In order to reduce the memory usage of large guests (as well
as improve performance), tell KVM about the number of interrupts
we require.
To avoid synchronization with the various device creation,
use a late_init callback to compute the GIC configuration.
[Andre: rename to gic__init_gic() to ease future expansion]
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
The ARM GIC emulation needs to be told the number of interrupts
it has to support. As commit 1c262fa1dc7bc ("kvm tools: irq: make
irq__alloc_line generic") made the interrupt counter private,
add a new accessor returning the number of interrupt lines we've
allocated so far.
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
As of 3.14, KVM/arm supports the creation/configuration of the GIC through
a more generic device API, which is now the preferred way to do so.
Plumb the new API in, and allow the old code to be used as a fallback.
[Andre: Rename some functions on the way to differentiate between
creation and initialisation more clearly and fix error path.]
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
On AArch64 system with a GICv2, the GICC range can be aligned
to the last 4k block of a 64k page, ending up straddling two
64k pages. In order not to conflict with the distributor mapping,
allocate two 64k pages to the CPU interface.
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
On registering the ioeventfds for the virtio-pci device we cover both
the I/O ports and the MMIO BAR.
But as the current code advertises both as PIO, the host kernel gets
the wrong bus number for the MMIO region.
Fix the issue by marking only the actual PIO area as PIO.
This fixes vhost-net on x86.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Two guest memory regions need to be defined and two "mem=" parameters
need to be passed to guest kernel to support more than 256 MB.
Signed-off-by: Andreas Herrmann <andreas.herrmann@caviumnetworks.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
In PCI config space there is an interrupt line field (offset 0x3f),
which is used to initially communicate the IRQ line number from
firmware to the OS. _Hardware_ should never use this information,
as the OS is free to write any information in there.
But kvmtool uses this number when it triggers IRQs in the guest,
which fails starting with Linux 3.19-rc1, where the PCI layer starts
writing the virtual IRQ number in there.
Fix that by storing the IRQ number in a separate field in
struct virtio_pci, which is independent from the PCI config space
and cannot be influenced by the guest.
This fixes ARM/ARM64 guests using PCI with newer kernels.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
We have always built kvmtool as 64-bit on powerpc, but mainly just out
of habit. There's not AFAIK any reason we *can't* build 32-bit.
So fix up a few places where we were assuming 64-bit, and drop the
Makefile logic that forces 64-bit.
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Now that we don't have the kernel header on hand, just define the
minimum set of hcall opcodes and return values we need in order to
build.
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Instead of referring to the Linux header including the barrier
macros, copy over the rather simple implementation for the PowerPC
barrier instructions kvmtool uses. This fixes build for powerpc.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
The reboot() call that guest/init used was an older version, the
manpage talks about libc4 and libc5.
Update it to the current version exported by glibc by using the right
include file and adjusting the parameter.
This also fixed GCC 5.1.0 compile, because linux/reboot.h misses the
actual prototype.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
GCC 5.1.0 complains about missing prototypes for waitpid() and
mkdir(), so add the appropriate #includes to get rid of the warning.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
VIRTIO_PCI_QUEUE_NOTIFY is 16-bit and iowrite16 is used in
drivers/virtio/virtio_pci.c to notify the other side.
If the size doesn't match notification via mmio write will fail.
Signed-off-by: Andreas Herrmann <andreas.herrmann@caviumnetworks.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Otherwise vhost does not work if a virtio descriptor is used that was
allocated from a guest memory bank not registered as
vhost_memory_region.
Signed-off-by: Andreas Herrmann <andreas.herrmann@caviumnetworks.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Since commit d2a7ddff4 (Add minimal support for macvtap) opening
of tap device might fail. lkvm shows
Warning: Config tap device error. Are you root?
virtio_net_request_tap passed wrong pointer for struct ifreq to
TUNSETIFF ioctl.
Signed-off-by: Andreas Herrmann <andreas.herrmann@caviumnetworks.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
When including asm-generic/types.h instead of asm/types.h it results
in conflicting types for __s64 et al (at least with my
toolchain). Other header files are including asm/types.h
(e.g. include/kvm/ioport.h) and types defined there don't necessarily
match the types defined in asm-generic/types.h.
Signed-off-by: Andreas Herrmann <andreas.herrmann@caviumnetworks.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Having ARCH defined to something unknown or trying to compile kvmtool
on an unsupported architecture barfs quite late, actually after
checking for libraries. This gives a lot of messages, but the only
useful one (architecture not supported) is easily overseen.
Abort early if ARCH contains an unknown architecture name.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
The Makefile checks for the libaries needed to display the guest's
framebuffer unconditionally on all architectures.
However currently only x86 supports a guest framebuffer.
Introduce ARCH_HAS_FRAMEBUFFER in the Makefile and check for the
needed libraries only on x86. If other architectures add support for
it later, we can add them easily.
This avoids pulling in unneeded libraries into the lkvm binary, which
breaks copying over binaries into systems with only minimal userland.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
kvmtool doesn't build with arbitrary flags, so let's clear CFLAGS and
LDFLAGS by default at the top of the Makefile, allowing people to add
additional options there if they really want to.
Reported by Dave Jones, who ended up passing -std=gnu99 by mistake.
Reported-by: Dave Jones <davej@codemonkey.org.uk>
Signed-off-by: Will Deacon <will.deacon@arm.com>
After building, there is a lot of clutter from the dependency system.
Let's clean this up by using dir/.file.d style dependencies, similar
to those used in the Linux kernel.
In order to support this, rearrange the dependency generation to
create the dependency files as we build rather than as a separate
step, and have make clean remove them.
Signed-off-by: Russell King <rmk@arm.linux.org.uk>
[will: added make target for builtin-help.static.o]
Signed-off-by: Will Deacon <will.deacon@arm.com>
Avoid accidential commits of vim editor temporary files.
Signed-off-by: Russell King <rmk@arm.linux.org.uk>
Signed-off-by: Will Deacon <will.deacon@arm.com>
dpkg in the guest fails when it tries to use fsync() on a directory:
openat(AT_FDCWD, "/var/lib/dpkg", O_RDONLY|O_NONBLOCK|O_LARGEFILE|O_DIRECTORY|O_CLOEXEC) = 4
fsync(4) = -1 EINVAL (Invalid argument)
stracing lkvm shows that this is converted to:
openat(AT_FDCWD, "/root/rootfs-32//var/lib/dpkg", O_RDONLY|O_NONBLOCK|O_DIRECTORY|O_CLOEXEC) = 368
fsync(0) = -1 EINVAL (Invalid argument)
In other words, we sync against the wrong file descriptor. This case
is not handled in the kvmtool code, let's add support for it.
Signed-off-by: Russell King <rmk@arm.linux.org.uk>
Signed-off-by: Will Deacon <will.deacon@arm.com>
The terminal handling thread and the virtio-net-ctrl don't
set their name, which ends up as follows up:
terminal => lkvm
virtio-net-ctrl => kvm-cpu-X !!
Set the thread name explicitly to term-poll and virtio-net-ctrl
respectively
Signed-off-by: Suzuki K. Poulose <suzuki.poulose@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
In util/iovec.c we reference EXPORT_SYMBOL, which is only useful
within the kernel. To get rid of the dummy include file, simply
remove those lines.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
include/kvm/types.h seemed to be in use once, but it does no longer
contain any useful definition. Remove it.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
This file contains a prototype solely used by the BIOS blob code
for x86 systems. Make this clear by moving the include file into
the x86 directory.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Two BIOS source files used a definition from <asm/processor-flags.h>,
which could not be found on x86_64 multiarch machines without the
compat headers installed.
Create a file with the single definition actually used in the kvmtool
source tree.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
This header just holds three trivial functions, but pulls in
linux/types.h, which calls for trouble when compiling BIOS files on
x86-64 multiarch machines.
Remove unnecessary includes and move the definitions into the files
where they are used.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
The x86 BIOS code included some definitions for the E820 table from
the Linux system headers, but those could not be found on x86_64
multiarch systems without the compat headers installed.
Pull the required definitions into the kvmtool source directory,
removing the reference to <asm/e820.h>.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
For x86 the files that become the BIOS blob are compiled with -m32.
Although we avoid including any system libraries, we use some headers
from /usr/include, which asks for trouble if compiling on a x86-64
multiarch system (Debian/Ubuntu). Without 32-bit compat headers
installed there, the compiler will not find the right files.
However as the BIOS code is actually a self-contained binary without
any relationship to the Linux userland, it should not use or rely
on Linux system headers.
Replace includes of linux/types.h in BIOS code with asm/bios/types.h,
which simply contains the u{8,16,32,64} data types needed by the code.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
At this point kvmtool compiles fines out of the Linux tree (at least
for x86 and arm/arm64), so update the README to reflect this.
Also add an INSTALL file to give more user friendly building
instructions.
Document the optional libraries and their common package names on the
way.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Currently there is a warning when the (highly optional) GTK3 support
cannot be compiled in. Actually many may consider GTK3 the least
interesting of the optional packages, so replace this rather verbose
warning with a brief list of not found optional packages: This will
give a much nicer output like this:
Makefile:291: Skipping optional libraries: bfd GTK3 vncserver aio
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Instead of hardcoding libfdt availability, check for it like for
the other libraries and fail (and warn) if it's not available.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
kvmtool used the in-kernel version of the device tree handling
library. Now that we are a proper userland tool, use the system's
library for that purpose. Actually this seems to fix some
long-standing warning generated by the Linux copy.
Also fix up a bogus x86 warning (no FDT needed here).
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
In config/utilities.mak we have two macros checking either
compile-ability or build-ability of a library. Actually they are
defined the same, with try-cc behaving like try-build for no reason.
Fix try-cc to really only test compilation (no linking) and replace
all calls to try-cc in the Makefile with try-build (as this is what
we need). Simplify the Makefile calls a bit on the way.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
The 9p kernel code uses separate types for uid_t and gid_t. To avoid
changing too much code needlessly, copy over the kernel definitions
from uidgid.h into 9p.h.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
The special Linux poison bits do not make sense in userland, so
just replace them with NULL to spot bogus accesses.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
The original Linux copy of rbtree.c used EXPORT_SYMBOL macros which
we don't need. Remove them and get rid of the now unneeded export.h.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
The ARM part requires psci.h, which is pretty recent, so older
distributions might not carry it. This pulls in sizes.h, which is not
an uapi header, but quite convenient and not really Linux specific.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Although those header files are technically Linux internal, they
hold definitions for x86 hardware related bits (APIC, MPSPEC, VESA).
We just reuse their definitions, so copy those files over.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Similarily to the generic uapi/linux/kvm.h, each architecture
carries a kvm.h header in its arch/*/include/uapi/asm directory.
These contain bits for the architecture specific interface.
Since we use many recent features in kvmtool, the system headers
provided by the distribution are usually not up-to-date.
Copy the Linux v4.1-rc6 versions of those files for all supported
architectures into the kvmtool tree to get access to the full glory.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Some distributions with an older uapi header copy do not have all the
definitions that we need (Ubuntu 12.04 comes to mind).
So copy the latest versions from the Linux tree to override the system
header versions.
Also include virtio_mmio.h, which is technically not an uapi header,
but is used in such a way. QEMU has actually a copy of these
definitions in its own header file, so our approach should even be
better.
Those files are from a Linux v4.1-rc6 source tree, but the virtio-1.0
updates have been left out. kvmtool implements only the older version
of virtio.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
This file includes 9P protocol definitions which we need.
Copy them from a recent Linux tree.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
The uapi copy of kvm.h is usually outdated, so copy over a fresh copy
from the Linux v4.1-rc6 source tree. This exercise should later be
repeated when new kvmtool code requires this.
It seems to be common practise in QEMU to do so.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Instead of moving the whole include file in, we just copy the needed
definitions from byteorder/generic.h into the kvmtool headers.
This avoids needless clutter of the include directory.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
This is not really Linux specific, just a neat implementation
used by kvmtool, too.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Copy some non-Linux specific header files over to the kvmtool
directory. This includes generic helpers like stringify and list
definitions.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
The Makefile used to reference ../.. on some occasions to point
to the Linux source tree. Since this is no longer there, remove
those bogus paths.
Since we cannot (and don't want to) determine the Linux kernel
version on this way, we hardcode v3.18.0 for now, which is the version
the in-kernel-tree version used at the point of forking.
This should be later revisited and adjusted to whatever versioning
scheme we will use.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Marc Zyngier