Use memdup_user to duplicate a memory region from user-space to
kernel-space, instead of open coding using kmalloc & copy_from_user.
Signed-off-by: Muhammad Falak R Wani <falakreyaz@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The MEN Chameleon specification states that a chameleon FPGA can include a
bridge descriptor, which then opens up a new bus behind this bridge. MCB
included subdevice handling code in the core, but no support for bus
descriptors in the parser, due to a lack of hardware access.
As this is technically dead code, but it gets executed on a device add,
I've decided to remove it.
Signed-off-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The to_mcb_{bus,device,driver}() macros lacked type safety, so convert them to
inline functions to enforce compile time type checking.
Signed-off-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Added support for the bar descriptor. This type is used for FPGAs
connect to the LPC or to a non PCI bus.
The Bar descriptor could have a maximum of 6 BARs. Each of the
devices within the FPGA could be mapped to a different BAR.
The BAR descriptor is comparable to the PCI header.
Signed-off-by: Andreas Werner <andreas.werner@men.de>
[ free bar descriptor in the non-error case ]
Signed-off-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add support for MCB bases FPGAs connected to the LPC or
non PCI Bus.
This driver currently supports the SC24 board. The FPGA
is connected to the LPC bus and is identified using the BIOS
DMI string.
Signed-off-by: Andreas Werner <andreas.werner@men.de>
Signed-off-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The variable i contains a total number of resources (including
IORESOURCE_IRQ). However, we want the dmem_region_start to point
after the last resource of type IORESOURCE_MEM. The original behaviour
leads (very likely) to skipping several UIO mapping regions and makes
them useless. Fix this by computing dmem_region_start from the uiomem
which points to the last used UIO mapping.
Fixes: 0a0c3b5a24 ("Add new uio device for dynamic memory allocation")
Signed-off-by: Jan Viktorin <viktorin@rehivetech.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
With this patch we add start/stop filtering as specified on
the perf cmd line. When the IP matches the start address
trace generation gets triggered. The stop condition is
achieved when the IP matches the stop address.
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This patch adds the capability to specify address ranges from
the perf cmd line using the --filter option. If the IP
falls within the range(s) program flow traces are generated.
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The include/exclude function of a tracer is applicable to address
range and start/stop filters. To avoid duplication and reuse code
moving the include/exclude configuration to a function of its own.
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Introducing a new function to do address range configuration
generic enough to work for any address range and any comparator.
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The default filter configuration was hard to read and included
some redundancy. This patch attempts to stream line configuration
and improve readability.
No change of functionality is included.
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Splitting the steps involved in the configuration of a tracer.
The first part is generic and can be reused for both sysFS and
Perf methods.
The second part pertains to the configuration of filters
themselves where the source of the information used to
configure the filters will vary depending on the access
methods.
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This patch implements the required API needed to access
and retrieve range and start/stop filters from the perf core.
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Both ETMv3 and ETMv4 drivers are declaring an 'enum etm_addr_type',
creating reduncancy.
This patch removes the enumeration from the driver files and adds
it to a common header.
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
With this commit [1] address range filter information is now found
in the struct hw_perf_event::addr_filters. As such pass the event
itself to the coresight_source::enable/disable() functions so that
both event attribute and filter can be accessible for configuration.
[1] 'commit 375637bc52 ("perf/core: Introduce address range filtering")'
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The ETM registers are classified into 2 categories: trace and management.
The core power domain contains most of the trace unit logic including
all(except TRCOSLAR and TRCOSLSR) the trace registers. The debug power
domain contains the external debugger interface including all management
registers.
This patch adds coresight unit specific function coresight_simple_func
which can be used for ETM trace registers by providing a ETM specific
read function which does smp cross call to ensure the trace core is
powered up before the register is accessed.
Cc: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The Coresight ETMv4 architecture provides a way to request to keep the
power to the trace unit. This might help to collect the traces without
the need to disable the CPU power management(entering/exiting deeper
idle states).
Trace PowerDown Control Register provides powerup request bit which when
set requests the system to retain power to the trace unit and emulate
the powerdown request.
Typically, a trace unit drives a signal to the power controller to
request that the trace unit core power domain is powered up. However,
if the trace unit and the CPU are in the same power domain then the
implementation might combine the trace unit power up status with a
signal from the CPU.
This patch requests to retain power to the trace unit when active and
to remove when inactive. Note this change will only request but the
behaviour depends on the implementation. However, it matches the
exact behaviour expected when the external debugger is connected with
respect to CPU power states.
Cc: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The kfree() function tests whether its argument is NULL and then
returns immediately. Thus the test around the call is not needed.
This issue was detected by using the Coccinelle software.
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Each coresight device prepares a description for coresight_register()
in struct coresight_desc. Once we register the device, the description is
useless and can be freed. The coresight_desc is small enough (48bytes on
64bit)i to be allocated on the stack. Hence use an automatic variable to
avoid a needless dynamic allocation and wasting the memory(which will only
be free'd when the device is destroyed).
Cc: Mathieu Poirier <mathieu.poirier@linaro.org>
Cc: Pratik Patel <pratikp@codeaurora.org>
Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
of_node_put needs to be called when the device node which is got
from of_parse_phandle has finished using.
Cc: linux-arm-kernel@lists.infradead.org
Cc: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Peter Chen <peter.chen@nxp.com>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
It is mandatory to enable a coresight block's power domain before
trying to access management registers. Otherwise the transaction
simply stalls, leading to a system hang.
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Reviewed-by: Sudeep Holla <sudeep.holla@arm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Depending on when CoreSight device are discovered it is possible
that some IP block may be referencing devices that have not been
added to the bus yet. The end result is missing nodes in the
CoreSight topology even when the devices are present and properly
initialised.
This patch solves the problem by asking the driver core to
try initialising the device at a later time when the children
of a CoreSight node are missing.
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When we encounter a timeout waiting for a status change via
coresight_timeout, the caller always print the offset which
was tried. This is pretty much useless as it doesn't specify
the bit position we wait for. Also, one needs to lookup the
TRM to figure out, what was wrong. This patch changes all
such error messages to print something more meaningful.
Cc: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Use the defined symbol rather than hardcoding the value to
check whether the TMC buffer is full.
Cc: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This patch cleans up the peripheral id table for different ETMv4
implementations.
As per Cortex-A53 TRM, the ETM has following id values:
Peripheral ID0 0x5D 0xFE0
Peripheral ID1 0xB9 0xFE4
Peripheral ID2 0x4B 0xFE8
Peripheral ID3 0x00 0xFEC
where, PID2: has the following format:
[7:4] Revision
[3] JEDEC 0b1 res1. Indicates a JEP106 identity code is used
[2:0] DES_1 0b011 ARM Limited. This is bits[6:4] of JEP106 ID code
The existing table entry checks only the bits [1:0], which is not
sufficient enough. Fix it to match bits [3:0], just like the other
entries do. While at it, correct the comment for A57 and the A53 entry.
Cc: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
At present the ETF or ETR gives out the entire device
buffer, even if there is less or even no trace data
available. This patch limits the trace data given out to
the actual trace data collected.
Cc: mathieu.poirier@linaro.org
Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This is a cleanup patch.
coresight_device->conns holds an array to point to the devices
connected to the OUT ports of a component. Sinks, e.g ETR, do not
have an OUT port (nr_outport = 0), as it streams the trace to
memory via AXI.
At coresight_register() we do :
conns = kcalloc(csdev->nr_outport, sizeof(*conns), GFP_KERNEL);
if (!conns) {
ret = -ENOMEM;
goto err_kzalloc_conns;
}
For ETR, since the total size requested for kcalloc is zero, the return
value is, ZERO_SIZE_PTR ( != NULL). Hence, csdev->conns = ZERO_SIZE_PTR
which cannot be verified later to contain a valid pointer. The code which
accesses the csdev->conns is bounded by the csdev->nr_outport check,
hence we don't try to dereference the ZERO_SIZE_PTR. This patch cleans
up the csdev->conns initialisation to make sure we initialise it
properly(i.e, either NULL or valid conns array).
Cc: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This patch cleans up the error handling path for tmc_probe
as a side effect of the removal of the spurious dma_free_coherent().
Cc: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit de5461970b ("coresight: tmc: allocating memory when needed")
removed the static allocation of buffer for the trace data in ETR mode in
tmc_probe. However it failed to remove the "devm_free_coherent" in
tmc_probe when the probe fails due to other reasons. This patch gets
rid of the incorrect dma_free_coherent() call.
Fixes: commit de5461970b ("coresight: tmc: allocating memory when needed")
Cc: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
etm4_trace_id is not guaranteed to be executed on the CPU whose ETM is
being accessed. This leads to exception similar to below one if the
CPU whose ETM is being accessed is in deeper idle states. So it must
be executed on the CPU whose ETM is being accessed.
Unhandled fault: synchronous external abort (0x96000210) at 0xffff000008db4040
Internal error: : 96000210 [#1] PREEMPT SMP
Modules linked in:
CPU: 5 PID: 5979 Comm: etm.sh Not tainted 4.7.0-rc3 #159
Hardware name: ARM Juno development board (r2) (DT)
task: ffff80096dd34b00 ti: ffff80096dfe4000 task.ti: ffff80096dfe4000
PC is at etm4_trace_id+0x5c/0x90
LR is at etm4_trace_id+0x3c/0x90
Call trace:
etm4_trace_id+0x5c/0x90
coresight_id_match+0x78/0xa8
bus_for_each_dev+0x60/0xa0
coresight_enable+0xc0/0x1b8
enable_source_store+0x3c/0x70
dev_attr_store+0x18/0x28
sysfs_kf_write+0x48/0x58
kernfs_fop_write+0x14c/0x1e0
__vfs_write+0x1c/0x100
vfs_write+0xa0/0x1b8
SyS_write+0x44/0xa0
el0_svc_naked+0x24/0x28
However, TRCTRACEIDR is not guaranteed to hold the previous programmed
trace id if it enters deeper idle states. Further, the trace id that is
computed in etm4_init_trace_id is programmed into TRCTRACEIDR only in
etm4_enable_hw which happens much later in the sequence after
coresight_id_match is executed from enable_source_store.
This patch simplifies etm4_trace_id by returning the stashed trace id
value similar to etm4_cpu_id.
Cc: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
CoreSight STM device allows direct mapping of the channel regions to
userspace for zero-copy writing. To support this ability, the STM
framework has provided a hook 'mmio_addr', this patch just implemented
this hook for CoreSight STM.
This patch also added an item into 'channel_space' to save the physical
base address of channel region which mmap operation needs to know.
Signed-off-by: Chunyan Zhang <zhang.chunyan@linaro.org>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If the addition of the coresight devices get deferred, then there's a
window before child_name is populated by of_get_coresight_platform_data
from the respective component driver's probe and the attempted to access
the same from coresight_orphan_match resulting in kernel NULL pointer
dereference as below:
Unable to handle kernel NULL pointer dereference at virtual address 0x0
Internal error: Oops: 96000004 [#1] PREEMPT SMP
Modules linked in:
CPU: 0 PID: 1038 Comm: kworker/0:1 Not tainted 4.7.0-rc3 #124
Hardware name: ARM Juno development board (r2) (DT)
Workqueue: events amba_deferred_retry_func
PC is at strcmp+0x1c/0x160
LR is at coresight_orphan_match+0x7c/0xd0
Call trace:
strcmp+0x1c/0x160
bus_for_each_dev+0x60/0xa0
coresight_register+0x264/0x2e0
tmc_probe+0x130/0x310
amba_probe+0xd4/0x1c8
driver_probe_device+0x22c/0x418
__device_attach_driver+0xbc/0x158
bus_for_each_drv+0x58/0x98
__device_attach+0xc4/0x160
device_initial_probe+0x10/0x18
bus_probe_device+0x94/0xa0
device_add+0x344/0x580
amba_device_try_add+0x194/0x238
amba_deferred_retry_func+0x48/0xd0
process_one_work+0x118/0x378
worker_thread+0x48/0x498
kthread+0xd0/0xe8
ret_from_fork+0x10/0x40
This patch adds a check for non-NULL conn->child_name before accessing
the same.
Cc: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reports for available memory should use the si_mem_available() value.
The previous freeram value does not include available page cache memory.
Signed-off-by: Alex Ng <alexng@messages.microsoft.com>
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
lockdep reports possible circular locking dependency when udev is used
for memory onlining:
systemd-udevd/3996 is trying to acquire lock:
((memory_chain).rwsem){++++.+}, at: [<ffffffff810d137e>] __blocking_notifier_call_chain+0x4e/0xc0
but task is already holding lock:
(&dm_device.ha_region_mutex){+.+.+.}, at: [<ffffffffa015382e>] hv_memory_notifier+0x5e/0xc0 [hv_balloon]
...
which is probably a false positive because we take and release
ha_region_mutex from memory notifier chain depending on the arg. No real
deadlocks were reported so far (though I'm not really sure about
preemptible kernels...) but we don't really need to hold the mutex
for so long. We use it to protect ha_region_list (and its members) and the
num_pages_onlined counter. None of these operations require us to sleep
and nothing is slow, switch to using spinlock with interrupts disabled.
While on it, replace list_for_each -> list_for_each_entry as we actually
need entries in all these cases, drop meaningless list_empty() checks.
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
With the recently introduced in-kernel memory onlining
(MEMORY_HOTPLUG_DEFAULT_ONLINE) these is no point in waiting for pages
to come online in the driver and we can get rid of the waiting.
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
I'm observing the following hot add requests from the WS2012 host:
hot_add_req: start_pfn = 0x108200 count = 330752
hot_add_req: start_pfn = 0x158e00 count = 193536
hot_add_req: start_pfn = 0x188400 count = 239616
As the host doesn't specify hot add regions we're trying to create
128Mb-aligned region covering the first request, we create the 0x108000 -
0x160000 region and we add 0x108000 - 0x158e00 memory. The second request
passes the pfn_covered() check, we enlarge the region to 0x108000 -
0x190000 and add 0x158e00 - 0x188200 memory. The problem emerges with the
third request as it starts at 0x188400 so there is a 0x200 gap which is
not covered. As the end of our region is 0x190000 now it again passes the
pfn_covered() check were we just adjust the covered_end_pfn and make it
0x188400 instead of 0x188200 which means that we'll try to online
0x188200-0x188400 pages but these pages were never assigned to us and we
crash.
We can't react to such requests by creating new hot add regions as it may
happen that the whole suggested range falls into the previously identified
128Mb-aligned area so we'll end up adding nothing or create intersecting
regions and our current logic doesn't allow that. Instead, create a list of
such 'gaps' and check for them in the page online callback.
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Windows 2012 (non-R2) does not specify hot add region in hot add requests
and the logic in hot_add_req() is trying to find a 128Mb-aligned region
covering the request. It may also happen that host's requests are not 128Mb
aligned and the created ha_region will start before the first specified
PFN. We can't online these non-present pages but we don't remember the real
start of the region.
This is a regression introduced by the commit 5abbbb75d7 ("Drivers: hv:
hv_balloon: don't lose memory when onlining order is not natural"). While
the idea of keeping the 'moving window' was wrong (as there is no guarantee
that hot add requests come ordered) we should still keep track of
covered_start_pfn. This is not a revert, the logic is different.
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
KVP daemon does fork()/exec() (with popen()) so we need to close our fds
to avoid sharing them with child processes. The immediate implication of
not doing so I see is SELinux complaining about 'ip' trying to access
'/dev/vmbus/hv_kvp'.
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
On Hyper-V, performance critical channels use the monitor
mechanism to signal the host when the guest posts mesages
for the host. This mechanism minimizes the hypervisor intercepts
and also makes the host more efficient in that each time the
host is woken up, it processes a batch of messages as opposed to
just one. The goal here is improve the throughput and this is at
the expense of increased latency.
Implement a mechanism to let the client driver decide if latency
is important.
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The current delay between retries is unnecessarily high and is negatively
affecting the time it takes to boot the system.
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
For synthetic NIC channels, enable explicit signaling policy as netvsc wants to
explicitly control when the host is to be signaled.
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There is a rare race when we remove an entry from the global list
hv_context.percpu_list[cpu] in hv_process_channel_removal() ->
percpu_channel_deq() -> list_del(): at this time, if vmbus_on_event() ->
process_chn_event() -> pcpu_relid2channel() is trying to query the list,
we can get the kernel fault.
Similarly, we also have the issue in the code path: vmbus_process_offer() ->
percpu_channel_enq().
We can resolve the issue by disabling the tasklet when updating the list.
The patch also moves vmbus_release_relid() to a later place where
the channel has been removed from the per-cpu and the global lists.
Reported-by: Rolf Neugebauer <rolf.neugebauer@docker.com>
Signed-off-by: Dexuan Cui <decui@microsoft.com>
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Background: userspace daemons registration protocol for Hyper-V utilities
drivers has two steps:
1) daemon writes its own version to kernel
2) kernel reads it and replies with module version
at this point we consider the handshake procedure being completed and we
do hv_poll_channel() transitioning the utility device to HVUTIL_READY
state. At this point we're ready to handle messages from kernel.
When hvutil_transport is in HVUTIL_TRANSPORT_CHARDEV mode we have a
single buffer for outgoing message. hvutil_transport_send() puts to this
buffer and till the buffer is cleared with hvt_op_read() returns -EFAULT
to all consequent calls. Host<->guest protocol guarantees there is no more
than one request at a time and we will not get new requests till we reply
to the previous one so this single message buffer is enough.
Now to the race. When we finish negotiation procedure and send kernel
module version to userspace with hvutil_transport_send() it goes into the
above mentioned buffer and if the daemon is slow enough to read it from
there we can get a collision when a request from the host comes, we won't
be able to put anything to the buffer so the request will be lost. To
solve the issue we need to know when the negotiation is really done (when
the version message is read by the daemon) and transition to HVUTIL_READY
state after this happens. Implement a callback on read to support this.
Old style netlink communication is not affected by the change, we don't
really know when these messages are delivered but we don't have a single
message buffer there.
Reported-by: Barry Davis <barry_davis@stormagic.com>
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
vmbus_teardown_gpadl() can result in infinite wait when it is called on 5
second timeout in vmbus_open(). The issue is caused by the fact that gpadl
teardown operation won't ever succeed for an opened channel and the timeout
isn't always enough. As a guest, we can always trust the host to respond to
our request (and there is nothing we can do if it doesn't).
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In some cases create_gpadl_header() allocates submessages but we never
free them.
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
We use messagecount only once in vmbus_establish_gpadl() to check if
it is safe to iterate through the submsglist. We can just initialize
the list header in all cases in create_gpadl_header() instead.
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When we crash from NMI context (e.g. after NMI injection from host when
'sysctl -w kernel.unknown_nmi_panic=1' is set) we hit
kernel BUG at mm/vmalloc.c:1530!
as vfree() is denied. While the issue could be solved with in_nmi() check
instead I opted for skipping vfree on all sorts of crashes to reduce the
amount of work which can cause consequent crashes. We don't really need to
free anything on crash.
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Joe Perches points out [1] that this pattern isn't currently safe.
This driver doesn't really need the zeroing semantic anyway;
by restructuring the code slightly we can initialize all the
fields of the structure up front instead.
[1] https://lkml.kernel.org/r/1469729491.3998.58.camel@perches.com
Signed-off-by: Chris Metcalf <cmetcalf@mellanox.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Muhammad Falak R Wani