Cédric Le Goater [Wed, 29 Mar 2017 13:53:23 +0000 (15:53 +0200)]
ppc/xics: introduce an 'intc' backlink under PowerPCCPU
Today, the ICPState array of the sPAPR machine is indexed with
'cpu_index' of the CPUState. This numbering of CPUs is internal to
QEMU and the guest only knows about what is exposed in the device
tree, that is the 'cpu_dt_id'. This is why sPAPR uses the helper
xics_get_cpu_index_by_dt_id() to do the mapping in a couple of places.
To provide a more generic XICS layer, we need to abstract the IRQ
'server' number and remove any assumption made on its nature. It
should not be used as a 'cpu_index' for lookups like xics_cpu_setup()
and xics_cpu_destroy() do.
To reach that goal, we choose to introduce a generic 'intc' backlink
under PowerPCCPU, and let the machine core init routine do the
ICPState lookup. The resulting object is passed on to xics_cpu_setup()
which does the store under PowerPCCPU. The IRQ 'server' number in XICS
is now generic. sPAPR uses 'cpu_dt_id' and PowerNV will use 'PIR'
number.
This also has the benefit of simplifying the sPAPR hcall routines
which do not need to do any ICPState lookups anymore.
target/ppc: Add ibm,processor-radix-AP-encodings for TCG
The ibm,processor-radix-AP-encodings device tree property of the cpu node
is used to specify the radix mode supported page sizes of the processor
to the guest os. Contained in the top 3 bits of the msb is the actual
page size (AP) encoding associated with the corresponding radix mode
supported page size. Add this property for a TCG guest, note the TCG code
is capable of translating any format so just add the 4 default page sizes.
The ibm,processor-radix-AP-encodings device tree property is defined as:
One to n cells in ascending order of radix mode supported page sizes
encoded as BE ints (32bit on ppc) in the form:
0bxxxyyyyyyyyyyyyyyyyyyyyyyyyyyyyy
- 0bxxx -> AP encoding
- 0byyyyyyyyyyyyyyyyyyyyyyyyyyyyy -> supported page size encoded as a shift
spapr_pci: Warn when RAM page size is not enabled in IOMMU page mask
If a page size used by QEMU is not enabled in the PHB IOMMU page mask,
in-kernel acceleration of TCE handling won't be enabled and performance
might be slower than expected.
This prints a warning if system page size is not enabled. This should
print a warning if huge pages are enabled but sphb.pgsz still uses
the default value of 4K|64K.
target-ppc/kvm: Enable in-kernel TCE acceleration for multi-tce
This enables in-kernel handling of H_PUT_TCE_INDIRECT and
H_STUFF_TCE hypercalls. The host kernel support is there since v4.6,
in particular d3695aa4f452
("KVM: PPC: Add support for multiple-TCE hcalls").
H_PUT_TCE is already accelerated and does not need any special enablement.
Sam Bobroff [Sun, 19 Mar 2017 23:46:49 +0000 (10:46 +1100)]
spapr: Workaround for broken radix guests
For a little while around 4.9, Linux kernels that saw the radix bit in
ibm,pa-features would attempt to set up the MMU as if they were a
hypervisor, even if they were a guest, which would cause them to
crash.
Work around this by detecting pre-ISA 3.0 guests by their lack of that
bit in option vector 1, and then removing the radix bit from
ibm,pa-features. Note: This now requires regeneration of that node
after CAS negotiation.
Sam Bobroff [Thu, 23 Mar 2017 03:46:00 +0000 (14:46 +1100)]
spapr: Enable ISA 3.0 MMU mode selection via CAS
Add the new node, /chosen/ibm,arch-vec-5-platform-support to the
device tree. This allows the guest to determine which modes are
supported by the hypervisor.
Update the option vector processing in h_client_architecture_support()
to handle the new MMU bits. This allows guests to request hash or
radix mode and QEMU to create the guest's HPT at this time if it is
necessary but hasn't yet been done. QEMU will terminate the guest if
it requests an unavailable mode, as required by the architecture.
Extend the ibm,pa-features node with the new ISA 3.0 values
and set the radix bit if KVM supports radix mode. This probably won't
be used directly by guests to determine the availability of radix mode
(that is indicated by the new node added above) but the architecture
requires that it be set when the hardware supports it.
If QEMU is using KVM, and KVM is capable of running in radix mode,
guests can be run in real-mode without allocating a HPT (because KVM
will use a minimal RPT). So in this case, we avoid creating the HPT
at reset time and later (during CAS) create it if it is necessary.
ISA 3.0 guests will now begin to call h_register_process_table(),
which has been added previously.
Signed-off-by: Sam Bobroff <[email protected]>
[dwg: Strip some unneeded prefix from error messages] Signed-off-by: David Gibson <[email protected]>
The H_REGISTER_PROCESS_TABLE H_CALL is used by a guest to indicate to the
hypervisor where in memory its process table is and how translation should
be performed using this process table.
Provide the implementation of this H_CALL for a guest.
We first check for invalid flags, then parse the flags to determine the
operation, and then check the other parameters for valid values based on
the operation (register new table/deregister table/maintain registration).
The process table is then stored in the appropriate location and registered
with the hypervisor (if running under KVM), and the LPCR_[UPRT/GTSE] bits
are updated as required.
Signed-off-by: Suraj Jitindar Singh <[email protected]> Signed-off-by: Sam Bobroff <[email protected]>
[dwg: Correct missing prototype and uninitialized variable] Signed-off-by: David Gibson <[email protected]>
target/ppc: Add new H-CALL shells for in memory table translation
The use of the new in memory tables introduced in ISAv3.00 for translation,
also referred to as process tables, requires the introduction of 3 new
H-CALLs; H_REGISTER_PROCESS_TABLE, H_CLEAN_SLB, and H_INVALIDATE_PID.
Add shells for each of these and register them as the hypercall handlers.
Currently they all log an unimplemented hypercall and return H_FUNCTION.
Sam Bobroff [Sun, 19 Mar 2017 23:46:43 +0000 (10:46 +1100)]
spapr: Add ibm,processor-radix-AP-encodings to the device tree
Use the new ioctl, KVM_PPC_GET_RMMU_INFO, to fetch radix MMU
information from KVM and present the page encodings in the device tree
under ibm,processor-radix-AP-encodings. This provides page size
information to the guest which is necessary for it to use radix mode.
Signed-off-by: Sam Bobroff <[email protected]>
[dwg: Compile fix for 32-bit targets, style nit fix] Signed-off-by: David Gibson <[email protected]>
target-ppc: kvm: make use of KVM_CREATE_SPAPR_TCE_64
KVM_CAP_SPAPR_TCE capability allows creating TCE tables in KVM which
allows having in-kernel acceleration for H_PUT_TCE_xxx hypercalls.
However it only supports 32bit DMA windows at zero bus offset.
There is a new KVM_CAP_SPAPR_TCE_64 capability which supports 64bit
window size, variable page size and bus offset.
This makes use of the new capability. The kernel headers are already
updated as the kernel support went in to v4.6.
Thomas Huth [Thu, 9 Mar 2017 18:37:53 +0000 (19:37 +0100)]
hw/ppc/pnv: Classify the "PowerNV Chip" devices as CPU devices
The devices that are derived from TYPE_PNV_CHIP currently show up
as "uncategorized" devices in the help text of "-device ?". Since
they obviously are related to the CPU, let's put them into the
CPU category instead.
Also use an 'sPAPRRTCState' attribute under the sPAPR machine to hold
the RTC object. Overall, these changes remove an unnecessary and
implicit dependency on SysBus.
Sam Bobroff [Wed, 29 Mar 2017 05:01:28 +0000 (16:01 +1100)]
target/ppc: Improve accuracy of guest HTM availability on P8s
On Power8 hosts it is currently theoretically possible for QEMU/KVM-HV guests
to receive a ibm,pa-features property indicating that HTM support is available
when it is not. The situation would occur if the platform firmware of
a Power8 host cleared the HTM bit of the ibm,pa-features property.
QEMU would query KVM for the availability of HTM, which will return no
support, but workaround code in kvm_arch_init_vcpu() would then
re-enable it because KVM_HV is in use and the processor is P8.
This patch adjusts the workaround in kvm_arch_init_vcpu() so that it does not
enable HTM (in the above case) unless the host kernel indicates to the QEMU
process, via the auxiliary vector, that userspace can use HTM (via the HWCAP2
bit KVM_FEATURE2_HTM).
The reason to use the value from the auxiliary vector is that it is
set based only on what the host kernel found in the ibm,pa-features
HTM bit at boot time.
Peter Maydell [Tue, 25 Apr 2017 13:14:17 +0000 (14:14 +0100)]
Merge remote-tracking branch 'remotes/jasowang/tags/net-pull-request' into staging
# gpg: Signature made Tue 25 Apr 2017 12:22:03 BST
# gpg: using RSA key 0xEF04965B398D6211
# gpg: Good signature from "Jason Wang (Jason Wang on RedHat) <[email protected]>"
# gpg: WARNING: This key is not certified with sufficiently trusted signatures!
# gpg: It is not certain that the signature belongs to the owner.
# Primary key fingerprint: 215D 46F4 8246 689E C77F 3562 EF04 965B 398D 6211
* remotes/jasowang/tags/net-pull-request:
COLO-compare: Optimize tcp compare trace event
COLO-compare: Optimize tcp compare for option field
slirp: add a fake NC-SI backend
aspeed: add a FTGMAC100 nic
net/ftgmac100: add a 'aspeed' property
net: add FTGMAC100 support
hw/net: add MII definitions
colo-compare: Fix old packet check bug.
Philipp Kern [Tue, 18 Aug 2015 11:50:55 +0000 (13:50 +0200)]
target-s390x: Mask the SIGP order_code to 8bit.
According to "CPU Signaling and Response", "Signal-Processor Orders",
the order field is bit position 56-63. Without this, the Linux
guest kernel is sometimes unable to stop emulation and enters
an infinite loop of "XXX unknown sigp: 0xffffffff00000005".
macOS 10.12 deprecated/replaced many AppKit constants to make naming
more consistent. Use the new constants, and #define them to the
old constants when compiling against a pre-10.12 SDK.
COLO-compare: Optimize tcp compare for option field
In this patch we support packet that have tcp options field.
Add tcp options field check, If the packet have options
field we just skip it and compare tcp payload,
Avoid unnecessary checkpoint, optimize performance.
NC-SI (Network Controller Sideband Interface) enables a BMC to manage
a set of NICs on a system. This model takes the simplest approach and
reverses the NC-SI packets to pretend a NIC is present and exercise
the Linux driver.
The NCSI header file <ncsi-pkt.h> comes from mainline Linux and was
untabified.
There is a second NIC but we do not use it for the moment. We use the
'aspeed' property to tune the definition of the end of ring buffer bit
for the Aspeed SoCs.
Exynos4210 has four SD/MMC controllers supporting:
- SD Standard Host Specification Version 2.0,
- MMC Specification Version 4.3,
- SDIO Card Specification Version 2.0,
- DMA and ADMA.
Add emulation of SDHCI devices which allows accessing storage through SD
cards. Differences from real hardware:
- Devices are shipped with eMMC memory, not SD card.
- The Exynos4210 SDHCI has few more registers, e.g. for
controlling the clocks, additional status (0x80, 0x84, 0x8c). These
are not implemented.
Testing on smdkc210 machine with "-drive file=FILE,if=sd,bus=0,index=2".
* remotes/cody/tags/block-pull-request:
qemu-iotests: _cleanup_qemu must be called on exit
block/rbd: Add support for reopen()
block/rbd - update variable names to more apt names
block: use bdrv_can_set_read_only() during reopen
block: introduce bdrv_can_set_read_only()
block: code movement
block: honor BDRV_O_ALLOW_RDWR when clearing bs->read_only
block: do not set BDS read_only if copy_on_read enabled
block: add bdrv_set_read_only() helper function
qemu-iotests: exclude vxhs from image creation via protocol
block/vxhs.c: Add qemu-iotests for new block device type "vxhs"
block/vxhs.c: Add support for a new block device type called "vxhs"
Jeff Cody [Fri, 7 Apr 2017 20:55:31 +0000 (16:55 -0400)]
block/rbd - update variable names to more apt names
Update 'clientname' to be 'user', which tracks better with both
the QAPI and rados variable naming.
Update 'name' to be 'image_name', as it indicates the rbd image.
Naming it 'image' would have been ideal, but we are using that for
the rados_image_t value returned by rbd_open().
Jeff Cody [Fri, 7 Apr 2017 20:55:27 +0000 (16:55 -0400)]
block: honor BDRV_O_ALLOW_RDWR when clearing bs->read_only
The BDRV_O_ALLOW_RDWR flag allows / prohibits the changing of
the BDS 'read_only' state, but there are a few places where it
is ignored. In the bdrv_set_read_only() helper, make sure to
honor the flag.
Jeff Cody [Fri, 7 Apr 2017 20:55:26 +0000 (16:55 -0400)]
block: do not set BDS read_only if copy_on_read enabled
A few block drivers will set the BDS read_only flag from their
.bdrv_open() function. This means the bs->read_only flag could
be set after we enable copy_on_read, as the BDRV_O_COPY_ON_READ
flag check occurs prior to the call to bdrv->bdrv_open().
This adds an error return to bdrv_set_read_only(), and an error will be
return if we try to set the BDS to read_only while copy_on_read is
enabled.
This patch also changes the behavior of vvfat. Before, vvfat could
override the drive 'readonly' flag with its own, internal 'rw' flag.
For instance, this -drive parameter would result in a writable image:
* remotes/armbru/tags/pull-error-2017-04-24:
error: Apply error_propagate_null.cocci again
qga: Make errp the last parameter of qga_vss_fsfreeze
migration: Make errp the last parameter of local functions
scsi: Make errp the last parameter of virtio_scsi_common_realize
fdc: Make errp the last parameter of fdctrl_connect_drives
nfs: Make errp the last parameter of nfs_client_open
block: Make errp the last parameter of commit_active_start
mirror: Make errp the last parameter of mirror_start_job
crypto: Make errp the last parameter of functions
block: Make errp the last parameter of bdrv_img_create
socket: Make errp the last parameter of vsock_connect_saddr
socket: Make errp the last parameter of unix_connect_saddr
socket: Make errp the last parameter of inet_connect_saddr
socket: Make errp the last parameter of socket_connect
util/error: Fix leak in error_vprepend()
ppc: Add SM501 device in ppc softmmu targets default configs
This is not used by default on any emulated machine yet but it is
still useful to have it compiled so it can be added from the command
line for clients that can use it (e.g. MorphOS has no driver for any
other emulated video cards but can output via SM501)
This is to allow clients to initialise these without failing as long
as no 2D engine function is called that would use the written value.
Saved values are not used yet (may get used when more of 2D engine is
added sometimes) and clients normally only write to most of these
registers, nothing is known to ever read them but they are documented
as read/write so also implement read for these.
Rework HWC handling to simplify it and fix cursor not updating on
screen as needed. Previously cursor was not updated because checking
for changes in a line overrode the update flag set for the cursor but
fixing this is not enough because the cursor should also be updated if
its shape or location changes. Introduce hwc_invalidate() function to
handle that similar to other display controller models.
We only emulate the sysbus device in its default LE mode and PCI is LE
as well so specify this for registers and framebuffer memory.
Note that though the Linux kernel driver has code which claims to
handle both big and little endian, it is obviously bogus for 16 bit
and cannot be trusted as a source of information on the framebuffer
pixel format. This is our best guess about device behaviour based on
the specs and testing with MorphOS that is known to work on real HW.
Do not use the base address to access data in local memory. This is in
preparation to allow chip connected via PCI where base address depends
on where the BAR is mapped so it will be unknown.
Adding vmstate saving is not in this patch because the state structure
will be changed in further patches, then another patch will add
vmstate descriptor after those changes.
Peter Maydell [Mon, 24 Apr 2017 08:43:03 +0000 (09:43 +0100)]
Merge remote-tracking branch 'remotes/mcayland/tags/qemu-sparc-signed' into staging
qemu-sparc update
# gpg: Signature made Fri 21 Apr 2017 20:09:35 BST
# gpg: using RSA key 0x5BC2C56FAE0F321F
# gpg: Good signature from "Mark Cave-Ayland <[email protected]>"
# Primary key fingerprint: CC62 1AB9 8E82 200D 915C C9C4 5BC2 C56F AE0F 321F
* remotes/mcayland/tags/qemu-sparc-signed:
tcx: switch to load_image_mr() and remove prom_addr hack
tcx: use tcx_set_dirty() for accelerated ops
tcx: remove primitives for non-32-bit surfaces
tcx: remove TARGET_PAGE_SIZE from tcx24_update_display()
tcx: remove TARGET_PAGE_SIZE from tcx_update_display()
tcx: remove page24 and cpage from tcx24_update_display()
tcx: alter tcx24_reset_dirty() to accept address and length parameters
tcx: alter tcx24_check_dirty() to accept address and length parameters
tcx: ensure tcx_set_dirty() also invalidates the 24-bit plane and cplane
tcx: alter tcx_set_dirty() to accept address and length parameters
cg3: switch to load_image_mr() and remove prom-addr hack
cg3: fix up size parameter for memory_region_get_dirty()
cg3: remove TARGET_PAGE_SIZE rounding on dirty page detection
Fix standard vga mode check: Both s->config and s->enabled must be set
to enable vmware command fifo processing.
Drop dirty tracking code from the fifo rendering code path, it isn't
used anyway because vmsvga turns off dirty tracking when leaving
standard vga mode.
The vga code clears the dirty bits *after* reading the framebuffer
memory. So if the guest framebuffer updates hits the race window
between vga reading the framebuffer and vga clearing the dirty bits
vga will miss that update
Fix it by using the new memory_region_copy_and_clear_dirty()
memory_region_copy_get_dirty() functions. That way we clear the
dirty bitmap before reading the framebuffer. Any guest display
updates happening in parallel will be properly tracked in the
dirty bitmap then and the next display refresh will pick them up.
Problem triggers with mttcg only. Before mttcg was merged tcg
never ran in parallel to vga emulation. Using kvm will hide the
problem too, due to qemu operating on a userspace copy of the
kernel's dirty bitmap.
Add vga_scanline_invalidated helper to check whenever a scanline was
invalidated. Add a sanity check to fix OOB read access for display
heights larger than 2048.
Only cirrus uses this, for hardware cursor rendering, so having this
work properly for the first 2048 scanlines only shouldn't be a problem
as the cirrus can't handle large resolutions anyway. Also changing the
invalidated_y_table size would break live migration.
memory: add support getting and using a dirty bitmap copy.
This patch adds support for getting and using a local copy of the dirty
bitmap.
memory_region_snapshot_and_clear_dirty() will create a snapshot of the
dirty bitmap for the specified range, clear the dirty bitmap and return
the copy. The returned bitmap can be a bit larger than requested, the
range is expanded so the code can copy unsigned longs from the bitmap
and avoid atomic bit update operations.
memory_region_snapshot_get_dirty() will return the dirty status of
pages, pretty much like memory_region_get_dirty(), but using the copy
returned by memory_region_copy_and_clear_dirty().
The FTGMAC100 device is an Ethernet controller with DMA function that
can be found on Aspeed SoCs (which include NCSI).
It is fully compliant with IEEE 802.3 specification for 10/100 Mbps
Ethernet and IEEE 802.3z specification for 1000 Mbps Ethernet and
includes Reduced Media Independent Interface (RMII) and Reduced
Gigabit Media Independent Interface (RGMII) interfaces. It adopts an
AHB bus interface and integrates a link list DMA engine with direct
M-Bus accesses for transmitting and receiving packets. It has
independent TX/RX fifos, supports half and full duplex (1000 Mbps mode
only supports full duplex), flow control for full duplex and
backpressure for half duplex.
The FTGMAC100 also implements IP, TCP, UDP checksum offloads and
supports IEEE 802.1Q VLAN tag insertion and removal. It offers
high-priority transmit queue for QoS and CoS applications
This model is backed with a RealTek 8211E PHY which is the chip found
on the AST2500 EVB. It is complete enough to satisfy two different
Linux drivers and a U-Boot driver. Not supported features are :
This adds comments on the Basic mode control and status registers bit
definitions. It also adds a couple of bits for 1000BASE-T and the
RealTek 8211E PHY for the FTGMAC100 model to use.
* remotes/juanquintela/tags/migration/20170421: (65 commits)
hmp: info migrate_parameters format tunes
hmp: info migrate_capability format tunes
migration: rename max_size to threshold_size
migration: set current_active_state once
virtio-rng: stop virtqueue while the CPU is stopped
migration: don't close a file descriptor while it can be in use
ram: Remove migration_bitmap_extend()
migration: Disable hotplug/unplug during migration
qdev: Move qdev_unplug() to qdev-monitor.c
qdev: Export qdev_hot_removed
qdev: qdev_hotplug is really a bool
migration: Remove MigrationState parameter from migration_is_idle()
ram: Use RAMBitmap type for coherence
ram: rename last_ram_offset() last_ram_pages()
ram: Use ramblock and page offset instead of absolute offset
ram: Change offset field in PageSearchStatus to page
ram: Remember last_page instead of last_offset
ram: Use page number instead of an address for the bitmap operations
ram: reorganize last_sent_block
ram: ram_discard_range() don't use the mis parameter
...
* remotes/stefanha/tags/tracing-pull-request:
simpletrace: document Analyzer method signatures
trace: Put all trace.o into libqemuutil.a
configure: eliminate Python dependency for --help
* remotes/stefanha/tags/block-pull-request:
MAINTAINERS: update my email address
MAINTAINERS: update Wen's email address
migration/block: use blk_pwrite_zeroes for each zero cluster
throttle: make throttle_config(throttle_get_config()) symmetric
throttle: do not use invalid config in test
qemu-options: explain disk I/O throttling options
* remotes/cohuck/tags/s390x-20170421:
s390x: Drop useless casts
s390x: register I/O adapters per ISC during init
s390x/flic: cache flic in s390_get_flic
s390x: initialize flic before I/O subsystems
s390x: use enum for adapter type and standardize its naming
s390x/css: consolidate the devno property for ccw devices
s390x/css: provide introspection for virtual subchannel and device busid
s390x/css: introduce read-only property type for device ids
s390x/pci: make printf always compile in debug output
s390x/kvm: make printf always compile in debug output
s390x: introduce 2.10 compat machine
Peter Xu [Sat, 1 Apr 2017 08:18:43 +0000 (16:18 +0800)]
migration: rename max_size to threshold_size
In migration codes (especially in migration_thread()), max_size is used
in many place for the threshold value that we will start to do the final
flush and jump to the next stage to dump the whole rest things to
destination. However its name is confusing to first readers. Let's
rename it to "threshold_size" when proper and add a comment for it. No
functional change is made.
virtio-rng: stop virtqueue while the CPU is stopped
If we modify the virtio-rng virqueue while the
vmstate is already migrated we can have some
inconsistencies between the virtqueue state and
the memory content.
To avoid this, stop the virtqueue while the CPU
is stopped.
projects / qemu.git / log