goldstein <[email protected]>
Signed-off-by: Mike Marshall <[email protected]>
|
|
It is completely unused.
Signed-off-by: David S. Miller <[email protected]>
|
|
Fix [-Wold-style-declaration] GCC warnings by moving the inline keyword
before the return type.
Signed-off-by: Tobias Klauser <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
Fix [-Wold-style-declaration] GCC warnings by moving the inline keyword
before the return type.
Signed-off-by: Tobias Klnuser <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
Individual scheduler domain should consist different hierarchy
consisting of cores sharing similar property. Currently, no
scheduler domain is defined separately for the cores that shares
the last level cache. As a result, the scheduler fails to take
advantage of cache locality while migrating tasks during load
balancing.
Here are the cpu masks currently present for sparc that are/can
be used in scheduler domain construction.
cpu_core_map : set based on the cores that shares l1 cache.
core_core_sib_map : is set based on the socket id.
The prior SPARC notion of socket was defined as highest level of
shared cache. However, the MD record on T7 platforms now describes
the CPUs that share the physical socket and this is no longer tied
to shared cache.
That's why a separate cpu mask needs to be created that truly
represent highest level of shared cache for all platforms.
Signed-off-by: Atish Patra <[email protected]>
Reviewed-by: Chris Hyser <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
SEC registers are not accessible when the TXE device is in low power
state, hence the SEC interrupt cannot be processed if device is not
awake.
In some rare cases entrance to low power state (aliveness off) and input
ready bits can be signaled at the same time, resulting in communication
stall as input ready won't be signaled again after waking up. To resolve
this IPC_HHIER_SEC bit in HHISR_REG should not be cleaned if the
interrupt is not processed.
Cc: [email protected]
Signed-off-by: Alexander Usyskin <[email protected]>
Signed-off-by: Tomas Winkler <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
|
|
struct
Prevents leaking pointers between processes
Signed-off-by: Arve Hjønnevåg <[email protected]>
Signed-off-by: Martijn Coenen <[email protected]>
Cc: stable <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
|
|
Prevent using a binder_ref with only weak references where a strong
reference is required.
Signed-off-by: Arve Hjønnevåg <[email protected]>
Signed-off-by: Martijn Coenen <[email protected]>
Cc: stable <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
|
|
Instantiated SPI device nodes are marked with OF_POPULATE. This was
introduced in bd6c164. On unloading, loaded device nodes will of course
be unmarked. The problem are nodes that fail during initialisation: If a
node fails, it won't be unloaded and hence not be unmarked.
If a SPI driver module is unloaded and reloaded, it will skip nodes that
failed before.
Skip device nodes that are already populated and mark them only in case
of success.
Note that the same issue exists for I2C.
Fixes: bd6c164 ("spi: Mark instantiated device nodes with OF_POPULATE")
Signed-off-by: Ralf Ramsauer <[email protected]>
Reviewed-by: Geert Uytterhoeven <[email protected]>
Acked-by: Pantelis Antoniou <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
Cc: [email protected]
|
|
if user disables IOC from debugger at startup (by clearing @ioc_enable),
@ioc_exists is cleared too. This means boot prints don't capture the
fact that IOC was present but disabled which could be misleading.
So invert how we use @ioc_enable and @ioc_exists and make it more
canonical. @ioc_exists represent whether hardware is present or not and
stays same whether enabled or not. @ioc_enable is still user driven,
but will be auto-disabled if IOC hardware is not present, i.e. if
@ioc_exist=0. This is opposite to what we were doing before, but much
clearer.
This means @ioc_enable is now the "exported" toggle in rest of code such
as dma mapping API.
Signed-off-by: Vineet Gupta <[email protected]>
|
|
Older ARC700 cores (ARC750 specifically) lack instructions to implement
atomic r-w-w. This is problematic for userspace libraries such as NPTL
which need atomic primitives. So enable them by providing kernel assist.
This is costly but really the only sane soluton (othern than tight
spinning using the otherwise availiable atomic exchange EX instruciton).
Good thing is there are only a few of these cores running Linux out in
the wild.
This only works on UP systems.
Reviewed-by: Colin Ian King <[email protected]>
Signed-off-by: Vineet Gupta <[email protected]>
|
|
dm_get_target_type() was previously called so any error returned from
dm_table_add_target() must first call dm_put_target_type(). Otherwise
the DM target module's reference count will leak and the associated
kernel module will be unable to be removed.
Also, leverage the fact that r is already -EINVAL and remove an extra
newline.
Fixes: 36a0456 ("dm table: add immutable feature")
Fixes: cc6cbe1 ("dm table: add always writeable feature")
Fixes: 3791e2f ("dm table: add singleton feature")
Cc: [email protected] # 3.2+
Signed-off-by: tang.junhui <[email protected]>
Signed-off-by: Mike Snitzer <[email protected]>
|
|
Don't ignore errors here: Set backend state to unknown when
unsuccessful.
Signed-off-by: Jan Beulich <[email protected]>
Signed-off-by: David Vrabel <[email protected]>
|
|
This is more efficient than list_for_each_safe() when list modification
is accompanied by breaking out of the loop.
Signed-off-by: Jan Beulich <[email protected]>
Reviewed-by: Juergen Gross <[email protected]>
Signed-off-by: David Vrabel <[email protected]>
|
|
Three newly introduced functions are not defined when CONFIG_XEN_PVHVM is
disabled, but are still being used:
arch/x86/xen/enlighten.c:141:12: warning: ‘xen_cpu_up_prepare’ used but never defined
arch/x86/xen/enlighten.c:142:12: warning: ‘xen_cpu_up_online’ used but never defined
arch/x86/xen/enlighten.c:143:12: warning: ‘xen_cpu_dead’ used but never defined
Fixes: 4d737042d6c4 ("xen/x86: Convert to hotplug state machine")
Signed-off-by: Arnd Bergmann <[email protected]>
Signed-off-by: David Vrabel <[email protected]>
|
|
The Xen docs specify several flags which a guest can set to advertise
which values of the xenstore control/shutdown key it will recognize.
This patch adds code to write all the relevant feature-flag keys.
Based-on-patch-by: Paul Durrant <[email protected]>
Signed-off-by: Juergen Gross <[email protected]>
Reviewed-by: David Vrabel <[email protected]>
Reviewed-by: Paul Durrant <[email protected]>
Signed-off-by: David Vrabel <[email protected]>
|
|
We should release ->interfaces[0] as well.
Fixes: b15d97d77017 ("greybus: core: add module abstraction")
Signed-off-by: Dan Carpenter <[email protected]>
Acked-by: Viresh Kumar <[email protected]>
Acked-by: Johan Hovold <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
|
|
Fix to return a negative error code from the es2_arpc_in_enable() error
handling case instead of 0, as done elsewhere in this function.
Fixes: 9d9d3777a9db ("greybus: es2: Add a new bulk in endpoint for APBridgeA RPC")
Signed-off-by: Wei Yongjun <[email protected]>
Acked-by: Viresh Kumar <[email protected]>
Reviewed-by: Johan Hovold <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
|
|
arche_platform_change_state()
This node pointer is returned by of_find_compatible_node() with
refcount incremented in this function. of_node_put() on it before
exitting this function.
This is detected by Coccinelle semantic patch.
Signed-off-by: Wei Yongjun <[email protected]>
Acked-by: Johan Hovold <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
|
|
If the copy_to_user() fails we should unlock and return directly without
updating "cnt". Also the return value should be -EFAULT instead of the
number of bytes remaining.
Fixes: 02b23803c6af ("staging: android: ion: Add ioctl to query available heaps")
Signed-off-by: Dan Carpenter <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
|
|
AR8031 and AR8035 have the same PHY ID mask of 0xffffffef.
So fix it and make it match with the PHY ID mask definition
at drivers/net/phy/at803x.c.
Signed-off-by: Fabio Estevam <[email protected]>
Signed-off-by: Shawn Guo <[email protected]>
|
|
Commit 4bcc595ccd80 (printk: reinstate KERN_CONT for printing
continuation lines) exposed a missing KERN_CONT from one of the
messages shown on entering suspend. With v4.9-rc1, the 'done.' shown
after syncing the filesystems no longer appears as a continuation but
a new message with its own timestamp.
[ 9.259566] PM: Syncing filesystems ... [ 9.264119] done.
Fix this by adding the KERN_CONT log level for the 'done.' part of the
message seen after syncing filesystems. While we are at it, convert
these suspend printks to pr_info and pr_cont, respectively.
Signed-off-by: Jon Hunter <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>
|
|
Since the controller on R-Car Gen3 doesn't have any status registers
to detect initialization (LPSTS.SUSPM = 1) and the initialization needs
up to 45 usec, this patch adds wait after the initialization. Otherwise,
writing other registers (e.g. INTENB0) will fail.
Fixes: de18757e272d ("usb: renesas_usbhs: add R-Car Gen3 power control")
Cc: <[email protected]> # v4.6+
Cc: <[email protected]>
Signed-off-by: Yoshihiro Shimoda <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
|
|
Although ghes_proc() tests for errors while reading the error status,
it always return success (0). Fix this by propagating the return
value.
Fixes: d334a49113a4a33 (ACPI, APEI, Generic Hardware Error Source memory error support)
Signed-of-by: Punit Agrawal <[email protected]>
Tested-by: Tyler Baicar <[email protected]>
Reviewed-by: Borislav Petkov <[email protected]>
[ rjw: Subject ]
Signed-off-by: Rafael J. Wysocki <[email protected]>
|
|
Increase ohci watchout delay to 275 ms. Previous delay was 250 ms
with 20 ms of slack, after removing slack time some ohci controllers don't
respond in time. Logs from systems with controllers that have the
issue would show "HcDoneHead not written back; disabled"
Signed-off-by: Bryan Paluch <[email protected]>
Acked-by: Alan Stern <[email protected]>
Cc: stable <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
|
|
If we're booting pandaboard using NFSroot over built-in g_ether, we
can get the following after booting once and doing a warm reset:
g_ether gadget: ecm_open
g_ether gadget: notify connect true
...
WARNING: CPU: 0 PID: 1 at drivers/bus/omap_l3_noc.c:147
l3_interrupt_handler+0x220/0x34c
44000000.ocp:L3 Custom Error: MASTER MPU TARGET L4CFG (Read):
Data Access in User mode du ring Functional access
...
Fix the issue by calling pm_runtime functions from
musb_gadget_queue.
Note that in the long run we should be able to queue the pending
transfers if pm_runtime is not active, and flush the queue from
pm_runtime_resume.
Reported-by: Laurent Pinchart <[email protected]>
Tested-by: Laurent Pinchart <[email protected]>
Signed-off-by: Tony Lindgren <[email protected]>
Signed-off-by: Bin Liu <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
|
|
If we configure musb with 2430 glue as a peripheral, and then rmmod
omap2430 module, we'll get the following error:
[ INFO: HARDIRQ-safe -> HARDIRQ-unsafe lock order detected ]
...
rmmod/413 [HC0[0]:SC0[0]:HE0:SE1] is trying to acquire:
(&phy->mutex){+.+.+.}, at: [<c04b9fd0>] phy_power_off+0x1c/0xb8
[ 204.678710]
and this task is already holding:
(&(&musb->lock)->rlock){-.-...}, at: [<bf3a482c>]
musb_gadget_stop+0x24/0xec [musb_hdrc]
which would create a new lock dependency:
(&(&musb->lock)->rlock){-.-...} -> (&phy->mutex){+.+.+.}
...
This is because some glue layers expect musb_platform_enable/disable
to be called with spinlock held, and 2430 glue layer has USB PHY on
the I2C bus using a mutex.
We could fix the glue layers to take the spinlock, but we still have
a problem of musb_plaform_enable/disable being called in an unbalanced
manner. So that would still lead into USB PHY enable/disable related
problems for omap2430 glue layer.
While it makes sense to only enable USB PHY when needed from PM point
of view, in this case we just can't do it yet without breaking things.
So let's just revert phy_enable/disable related changes instead and
reconsider this after we have fixed musb_platform_enable/disable to
be balanced.
Fixes: a83e17d0f73b ("usb: musb: Improve PM runtime and phy handling for 2430 glue layer")
Reviewed-by: Laurent Pinchart <[email protected]>
Signed-off-by: Tony Lindgren <[email protected]>
Signed-off-by: Bin Liu <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
|
|
Socionext LD11 SoC (arch/arm64/boot/dts/socionext/uniphier-ld11.dtsi)
needs to handle 4 reset lines for EHCI.
Signed-off-by: Masahiro Yamada <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
|
|
The reset value of RWC is 0, set RemoteWakeupConnected bit explicitly
before calling ohci_run, it also fixes the issue that the mass
storage stick connected wasn't suspended when the system suspend.
Signed-off-by: Wenyou Yang <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
|
|
Commit 103544d86976 ("ACPI,PCI,IRQ: reduce resource requirements")
replaced the addition of PIRQ_PENALTY_PCI_USING in acpi_pci_link_allocate()
with an addition in acpi_irq_pci_sharing_penalty(), but f7eca374f000
("ACPI,PCI,IRQ: separate ISA penalty calculation") removed the use
of acpi_irq_pci_sharing_penalty() for ISA IRQs.
Therefore, PIRQ_PENALTY_PCI_USING is missing from ISA IRQs used by
interrupt links. Include that penalty by adding it in the
acpi_pci_link_allocate() path.
Fixes: f7eca374f000 (ACPI,PCI,IRQ: separate ISA penalty calculation)
Signed-off-by: Sinan Kaya <[email protected]>
Acked-by: Bjorn Helgaas <[email protected]>
Tested-by: Jonathan Liu <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>
|
|
Ondrej reported that IRQs stopped working in v4.7 on several
platforms. A typical scenario, from Ondrej's VT82C694X/694X, is:
ACPI: Using PIC for interrupt routing
ACPI: PCI Interrupt Link [LNKA] (IRQs 1 3 4 5 6 7 10 *11 12 14 15)
ACPI: No IRQ available for PCI Interrupt Link [LNKA]
8139too 0000:00:0f.0: PCI INT A: no GSI
We're using PIC routing, so acpi_irq_balance == 0, and LNKA is already
active at IRQ 11. In that case, acpi_pci_link_allocate() only tries
to use the active IRQ (IRQ 11) which also happens to be the SCI.
We should penalize the SCI by PIRQ_PENALTY_PCI_USING, but
irq_get_trigger_type(11) returns something other than
IRQ_TYPE_LEVEL_LOW, so we penalize it by PIRQ_PENALTY_ISA_ALWAYS
instead, which makes acpi_pci_link_allocate() assume the IRQ isn't
available and give up.
Add acpi_penalize_sci_irq() so platforms can tell us the SCI IRQ,
trigger, and polarity directly and we don't have to depend on
irq_get_trigger_type().
Fixes: 103544d86976 (ACPI,PCI,IRQ: reduce resource requirements)
Link: http://lkml.kernel.org/r/[email protected]
Reported-by: Ondrej Zary <[email protected]>
Acked-by: Bjorn Helgaas <[email protected]>
Signed-off-by: Sinan Kaya <[email protected]>
Tested-by: Jonathan Liu <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>
|
|
We do not want to store the SCI penalty in the acpi_isa_irq_penalty[]
table because acpi_isa_irq_penalty[] only holds ISA IRQ penalties and
there's no guarantee that the SCI is an ISA IRQ. We add in the SCI
penalty as a special case in acpi_irq_get_penalty().
But if we called acpi_penalize_isa_irq() or acpi_irq_penalty_update()
for an SCI that happened to be an ISA IRQ, they stored the SCI
penalty (part of the acpi_irq_get_penalty() return value) in
acpi_isa_irq_penalty[]. Subsequent calls to acpi_irq_get_penalty()
returned a penalty that included *two* SCI penalties.
Fixes: 103544d86976 (ACPI,PCI,IRQ: reduce resource requirements)
Signed-off-by: Sinan Kaya <[email protected]>
Acked-by: Bjorn Helgaas <[email protected]>
Tested-by: Jonathan Liu <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>
|
|
The global timer IRQ (PPI[0], PPI 11 in device tree terms) is a
rising edge interrupt. The ARM Cortex-A5 MPCore TRM in Chapter
10.1.2. Interrupt types and sources says:
"Interrupt is rising-edge sensitive."
The bits seem to be read-only, hence this missconfiguration had
no negative effect. However, with commit 992345a58e0c
("irqchip/gic: WARN if setting the interrupt type for a PPI fails")
warnings such as this get printed:
GIC: PPI11 is secure or misconfigured
With this change the new configuration matches the default
configuration and no warning is printed anymore.
Signed-off-by: Stefan Agner <[email protected]>
Signed-off-by: Shawn Guo <[email protected]>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio into staging-linus
Jonathan writes:
First set of IIO fixes for the 4.9 cycle.
* atlas chemical
- Fix alignment of big endian values in a larger storage (by using the right
size storage)
* maxim thermocouple
- Fix alignment of big endian values in larger (by using the correct
sized storage).
* sca3000
- Handle unexpected mode values.
* ti-adc081
- Select IIO_TRIGGERED_BUFFER to avoid build errors
|
|
This fixes a race condition where one thread that is entering or
leaving a power-saving state can inadvertently ignore the lock bit
that was set by another thread, and potentially also clear it.
The core_idle_lock_held function is called when the lock bit is
seen to be set. It polls the lock bit until it is clear, then
does a lwarx to load the word containing the lock bit and thread
idle bits so it can be updated. However, it is possible that the
value loaded with the lwarx has the lock bit set, even though an
immediately preceding lwz loaded a value with the lock bit clear.
If this happens then we go ahead and update the word despite the
lock bit being set, and when called from pnv_enter_arch207_idle_mode,
we will subsequently clear the lock bit.
No identifiable misbehaviour has been attributed to this race.
This fixes it by checking the lock bit in the value loaded by the
lwarx. If it is set then we just go back and keep on polling.
Fixes: b32aadc1a8ed ("powerpc/powernv: Fix race in updating core_idle_state")
Cc: [email protected] # v4.2+
Signed-off-by: Paul Mackerras <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
|
|
Commit 8117ac6a6c2f ("powerpc/powernv: Switch off MMU before entering
nap/sleep/rvwinkle mode", 2014-12-10) fixed a race condition where one
thread entering a KVM guest could switch the MMU context to the guest
while another thread was still in host kernel context with the MMU on.
That commit moved the point where a thread entering a power-saving
mode set its kvm_hstate.hwthread_state field in its PACA to
KVM_HWTHREAD_IN_IDLE from a point where the MMU was on to after the
MMU had been switched off. That commit also added a comment
explaining that we have to switch to real mode before setting
hwthread_state to avoid this race.
Nevertheless, commit 4eae2c9ae54a ("powerpc/powernv: Make
pnv_powersave_common more generic", 2016-07-08) subsequently moved
the setting of hwthread_state back to a point where the MMU is on,
thus reintroducing the race, despite the comment saying that this
should not be done being included in full in the context lines of
the patch that did it.
This fixes the race again and adds a bigger and shoutier comment
explaining the potential race condition.
Fixes: 4eae2c9ae54a ("powerpc/powernv: Make pnv_powersave_common more generic")
Cc: [email protected] # v4.8+
Signed-off-by: Paul Mackerras <[email protected]>
Reviewed-by: Shreyas B. Prabhu <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
|
|
Standby (hotplug) memory should be added to ZONE_MOVABLE on s390. After
commit 199071f1 "s390/mm: make arch_add_memory() NUMA aware",
arch_add_memory() used memblock_end_of_DRAM() to find out the end of
ZONE_NORMAL and the beginning of ZONE_MOVABLE. However, commit 7f36e3e5
"memory-hotplug: add hot-added memory ranges to memblock before allocate
node_data for a node." moved the call of memblock_add_node() before
the call of arch_add_memory() in add_memory_resource(), and thus changed
the return value of memblock_end_of_DRAM() when called in
arch_add_memory(). As a result, arch_add_memory() will think that all
memory blocks should be added to ZONE_NORMAL.
Fix this by changing the logic in arch_add_memory() so that it will
manually iterate over all zones of a given node to find out which zone
a memory block should be added to.
Reviewed-by: Heiko Carstens <[email protected]>
Signed-off-by: Gerald Schaefer <[email protected]>
Signed-off-by: Martin Schwidefsky <[email protected]>
|
|
Use pr_cont instead of printk calls also within show_stack and
die in order to avoid extra line breaks.
Signed-off-by: Heiko Carstens <[email protected]>
Signed-off-by: Martin Schwidefsky <[email protected]>
|
|
If of_genpd_add_provider_onecell() fails the following kernel crash is
observed on a kernel built with multi_v7_defconfig:
[ 1.739301] [00000040] *pgd=00000000
[ 1.739310] Internal error: Oops: 5 [#1] SMP ARM
[ 1.739319] Modules linked in:
[ 1.739328] CPU: 1 PID: 95 Comm: kworker/1:4 Not tainted 4.8.0-11897-g6b5e09a #1
[ 1.739331] Hardware name: Freescale i.MX6 Quad/DualLite (Device Tree)
[ 1.739352] Workqueue: pm genpd_power_off_work_fn
[ 1.739356] task: ee63d400 task.stack: ee70a000
[ 1.739365] PC is at mutex_lock+0xc/0x4c
[ 1.739374] LR is at regulator_disable+0x2c/0x60
[ 1.739379] pc : [<c0bc0da0>] lr : [<c06e4b10>] psr: 60000013
[ 1.739379] sp : ee70beb0 ip : 10624dd3 fp : ee6e6280
[ 1.739382] r10: eefb0900 r9 : 00000000 r8 : c1309918
[ 1.739385] r7 : 00000000 r6 : 00000040 r5 : 00000000 r4 : 00000040
[ 1.739390] r3 : 0000004c r2 : 7fffd540 r1 : 000001e4 r0 : 00000040
Instead of returning of_genpd_add_provider_onecell() directly,
we should check its return value and in the case of error we
should unwind the previously taken actions, which in these case are:
- Call imx6q_pm_pu_power_off()
- Set imx6q_pu_domain.reg back to NULL
Setting imx6q_pu_domain.reg to NULL in the error case is important
as it will prevent further operations in the pu_reg regulator.
This kernel crash is not observed with imx_v6_v7_defconfig because
it selects GPU and VPU drivers, which are consumers of the GPC block
and thus change the refcount of the pu_reg regulator.
Signed-off-by: Fabio Estevam <[email protected]>
Signed-off-by: Shawn Guo <[email protected]>
|
|
Since commit 0159ec670763dd ("PM / Domains: Verify the PM domain is present
when adding a provider") the following regression is observed on imx6:
imx-gpc: probe of 20dc000.gpc failed with error -22
The gpc probe fails because of_genpd_add_provider_onecell() now checks
if all the domains are initialized via pm_genpd_present() function
and it fails because not all the power domains are initialized.
In order to fix this error, initialize all the power domains from
imx_gpc_domains[], not only the imx6q_pu_domain.base one.
Reported-by: Olof's autobooter <[email protected]>
Signed-off-by: Fabio Estevam <[email protected]>
Signed-off-by: Shawn Guo <[email protected]>
|