On the interrupt path, we repeatedly establish the pointer to the
storvsc_device. While the compiler does inline get_in_stor_device() (and
other static functions) in the call chain in the interrupt path, the
compiler is repeatedly inlining the call to get_in_stor_device() each
time it is invoked. The return value of get_in_stor_device() can be
cached in the interrupt path since there is higher level serialization
in place to ensure correct handling when the module unload races with
the processing of an incoming message from the host. Optimize this code
path by caching the pointer to storvsc_device and passing it as an
argument.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reviewed-by: Long Li <[email protected]>
Reviewed-by: Johannes Thumshirn <[email protected]>
Reviewed-by: Hannes Reinecke <[email protected]>
Tested-by: Alex Ng <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
The function storvsc_channel_init() repeatedly interacts with the host
to extract various channel properties. Refactor this code to eliminate
code repetition.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reviewed-by: Long Li <[email protected]>
Reviewed-by: Johannes Thumshirn <[email protected]>
Reviewed-by: Hannes Reinecke <[email protected]>
Tested-by: Alex Ng <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
For FC devices managed by this driver, atttach the appropriate transport
template. This will allow us to create the appropriate sysfs files for
these devices. With this we can publish the wwn for both the port and the node.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reviewed-by: Long Li <[email protected]>
Tested-by: Alex Ng <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
For performance critical devices, we distribute the incoming
channel interrupt load across available CPUs in the guest.
Include Fibre channel devices in the set of devices for which
we would distribute the interrupt load.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
When the handshake with daemon is complete, we should poll the channel since
during the handshake, we will not be processing any messages. This is a
potential bug if the host is waiting for a response from the guest.
I would like to thank Dexuan for pointing this out.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Force all channel messages to be delivered on CPU0. These messages are not
performance critical and are used during the setup and teardown of the
channel.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Currently we have two policies for deciding when to signal the host:
One based on the ring buffer state and the other based on what the
VMBUS client driver wants to do. Consider the case when the client
wants to explicitly control when to signal the host. In this case,
if the client were to defer signaling, we will not be able to signal
the host subsequently when the client does want to signal since the
ring buffer state will prevent the signaling. Implement logic to
have only one signaling policy in force for a given channel.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reviewed-by: Haiyang Zhang <[email protected]>
Tested-by: Haiyang Zhang <[email protected]>
Cc: <[email protected]> # v4.2+
Signed-off-by: Greg Kroah-Hartman <[email protected]>
The irq we extract from ACPI is not used - we deliver hypervisor
interrupts on a special vector. Make the necessary adjustments.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
The macro VMBUS_DEVICE() is unused; get rid of it.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Consistently use uuid_le type in the Hyper-V driver code.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Util services such as KVP and FCOPY need assistance from daemon's running
in user space. Increase the timeout so we don't prematurely terminate
the transaction in the kernel. Host sets up a 60 second timeout for
all util driver transactions. The host will retry the transaction if it
times out. Set the guest timeout at 30 seconds.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Eliminate vlan_tci from struct hv_netvsc_packet.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Eliminate xmit_more from struct hv_netvsc_packet.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Eliminate completion_func from struct hv_netvsc_packet.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Eliminate is_data_pkt from struct hv_netvsc_packet.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Eliminate send_completion_tid from struct hv_netvsc_packet.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
The rndis header is 116 bytes big and can be placed in the default
head room that will be available in the skb. Since the netvsc packet
is less than 48 bytes, we can use the skb control buffer
for the netvsc packet. With these changes we don't need to
ask for additional head room.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reviewed-by: Haiyang Zhang <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Eliminate send_completion_ctx from struct hv_netvsc_packet.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reviewed-by: Haiyang Zhang <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Eliminate send_completion from struct hv_netvsc_packet.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reviewed-by: Haiyang Zhang <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Eliminatte the data field from struct hv_netvsc_packet.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reviewed-by: Haiyang Zhang <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Eliminate rndis_msg pointer from hv_netvsc_packet structure.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reviewed-by: Haiyang Zhang <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Eliminate the channel field in hv_netvsc_packet structure.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reviewed-by: Haiyang Zhang <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Rearrange the elements of struct hv_negtvsc_packet for optimal layout -
eliminate unnecessary padding.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reviewed-by: Haiyang Zhang <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
As part of reducing the size of the hv_netvsc_packet, resize some of the
variables based on their usage.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reviewed-by: Haiyang Zhang <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
SRB status can have additional information. Mask these out before processing
SRB status.
This patch was sent as part of a collection of patches more than a year ago.
While the rest of the patches in the set were comitted, this patch was not.
I woulod like to thank Olaf for noticing that this patch was not committed
upstream.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reviewed-by: Long Li <[email protected]>
Signed-off-by: James Bottomley <[email protected]>
Don't set the SRB_FLAGS_QUEUE_ACTION_ENABLE flag since we are not specifying
tags. Without this, the qlogic driver doesn't work properly with storvsc.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Cc: [email protected]
Signed-off-by: James Bottomley <[email protected]>
In the function storvsc_channel_init(), error code was not getting
set correctly in some of the failure cases. Fix this issue.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reported-by: Dan Carpenter <[email protected]>
Signed-off-by: James Bottomley <[email protected]>
The current Hyper-V clock source is based on the per-partition reference counter
and this counter is being accessed via s synthetic MSR - HV_X64_MSR_TIME_REF_COUNT.
Hyper-V has a more efficient way of computing the per-partition reference
counter value that does not involve reading a synthetic MSR. We implement
a time source based on this mechanism.
Tested-by: Vivek Yadav <[email protected]>
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
The current code tracks the assigned CPUs within a NUMA node in the context of
the primary channel. So, if we have a VM with a single NUMA node with 8 VCPUs, we may
end up unevenly distributing the channel load. Fix the issue by tracking affiliations
globally.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
We cycle through all the "high performance" channels to distribute
load across the available CPUs. Process the NetworkDirect as a
high performance device.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
The guest may have to send a completion packet back to the host.
To support this usage, permit sending a packet without a payload -
we would be only sending the descriptor in this case.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
The current code returns from probe without waiting for the proper handling
of subchannels that may be requested. If the netvsc driver were to be rapidly
loaded/unloaded, we can trigger a panic as the unload will be tearing
down state that may not have been fully setup yet. We fix this issue by making
sure that we return from the probe call only after ensuring that the
sub-channel offers in flight are properly handled.
Reviewed-and-tested-by: Haiyang Zhang <[email protected]
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Allocate ring buffer memory from the NUMA node assigned to the channel.
Since this is a performance and not a correctness issue, if the node specific
allocation were to fail, fall back and allocate without specifying the node.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Channels/sub-channels can be affinitized to VCPUs in the guest. Implement
this affinity in a way that is NUMA aware. The current protocol distributed
the primary channels uniformly across all available CPUs. The new protocol
is NUMA aware: primary channels are distributed across the available NUMA
nodes while the sub-channels within a primary channel are distributed amongst
CPUs within the NUMA node assigned to the primary channel.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Map target_cpu to target_vcpu using the mapping table.
We should use the mapping table to transform guest CPU ID to VP Index
as is done for the non-performance critical channels.
While the value CPU 0 is special and will
map to VP index 0, it is good to be consistent.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Allocate the receive bufer from the NUMA node assigned to the primary
channel.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
The current algorithm for deciding on the number of VRSS channels is
not optimal since we open up the min of number of CPUs online and the
number of VRSS channels the host is offering. So on a 32 VCPU guest
we could potentially open 32 VRSS subchannels. Experimentation has
shown that it is best to limit the number of VRSS channels to the number
of CPUs within a NUMA node.
Here is the new algorithm for deciding on the number of sub-channels we
would open up:
1) Pick the minimum of what the host is offering and what the driver
in the guest is specifying as the default value.
2) Pick the minimum of (1) and the numbers of CPUs in the NUMA
node the primary channel is bound to.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Implement the protocol for tearing down the monitor state established with
the host.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Tested-by: Vitaly Kuznetsov <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Based on the information given to this driver (via the xmit_more skb flag),
we can defer signaling the host if more packets are on the way. This will help
make the host more efficient since it can potentially process a larger batch of
packets. Implement this optimization.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Set the SRB flags correctly when there is no data transfer. Without this
change some IHV drivers will fail valid commands such as TEST_UNIT_READY.
Cc: <[email protected]>
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reviewed-by: Long Li <[email protected]>
Signed-off-by: James Bottomley <[email protected]>
Commit b08cc79155 eliminated memory
allocation in the packet send path:
"hv_netvsc: Eliminate memory allocation in the packet send path
The network protocol used to communicate with the host is the remote ndis (rndis)
protocol. We need to decorate each outgoing packet with a rndis header and
additional rndis state (rndis per-packet state). To manage this state, we
currently allocate memory in the transmit path. Eliminate this allocation by
requesting additional head room in the skb."
This commit introduced a bug since it did not account for the case if the skb
was cloned. Fix this bug.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Tested-by: Dexuan Cui <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
The current code assumes that the scatterlists presented are not chained.
Fix the code to not make this assumption.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reviewed-by: Long Li <[email protected]>
Signed-off-by: James Bottomley <[email protected]>
The storage protocol informs the guest of the I/O capabilities of the storage
stack. Retrieve this information and use it in the guest.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reviewed-by: Long Li <[email protected]>
Signed-off-by: James Bottomley <[email protected]>
The current code always sent packets without data on the primary channel.
Properly distribute sending of packets with no data amongst all available
channels. I would like to thank Long Li for noticing this problem.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reviewed-by: Long Li <[email protected]>
Signed-off-by: James Bottomley <[email protected]>
Size the queue depth based on the ringbuffer size. Also accommodate for the
fact that we could have multiple channels (ringbuffers) per adaptor.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reviewed-by: Long Li <[email protected]>
Signed-off-by: James Bottomley <[email protected]>
Increase the default ring buffer size as this can significantly
improve performance especially on high latency storage back-ends.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reviewed-by: Long Li <[email protected]>
Signed-off-by: James Bottomley <[email protected]>
The network protocol used to communicate with the host is the remote ndis (rndis)
protocol. We need to decorate each outgoing packet with a rndis header and
additional rndis state (rndis per-packet state). To manage this state, we
currently allocate memory in the transmit path. Eliminate this allocation by
requesting additional head room in the skb.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
In preparation for embedding the rndis state and other packet state into
the skb, cleanup the test for freeing the skb.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Don't wait after sending request for offers to the host. This wait is
unnecessary and simply adds 5 seconds to the boot time.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Cc: <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Handle the case when the write to the ringbuffer fails. In this case,
unconditionally signal the host. Since we may have deferred signalling
the host based on the kick_q parameter, signalling the host
unconditionally in this case deals with the issue.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Export the vmbus_sendpacket_pagebuffer_ctl() interface. This export will be
used by the netvsc driver.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
When a channel has been rescinded, the close operation is a noop.
Restructure the code so we deal with the rescind condition after
we properly cleanup the channel. I would like to thank
Dexuan Cui <[email protected]> for observing this problem.
The current code leaks memory when the channel is rescinded.
The current char-next branch is broken and this patch fixes
the bug.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch is a continuation of the rescind handling cleanup work. We cannot
block in the global message handling work context especially if we are blocking
waiting for the host to wake us up. I would like to thank
Dexuan Cui <[email protected]> for observing this problem.
The current char-next branch is broken and this patch fixes
the bug.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Implement an API that gives additional control on the what VMBUS flags will be
set as well as if the host needs to be signalled. This API will be
useful for clients that want to batch up requests to the host.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Implement an API for sending pagebuffers that gives more control to the client
in terms of setting the vmbus flags as well as deciding when to
notify the host. This will be useful for enabling batch processing.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
The current algorithm for picking an outgoing channel was not distributing
the load well. Implement a simple round-robin scheme to ensure good
distribution of the outgoing traffic.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reviewed-by: Long Li <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Currently we log messages when either we are not able to map an ID to a
channel or when the channel does not have a callback associated
(in the channel interrupt handling path). These messages don't add
any value, get rid of them.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
When the offer is rescinded, vmbus_close() can free up the channel;
deinitialize the service before closing the channel.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Properly rollback state in vmbus_pocess_offer() in the failure paths.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Execute both ressind and offer messages in the same work context. This serializes these
operations and naturally addresses the various corner cases.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
In response to a rescind message, we need to remove the channel and the
corresponding device. Cleanup this code path by factoring out the code
to remove a channel.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Handle the case when the device may be removed when the device has no driver
attached to it.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
NetworkDirect is a service that supports guest RDMA.
Define the GUID for this service.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Correctly rollback state if the failure occurs after we have handed over
the ownership of the buffer to the host.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Cc: [email protected]
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Currently, the API for sending a multi-page buffer over VMBUS is limited to
a maximum pfn array of MAX_MULTIPAGE_BUFFER_COUNT. This limitation is
not imposed by the host and unnecessarily limits the maximum payload
that can be sent. Implement an API that does not have this restriction.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Correctly compute the local (gpadl) handle.
I would like to thank Michael Brown <[email protected]> for seeing this bug.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reported-by: Michael Brown <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Implement a clockevent device based on the timer support available on
Hyper-V.
In this version of the patch I have addressed Jason's review comments.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reviewed-by: Jason Wang <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
We currently release memory (balloon down) in the interrupt context and we also
post memory status while releasing memory. Rather than posting the status
in the interrupt context, wakeup the status posting thread to post the status.
This will address the inconsistent lock state that Sitsofe Wheeler <[email protected]>
reported:
http://lkml.iu.edu/hypermail/linux/kernel/1411.1/00075.html
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reported-by: Sitsofe Wheeler <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
We support memory hot-add in the Hyper-V balloon driver by hot adding an appropriately
sized and aligned region and controlling the on-lining of pages within that region
based on the pages that the host wants us to online. We do this because the
granularity and alignment requirements in Linux are different from what Windows
expects. The state to manage the onlining of pages needs to be correctly
protected. Fix this bug.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Make adjustments in computing the balloon floor. The current computation
of the balloon floor was not appropriate for virtual machines with more than
10 GB of assigned memory - we would get into situations where the host would
agressively balloon down the guest and leave the guest in an unusable state.
This patch fixes the issue by raising the floor.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Replace calls for smp_processor_id() to get_cpu() to get the CPU ID of
the current CPU. In these instances, there is no correctness issue with
regards to preemption, we just need the current CPU ID.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
On win8 and win8 r2 hosts force SPC-3 compliance for MSFT virtual disks.
Ubuntu has been carrying a similar patch outside the tree for a while now.
Starting with win10, the host will support SPC-3 compliance. Based on all
the testing that has been done on win8 and win8 r2 hosts, we are comfortable
claiming SPC-3 compliance on these hosts as well. This will enable TRIM
support on these hosts.
Suggested by: James Bottomley <[email protected]>
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reviewed-by: Long Li <[email protected]>
Signed-off-by: Christoph Hellwig <[email protected]>
Commit 4cd83ecdac changed the limits to
reflect the values on the host. It turns out that WS2008R2 cannot
correctly handle these new limits. Fix this bug by setting the limits
based on the host.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reviewed-by: Long Li <[email protected]>
Signed-off-by: Christoph Hellwig <[email protected]>
The host asks the guest to scan when a LUN is removed or added.
The only way a guest can identify the removed LUN is when an I/O is
attempted on a removed LUN - the SRB status code indicates that the LUN
is invalid. We currently handle this SRB status and remove the device.
Rather than waiting for an I/O to remove the device, force the discovery of
LUNs that may have been removed prior to discovering LUNs that may have
been added.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reviewed-by: Long Li <[email protected]>
Signed-off-by: Christoph Hellwig <[email protected]>
The virtual HBA that storvsc implements can support multiple channels and
targets. So, scan the host when the host notifies that a scan is needed.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reviewed-by: Long Li <[email protected]>
Signed-off-by: Christoph Hellwig <[email protected]>
After the packet is successfully sent, we should not touch the packet
as it may have been freed. This patch is based on the work done by
Long Li <[email protected]>.
David, please queue this up for stable.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reported-by: Sitsofe Wheeler <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
After the packet is successfully sent, we should not touch the skb
as it may have been freed. This patch is based on the work done by
Long Li <[email protected]>.
In this version of the patch I have fixed issues pointed out by David.
David, please queue this up for stable.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Tested-by: Long Li <[email protected]>
Tested-by: Sitsofe Wheeler <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
In win8 we have a feature that allows for interrupt driven flow management
for host/guest communication. For instance, if the host were blocked because
there was no space available in the ringbuffer, the host could request that the
guest send an interrupt when space becomes available in the ringbuffer (when
the guest drains the ringbuffer).
While this feature was implemented in the guest a while ago, we had not
advertised that the guest supported this feature. This patch advertises
the support to the host.
For pre-win8 hosts, this has no effect since the size of the ringbuffer
control structure has not changed and all changes have been backward
compatible - unused/reserved space has been used to implement this
feature.
In this version of the patch I have cleaned up the commit log based on
feedback from Greg KH.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Disable preemption when sampling current processor ID when preemption
is otherwise possible.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Tested-by: Sitsofe Wheeler <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Minimize failures in this function by pre-allocating the buffer
for posting messages. The hypercall for posting the message can fail
for a number of reasons:
1. Transient resource related issues
2. Buffer alignment
3. Buffer cannot span a page boundry
We address issues 2 and 3 by preallocating a per-cpu page for the buffer.
Transient resource related failures are handled by retrying by the callers
of this function.
This patch is based on the investigation
done by Dexuan Cui <[email protected]>.
I would like to thank Sitsofe Wheeler <[email protected]>
for reporting the issue and helping in debuggging.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reported-by: Sitsofe Wheeler <[email protected]>
Cc: <[email protected]>
Tested-by: Sitsofe Wheeler <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Eliminate calls to BUG_ON() in vmbus_close_internal().
We have chosen to potentially leak memory, than crash the guest
in case of failures.
In this version of the patch I have addressed comments from
Dan Carpenter ([email protected]).
Signed-off-by: K. Y. Srinivasan <[email protected]>
Cc: <[email protected]>
Tested-by: Sitsofe Wheeler <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Eliminate the call to BUG_ON() by waiting for the host to respond. We are
trying to reclaim the ownership of memory that was given to the host and so
we will have to wait until the host responds.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Cc: <[email protected]>
Tested-by: Sitsofe Wheeler <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Eliminate calls to BUG_ON() by properly handling errors. In cases where
rollback is possible, we will return the appropriate error to have the
calling code decide how to rollback state. In the case where we are
transferring ownership of the guest physical pages to the host,
we will wait for the host to respond.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Cc: <[email protected]>
Tested-by: Sitsofe Wheeler <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Posting messages to the host can fail because of transient resource
related failures. Correctly deal with these failures and increase the
number of attempts to post the message before giving up.
In this version of the patch, I have normalized the error code to
Linux error code.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Cc: <[email protected]>
Tested-by: Sitsofe Wheeler <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Get rid of the warning messages since they will clutter up various system logs
and are of questionable value to the end user. For debugging purposes, this
information can be gotten by setting the scsi log level appropriately.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reviewed-by: Olaf Hering <[email protected]>
Signed-off-by: Christoph Hellwig <[email protected]>
WS2008R2 is a supported platform and it turns out that the maximum sendbuf
size that ws2008R2 can support is only 15MB. Make the necessary
adjustment.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Intel did some benchmarking on our network throughput when Linux on Hyper-V
is as used as a gateway. This fix gave us almost a 1 Gbps additional throughput
on about 5Gbps base throughput we hadi, prior to increasing the sendbuf size.
The sendbuf mechanism is a copy based transport that we have which is clearly
more optimal than the copy-free page flipping mechanism (for small packets).
In the forwarding scenario, we deal only with MTU sized packets,
and increasing the size of the senbuf area gave us the additional performance.
For what it is worth, Windows guests on Hyper-V, I am told use similar sendbuf
size as well.
The exact value of sendbuf I think is less important than the fact that it needs
to be larger than what Linux can allocate as physically contiguous memory.
Thus the change over to allocating via vmalloc().
We currently allocate 16MB receive buffer and we use vmalloc there for allocation.
Also the low level channel code has already been modified to deal with physically
dis-contiguous memory in the ringbuffer setup.
Based on experimentation Intel did, they say there was some improvement in throughput
as the sendbuf size was increased up to 16MB and there was no effect on throughput
beyond 16MB. Thus I have chosen 16MB here.
Increasing the sendbuf value makes a material difference in small packet handling
In this version of the patch, based on David's feedback, I have added
additional details in the commit log.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Add blist flags to permit the reading of the VPD pages even when
the target may claim SPC-2 compliance. MSFT targets currently
claim SPC-2 compliance while they implement post SPC-2 features.
With this patch we can correctly handle WRITE_SAME_16 issues.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reviewed-by: Hannes Reinecke <[email protected]>
Signed-off-by: Christoph Hellwig <[email protected]>
On some Windows hosts on FC SANs, TEST_UNIT_READY can return SRB_STATUS_ERROR.
Correctly handle this. Note that there is sufficient sense information to
support scsi error handling even in this case.
Signed-off-by: K. Y. Srinivasan <[email protected]>
Reviewed-by: Hannes Reinecke <[email protected]>
Cc: <[email protected]>
Signed-off-by: Christoph Hellwig <[email protected]>
K. Y. Srinivasan