Fix race in device resume path - need to ensure we add the watch only
after the device-specific resume code has run. Also removed some unused /
usless code from netfront.c.
[LINUX] Fix interaction between idle loop and RCU subsystem.
There is a problem with the current implementation of stop_hz_timer in
arch/i386/kernel/time-xen.c where the hz timer can be stopped on a CPU
which has RCU callbacks pending.
This patch backports a new RCU API created to fix this problem for the
s390 implementation of stop_hz_timer and also updates the time-xen.c
implementation of stop_hz_timer to call the new API.
Signed-off-by: Harry Butterworth <butterwo@uk.ibm.com>
[HVM][VMX] Fix the build with certain versions of gcc.
Restrict register choices to this which are byte addressable. Signed-off-by: Keir Fraser <keir@xensource.com>
[HVM][VMX] Provide right view of cpuid to the HVM guests.
Some of the CPU features such as APIC, PAE, MTRR, HT are virtualized;
while others are not virtualized yet such as TM1, TM2, MCA and there are
some features which do not need virtualization such as MMX. With the
patch Guest sees only those processor features in the cpuid which are
virtualized in the hyper visor, or do not need any virtualization in
hypervisor.
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com> Signed-off-by: Jun Nakajima <jun.nakajima@intel.com> Signed-off-by: Asit Mallick <asit.k.mallick@intel.com>
[HVM][DM] Backport the USB support from QEMU 0.8.1 to the current Xen
device model. To support USB there are two new optional config lines
that can be added to the guest config file:
usb=1
This will enable USB without defining a specific USB device.
This option is assumed and not needed if the `usbdevice' option
is given.
usbdevice='device'
This will enable USB and also enable support for the given device.
Currently, the only two devices are `mouse' (a PS/2 mouse) and
`tablet' (an absolute pointing device). The advantage of `tablet'
is that Windows guests will automatically recognize and support this
device so specifying the config line:
usbdevice='tablet'
will create a mouse that works transparently with Windows guests
under VNC. (Linux doesn't recognize the USB tablet yet so Linux
guests under VNC will still need the Summagraphics emulation.)
Signed-off-by: Don Dugger <donald.d.dugger@intel.com>
[HVM][VMX] Cleanups and fixes to VMCS lifecycle.
1. Maintain a 'launched' software flag to select between
VMLAUNCH and VMRESUME.
2. Take more care with VMPTRLD/VMCLEAR.
Also various other fixes (e.g., safe testing of
condition codes after executing a VMX instruction).
[LINUX] Fix IRQ SMP affinity logic for event channels.
The logic now mimics native x86 behaviour: a request to change
affinity via /proc is held until the next interrupt on that
event channel. So /proc/irq/n/smp_affinity may not change
immediately!
Other notes:
1. CPU-specific interrupts silently ignore requests to change
affinity. For example, resched0, timer0, callfunc0, ...
2. Reading smp_affinity always returns a cpumask containing
a single cpu. An event channel can only be bound to a single
cpu at a time. Neither Xen nor XenLinux implement IRQ
balancing: requires a user-space balancing daemon.
[XEN] Replace direct common-code access of evtchn_upcall_mask
with local_event_delivery_* accessors.
Notes:
1. Still some (read-only, debug) use in keyhandler.c
2. Still accesses through current->vcpu_info.
Both above may need to be compiled only for architectures
that use event channels.
[LINUX] Remove the only use of evtchn_upcall_mask in common code.
Replace with an assertion of irq_disabled(). Signed-off-by: Keir Fraser <keir@xensource.com>
[XEN] VCPU operation register_runstate should work even when
called from non-local VCPU. Necessary as Linux always calls
this operation from VCPU0 during secondary-VCPU bringup. Signed-off-by: Keir Fraser <keir@xensource.com>
[ACM] Replace enumerations with macros with qualified names.
Old unqualified enumeration names polluted the global namespace. Signed-off-by: Keir Fraser <keir@xensource.com>
[LINUX] Simplify /proc/iomem initialisation -- no need to fake a
domU memory map. The main e820 map already terminates at max_pfn. Signed-off-by: Keir Fraser <keir@xensource.com>
[LINUX][X86/64] Initialise pages outside initial allocation so that
they are picked up by the balloon driver.
From: Jan Beulich Signed-off-by: Keir Fraser <keir@xensource.com>
[LINUX][XENBUS] replace xenbus_transaction_t with an opaque transaction ID.
xenbus_transaction_t -> struct xenbus_transaction, which just contains
a single u32. Also renamed XBT_NULL to XBT_NIL to emphasize that it
isn't a NULL pointer, but a NIL transaction ID. Compile and boot tested.
Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
[chrisw: updated for xen-unstable from patchqueue] Signed-off-by: Chris Wright <chrisw@sous-sol.org>
[LINUX][XENBUS] Drop (struct xenbus_device)->data, and simply use ->dev.driver_data
instead, since that's what it is there for. Similar to patch
in patchqueue started by Jeremy, and expanded by me to include all
relevant drivers in xen-unstable.
Cc: Jeremy Fitzhardinge <jeremy@xensource.com> Signed-off-by: Chris Wright <chrisw@sous-sol.org>
[LINUX] Use structs not typedefs, remove trailing whitespaces, place symbol
EXPORTs with functions, order includes, eliminate unsed macros, don't
initialize static data to NULL, etc. Cleanups from patchqueue.
[NET] front: There's a small leak on a couple error paths in setup_device().
While there rearrange the ring setup order slightly to simplify error
path since netif_free() will cleanup once ring_ref is valid. And use
get_zeroed_page() instead of __get_free_page()/memset(). Handle error if
bind_evtchn_to_irqhandler() fails, as bad info->irq value is likely to
cause oops later. In create_device(), gnttab_free_grant_references()
is accidentally called twice on tx_head during cleanup from failed
gnttab_alloc_grant_references() on rx_head, which could corrupt
gnttab_free_count.
[XEN] Reduce default maximum allocation order from 2^20 pages to 2^11.
On x86 this corresponds to a maximum aligned contiguous allocation of 8MB.
This can be overridden by architectures if need be. Signed-off-by: Keir Fraser <keir@xensource.com>
[LINUX] Set up /proc/iomem in a sensibel way to indicate
code & data resources, and a RAM area on domU.
From: Gerd Hoffmann <kraxel@suse.de> Signed-off-by: Keir Fraser <keir@xensource.com>
[LINUX] Only trigger unhandled irq path if irq is not shared across
multiple guests (another guest may have handled the interrupt). Signed-off-by: Keir Fraser <keir@xensource.com>
[TOOLS] Fix domain builder to carefully check that mapped memory area
does not overflow and wrap to zero. Signed-off-by: Keir Fraser <keir@xensource.com>
[HVM][MMIO] Support decode of 0x83 opcode (or imm8,m32/64').
This instruction is used by both WinXP during an SMP installation and by
Sun Solaris. With this patch Sun Solaris UP is able to boot. Windows
SMP gets a bit further but still has issues.
Signed-off-by: Tom Woller <thomas.woller@amd.com> Signed-off-by: Travis Betak <tbetak.woller@amd.com>
[HVM][DM] Cleaner way of clearing VGA memory on mode changes (thanks to
Fabrice Bellard on the QEMU project). This patch clears the
memory in the BIOS call rather than when the Cirrus Logic register changes,
which more closely matches what happens on the real hardware.
Signed-off-by: Don Dugger <donald.d.dugger@intel.com>
[LIBXC] Convert all printf/fprintf uses to use a macro instead.
Thus all can be disabled at compile time. It would be easy to
make enabling/disabling a run-time option too. Signed-off-by: Keir Fraser <keir@xensource.com>
[HVMLOADER] Revert to uniprocessor MPS table in rombios.
Advertising more CPUs than the guest is allocated hangs boot
of Windows XP (selects wrong HAL). The correct fix here is
to dynamically generate the MPS table (at least the CPU
entries) from hvmloader binary. Signed-off-by: Keir Fraser <keir@xensource.com>
Use explicitly-sized types in the dom0_ops and privcmd structures.
As discussed previously, the these operations are not performance-sensitive, so
the additional cache footprint shouldn't be an issue.
Represent PFNs with their own type, rather than 'unsigned long'.
('long' changes size and alignment between 32- and 64-bit ABIs.) Signed-off-by: Hollis Blanchard <hollisb@us.ibm.com>
[NET] front: Turn grant-ref error into a fatal bug. That's the best we
can do until we have a backend driver recovery method. Signed-off-by: Keir Fraser <keir@xensource.com>
[X86EMUL] Mark MOV instruction as not needing writeback.
Fix the test harness for x86/64 -- map emulated addresses
to low 4GB of address space. Signed-off-by: Keir Fraser <keir@xensource.com>
[HVM][VMX] Interrupts must be kept disabled when entering Xen for
external interrupt processing. Remove code that immediately
reenabled interrupt delivery on VMEXIT. Signed-off-by: Seteven Smith <sos22@cam.ac.uk> Signed-off-by: Keir Fraser <keir@xensource.com>
This patch adds scatter-and-gather transmission support to the frontend.
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 backend. It also
advertises this fact through xenbus so that the frontend 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>
[VMXASSIST] Extensions to vmxassist to handle additional mode switch cases.
Extends vmxassist to handle two cases related to mode switching found
while experimenting with different boot loaders.
The first case is use of the JMP instruction with memory location
operands to complete a switch to protected mode. This patch adds
emulation for this form of the JMP instruction to vmxassist.
The second case is where boot loader code does not save/restore a
non-zero SS register across a protected mode traversal.
Zeroing the SS register in vmxassist results in all sorts of problems
in the domU after returning back to real mode. This patch stores
segment register values before entering protected mode and correctly
restores the old values (instead of an incorrect zero value) when
reentering real mode.
Signed-off-by: Kevin Tronkowski <ktronkowski@virtualiron.com>
[NET] back: fix synchronisation of access to deallocation buffer ring.
Must ensure ring is written to before producer index is incremented.
Bug diagnosed by Ky Srinivasan <ksrinivasan@novell.com> Signed-off-by: Keir Fraser <keir@xensource.com>
Extend callback-registration hypercall to take a flags argument.
This can currently be used to request that event delivery be
disabled during callback processing. Signed-off-by: Keir Fraser <keir@xensource.com>
Create new vmassist type 'pae_extended_cr3'. Only advertise
pae_pgdir_above_4gb tp guests that have enabled this vmassist.
Control tools ensure all PAE page directories are below 4GB
unless the vmassist is enabled (triggered via an extended-cr3
option in guest Elf header). Signed-off-by: Keir Fraser <keir@xensource.com>
[LINUX][X86_64] Destroy initial page-table mappings to avoid overlap with modules.
The temporary mappings needed to set up the 1:1 mappings must be torn
down after use; otherwise they may trigger the
WARN_ON() in vmap_pte_range() (namely if the chunk allocated to hold
kernel and initial page tables gets close to or
exceeds 128Mb, or if a sufficiently high mem= argument causes the
static allocations to grow beyond 128Mb, which in
either case means these mappings extend into the modules area).
Just like the standard loopback device, SG support here is innate.
So all we need to do is mark it as such and zero the change_mtu method
so that the MTU can be changed at will.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Add backing support for HDIO_GETGEO ioctl to blkfront.
Inspired by an earlier patch from Charles Coffing. Signed-Off-By: Jan Beulich <jbeulich@novell.com>
[PAE] Fix support for pgdirs above 4GB. Requires an interface change so that %cr3
is extended to include high-order address bits at bottom of %cr3 value.
Guests who understand this interface change publish the fact by setting the
option 'PAE=yes[extended-cr3]' in their __xen_guest section. Signed-off-by: Keir Fraser <keir@xensource.com>
[IA64] Fix Xen running on IA64. Some function prototypes were
moved to domain.h and hence undefined in some source files.
This meant that some return types defaulted to int, resulting
in corruption. Signed-off-by: Kevin Tian <kevin.tian@intel.com>
Fix pagetable accessor macros in Xen to have better names and to
preserve top bits of PAE pgdirs situated above 4GB. Signed-off-by: Keir Fraser <keir@xensource.com>
Allow all trace-buffer hypercalls to be executed even when tracing is currently disabled.
From: George Dunlap Signed-off-by: Keir Fraser <keir@xensource.com>
The temporary mappings needed to set up the 1:1 mappings must be torn down after use; otherwise they may trigger the
WARN_ON() in vmap_pte_range() (namely if the chunk allocated to hold kernel and initial page tables gets close to or
exceeds 128Mb, or if a sufficiently high mem= argument causes the static allocations to grow beyond 128Mb, which in
either case means these mappings extend into the modules area).
[HVM] Allow more dynamic use of the qemu serial lines by allowing the Summagraphics port to float.
qemu will emulate up to 4 serial lines (COM1 - COM4). One of these
lines supports a Summagraphics protocol. This has changed recently,
and it's now always COM2. This patch allows the user to specify
to qemu which lines should be serial lines, and the code will place
the Summagraphics compatible line at just above the last specified
serial line. qemu defaults serial line 0 (COM1) to be "vc", resulting
in the Summagraphics support residing on COM2. This is in keeping
with the current behavior.
The results, via qemu command line, are:
<none> COM2
-serial x COM2
-serial x -serial x COM3
-serial x -serial x -serial x COM4
-serial x -serial x -serial x -serial x <none>
The results are deterministic, and allow the use of up to 4 serial
lines including none with Summagraphics support. There is no
change from current behavior, unless requested.
[SECURITY] Allow domid and ssidref args to get_decision in
util/security.py to be either strings or ints (previously only strings
would work due to the underlying call to lowlevel.acm).
Signed-off-by: Bryan D. Payne <bdpayne@us.ibm.com> Signed-off-by: Reiner Sailer <sailer@us.ibm.com>
[LINUX] Converge ia64 versions of xen/drivers Makefiles with general versions.
This patch converges the ia64-specific and general versions of
xen/drivers/Makefile and xen/drivers/core/Makefile, using Kconfig
settings instead to control whether specific pieces are built.
Fix negation of unsigned quantities in the Xen x86 emulator.
This fixes the problems left behind by c/s 10171. Again pointed
out by Jan Beulich; and again different from his suggested patch.
Hopefully this one will be less embarrassing.
This test imports a ram disk device as a physical device into a domU.
The domU initialises the ram disk with data from /dev/urandom and
calculates the md5 checksum of the data (using tee as it is written so as to
avoid reading it back from the device which might potentially mask
problems).
The domU is stopped and the md5 checksum of the data on the device is
calculated by dom0. The test succeeds if the checksums match, indicating
that all the data written by domU was sucessfully committed to the
device.
This patch also enables tee in BusyBox on the ramdisk and increments the
xm-test version number to 0.8.0.
The patch also installs the block-integrity tests in the default test
set so they get executed.
Signed-off-by: Harry Butterworth <butterwo@uk.ibm.com> Signed-off-by: Ewan Mellor <ewan@xensource.com>
network-bridge: Fall back to eth0/xenbr0 if no default route.
From: Dave Lively <dave.lively@gmail.com> Signed-off-by: Keir Fraser <keir@xensource.com>
projects / xen.git / log