The multiboot spec (https://www.gnu.org/software/grub/manual/multiboot/),
section 3.1.3, allows for bss_end_addr to be zero.

A zero bss_end_addr signifies there is no .bss section.
Suggested-by: NDaniel Kiper <daniel.kiper@oracle.com>
Signed-off-by: NJack Schwartz <jack.schwartz@oracle.com>
Reviewed-by: NDaniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: NPrasad J Pandit <pjp@fedoraproject.org>
Signed-off-by: NKevin Wolf <kwolf@redhat.com>

Use types that are defined by QEMU in trace events caused build failures
for the UST trace backend:

  In file included from trace-ust-all.c:13:0:
  trace-ust-all.h:11844:206: error: unknown type name ‘hwaddr’

It only knows about C built-in types, and any types that are pulled in
from includs of qemu-common.h and lttng/tracepoint.h. This does not
include the 'hwaddr' type, so replace it with a uint64_t which is what
exec/hwaddr.h defines 'hwaddr' as. This fixes the build failure
introduced by

  commit 9eb8040c
  Author: Peter Maydell <peter.maydell@linaro.org>
  Date:   Fri Mar 2 10:45:39 2018 +0000

    hw/misc/tz-ppc: Model TrustZone peripheral protection controller
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>
Reviewed-by: NPeter Maydell <peter.maydell@linaro.org>
Message-id: 20180306134317.836-1-berrange@redhat.com
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

basename(3) and dirname(3) modify their argument and may return
pointers to statically allocated memory which may be overwritten by
subsequent calls.
g_path_get_basename and g_path_get_dirname have no such issues, and
therefore more preferable.
Signed-off-by: NJulia Suvorova <jusual@mail.ru>
Message-Id: <1519888086-4207-1-git-send-email-jusual@mail.ru>
Reviewed-by: NMarc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: NCornelia Huck <cohuck@redhat.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

The MemoryListener is registered on address_space_memory, there is
not much to assert.  This currently works because the callback
is invoked only once when the listener is registered, but section->fv
is the _new_ FlatView, not the old one on later calls and that
would break.

This confines address_space_to_flatview to exec.c and memory.c.
Acked-by: NDavid Gibson <david@gibson.dropbear.id.au>
Reviewed-by: NAlexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Automatic creation of SCSI controllers for "-drive if=scsi" for x86
machines was quite a bad idea (see description of commit f778a82f
for details). This is marked as deprecated since QEMU v2.9.0, and as
far as I know, nobody complained that this is still urgently required
anymore. Time to remove this now.
Suggested-by: NMarkus Armbruster <armbru@redhat.com>
Signed-off-by: NThomas Huth <thuth@redhat.com>
Message-Id: <1519123357-13225-1-git-send-email-thuth@redhat.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

This is useful to help diagnose problems related to address clashes during
MacOS 9 boot.
Signed-off-by: NMark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>

This patch fixes an incorrect behavior when the -kernel argument has been
specified without -bios. In this case the kernel was loaded twice. At address
32M as a raw image and afterwards by load_elf/load_uimage at the
corresponding load address. In this case the region for the device tree and
the raw kernel image may overlap.

The patch fixes the behavior by loading the kernel image once with
load_elf/load_uimage and skips loading the raw image.

When here do not use bios_name/size for the kernel and use a more generic
name called payload_name/size.

New in v3: dtb must be stored between kernel and initrd because Linux can
           handle the dtb only within the first 64MB. Add a comment to
           clarify the behavior.
Signed-off-by: NDavid Engraf <david.engraf@sysgo.com>
Reviewed-by: NDavid Gibson <david@gibson.dropbear.id.au>
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>

Linux kernel commit 2a9d832cc9aae21ea827520fef635b6c49a06c6d
(of: Add bindings for chosen node, stdout-path) deprecated chosen property
"linux,stdout-path" and "stdout".

Introduce the new property "stdout-path" and continue supporting the older
property to remain compatible with existing/older firmware. This older property
can be deprecated after 5 years.
Signed-off-by: NNikunj A Dadhania <nikunj@linux.vnet.ibm.com>
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>

The sxxm (speculative execution exploit mitigation) machine type is a
variant of the 2.12 machine type with workarounds for speculative
execution vulnerabilities enabled by default.
Signed-off-by: NSuraj Jitindar Singh <sjitindarsingh@gmail.com>
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>

Convert cap-ibs (indirect branch speculation) to a custom spapr-cap
type.

All tristate caps have now been converted to custom spapr-caps, so
remove the remaining support for them.
Signed-off-by: NSuraj Jitindar Singh <sjitindarsingh@gmail.com>
[dwg: Don't explicitly list "?"/help option, trust convention]
[dwg: Fold tristate removal into here, to not break bisect]
[dwg: Fix minor style problems]
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>

Convert cap-sbbc (speculation barrier bounds checking) to a custom
spapr-cap type.
Signed-off-by: NSuraj Jitindar Singh <sjitindarsingh@gmail.com>
[dwg: Removed trailing whitespace]
[dwg: Don't explicitly list "?"/help option, trust convention]
[dwg: Fix some minor style problems]
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>

Convert cap-cfpc (cache flush on privilege change) to a custom spapr-cap
type.
Signed-off-by: NSuraj Jitindar Singh <sjitindarsingh@gmail.com>
[dwg: Don't explicitly list "?"/help option, trusting convention]
[dwg: Strip no-longer-necessary ATTRIBUTE_UNUSED back off]
[dwg: Fix some minor style problems]
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>

There are currently 2 implemented types of spapr-caps, boolean and
tristate. However there may be a need for caps which don't fit either of
these options. Add a custom capability type for which a list of custom
valid strings can be specified and implement the get/set functions for
these. Also add a field for help text to describe the available options.
Signed-off-by: NSuraj Jitindar Singh <sjitindarsingh@gmail.com>
[dwg: Change "help" option to "?" matching qemu conventions]
[dwg: Add ATTRIBUTE_UNUSED to avoid breaking bisect]
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>

Move the remaining comment into macio.c for reference, then remove the
macio_init() function and instantiate the macio devices for both Old World
and New World machines via qdev_init_nofail() directly.
Signed-off-by: NMark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>

This removes the last of the functionality from macio_init() in preparation
for its subsequent removal.
Signed-off-by: NMark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: NDavid Gibson <david@gibson.dropbear.id.au>
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>

Also switch macio_newworld_realize() over to use it rather than using the pic_mem
memory region directly.

Now that both Old World and New World macio devices no longer make use of the
pic_mem memory region directly, we can remove it.
Signed-off-by: NMark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: NDavid Gibson <david@gibson.dropbear.id.au>
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>

This is to faciliate access to OpenPICState when wiring up the PIC to the macio
controller.
Signed-off-by: NMark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: NDavid Gibson <david@gibson.dropbear.id.au>
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>

This is needed before the next patch because the target-dependent kvm stub
uses the existing kvm_openpic_connect_vcpu() declaration, making it impossible
to move the device-specific declarations into the same file without breaking
ppc-linux-user compilation.
Signed-off-by: NMark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>

Also switch macio_oldworld_realize() over to use it rather than using the pic_mem
memory region directly.
Signed-off-by: NMark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: NDavid Gibson <david@gibson.dropbear.id.au>
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>

Signed-off-by: NMark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: NDavid Gibson <david@gibson.dropbear.id.au>
Reviewed-by: NPhilippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>

This enables the device to be made available during the setup of the Old World
machine. In order to pass back the previous set of IRQs we temporarily introduce
a new pic_irqs parameter until it can be removed.

An additional benefit of this change is that it is also possible to remove the
pic_mem pointer used for macio by accessing the memory region via sysbus.
Signed-off-by: NMark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: NDavid Gibson <david@gibson.dropbear.id.au>
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>

Signed-off-by: NMark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>

Signed-off-by: NMark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>

Now that the ESCC device is instantiated directly via qdev, move it to within
the macio device and wire up the IRQs and memory regions using the sysbus API.

This enables to remove the now-obsolete escc_mem parameter to the macio_init()
function.

(Note this patch also contains small touch-ups to the formatting in
macio_escc_legacy_setup() and ppc_heathrow_init() in order to keep checkpatch
happy)
Signed-off-by: NMark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: NPhilippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>

The current recommendation is to embed subdevices directly within their container
device, so do this for the DBDMA device.
Signed-off-by: NMark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>

VSMT must be set in order to compute VCPU ids. This means that the
following functions must not be called before spapr_set_vsmt_mode()
was called:
- spapr_vcpu_id()
- spapr_is_thread0_in_vcore()
- xics_max_server_number()

We had a recent regression where the latter would be called before VSMT
was set, and broke migration of some old machine types. This patch
adds assert() in the above functions to avoid problems in the future.

Also, since VSMT is really a CPU related thing, spapr_set_vsmt_mode() is
now called from spapr_init_cpus(), just before the first VSMT user.
Signed-off-by: NGreg Kurz <groug@kaod.org>
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>

Some older machine types create more ICPs than needed. We hence
need to register up to xics_max_server_number() dummy ICPs to
accomodate the migration of these machine types.

Recent VSMT rework changed xics_max_server_number() to return

    DIV_ROUND_UP(max_cpus * spapr->vsmt, smp_threads)

instead of

    DIV_ROUND_UP(max_cpus * kvmppc_smt_threads(), smp_threads);

The change is okay but it requires spapr->vsmt to be set, which
isn't the case with the current code. This causes the formula to
return zero and we don't create dummy ICPs. This breaks migration
of older guests as reported here:

    https://bugzilla.redhat.com/show_bug.cgi?id=1549087

The dummy ICP workaround doesn't really have a dependency on XICS
itself. But it does depend on proper VCPU id numbering and it must
be applied before creating vCPUs (ie, creating real ICPs). So this
patch moves the workaround to spapr_init_cpus(), which already
assumes VSMT to be set.

Fixes: 72194664 ("spapr: use spapr->vsmt to compute VCPU ids")
Reported-by: NLukas Doktor <ldoktor@redhat.com>
Signed-off-by: NGreg Kurz <groug@kaod.org>
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>

Add emulation of aCube Sam460ex board based on AMCC 460EX embedded SoC.
This is not a complete implementation yet with a lot of components
still missing but enough for the U-Boot firmware to start and to boot
a Linux kernel or AROS.
Signed-off-by: NFrançois Revol <revol@free.fr>
Signed-off-by: NBALATON Zoltan <balaton@eik.bme.hu>
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>

This is the PCIX controller found in newer 440 core SoCs e.g. the
AMMC 460EX. The device tree refers to this as plb-pcix compared to
the plb-pci controller in older 440 SoCs.
Signed-off-by: NBALATON Zoltan <balaton@eik.bme.hu>
[dwg: Remove hwaddr from trace-events, that doesn't work with some
 trace backends]
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>

Signed-off-by: NBALATON Zoltan <balaton@eik.bme.hu>
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>

Commit 5d0fb150 "spapr: consolidate the VCPU id numbering logic
in a single place" introduced a helper to detect thread0 of a virtual
core based on its VCPU id. This is used to create CPU core nodes in
the DT, but it is broken in TCG.

$ qemu-system-ppc64 -nographic -accel tcg -machine dumpdtb=dtb.bin \
                    -smp cores=16,maxcpus=16,threads=1
$ dtc -f -O dts dtb.bin | grep POWER8
                PowerPC,POWER8@0 {
                PowerPC,POWER8@8 {

instead of the expected 16 cores that we get with KVM:

$ dtc -f -O dts dtb.bin | grep POWER8
                PowerPC,POWER8@0 {
                PowerPC,POWER8@8 {
                PowerPC,POWER8@10 {
                PowerPC,POWER8@18 {
                PowerPC,POWER8@20 {
                PowerPC,POWER8@28 {
                PowerPC,POWER8@30 {
                PowerPC,POWER8@38 {
                PowerPC,POWER8@40 {
                PowerPC,POWER8@48 {
                PowerPC,POWER8@50 {
                PowerPC,POWER8@58 {
                PowerPC,POWER8@60 {
                PowerPC,POWER8@68 {
                PowerPC,POWER8@70 {
                PowerPC,POWER8@78 {

This happens because spapr_get_vcpu_id() maps VCPU ids to
cs->cpu_index in TCG mode. This confuses the code in
spapr_is_thread0_in_vcore(), since it assumes thread0 VCPU
ids to have a spapr->vsmt spacing.

    spapr_get_vcpu_id(cpu) % spapr->vsmt == 0

Actually, there's no real reason to expose cs->cpu_index instead
of the VCPU id, since we also generate it with TCG. Also we already
set it explicitly in spapr_set_vcpu_id(), so there's no real reason
either to call kvm_arch_vcpu_id() with KVM.

This patch unifies spapr_get_vcpu_id() to always return the computed
VCPU id both in TCG and KVM. This is one step forward towards KVM<->TCG
migration.

Fixes: 5d0fb150Reported-by: NCédric Le Goater <clg@kaod.org>
Signed-off-by: NGreg Kurz <groug@kaod.org>
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>

Headers like "hw/loader.h" and "qemu/sockets.h" are not needed in
the hw/net/*.c files. And Some other headers are included via other
headers already, so we can drop them, too.
Signed-off-by: NThomas Huth <thuth@redhat.com>
Signed-off-by: NJason Wang <jasowang@redhat.com>

The previous commit improved compile time by including less of the
generated QAPI headers.  This is impossible for stuff defined directly
in qapi-schema.json, because that ends up in headers that that pull in
everything.

Move everything but include directives from qapi-schema.json to new
sub-module qapi/misc.json, then include just the "misc" shard where
possible.

It's possible everywhere, except:

* monitor.c needs qmp-command.h to get qmp_init_marshal()

* monitor.c, ui/vnc.c and the generated qapi-event-FOO.c need
  qapi-event.h to get enum QAPIEvent

Perhaps we'll get rid of those some other day.

Adding a type to qapi/migration.json now recompiles some 120 instead
of 2300 out of 5100 objects.
Signed-off-by: NMarkus Armbruster <armbru@redhat.com>
Message-Id: <20180211093607.27351-25-armbru@redhat.com>
[eblake: rebase to master]
Signed-off-by: NEric Blake <eblake@redhat.com>

In my "build everything" tree, a change to the types in
qapi-schema.json triggers a recompile of about 4800 out of 5100
objects.

The previous commit split up qmp-commands.h, qmp-event.h, qmp-visit.h,
qapi-types.h.  Each of these headers still includes all its shards.
Reduce compile time by including just the shards we actually need.

To illustrate the benefits: adding a type to qapi/migration.json now
recompiles some 2300 instead of 4800 objects.  The next commit will
improve it further.
Signed-off-by: NMarkus Armbruster <armbru@redhat.com>
Message-Id: <20180211093607.27351-24-armbru@redhat.com>
Reviewed-by: NEric Blake <eblake@redhat.com>
Reviewed-by: NMarc-André Lureau <marcandre.lureau@redhat.com>
[eblake: rebase to master]
Signed-off-by: NEric Blake <eblake@redhat.com>

Fixes the following ASAN warnings:

/home/elmarco/src/qemu/hw/net/net_tx_pkt.c:201:27: runtime error: member access within misaligned address 0x631000028846 for type 'struct ip_header', which requires 4 byte alignment
0x631000028846: note: pointer points here
 01 00 00 00 45 00  01 a9 01 00 00 00 40 11  78 45 00 00 00 00 ff ff  ff ff 00 00 00 00 00 00  00 00
             ^
/home/elmarco/src/qemu/hw/net/net_tx_pkt.c:208:63: runtime error: member access within misaligned address 0x631000028846 for type 'struct ip_header', which requires 4 byte alignment
0x631000028846: note: pointer points here
 01 00 00 00 45 00  01 a9 01 00 00 00 40 11  78 45 00 00 00 00 ff ff  ff ff 00 00 00 00 00 00  00 00
             ^
/home/elmarco/src/qemu/hw/net/net_tx_pkt.c:210:13: runtime error: member access within misaligned address 0x631000028846 for type 'struct ip_header', which requires 4 byte alignment
0x631000028846: note: pointer points here
 01 00 00 00 45 00  01 a9 01 00 00 00 40 11  78 45 00 00 00 00 ff ff  ff ff 00 00 00 00 00 00  00 00
Signed-off-by: NMarc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: NPeter Maydell <peter.maydell@linaro.org>
Message-id: 20180209190340.19516-1-marcandre.lureau@redhat.com
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

This reverts commit 4da97120.

blk_aio_flush() now handles the blk->root == NULL case, so we no longer
need this workaround.

Cc: John Snow <jsnow@redhat.com>
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: NEric Blake <eblake@redhat.com>
Signed-off-by: NKevin Wolf <kwolf@redhat.com>

Define a new board model for the MPS2 with an AN505 FPGA image
containing a Cortex-M33. Since the FPGA images for TrustZone
cores (AN505, and the similar AN519 for Cortex-M23) have a
significantly different layout of devices to the non-TrustZone
images, we use a new source file rather than shoehorning them
into the existing mps2.c.
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>
Reviewed-by: NRichard Henderson <richard.henderson@linaro.org>
Message-id: 20180220180325.29818-20-peter.maydell@linaro.org

Model the Arm IoT Kit documented in
http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.ecm0601256/index.html

The Arm IoT Kit is a subsystem which includes a CPU and some devices,
and is intended be extended by adding extra devices to form a
complete system.  It is used in the MPS2 board's AN505 image for the
Cortex-M33.
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>
Reviewed-by: NRichard Henderson <richard.henderson@linaro.org>
Message-id: 20180220180325.29818-19-peter.maydell@linaro.org

Add remaining easy registers to iotkit-secctl:
 * NSCCFG just routes its two bits out to external GPIO lines
 * BRGINSTAT/BRGINTCLR/BRGINTEN can be dummies, because QEMU's
   bus fabric can never report errors
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>
Message-id: 20180220180325.29818-18-peter.maydell@linaro.org