This patch fixes the copy_from/to_guest problem.
As Akio reported, modularised netback causes dom0's down.
The following process is happened in gnttab_transfer()@
xen/common/grant_table.c:
gnttab_transfer()
=> steal_page()
=> assign_domain_page_cmpxchg_rel()
=> domain_page_flush()
=> domain_flush_vtlb_all() // all TLBs are flushed
...
=> __copy_to_guest_offset() // always fail to copy
The embedded netback module has no problem because it uses TR pinned
data. But modularised one is out of TR. So copy_from/to_guest issue
must be solved in order to modularise drivers.
Signed-off-by: Kouya SHIMURA <kouya@jp.fujitsu.com>
This fixes
- one instance of using XEN_BLOCK_EVENTS where XEN_UNBLOCK_EVENTS was
meant
- four instances of left-over cli-s that should have been replaced
with XEN_BLOCK_EVENTS
- one instance of slightly wrong placement (without resulting in ill
behavior, but needlessly making the diff to native more complex)
Signed-off-by: Jan Beulich <jbeulich@novell.com>
This changeset is from upstream Linux. We need it so Xen users
using NAT are not unnecessarily alarmed by these WARN_ON errors
especially since they're not in a position to do anything about
it. Original changelog:
[NET]: Kill the WARN_ON() calls for checksum fixups.
We have a more complete solution in the works, involving
the seperation of CHECKSUM_HW on input vs. output, and
having netfilter properly do incremental checksums.
But that is a very involved patch and is thus 2.6.19
material.
What we have now is infinitely better than the past,
wherein all TSO packets were dropped due to corrupt
checksums as soon at the NAT module was loaded. At
least now, the checksums do get fixed up, it just
isn't the cleanest nor most optimal solution.
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Split the bits of page.h which are needed by PV-on-HVM drivers out
into a separate header file, so that it's a bit easier to get the xenolinux
machine address functions when compiled against a native kernel.
Signed-off-by: Steven Smith <ssmith@xensource.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
By replacing dev_alloc_skb with alloc_skb we need to do the 16-byte
reservation ourselves. Also, it's unnecessary to align the size
when calling alloc_skb since the alignment is an implementation
detail of alloc_skb and it already takes care of it for us.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Thsi fixes problems where a driver is loaded from a script, and the
following parts of the script depend on devices having connected
by the time the modprobe/insmod returns.
Based on a tested patch from Rik van Riel at Red Hat.
Signed-off-by: Keir Fraser <keir@xensource.com>
I cannot think of any circumstance that the Xen console device would
not have CON_ENABLED. With out this, DomUs that panic() before the
init process has started (and maybe even after) do not get output hi
priority log_levels (like panic()) from LOG_BUF().
Signed-off-by: Jimi Xenidis <jimix@watson.ibm.com>
This backend use a slot per pci device. Contrary to vpci two functions from
one slot appear as two slots. This is useful on ia64 because it doesn't scan
functions > 0 if function 0 does not exist.
Signed-off-by: Tristan Gingold <tristan.gingold@bull.net>
Because the backend brings up the interface long before the frontend
has booted up, it is possible that by the time we get here we already
have packets queued up for processing.
If we don't process them here, we may delay them more than what is
necessary. Worse yet, it is possible to miss the notification
interrupt from the backend in such a way that we never get another
one until we bring the interface down and up.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Each HVM domain has a space of HVM parameters associated with it,
and these can be manipulated via a new hvm_op hypercall. This means
that the hypervisor no longer needs to parse the hvm_info table, so
remove that code.
Signed-off-by: Steven Smith <ssmith@xensource.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
It does this by intercepting XS_WATCH requests written
to /proc/xen/xenbus and then re-submitting the request to the
in-kernel xenstore interface, in
linux-2.6-xen-sparse/drivers/xen/xenbus/xenbus_xs.c. When a callback
occurs, an in-kernel function is invoked, which then reconstructs a
response in the format expected by userspace, and sends this response
through /proc/xen/xenbus.
It was necessary to add some supporting infrastructure to
linux-2.6-xen-sparse/drivers/xen/xenbus/xenbus_dev.c, such as an
additional mutex to protect the response queue and a list of active
watches associated with each connection.
Signed-off-by: Michael LeMay <mdlemay@epoch.ncsc.mil>
The for loop in pcistub_match_one will loop forever if the
dev->bus->self links to itself at the uppermost bridge. Adding a
check to prevent linking back in on itself prevents this.
Signed-off-by: Jon Mason <jdmason@us.ibm.com>
This fixes a bug whereby foreign pages were freed by the unpin, which
then become owned by the local domain before it destroys its ptes.
It therefore (erroneously) detects the mappings as local and so
updates reference counts, leading to crashes.
Signed-off-by: Keir Fraser <keir@xensource.com>
Rewrite utils.c patch to be less ugly:
do not set area->phys_addr, recompute phys_addr in free_vm_area.
Signed-off-by: Tristan Gingold <tristan.gingold@bull.net>
This patch adds TSO transmission support to the backend.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Disable for now, as domU->dom0 direction.
Signed-off-by: Keir Fraser <keir@xensource.com>
This patch adds TCP Segmentation Offload (TSO) support to the
frontend.
It also advertises this fact through xenbus so that the frontend can
detect this and send through TSO requests only if it is supported.
This is done using an extra request slot which is indicated by a flag
in the first slot. In future checksum offload can be done in the same
way.
Even though only TSO is supported for now the code actually supports
GSO so it can be applied to any other protocol. The only missing bit
is the detection of host support for a specific GSO protocol. Once
that is added we can advertise all supported protocols to the guest.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Disable for now, as in domU->dom0 direction.
Signed-off-by: Keir Fraser <keir@xensource.com>
This patch adds scatter-and-gather transmission support to the
backend. This allows the MTU to be raised right now and the potential
for TSO in future.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This patch adds scatter-and-gather support to the frontend. It also
advertises this fact through xenbus so that the backend can detect
this and send through SG requests only if it is supported.
SG support is required to support skb's larger than one page. This
in turn is needed for either jumbo MTU or TSO. One of these is
required to bring local networking performance up to a level that
is acceptable.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This patch adds a tx queue to the backend if the frontend supports rx
refill notification. A queue is needed because SG/TSO greatly reduces
the number of packets that can be stored in the rx ring. Given an rx
ring with 256 entries, a maximum TSO packet can occupy as many as 18
entries, meaning that the entire ring can only hold 14 packets. This
is too small at high bandwidths with large TCP RX windows.
Having a tx queue does not present a new security risk as the queue is
a fixed size buffer just like the rx ring. So each guest can only
hold a
fixed amount of memory (proportional to the tx queue length) on the
host.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Cannot BUG_ON netbk_queue_full yet !netbk_can_queue, as this can be
triggered by a misbehaving client. Set req_event appropriately when
stopping the packet queue, or we will not receive a notification.
Signed-off-by: Keir Fraser <keir@xensource.com>
used to be used for sending of vTPM control commands, but now this is
all done with the hotplug scripts.
Signed-off-by: Stefan Berger <stefanb@us.ibm.com>
Added in Kconfig.
On ia64, use ia64 sysdata instead of pcifront genuine one.
Fix a warning on xenbus.c
Signed-off-by: Tristan Gingold <tristan.gingold@bull.net>
This patch adds support to the frontend for notifying the backend whenever
the rx ring is refilled. This is required in order for the backend to
get a tx queue.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Modified to only send notification if req_event index is set
appropriately.
Signed-off-by: Keir Fraser <keir@xensource.com>
The connection status to the frontend can be represented using
netif_carrier_ok instead of netif->status. As a result, we delay
the construction of the dev qdisc until the carrier comes on. This
is a prerequisite for adding a tx queue.
By the same token, netif->active is now simply the conjunction of
netif_running and netif_carrier_ok so it too can be removed.
Because netif_carrier_off/netif_carrier_on and rtnl_lock all entail
memory barriers, there is no need to have extra memory barriers around
them.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This patch moves all rx request pushing to network_alloc_rx_buffers.
This is needed to reduce churn for TSO. More importantly, this makes
it easier to send notifications when adding rx requests which is
required for having a queue in dom0.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Keir Fraser <keir@xensource.com>
This is an update to c/s
10855:03c8002068d9d60c7bbfc2f41af975e09b2e2211
which should have contained the following changeset message
(rather than 'Merge.').
[NET] front: Stop using rx->id
With the current protocol for transferring packets from dom0 to domU,
the
rx->id field is useless because it can be derived from the rx request
ring
ID. In particular,
rx->id = (ring_id & NET_RX_RING_SIZE - 1) + 1;
This formula works because the rx response to each request always
occupies
the same slot that the request arrived in. This in turn is a
consequence
of the fact that each packet only occupies one slot.
The other important reason that this works for dom0=>domU but not
domU=>dom0
is that the resource associated with the rx->id is freed immediately
while
in the domU=>dom0 case the resource is held until the skb is liberated
by
dom0.
Using this formula we can essentially remove rx->id from the protocol,
freeing up space that could be instead be used by things like TSO.
The
only constraint is that the backend must obey the rule that each id
must
only be used in the response that occupies the same slot as the
request.
The actual field of rx->id is still maintained for compatibility with
older backends.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
