Gerd Hoffmann [Thu, 10 May 2012 12:12:38 +0000 (14:12 +0200)]
ehci: add queuing support
Add packet queuing. Follow the qTD chain to see if there are more
packets we can submit. Improves performance on larger transfers,
especially with usb-host, as we don't have to wait for a packet to
finish before sending the next one to the host for processing.
Gerd Hoffmann [Wed, 9 May 2012 15:06:36 +0000 (17:06 +0200)]
ehci: add EHCIPacket
Add a separate EHCIPacket struct and move fields over from EHCIQueue.
Preparing for supporting multiple packets per queue being in flight at
the same time. No functional changes yet.
Gerd Hoffmann [Wed, 16 May 2012 12:20:03 +0000 (14:20 +0200)]
scsi: prepare migration code for usb-storage support
usb-storage can't handle requests in one go as the data transfer can be
splitted into lots of usb packets. Because of that there can be
normal in-flight requests at savevm time and we need to handle that.
With other scsi hba's this happens only in case i/o is stopped due to
errors and there are pending requests which need to be restarted
(req->retry = true).
So, first we need to save req->retry and then handle the req->retry =
false case. Write requests are handled fine already. For read requests
we have to save the buffer as we will not restart the request (and thus
not refill the buffer) on the target host.
Gerd Hoffmann [Fri, 25 May 2012 10:53:47 +0000 (12:53 +0200)]
uhci: fix irq routing
The multifunction ich9 ehci controller with uhci companions uses a
different interrupt pin for each function. The three uhci devices
get pins A, B and C, whereas ehci uses pin D. This way the guest
can assign different IRQ lines to each controller.
Gerd Hoffmann [Mon, 14 May 2012 12:02:14 +0000 (14:02 +0200)]
uhci: zap uhci_pre_save
Cancel transactions before saving vmstate is pretty pointless and just
causes disruptions. We need to cancel them before *loading* vmstate,
but in that case uhci_reset() handles it already and no special action
is needed.
Gerd Hoffmann [Fri, 11 May 2012 07:33:07 +0000 (09:33 +0200)]
uhci: use bottom half
Schedule bottom half on completion of async packets instead of calling
uhci_process_frame directly. This way we run uhci_process_frame only
once in case multiple packets finish in a row. Also check whenever
there is bandwidth left before scheduling uhci_process_frame.
Gerd Hoffmann [Fri, 11 May 2012 07:18:05 +0000 (09:18 +0200)]
uhci: fix bandwidth management
uhci_process_frame() can be invoked multiple times per frame, so
accounting usb bandwith in a local variable doesn't fly, use a variable
in UHCIState instead. Also check the limit more frequently.
Paolo Bonzini [Tue, 29 May 2012 09:49:34 +0000 (11:49 +0200)]
build: move per-target hw/ objects to nested Makefile.objs
This completes the move to nested Makefiles for virtio and a few
other files that were not part of obj-TARGET-y, but still were
compiled separately for each target.
Paolo Bonzini [Tue, 29 May 2012 09:10:55 +0000 (11:10 +0200)]
build: convert libhw to nested Makefile.objs
After this patch, the libhw* directories will have a hierarchy
that mimics the source tree. This is useful because we do have
a couple of files there that are in the top source directory.
Paolo Bonzini [Tue, 29 May 2012 09:08:47 +0000 (11:08 +0200)]
build: move target-independent hw/ objects to nested Makefile.objs
This patch starts converting the hw/ directory. Some files in hw/
are compiled once, some twice (32-/64-bit), some once per target.
Each category is moved in a separate patch.
After this patch, the files that are compiled once will show the
same hierarchy in the build tree as they do in the source tree,
for example hw/qdev.o instead of just qdev.o.
Paolo Bonzini [Tue, 22 May 2012 11:42:53 +0000 (13:42 +0200)]
build: adapt qom/Makefile and move it to Makefile.objs
qom/ already used a separate makefile. Convert it to use relative
paths, and make it declare both common-obj-y and user-obj-y. This
way, the upper makefiles do not need to know that some QOM files
are compiled twice.
Paolo Bonzini [Tue, 22 May 2012 11:41:27 +0000 (13:41 +0200)]
build: move rules for nesting to Makefile.objs
At this point we will start adding nesting behavior to other files
than Makefile.target. Because Makefile.objs is included by
Makefile.target, it is simpler to move the processing of
subdirectories there.
To enable this, only add per-target files to obj-y. Use a separate
variable for the linker dependencies, all-obj-y. This variable includes
obj-y and also all objects that are taken from other directories.
Paolo Bonzini [Thu, 7 Jun 2012 06:22:56 +0000 (08:22 +0200)]
dump: remove dumping stuff from cpu-all.h
This simplifies things, because they will only be included for softmmu
targets and because the stubs are taken out-of-line in separate files,
which in the future could even be compiled only once.
Paolo Bonzini [Tue, 22 May 2012 11:41:27 +0000 (13:41 +0200)]
build: add rules for nesting
This adds the 'magic' rules that take care of subdirectories.
The subdirectory makefiles in the source tree are not complete; they
only define some variables (listed in nested-vars) according to the
configuration.
The magic rules descend into subdirectory makefiles and gather the
evaluated values of those variables. The values from all subdirectories
are joined together, each prefixed with the subdirectory name, and used
by the "real" makefiles.
Paolo Bonzini [Tue, 29 May 2012 10:41:34 +0000 (12:41 +0200)]
build: do not sprinkle around GENERATED_HEADERS dependencies
Keeping GENERATED_HEADERS dependencies up-to-date everywhere is complex.
We can simply make the Makefile depend on them, and they will be built
before all other targets.
Anthony Liguori [Wed, 6 Jun 2012 12:57:56 +0000 (20:57 +0800)]
Merge remote-tracking branch 'qmp/queue/qmp' into staging
* qmp/queue/qmp: (29 commits)
Add 'query-events' command to QMP to query async events
qapi: convert netdev_del
qapi: convert netdev_add
net: net_client_init(): use error_set()
net: purge the monitor object from all init functions
qemu-config: introduce qemu_find_opts_err()
qemu-config: find_list(): use error_set()
qerror: introduce QERR_INVALID_OPTION_GROUP
qemu-option: qemu_opts_from_qdict(): use error_set()
qemu-option: introduce qemu_opt_set_err()
qemu-option: opt_set(): use error_set()
qemu-option: qemu_opts_validate(): use error_set()
qemu-option: qemu_opt_parse(): use error_set()
qemu-option: parse_option_size(): use error_set()
qemu-option: parse_option_bool(): use error_set()
qemu-option: parse_option_number(): use error_set()
qemu-option: qemu_opts_create(): use error_set()
introduce a new monitor command 'dump-guest-memory' to dump guest's memory
make gdb_id() generally avialable and rename it to cpu_index()
target-i386: Add API to get note's size
...
Anthony Liguori [Wed, 6 Jun 2012 09:55:07 +0000 (17:55 +0800)]
Merge remote-tracking branch 'afaerber-or/qom-cpu-3' into staging
* afaerber-or/qom-cpu-3: (74 commits)
Kill off cpu_state_reset()
linux-user: Use cpu_reset() after cpu_init() / cpu_copy()
bsd-user: Use cpu_reset() in after cpu_init()
leon3: Store SPARCCPU in ResetData
leon3: Use cpu_sparc_init() to obtain SPARCCPU
sun4u: Store SPARCCPU in ResetData
sun4u: Let cpu_devinit() return SPARCCPU
sun4u: Use cpu_sparc_init() to obtain SPARCCPU
sun4m: Pass SPARCCPU to {main,secondary}_cpu_reset()
sun4m: Use cpu_sparc_init() to obtain SPARCCPU
target-sparc: Let cpu_sparc_init() return SPARCCPU
cpu-exec: Use cpu_reset() in cpu_exec() for TARGET_PPC
virtex_ml507: Pass PowerPCCPU to main_cpu_reset()
virtex_ml507: Let ppc440_init_xilinx() return PowerPCCPU
virtex_ml507: Use cpu_ppc_init() to obtain PowerPCCPU
ppc_prep: Pass PowerPCCPU to ppc_prep_reset()
ppc_prep: Use cpu_ppc_init() to obtain PowerPCCPU
ppc_oldworld: Pass PowerPCCPU to ppc_heathrow_reset()
ppc_oldworld: Use cpu_ppc_init() to obtain PowerPCCPU
ppc_newworld: Pass PowerPCCPU to ppc_core99_reset()
...
Jan Kiszka [Tue, 5 Jun 2012 19:03:57 +0000 (21:03 +0200)]
kvm: Fix build for non-CAP_IRQ_ROUTING targets
A type definition and a KVMState field initialization escaped the
required wrapping with KVM_CAP_IRQ_ROUTING. Also, we need to provide a
dummy kvm_irqchip_release_virq as virtio-pci references (but does not
use) it.
Andreas Färber [Sat, 5 May 2012 13:43:31 +0000 (15:43 +0200)]
Kill off cpu_state_reset()
In commit 1bba0dc932e8826a7d030df3767daf0bc339f9a2 cpu_reset()
was renamed to cpu_state_reset(), to allow introducing a new cpu_reset()
that would operate on QOM objects.
All callers have been updated except for one in target-mips, so drop all
implementations except for the one in target-mips and move the
declaration there until MIPSCPU reset can be fully QOM'ified.
Andreas Färber [Sat, 5 May 2012 11:06:10 +0000 (13:06 +0200)]
target-mips: Use cpu_reset() in cpu_mips_init()
Commit 0f71a7095db6bc055bc5bb520d85ea650cca8a33 (target-mips: QOM'ify
CPU) hooked up cpu_state_reset() to CPUClass::reset(). Dropping the
introduction of subclasses for 1.1, due to mips_def_t the reset code
could not be QOM'ified yet, i.e. cpu_state_reset() will not forward to
CPUClass::reset().
Update cpu_mips_init() with cpu_reset() nontheless, indirectly calling
cpu_state_reset(), so that generic code can be converted to call
cpu_reset().
projects / qemu.git / log