hw/arm/virt: allow pmu instantiation with userspace irqchip

Move the in-kernel-irqchip test to only guard the set-irq
stage, not the init stage of the PMU.  Also add the PMU to
the KVM device irq line synchronization to enable its use.

Signed-off-by: Andrew Jones <drjones@redhat.com>
Reviewed-by: Christoffer Dall <cdall@linaro.org>
Message-id: 1500471597-2517-4-git-send-email-drjones@redhat.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

target/arm/kvm: pmu: split init and set-irq stages

When adding a PMU with a userspace irqchip we skip the set-irq
stage of device creation. Split the 'create' function into two
functions 'init' and 'set-irq' so they may be called separately.

Signed-off-by: Andrew Jones <drjones@redhat.com>
Reviewed-by: Christoffer Dall <cdall@linaro.org>
Message-id: 1500471597-2517-3-git-send-email-drjones@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/arm/virt: add pmu interrupt state

Mimicking gicv3-maintenance-interrupt, add the PMU's interrupt to
CPU state.

Signed-off-by: Andrew Jones <drjones@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1500471597-2517-2-git-send-email-drjones@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/arm: use defined type name instead of hard-coded string

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

loader: Ignore zero-sized ELF segments

Some ELF files have program headers that specify segments that
are of zero size. Ignore them, rather than trying to create
zero-length ROM blobs for them, because the zero-length blob
can falsely trigger the overlapping-ROM-blobs check.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Tested-by: Hua Yanghao <huayanghao@gmail.com>
Message-id: 1502116754-18867-3-git-send-email-peter.maydell@linaro.org

loader: Handle ELF files with overlapping zero-initialized data

For embedded systems, notably ARM, one common use of ELF
file segments is that the 'physical addresses' represent load addresses
and the 'virtual addresses' execution addresses, such that
the load addresses are packed into ROM or flash, and the
relocation and zero-initialization of data is done at runtime.
This means that the 'memsz' in the segment header represents
the runtime size of the segment, but the size that needs to
be loaded is only the 'filesz'. In particular, paddr+memsz
may overlap with the next segment to be loaded, as in this
example:

0x70000001 off    0x00007f68 vaddr 0x00008150 paddr 0x00008150 align 2**2
         filesz 0x00000008 memsz 0x00000008 flags r--
    LOAD off    0x000000f4 vaddr 0x00000000 paddr 0x00000000 align 2**2
         filesz 0x00000124 memsz 0x00000124 flags r--
    LOAD off    0x00000218 vaddr 0x00000400 paddr 0x00000400 align 2**3
         filesz 0x00007d58 memsz 0x00007d58 flags r-x
    LOAD off    0x00007f70 vaddr 0x20000140 paddr 0x00008158 align 2**3
         filesz 0x00000a80 memsz 0x000022f8 flags rw-
    LOAD off    0x000089f0 vaddr 0x20002438 paddr 0x00008bd8 align 2**0
         filesz 0x00000000 memsz 0x00004000 flags rw-
    LOAD off    0x000089f0 vaddr 0x20000000 paddr 0x20000000 align 2**0
         filesz 0x00000000 memsz 0x00000140 flags rw-

where the segment at paddr 0x8158 has a memsz of 0x2258 and
would overlap with the segment at paddr 0x8bd8 if QEMU's loader
tried to honour it. (At runtime the segments will not overlap
since their vaddrs are more widely spaced than their paddrs.)

Currently if you try to load an ELF file like this with QEMU then
it will fail with an error "rom: requested regions overlap",
because we create a ROM image for each segment using the memsz
as the size.

Support ELF files using this scheme, by truncating the
zero-initialized part of the segment if it would overlap another
segment. This will retain the existing loader behaviour for
all ELF files we currently accept, and also accept ELF files
which only need 'filesz' bytes to be loaded.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 1502116754-18867-2-git-send-email-peter.maydell@linaro.org

nvic: Implement "user accesses BusFault" SCS region behaviour

The ARMv7M architecture specifies that most of the addresses in the
PPB region (which includes the NVIC, systick and system registers)
are not accessible to unprivileged accesses, which should
BusFault with a few exceptions:
 * the STIR is configurably user-accessible
 * the ITM (which we don't implement at all) is always
   user-accessible

Implement this by switching the register access functions
to the _with_attrs scheme that lets us distinguish user
mode accesses.

This allows us to pull the handling of the CCR.USERSETMPEND
flag up to the level where we can make it generate a BusFault
as it should for non-permitted accesses.

Note that until the core ARM CPU code implements turning
MEMTX_ERROR into a BusFault the registers will continue to
act as RAZ/WI to user accesses.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 1501692241-23310-16-git-send-email-peter.maydell@linaro.org

armv7m_nvic.h: Move from include/hw/arm to include/hw/intc

The armv7m_nvic.h header file was accidentally placed in
include/hw/arm; move it to include/hw/intc to match where
its corresponding .c file lives.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 1501692241-23310-15-git-send-email-peter.maydell@linaro.org

target/arm: Create and use new function arm_v7m_is_handler_mode()

Add a utility function for testing whether the CPU is in Handler
mode; this is just a check whether v7m.exception is non-zero, but
we do it in several places and it makes the code a bit easier
to read to not have to mentally figure out what the test is testing.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 1501692241-23310-14-git-send-email-peter.maydell@linaro.org

target/arm: Don't calculate lr in arm_v7m_cpu_do_interrupt() until needed

Move the code in arm_v7m_cpu_do_interrupt() that calculates the
magic LR value down to when we're actually going to use it.
Having the calculation and use so far apart makes the code
a little harder to understand than it needs to be.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 1501692241-23310-13-git-send-email-peter.maydell@linaro.org

target/arm: Make arm_cpu_dump_state() handle the M-profile XPSR

Make the arm_cpu_dump_state() debug logging handle the M-profile XPSR
rather than assuming it's an A-profile CPSR.  On M profile the PSR
line of a register dump will now look like this:

XPSR=41000000 -Z-- T priv-thread

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 1501692241-23310-12-git-send-email-peter.maydell@linaro.org

target/arm: Don't use cpsr_write/cpsr_read to transfer M profile XPSR

For M profile the XPSR is a similar but not identical format to the
A profile CPSR/SPSR. (For instance the Thumb bit is in a different
place.) For guest accesses we make the M profile code go through
xpsr_read() and xpsr_write() which handle the different layout.
However for migration we use cpsr_read() and cpsr_write() to
marshal state into and out of the migration data stream. This
is pretty confusing and works more by luck than anything else.
Make M profile migration use xpsr_read() and xpsr_write() instead.

The most complicated part of this is handling the possibility
that the migration source is an older QEMU which hands us a
CPSR format value; helpfully we can always tell the two apart.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 1501692241-23310-11-git-send-email-peter.maydell@linaro.org

target/arm: Don't store M profile PRIMASK and FAULTMASK in daif

We currently store the M profile CPU register state PRIMASK and
FAULTMASK in the daif field of the CPU state in its I and F
bits. This is a legacy from the original implementation, which
tried to share the cpu_exec_interrupt code between A profile
and M profile. We've since separated out the two cases because
they are significantly different, so now there is no common
code between M and A profile which looks at env->daif: all the
uses are either in A-only or M-only code paths. Sharing the state
fields now is just confusing, and will make things awkward
when we implement v8M, where the PRIMASK and FAULTMASK
registers are banked between security states.

Switch M profile over to using v7m.faultmask and v7m.primask
fields for these registers.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 1501692241-23310-10-git-send-email-peter.maydell@linaro.org

target/arm: Define and use XPSR bit masks

The M profile XPSR is almost the same format as the A profile CPSR,
but not quite. Define some XPSR_* macros and use them where we
definitely dealing with an XPSR rather than reusing the CPSR ones.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 1501692241-23310-9-git-send-email-peter.maydell@linaro.org

target/arm: Fix outdated comment about exception exit

When we switched our handling of exception exit to detect
the magic addresses at translate time rather than via
a do_unassigned_access hook, we forgot to update a
comment; correct the omission.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 1501692241-23310-8-git-send-email-peter.maydell@linaro.org

target/arm: Remove incorrect comment about MPU_CTRL

Remove the comment that claims that some MPU_CTRL bits are stored
in sctlr_el[1]. This has never been true since MPU_CTRL was added
in commit 29c483a50607 -- the comment is a leftover from
Michael Davidsaver's original implementation, which I modified
not to use sctlr_el[1]; I forgot to delete the comment then.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 1501692241-23310-7-git-send-email-peter.maydell@linaro.org

hw/intc/armv7m_nvic.c: Remove out of date comment

Remove an out of date comment which says there's only one
item in the NVIC container region -- we put systick into its
own device object a while back and so now there are two
things in the container.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 1501692241-23310-6-git-send-email-peter.maydell@linaro.org

target/arm: Tighten up Thumb decode where new v8M insns will be

Tighten up the T32 decoder in the places where new v8M instructions
will be:
 * TT/TTT/TTA/TTAT are in what was nominally LDREX/STREX r15, ...
   which is UNPREDICTABLE:
   make the UNPREDICTABLE behaviour be to UNDEF
 * BXNS/BLXNS are distinguished from BX/BLX via the low 3 bits,
   which in previous architectural versions are SBZ:
   enforce the SBZ via UNDEF rather than ignoring it, and move
   the "ARCH(5)" UNDEF case up so we don't leak a TCG temporary
 * SG is in the encoding which would be LDRD/STRD with rn = r15;
   this is UNPREDICTABLE and we currently UNDEF:
   move this check further up the code so that we don't leak
   TCG temporaries in the UNDEF case and have a better place
   to put the SG decode.

This means that if a v8M binary is accidentally run on v7M
or if a test case hits something that we haven't implemented
yet the behaviour will be obvious (UNDEF) rather than obscure
(plough on treating it as a different instruction).

In the process, add some comments about the instruction patterns
at these points in the decode. Our Thumb and ARM decoders are
very difficult to understand currently, but gradually adding
comments like this should help to clarify what exactly has
been decoded when.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 1501692241-23310-5-git-send-email-peter.maydell@linaro.org

target/arm: Consolidate PMSA handling in get_phys_addr()

Currently get_phys_addr() has PMSAv7 handling before the
"is translation disabled?" check, and then PMSAv5 after it.
Tidy this up by making the PMSAv5 code handle the "MPU disabled"
case itself, so that we have all the PMSA code in one place.
This will make adding the PMSAv8 code slightly cleaner, and
also means that pre-v7 PMSA cores benefit from the MPU lookup
logging that the PMSAv7 codepath had.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 1501692241-23310-4-git-send-email-peter.maydell@linaro.org

target/arm: Don't trap WFI/WFE for M profile

M profile cores can never trap on WFI or WFE instructions. Check for
M profile in check_wfx_trap() to ensure this.

The existing code will do the right thing for v7M cores because
the hcr_el2 and scr_el3 registers will be all-zeroes and so we
won't attempt to trap, but when we start setting ARM_FEATURE_V8
for v8M cores the v8A handling of SCTLR.nTWE and .nTWI will not
give the right results.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 1501692241-23310-3-git-send-email-peter.maydell@linaro.org

target/arm: Use MMUAccessType enum rather than int

In the ARM get_phys_addr() code, switch to using the MMUAccessType
enum and its MMU_* values rather than int and literal 0/1/2.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 1501692241-23310-2-git-send-email-peter.maydell@linaro.org

Merge remote-tracking branch 'remotes/armbru/tags/pull-qapi-2017-09-01-v3' into staging

QAPI patches for 2017-09-01

# gpg: Signature made Mon 04 Sep 2017 12:30:31 BST
# gpg:                using RSA key 0x3870B400EB918653
# gpg: Good signature from "Markus Armbruster <armbru@redhat.com>"
# gpg:                 aka "Markus Armbruster <armbru@pond.sub.org>"
# Primary key fingerprint: 354B C8B3 D7EB 2A6B 6867  4E5F 3870 B400 EB91 8653

* remotes/armbru/tags/pull-qapi-2017-09-01-v3: (47 commits)
  qapi: drop the sentinel in enum array
  qapi: Change data type of the FOO_lookup generated for enum FOO
  qapi: Convert indirect uses of FOO_lookup[...] to qapi_enum_lookup()
  qapi: Mechanically convert FOO_lookup[...] to FOO_str(...)
  qapi: Generate FOO_str() macro for QAPI enum FOO
  qapi: Avoid unnecessary use of enum lookup table's sentinel
  qapi: Use qapi_enum_parse() in input_type_enum()
  crypto: Use qapi_enum_parse() in qcrypto_block_luks_name_lookup()
  quorum: Use qapi_enum_parse() in quorum_open()
  block: Use qemu_enum_parse() in blkdebug_debug_breakpoint()
  hmp: Use qapi_enum_parse() in hmp_migrate_set_parameter()
  hmp: Use qapi_enum_parse() in hmp_migrate_set_capability()
  tpm: Clean up model registration & lookup
  tpm: Clean up driver registration & lookup
  qapi: Drop superfluous qapi_enum_parse() parameter max
  qapi: Update qapi-code-gen.txt examples to match current code
  qapi-schema: Improve section headings
  qapi-schema: Move queries from common.json to qapi-schema.json
  qapi-schema: Make block-core.json self-contained
  qapi-schema: Fold event.json back into qapi-schema.json
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

qapi: drop the sentinel in enum array

Now that all usages have been converted to user lookup helpers.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20170822132255.23945-14-marcandre.lureau@redhat.com>
[Rebased, superfluous local variable dropped, missing
check-qom-proplist.c update added]
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503564371-26090-17-git-send-email-armbru@redhat.com>

qapi: Change data type of the FOO_lookup generated for enum FOO

Currently, a FOO_lookup is an array of strings terminated by a NULL
sentinel.

A future patch will generate enums with "holes".  NULL-termination
will cease to work then.

To prepare for that, store the length in the FOO_lookup by wrapping it
in a struct and adding a member for the length.

The sentinel will be dropped next.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20170822132255.23945-13-marcandre.lureau@redhat.com>
[Basically redone]
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503564371-26090-16-git-send-email-armbru@redhat.com>
[Rebased]

qapi: Convert indirect uses of FOO_lookup[...] to qapi_enum_lookup()

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503564371-26090-15-git-send-email-armbru@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>

qapi: Mechanically convert FOO_lookup[...] to FOO_str(...)

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503564371-26090-14-git-send-email-armbru@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>

qapi: Generate FOO_str() macro for QAPI enum FOO

The next commit will put it to use.  May look pointless now, but we're
going to change the FOO_lookup's type, and then it'll help.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503564371-26090-13-git-send-email-armbru@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>

qapi: Avoid unnecessary use of enum lookup table's sentinel

Currently, the FOO_lookup[] generated for QAPI enum types are
terminated by a NULL sentinel.

A future patch will generate enums with "holes".  NULL-termination
will cease to work then.

To prepare for that, replace "have we reached the sentinel?"
predicates by "have we reached the FOO__MAX value?" predicates.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503564371-26090-12-git-send-email-armbru@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>

qapi: Use qapi_enum_parse() in input_type_enum()

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503564371-26090-11-git-send-email-armbru@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>

crypto: Use qapi_enum_parse() in qcrypto_block_luks_name_lookup()

Cc: "Daniel P. Berrange" <berrange@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503564371-26090-10-git-send-email-armbru@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Acked-by: Daniel P. Berrange <berrange@redhat.com>

quorum: Use qapi_enum_parse() in quorum_open()

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20170822132255.23945-12-marcandre.lureau@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
[Rebased, qemu_opt_get() factored out, commit message tweaked]
Cc: Alberto Garcia <berto@igalia.com>
Reviewed-by: Alberto Garcia <berto@igalia.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503564371-26090-9-git-send-email-armbru@redhat.com>

block: Use qemu_enum_parse() in blkdebug_debug_breakpoint()

The error message on invalid blkdebug events changes from

    qemu-system-x86_64: LOCATION: Invalid event name "VALUE"

to

    qemu-system-x86_64: LOCATION: invalid parameter value: VALUE

Slight degradation, but the message is sub-par even before the patch.
When complaining about a parameter value, both parameter name and
value should be mentioned, as the value may well not be unique.  Left
for another day.

Also left is the error message's unhelpful location: it points to the
config=FILENAME rather than into that file.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20170822132255.23945-11-marcandre.lureau@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
[Rebased, commit message rewritten]
Cc: Kevin Wolf <kwolf@redhat.com>
Cc: Max Reitz <mreitz@redhat.com>
Cc: qemu-block@nongnu.org
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503564371-26090-8-git-send-email-armbru@redhat.com>

hmp: Use qapi_enum_parse() in hmp_migrate_set_parameter()

The error message on invalid parameter name changes from

    Invalid parameter "NAME"

to

    invalid parameter value: NAME

Slight degratation, perhaps.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20170822132255.23945-10-marcandre.lureau@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
[Rebased, assertion added, commit message rewritten]
Cc: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503564371-26090-7-git-send-email-armbru@redhat.com>

hmp: Use qapi_enum_parse() in hmp_migrate_set_capability()

The error message on invalid capability name changes from

    Invalid parameter "NAME"

to

    invalid parameter value: NAME

No worse than before.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20170822132255.23945-9-marcandre.lureau@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
[Rebased, commit message rewritten]
Cc: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503564371-26090-6-git-send-email-armbru@redhat.com>

tpm: Clean up model registration & lookup

We have a strict separation between enum TpmModel and tpm_models[]:

* TpmModel may have any number of members.  It just happens to have one.

* tpm_register_model() uses the first empty slot in tpm_models[].

  If you register more than tpm_models[] has space,
  tpn_register_model() fails.  Its caller silently ignores the
  failure.

  Register the same TpmModel more than once has no effect other than
  wasting tpm_models[] slots: tpm_model_is_registered() is happy with
  the first one it finds.

Since we only ever register one model, and tpm_models[] has space for
just that one, this contraption even works.

Turn tpm_models[] into a straight map from enum TpmType to bool.  Much
simpler.

Cc: Stefan Berger <stefanb@us.ibm.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503564371-26090-5-git-send-email-armbru@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
[Commit message typo fixed]

tpm: Clean up driver registration & lookup

We have a strict separation between enum TpmType and be_drivers[]:

* TpmType may have any number of members.  It just happens to have one.

* tpm_register_driver() uses the first empty slot in be_drivers[].

  If you register more than tpm_models[] has space,
  tpm_register_driver() fails.  Its caller silently ignores the
  failure.

  If you register more than one with a given TpmType,
  tpm_display_backend_drivers() will shows all of them, but
  tpm_driver_find_by_type() and tpm_get_backend_driver() will find
  only the one one that registered first.

Since we only ever register one driver, and be_drivers[] has space for
just that one, this contraption even works.

Turn be_drivers[] into a straight map from enum TpmType to driver.
Much simpler, and has a decent chance to actually work should we ever
acquire additional drivers.

While there, use qapi_enum_parse() in tpm_get_backend_driver().

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20170822132255.23945-8-marcandre.lureau@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
[Rebased, superfluous initializer dropped, commit message rewritten]
Cc: Stefan Berger <stefanb@us.ibm.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503564371-26090-4-git-send-email-armbru@redhat.com>

qapi: Drop superfluous qapi_enum_parse() parameter max

The lookup tables have a sentinel, no need to make callers pass their
size.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503564371-26090-3-git-send-email-armbru@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
[Rebased, commit message corrected]

qapi: Update qapi-code-gen.txt examples to match current code

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503564371-26090-2-git-send-email-armbru@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>

qapi-schema: Improve section headings

The generated QEMU QMP reference is now structured as follows:

    1.1 Introduction
    1.2 Stability Considerations
    1.3 Common data types
    1.4 Socket data types
    1.5 VM run state
    1.6 Cryptography
    1.7 Block devices
    1.7.1 Block core (VM unrelated)
    1.7.2 QAPI block definitions (vm unrelated)
    1.8 Character devices
    1.9 Net devices
    1.10 Rocker switch device
    1.11 TPM (trusted platform module) devices
    1.12 Remote desktop
    1.12.1 Spice
    1.12.2 VNC
    1.13 Input
    1.14 Migration
    1.15 Transactions
    1.16 Tracing
    1.17 QMP introspection
    1.18 Miscellanea

Section "1.18 Miscellanea" is still too big: it documents 134 symbols.
Section "1.7.1 Block core (VM unrelated)" is also rather big: 128
symbols.  All the others are of reasonable size.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503602048-12268-17-git-send-email-armbru@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>

qapi-schema: Move queries from common.json to qapi-schema.json

query-version and query-commands are in common.json for no good
reason.  Several similar queries are in qapi-schema.json.  Move them
there.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503602048-12268-16-git-send-email-armbru@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>

qapi-schema: Make block-core.json self-contained

Except for block-core.json, the sub-schemas are self-contained: if
they use a symbol defined in another sub-schema, they include that
sub-schema.  To check, feed the sub-schema to qapi2texi (or any other
QAPI generator) along with the pragma from qapi-schema.json.

Fix up things to make block-core.json self-contained, too.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503602048-12268-15-git-send-email-armbru@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>

qapi-schema: Fold event.json back into qapi-schema.json

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503602048-12268-14-git-send-email-armbru@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>

qapi-schema: Move block events from event.json to block.json

Cc: Alberto Garcia <berto@igalia.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503602048-12268-13-git-send-email-armbru@redhat.com>
Reviewed-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>

qapi-schema: Collect TPM stuff in qapi/tpm.json

Sadly, we don't have a TPM maintainer, not even a MAINTAINERS entry.
Create one, and mark it orphaned.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503602048-12268-12-git-send-email-armbru@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>

qapi-schema: Collect transaction stuff in qapi/transaction.json

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503602048-12268-11-git-send-email-armbru@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>

qapi-schema: Collect migration stuff in qapi/migration.json

Cc: Juan Quintela <quintela@redhat.com>
Cc: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503602048-12268-10-git-send-email-armbru@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>

qapi-schema: Collect UI stuff in qapi/ui.json

UI stuff is remote desktop stuff (Spice, VNC) and input stuff (mouse,
keyboard).

Cc: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503602048-12268-9-git-send-email-armbru@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Gerd Hoffmann <kraxel@redhat.com>

qapi-schema: Collect net device stuff in qapi/net.json

Cc: Jason Wang <jasowang@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503602048-12268-8-git-send-email-armbru@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>

qapi-schema: Collect char device stuff in qapi/char.json

Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Marc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503602048-12268-7-git-send-email-armbru@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>

qapi-schema: Collect run state stuff in qapi/run-state.json

Cc: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503602048-12268-6-git-send-email-armbru@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>

qapi-schema: Collect sockets stuff in qapi/sockets.json

Cc: "Daniel P. Berrange" <berrange@redhat.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503602048-12268-5-git-send-email-armbru@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>

qapi-schema: Rocker doc section contains unrelated stuff, fix

Bug: section "Rocker switch device" starts with the rocker stuff, but
then has unrelated stuff, like ReplayMode, xen-load-devices-state, ...

Cause: rocker.json is included in the middle of section "QMP commands".

Fix: include it in a sane place, namely next to the other sub-schemas.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <1503602048-12268-4-git-send-email-armbru@redhat.com>

qapi-schema: Introspection doc is in the wrong section, fix

Bug: introspection documentation is in section "Tracing commands".

Cause: sub-schema qapi/introspect.json lacks a section header, and
therefore goes into whatever section precedes its include.

Fix: add a section header.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <1503602048-12268-3-git-send-email-armbru@redhat.com>

qapi-schema: Document how generated documentation is ordered

Documentation generated with qapi2texi.py is in source order, with
included sub-schemas inserted at the first include directive
(subsequent include directives have no effect).  To get a sane and
stable order, it's best to include each sub-schema just once, or
include it first in qapi-schema.json.  Document that.

While there, drop a few redundant comments.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <1503602048-12268-2-git-send-email-armbru@redhat.com>

qlit: Tighten QLit list vs QList comparison

We check that all members of the QLit list are also in the QList.  We
neglect to check the other direction.  Fix that.

While there, use QLIST_FOREACH_ENTRY() to simplify the code and break
the loop on the first mismatch.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20170825105913.4060-13-marcandre.lureau@redhat.com>
[Commit message improved]
Signed-off-by: Markus Armbruster <armbru@redhat.com>

qlit: Tighten QLit dict vs QDict comparison

We check that all members of the QLit dictionary are also in the
QDict.  We neglect to check the other direction.

Comparing the number of members suffices, because QDict can't
contain duplicate members, and putting duplicates in a QLit is a
programming error.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20170825105913.4060-12-marcandre.lureau@redhat.com>
[Commit message improved]
Signed-off-by: Markus Armbruster <armbru@redhat.com>

tests/check-qlit: New, covering qobject/qlit.c

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20170825105913.4060-11-marcandre.lureau@redhat.com>
[Copyright notice correction squashed in, commit message tweaked]
Signed-off-by: Markus Armbruster <armbru@redhat.com>

qlit: Replace open-coded qnum_get_int() by call

Bonus: rids us of a side effect in an assertion.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20170825105913.4060-10-marcandre.lureau@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>

qlit: add QLIT_QNULL and QLIT_BOOL

As they are going to be used in the following patches.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20170825105913.4060-9-marcandre.lureau@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>

qlit: make qlit_equal_qobject() take const arguments

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20170825105913.4060-8-marcandre.lureau@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>

qlit: make qlit_equal_qobject return a bool

Make it more obvious about the expected return values.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20170825105913.4060-7-marcandre.lureau@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>

qlit: rename compare_litqobj_to_qobj() to qlit_equal_qobject()

compare_litqobj_to_qobj() lacks a qlit_ prefix.  Moreover, "compare"
suggests -1, 0, +1 for less than, equal and greater than.  The
function actually returns non-zero for equal, zero for unequal.
Rename to qlit_equal_qobject().

Its return type will be cleaned up in the next patch.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20170825105913.4060-6-marcandre.lureau@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>

qlit: Change compound literals to initializers

The QLIT_QFOO() macros expand into compound literals.  Sadly, gcc
doesn't recognizes these as constant expressions (clang does), which
makes the macros useless for initializing objects with static storage
duration.

There is a gcc bug about it:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71713

Change the macros to expand into initializers.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20170825105913.4060-5-marcandre.lureau@redhat.com>
[Commit message improved]
Signed-off-by: Markus Armbruster <armbru@redhat.com>

qlit: use QLit prefix consistently

Rename from LiteralQ to QLit.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20170825105913.4060-4-marcandre.lureau@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>

qlit: move qlit from check-qjson to qobject/

Fix code style issues while at it, to please checkpatch.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20170825105913.4060-3-marcandre.lureau@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>

qdict: Add qdict_put_null() helper, and put it to use

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20170825105913.4060-2-marcandre.lureau@redhat.com>
[Update to qobject.cocci squashed in, commit message tweaked]
Signed-off-by: Markus Armbruster <armbru@redhat.com>

qobject: Explain how QNum works, and why

Suggested-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1503384739-17207-1-git-send-email-armbru@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
[Comment typos fixed]

tests/qmp-test: Add generic, basic test of query commands

A command is a query if it has no side effect and yields a result.
Such commands are typically named query-FOO, but there are exceptions.

The basic idea is to find candidates with query-qmp-schema, filter out
the ones that aren't queries with an explicit blacklist, and test the
remaining ones against a QEMU with no special arguments.

The current blacklist is just add-fd.

The test can't do queries with arguments, because it knows nothing
about the arguments.  No coverage for query-cpu-model-baseline,
query-cpu-model-comparison, query-cpu-model-expansion, query-rocker,
query-rocker-ports, query-rocker-of-dpa-flows, and
query-rocker-of-dpa-groups.

Most tested commands are expected to succeed.  The test does not check
the return value then.

query-balloon and query-vm-generation-id are expected to fail because
they need a virtio-balloon / vmgenid device to succeed, and this test
is too dumb to set one up.  Could be addressed later.

query-acpi-ospm-status and query-hotpluggable-cpus are expected to
fail because they require features provided only by special machine
types, and this test is too dumb to set that up.  Could also be
addressed later.

Several commands may either be functional or stubs that always fail,
depending on build configuration.  Ideally, the stubs shouldn't be in
query-qmp-schema, but that requires QAPI schema compile-time
configuration, which we don't have, yet.  Until we do, we need to
figure out whether a command is a stub.  When we have a suitable
CONFIG_FOO preprocessor symbol is available, use that.  Else,
simply blacklist the command for now.

We get basic test coverage for the following commands, except as
noted:

    qom-list-types
    query-acpi-ospm-status      (expected to fail)
    query-balloon               (expected to fail)
    query-block
    query-block-jobs
    query-blockstats
    query-chardev
    query-chardev-backends
    query-command-line-options
    query-commands
    query-cpu-definitions       (blacklisted for now)
    query-cpus
    query-dump
    query-dump-guest-memory-capability
    query-events
    query-fdsets
    query-gic-capabilities      (blacklisted for now)
    query-hotpluggable-cpus     (expected to fail)
    query-iothreads
    query-kvm
    query-machines
    query-memdev
    query-memory-devices
    query-mice
    query-migrate
    query-migrate-cache-size
    query-migrate-capabilities
    query-migrate-parameters
    query-name
    query-named-block-nodes
    query-pci                   (blacklisted for now)
    query-qmp-schema
    query-rx-filter
    query-spice
    query-status
    query-target
    query-tpm
    query-tpm-models
    query-tpm-types
    query-uuid
    query-version
    query-vm-generation-id      (expected to fail)
    query-vnc
    query-vnc-servers
    query-xen-replication-status

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <1502461148-10154-1-git-send-email-armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
[Typos in code under #ifndef and in the commit message fixed]

Merge remote-tracking branch 'remotes/ehabkost/tags/x86-and-machine-pull-request' into staging

x86, cpu queue, 2017-09-01

CPU creation refactor plus x86 patches.

# gpg: Signature made Fri 01 Sep 2017 15:59:02 BST
# gpg:                using RSA key 0x2807936F984DC5A6
# gpg: Good signature from "Eduardo Habkost <ehabkost@redhat.com>"
# Primary key fingerprint: 5A32 2FD5 ABC4 D3DB ACCF  D1AA 2807 936F 984D C5A6

* remotes/ehabkost/tags/x86-and-machine-pull-request: (29 commits)
  ppc: replace cpu_ppc_init() with cpu_generic_init()
  unicore32: replace uc32_cpu_init() with cpu_generic_init()
  openrisc: replace cpu_openrisc_init() with cpu_generic_init()
  moxie: replace cpu_moxie_init() with cpu_generic_init()
  lm32: replace cpu_lm32_init() with cpu_generic_init()
  x86: replace cpu_x86_init() with cpu_generic_init()
  cris: replace cpu_cris_init() with cpu_generic_init()
  arm: replace cpu_arm_init() with cpu_generic_init()
  sh4: replace cpu_sh4_init() with cpu_generic_init()
  tricore: replace cpu_tricore_init() with cpu_generic_init()
  xtensa: replace cpu_xtensa_init() with cpu_generic_init()
  tilegx: replace cpu_tilegx_init() with cpu_generic_init()
  nios2: replace cpu_nios2_init() with cpu_generic_init()
  microblaze: replace cpu_mb_init() with cpu_generic_init()
  m68k: replace cpu_m68k_init() with cpu_generic_init()
  hppa: replace cpu_hppa_init() with cpu_generic_init()
  alpha: replace cpu_alpha_init() with cpu_generic_init()
  s390x: replace cpu_s390x_init() with cpu_generic_init()
  sparc: replace cpu_sparc_init() with cpu_generic_init()
  sparc: make cpu feature parsing property based
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

ppc: replace cpu_ppc_init() with cpu_generic_init()

it's just a wrapper, drop it and use cpu_generic_init() directly

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Acked-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Hervé Poussineau <hpoussin@reactos.org>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <1503592308-93913-26-git-send-email-imammedo@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

unicore32: replace uc32_cpu_init() with cpu_generic_init()

it's just a wrapper, drop it and use cpu_generic_init() directly

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <1503592308-93913-25-git-send-email-imammedo@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

openrisc: replace cpu_openrisc_init() with cpu_generic_init()

it's just a wrapper, drop it and use cpu_generic_init() directly

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Acked-by: Stafford Horne <shorne@gmail.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <1503592308-93913-24-git-send-email-imammedo@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

moxie: replace cpu_moxie_init() with cpu_generic_init()

it's just a wrapper, drop it and use cpu_generic_init() directly

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <1503592308-93913-23-git-send-email-imammedo@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

lm32: replace cpu_lm32_init() with cpu_generic_init()

it's just a wrapper, drop it and use cpu_generic_init() directly

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Acked-by: Michael Walle <michael@walle.cc>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <1503592308-93913-22-git-send-email-imammedo@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

x86: replace cpu_x86_init() with cpu_generic_init()

it's just a wrapper, drop it and use cpu_generic_init() directly

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <1503592308-93913-21-git-send-email-imammedo@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

cris: replace cpu_cris_init() with cpu_generic_init()

it's just a wrapper, drop it and use cpu_generic_init() directly

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <1503592308-93913-20-git-send-email-imammedo@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

arm: replace cpu_arm_init() with cpu_generic_init()

it's just a wrapper, drop it and use cpu_generic_init() directly

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <1503592308-93913-19-git-send-email-imammedo@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

sh4: replace cpu_sh4_init() with cpu_generic_init()

it's just a wrapper, drop it and use cpu_generic_init() directly

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <1503592308-93913-18-git-send-email-imammedo@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

tricore: replace cpu_tricore_init() with cpu_generic_init()

it's just a wrapper, drop it and use cpu_generic_init() directly

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <1503592308-93913-17-git-send-email-imammedo@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

xtensa: replace cpu_xtensa_init() with cpu_generic_init()

call xtensa_irq_init() at realize time which makes
cpu_xtensa_init() like generic cpu creation function.
As result we can replace it with cpu_generic_init()
which does the same job, reducing code duplication a bit.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <1503592308-93913-16-git-send-email-imammedo@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

tilegx: replace cpu_tilegx_init() with cpu_generic_init()

cpu_tilegx_init() always falls back to TYPE_TILEGX_CPU object
regardless of cpu_model. Put fallback logic into
tilegx_cpu_class_by_name() which would translate any cpu_model
into TYPE_TILEGX_CPU class and replace cpu_tilegx_init()
with cpu_generic_init().

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <1503592308-93913-15-git-send-email-imammedo@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

nios2: replace cpu_nios2_init() with cpu_generic_init()

cpu_nios2_init() always falls back to TYPE_NIOS2_CPU object
regardless of cpu_model. Put fallback logic into
nios2_cpu_class_by_name() which would translate any cpu_model
into TYPE_NIOS2_CPU class and replace cpu_nios2_init()
with cpu_generic_init()

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <1503592308-93913-14-git-send-email-imammedo@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

microblaze: replace cpu_mb_init() with cpu_generic_init()

cpu_mb_init() always falls back to TYPE_MICROBLAZE_CPU object
regardless of cpu_model. Put fallback logic into
mb_cpu_class_by_name() which would translate any cpu_model
into TYPE_MICROBLAZE_CPU class and replace cpu_mb_init()
with cpu_generic_init().

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <1503592308-93913-13-git-send-email-imammedo@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

m68k: replace cpu_m68k_init() with cpu_generic_init()

call register_m68k_insns() at realize time which makes
cpu_m68k_init() typical object creation function.
As result we can replace it with cpu_generic_init()
which does the same job, reducing code duplication a bit.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Tested-by: Thomas Huth <huth@tuxfamily.org>
Acked-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <1503592308-93913-12-git-send-email-imammedo@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

hppa: replace cpu_hppa_init() with cpu_generic_init()

drop custom cpu_hppa_init() in favor of cpu_generic_init(),
to make cpu_generic_init() work all we need is to provide
cc->class_by_name callback that would resolve any cpu_model
to the sole TYPE_HPPA_CPU to match current behaviour.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Acked-by: Richard Henderson <rth@twiddle.net>
Message-Id: <1503592308-93913-11-git-send-email-imammedo@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

alpha: replace cpu_alpha_init() with cpu_generic_init()

cpu_alpha_init() used to provide default fallback if invalid
(i.e. non existent) cpu_model were provided.

dp264 machine provides its own default so sole user of fallback
is [bsd|linux]-user targets which specifies 'any' cpu model that
fallbacks to "ev67" in cpu_alpha_init(). Push fallback handling
into alpha_cpu_class_by_name() and replace cpu_alpha_init() with
cpu_generic_init().

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Acked-by: Richard Henderson <rth@twiddle.net>
Message-Id: <1503592308-93913-10-git-send-email-imammedo@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

s390x: replace cpu_s390x_init() with cpu_generic_init()

cpu_s390x_init() is used only *-user targets indirectly
via cpu_init() macro and has a hack to assign ids to created
cpus (I'm not sure if 'id' really matters to *-user emulation).

So to on safe side, instead of having custom wrapper to do numbering
replace it with cpu_generic_init() and use S390CPUClass::next_cpu_id
which could serve the same purpose as static variable and move cpu->id
initialization to s390_cpu_initfn for CONFIG_USER_ONLY use-case.

PS:
ifdef is ugly but it allows us to hide s390x detail that isn't
set by *-user targets and reuse generic cpu creation utility
for btoh machine and user emulation.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Acked-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <1504185578-80843-1-git-send-email-imammedo@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

sparc: replace cpu_sparc_init() with cpu_generic_init()

it's just a wrapper, drop it and use cpu_generic_init() directly

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <1503592308-93913-8-git-send-email-imammedo@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

sparc: make cpu feature parsing property based

with features converted to properties we can use the same
approach as x86 for features parsing and drop legacy
approach that manipulated CPU instance directly.
New sparc_cpu_parse_features() will allow only +-feat
and explicitly disable feat=on|off syntax for now.

With that in place and sparc_cpu_parse_features() providing
generic CPUClass::parse_features callback, the cpu_sparc_init()
will do the same job as cpu_generic_init() so replace content
of cpu_sparc_init() with it.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Acked-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <1503672460-109436-1-git-send-email-imammedo@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

sparc: move adhoc CPUSPARCState initialization to realize time

SPARCCPU::env was initialized from previously set properties
(with help of sparc_cpu_parse_features) in cpu_sparc_register().
However there is not reason to keep it there as this task is
typically done at realize time. So move post properties
initialization into sparc_cpu_realizefn, which brings
cpu_sparc_init() closer to cpu_generic_init().

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <1503592308-93913-6-git-send-email-imammedo@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

sparc: convert cpu features to qdev properties

SPARC is the last target that uses legacy way of parsing
and initializing cpu features, drop legacy approach and
convert features to properties so that SPARC could as minimum
benefit from generic cpu_generic_init(), common with
x86 +-feat parser

PS:
the main purpose is to remove legacy way of cpu creation as
a blocker for unifying cpu creation code across targets.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <1503592308-93913-5-git-send-email-imammedo@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

sparc: embed sparc_def_t into CPUSPARCState

Make CPUSPARCState::def embedded so it would be allocated as part
of cpu instance and we won't have to worry about cleaning def pointer
up mannualy on cpu destruction.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <1503592308-93913-4-git-send-email-imammedo@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

sparc: convert cpu models to SPARC cpu subclasses

QOMfy cpu models handling introducing propper cpu types
for each cpu model.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <1503592308-93913-3-git-send-email-imammedo@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

qom: cpu: fix parsed feature string length

since commit ( 9262685b cpu: Factor out cpu_generic_init() )
features parsed by it were truncated only to the 1st feature
after CPU name due to fact that

   featurestr = strtok(NULL, ",");
   cc->parse_features(cpu, featurestr, &err);

would extract exactly one feature and parse_features() callback
would parse it and only it leaving the rest of features ignored.

Reuse approach from x86 custom impl. i.e. replace strtok() token
parsing with g_strsplit(), which would split feature string in
2 parts name and features list and pass the later to
parse_features() callback.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Message-Id: <1503592308-93913-2-git-send-email-imammedo@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

target-i386/cpu: Add new EPYC CPU model

Add a new base CPU model called 'EPYC' to model processors from AMD EPYC
family (which includes EPYC 76xx,75xx,74xx, 73xx and 72xx).

The following features bits have been added/removed compare to Opteron_G5

Added: monitor, movbe, rdrand, mmxext, ffxsr, rdtscp, cr8legacy, osvw,
       fsgsbase, bmi1, avx2, smep, bmi2, rdseed, adx, smap, clfshopt, sha
       xsaveopt, xsavec, xgetbv1, arat

Removed: xop, fma4, tbm

Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Richard Henderson <rth@twiddle.net>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Tom Lendacky <Thomas.Lendacky@amd.com>
Signed-off-by: Brijesh Singh <brijesh.singh@amd.com>
Message-Id: <20170815170051.127257-1-brijesh.singh@amd.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

hmp: allow apic-id for "info lapic"

Add [apic-id] support for hmp command "info lapic", which is
useful when debugging ipi and so on. Current behavior is not
changed when the parameter isn't specified.

Signed-off-by: Yi Wang <wang.yi59@zte.com.cn>
Signed-off-by: Yun Liu <liu.yunh@zte.com.cn>
Message-Id: <1501049917-4701-3-git-send-email-wang.yi59@zte.com.cn>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

cpu: cpu_by_arch_id() helper

The helper can be used for CPU object lookup using the CPU's
arch-specific ID (the one returned by CPUClass::get_arch_id()).

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
[Yi Wang: Added documentation comments]
Signed-off-by: Yi Wang <wang.yi59@zte.com.cn>
Signed-off-by: Yun Liu <liu.yunh@zte.com.cn>
[ehabkost: extracted cpu_by_arch_id() to a separate patch]
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

qom: Remove unused errp parameter from can_be_deleted()

The errp argument is ignored by all implementations of the
method, and user_creatable_del() would break if any
implementation set an error (because it calls error_setg(errp) if
the function returns false).  Remove the unused parameter.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <20170829220337.23427-1-ehabkost@redhat.com>
Reviewed-by: Gonglei <arei.gonglei@huawei.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

qapi: Fix error handling code on alternate conflict

The conflict check added by commit c0644771 ("qapi: Reject
alternates that can't work with keyval_parse()") doesn't work
with the following declaration:

  { 'alternate': 'Alt',
    'data': { 'one': 'bool',
              'two': 'str' } }

It crashes with:

  Traceback (most recent call last):
    File "./scripts/qapi-types.py", line 295, in <module>
      schema = QAPISchema(input_file)
    File "/home/ehabkost/rh/proj/virt/qemu/scripts/qapi.py", line 1468, in __init__
      self.exprs = check_exprs(parser.exprs)
    File "/home/ehabkost/rh/proj/virt/qemu/scripts/qapi.py", line 958, in check_exprs
      check_alternate(expr, info)
    File "/home/ehabkost/rh/proj/virt/qemu/scripts/qapi.py", line 830, in check_alternate
      % (name, key, types_seen[qtype]))
  KeyError: 'QTYPE_QSTRING'

This happens because the previously-seen conflicting member
('one') can't be found at types_seen[qtype], but at
types_seen['QTYPE_BOOL'].

Fix the bug by moving the error check to the same loop that adds
new items to types_seen, raising an exception if types_seen[qt]
is already set.

Add two additional test cases that can detect the bug.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <20170717180926.14924-1-ehabkost@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>

Merge remote-tracking branch 'remotes/elmarco/tags/tidy-pull-request' into staging

# gpg: Signature made Thu 31 Aug 2017 11:29:33 BST
# gpg:                using RSA key 0xDAE8E10975969CE5
# gpg: Good signature from "Marc-André Lureau <marcandre.lureau@redhat.com>"
# gpg:                 aka "Marc-André Lureau <marcandre.lureau@gmail.com>"
# 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: 87A9 BD93 3F87 C606 D276  F62D DAE8 E109 7596 9CE5

* remotes/elmarco/tags/tidy-pull-request: (29 commits)
  eepro100: replace g_malloc()+memcpy() with g_memdup()
  test-iov: replace g_malloc()+memcpy() with g_memdup()
  i386: replace g_malloc()+memcpy() with g_memdup()
  i386: introduce ELF_NOTE_SIZE macro
  decnumber: use DIV_ROUND_UP
  kvm: use DIV_ROUND_UP
  i386/dump: use DIV_ROUND_UP
  ppc: use DIV_ROUND_UP
  msix: use DIV_ROUND_UP
  usb-hub: use DIV_ROUND_UP
  q35: use DIV_ROUND_UP
  piix: use DIV_ROUND_UP
  virtio-serial: use DIV_ROUND_UP
  console: use DIV_ROUND_UP
  monitor: use DIV_ROUND_UP
  virtio-gpu: use DIV_ROUND_UP
  vga: use DIV_ROUND_UP
  ui: use DIV_ROUND_UP
  vnc: use DIV_ROUND_UP
  vvfat: use DIV_ROUND_UP
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>