Commit 8118f0950f "migration: Append JSON description of migration
stream" needs a JSON writer. The existing qobject_to_json() wasn't a
good fit, because it requires building a QObject to convert. Instead,
migration got its very own JSON writer, in commit 190c882ce2 "QJSON:
Add JSON writer". It tacitly limits numbers to int64_t, and strings
contents to characters that don't need escaping, unlike
qobject_to_json().
The previous commit factored the JSON writer out of qobject_to_json().
Replace migration's JSON writer by it.
Cc: Juan Quintela <quintela@redhat.com>
Cc: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201211171152.146877-17-armbru@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
We have two JSON writers written in C: qobject/qjson.c provides
qobject_to_json(), and migration/qjson.c provides a more low level
imperative interface. They don't share code. The latter tacitly
limits numbers to int64_t, and strings contents to characters that
don't need escaping.
Factor out qobject_to_json()'s JSON writer as qobject/json-writer.c.
Straightforward, except for numbers: since the writer is to be
independent of QObject, it can't use qnum_to_string(). Open-code it
instead. This is actually an improvement of sorts, because it
liberates qnum_to_string() from JSON's needs: its JSON-related FIXMEs
move to the JSON writer, where they belong.
The next commit will replace migration/qjson.c.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201211171152.146877-16-armbru@redhat.com>
No users left outside tests/, and the ones in tests/ can just as well
use qstring_get_str(). Do that, and drop the function.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201211171152.146877-14-armbru@redhat.com>
Commit aafb21a0b9 "qobject: let object_property_get_str() use new API"
isn't much of a simplification. Not worth having
object_property_get_str() differ from the other
object_property_get_FOO(). Revert.
This reverts commit aafb21a0b9.
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Daniel P. Berrangé <berrange@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201211171152.146877-12-armbru@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
I'm about to remove qobject_get_try_str(). Use qstring_get_str()
instead. Safe because the argument is known to be a QString here.
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: <20201211171152.146877-11-armbru@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
RdmaProtectedQList provides a thread-safe queue of int64_t on top of a
QList.
rdma_protected_qlist_destroy() calls qlist_destroy_obj() directly.
qlist_destroy_obj() is actually for use by qobject_destroy() only.
The next commit will make that obvious.
The minimal fix would be calling qobject_unref() instead. But QList
is actually a bad fit here. It's designed for representing JSON
arrays. We're better off with a GQueue here. Replace.
Cc: Yuval Shaia <yuval.shaia.ml@gmail.com>
Cc: Marcel Apfelbaum <marcel.apfelbaum@gmail.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201211171152.146877-8-armbru@redhat.com>
This reverts commit 164c374b75.
A free function for a reference-counted object is in bad taste.
Fortunately, this one is now also unused. Drop it.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201211171152.146877-7-armbru@redhat.com>
qobject_to_json() and qobject_to_json_pretty() build a GString, then
covert it to QString. Just one of the callers actually needs a
QString: qemu_rbd_parse_filename(). A few others need a string they
can modify: qmp_send_response(), qga's send_response(), to_json_str(),
and qmp_fd_vsend_fds(). The remainder just need a string.
Change qobject_to_json() and qobject_to_json_pretty() to return the
GString.
qemu_rbd_parse_filename() now has to convert to QString. All others
save a QString temporary. to_json_str() actually becomes a bit
simpler, because GString provides more convenient modification
functions.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201211171152.146877-6-armbru@redhat.com>
QString supports modifying its string, but it's quite limited: you can
only append. The remaining callers use it for building an initial
string, never for modifying it later.
Use of GString for building the initial string is actually more
convenient here. Change qobject_to_json() & friends to do that.
Once all such uses are replaced this way, QString can become immutable.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201211171152.146877-5-armbru@redhat.com>
GString has a richer set of string operations than QString. It should
be preferred to QString except where we need a QObject or reference
counting. We don't here. Switch to GString, and put its richer
interface to use.
Cc: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201211171152.146877-3-armbru@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Commit 48c043d0d1 "hmp: human-monitor-command: stop using the Memory
chardev driver" left us "if string is non-empty, duplicate it, else
duplicate the empty string". Meh. Duplicate it unconditionally.
Cc: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201211171152.146877-2-armbru@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
test_primitives() uses union member intmax_t max to compare the
integer members. Unspecified behavior. Has worked fine for many
years, though. Clean it up.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201210161452.2813491-11-armbru@redhat.com>
The string output visitor should serialize numbers so that the string
input visitor deserializes them back to the same number. It fails to
do so.
print_type_number() uses format %f. This is prone to nasty rounding
errors. For instance, numbers between 0 and 0.0000005 get flushed to
zero.
We currently use this visitor only for HMP info migrate, info network,
info qtree, and info memdev. No double values occur there as far as I
can tell.
Fix anyway by formatting with %.17g. 17 decimal digits always suffice
for IEEE double.
See also recent commit "qobject: Fix qnum_to_string() to use
sufficient precision".
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201210161452.2813491-9-armbru@redhat.com>
This demonstrates rounding error due to insufficient precision: double
3.1415926535897932 gets converted to JSON 3.141593.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201210161452.2813491-8-armbru@redhat.com>
We should serialize numbers to JSON so that they deserialize back to
the same number. We fail to do so.
The culprit is qnum_to_string(): it uses format %f with trailing '0'
trimmed. Results in pretty output for "nice" numbers, but is prone to
nasty rounding errors. For instance, numbers between 0 and 0.0000005
get flushed to zero.
Where exactly the incorrect rounding can bite is tiresome to gauge.
Here's my take.
* In QMP output, type 'number':
- query-blockstats value avg_rd_queue_depth
- QMP query-migrate values mbps, cache-miss-rate, encoding-rate,
busy-rate, compression-rate.
Relatively harmless, I guess.
* In tracing QMP input. Harmless.
* In qemu-ga output, type 'number': guest-get-users value login-time.
Harmless.
* In output of HMP qom-get. Harmless.
Not affected, because double values don't actually occur there (I
think):
* QMP output, type 'any':
* qom-get value
* qom-list, qom-list-properties value default-value
* query-cpu-model-comparison, query-cpu-model-baseline,
query-cpu-model-expansion value props.
* qemu-img --output json output.
* "json:" pseudo-filenames generated by bdrv_refresh_filename().
* The rbd block driver's "=keyvalue-pairs" hack.
* In -object help on property default values. Aside: use of JSON
feels inappropriate here.
* Output of HMP qom-get.
* Argument conversion to QemuOpts for qdev_device_add() and HMP with
qemu_opts_from_qdict()
QMP and HMP device_add, virtio-net failover primary creation,
xen-usb "usb-host" creation, HMP netdev_add, object_add.
* The uses of qobject_input_visitor_new_flat_confused()
As far as I can tell, none of the visited types contain double
values.
* Dumping ImageInfoSpecific with dump_qobject()
Fix by formatting with %.17g. 17 decimal digits always suffice for
IEEE double.
The change to expected test output illustrates the effect: the
rounding errors are gone, but some seemingly "nice" numbers now get
converted to not so nice strings, e.g. 0.42 to "0.41999999999999998".
This is because 0.42 is not representable exactly in double. It's
more accurate in this example than strictly necessary, though.
If ugly accuracy bothers us, we can we can try using the least number
of digits that still converts back to the same double. In this
example, "0.42" would do.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201210161452.2813491-7-armbru@redhat.com>
qnum_to_string() has a FIXME comment about rounding errors due to
insufficient precision. Cover it: 2.718281828459045 gets converted to
"2.718282". The next commit will fix it.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201210161452.2813491-6-armbru@redhat.com>
Move one of large_number()'s three checks to uint_number(), and the
other two to float_number().
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201210161452.2813491-5-armbru@redhat.com>
simple_number() checks only qnum_get_try_int(). Also check
qnum_get_try_uint() and qnum_get_double().
float_number() checks only qnum_get_double(). Also check
qnum_get_try_int() and qnum_get_try_uint().
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201210161452.2813491-3-armbru@redhat.com>
simple_number() and float_number() convert from JSON to QNumber and
back.
simple_number() tests "-0", but skips the conversion back to JSON,
because it yields "0", not "-0". Works as intended, so better cover
it: don't skip, but expect the funny result.
float_number() tests "-32.20e-10", but skips the conversion back to
JSON, because it yields "-0". This is a known bug in
qnum_to_string(), marked FIXME there. Cover the bug: don't skip, but
expect the funny result.
While there, switch from g_assert() to g_assert_cmpstr() & friends for
friendlier test failures.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201210161452.2813491-2-armbru@redhat.com>
Anywhere we create a list of just one item or by prepending items
(typically because order doesn't matter), we can use
QAPI_LIST_PREPEND(). But places where we must keep the list in order
by appending remain open-coded until later patches.
Note that as a side effect, this also performs a cleanup of two minor
issues in qga/commands-posix.c: the old code was performing
new = g_malloc0(sizeof(*ret));
which 1) is confusing because you have to verify whether 'new' and
'ret' are variables with the same type, and 2) would conflict with C++
compilation (not an actual problem for this file, but makes
copy-and-paste harder).
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <20201113011340.463563-5-eblake@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
[Straightforward conflicts due to commit a8aa94b5f8 "qga: update
schema for guest-get-disks 'dependents' field" and commit a10b453a52
"target/mips: Move mips_cpu_add_definition() from helper.c to cpu.c"
resolved. Commit message tweaked.]
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Instead of taking a list parameter and returning a new head at a
distance, just return the new item for the caller to insert into a
list via QAPI_LIST_PREPEND.
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <20201113011340.463563-4-eblake@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Instead of modifying the value member of a list element passed as a
parameter, and open-coding the manipulation of that list, it's nicer
to just return a freshly allocated value to be prepended to a list
using QAPI_LIST_PREPEND.
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <20201113011340.463563-3-eblake@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
- Improve the sifive_u DTB generation
- Add QSPI NOR flash to Microchip PFSoC
- Fix a bug in the Hypervisor HLVX/HLV/HSV instructions
- Fix some mstatus mask defines
- Ibex PLIC improvements
- OpenTitan memory layout update
- Initial steps towards support for 32-bit CPUs on 64-bit builds
-----BEGIN PGP SIGNATURE-----
iQEzBAABCAAdFiEE9sSsRtSTSGjTuM6PIeENKd+XcFQFAl/cRU4ACgkQIeENKd+X
cFTAPgf/dHOYiBeSZr0eg03LwpqiJ5ziVrvE9nvAjml8CsDvwlx6roEMT1Ynyquq
zs8sPb4a1Ro7rHBofHFqgHp8TO6wAiw2nDrT8YEt1iARO5Oh5IuHqs/wi8SNB2QF
d1Dv8/zIBOkK5+Fg/DQHTrPgq4fJZwY2jnVZAyUBuMW5UkvCVlJI4zGPwYyh+4ZS
xTWogMzSbyer3evfTg8f8AhvCGQMITwLo6Nxc4wj3bf1ZE8Br9UxorqPme4UwJ+r
Ip9/jXDlKI9BeE85XoOrQJNLR7OzLgdQ1S/LjeBYLQmsltOD49YcH6a6AX3YjDwW
Jj6GgXBTFGIUXbxc3ADpoMJQp+xDSA==
=Vj2m
-----END PGP SIGNATURE-----
Merge remote-tracking branch 'remotes/alistair/tags/pull-riscv-to-apply-20201217-1' into staging
A collection of RISC-V improvements:
- Improve the sifive_u DTB generation
- Add QSPI NOR flash to Microchip PFSoC
- Fix a bug in the Hypervisor HLVX/HLV/HSV instructions
- Fix some mstatus mask defines
- Ibex PLIC improvements
- OpenTitan memory layout update
- Initial steps towards support for 32-bit CPUs on 64-bit builds
# gpg: Signature made Fri 18 Dec 2020 05:59:42 GMT
# gpg: using RSA key F6C4AC46D4934868D3B8CE8F21E10D29DF977054
# gpg: Good signature from "Alistair Francis <alistair@alistair23.me>" [full]
# Primary key fingerprint: F6C4 AC46 D493 4868 D3B8 CE8F 21E1 0D29 DF97 7054
* remotes/alistair/tags/pull-riscv-to-apply-20201217-1: (23 commits)
riscv/opentitan: Update the OpenTitan memory layout
hw/riscv: Use the CPU to determine if 32-bit
target/riscv: cpu: Set XLEN independently from target
target/riscv: csr: Remove compile time XLEN checks
target/riscv: cpu_helper: Remove compile time XLEN checks
target/riscv: cpu: Remove compile time XLEN checks
target/riscv: Specify the XLEN for CPUs
target/riscv: Add a riscv_cpu_is_32bit() helper function
target/riscv: fpu_helper: Match function defs in HELPER macros
hw/riscv: sifive_u: Remove compile time XLEN checks
hw/riscv: spike: Remove compile time XLEN checks
hw/riscv: virt: Remove compile time XLEN checks
hw/riscv: boot: Remove compile time XLEN checks
riscv: virt: Remove target macro conditionals
riscv: spike: Remove target macro conditionals
target/riscv: Add a TYPE_RISCV_CPU_BASE CPU
hw/riscv: Expand the is 32-bit check to support more CPUs
intc/ibex_plic: Clear interrupts that occur during claim process
target/riscv: Fix definition of MSTATUS_TW and MSTATUS_TSR
target/riscv: Fix the bug of HLVX/HLV/HSV
...
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Instead of using string compares to determine if a RISC-V machine is
using 32-bit or 64-bit CPUs we can use the initalised CPUs. This avoids
us having to maintain a list of CPU names to compare against.
This commit also fixes the name of the function to match the
riscv_cpu_is_32bit() function.
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 8ab7614e5df93ab5267788b73dcd75f9f5615e82.1608142916.git.alistair.francis@wdc.com
Currently the riscv_is_32_bit() function only supports the generic rv32
CPUs. Extend the function to support the SiFive and LowRISC CPUs as
well.
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Palmer Dabbelt <palmerdabbelt@google.com>
Acked-by: Palmer Dabbelt <palmerdabbelt@google.com>
Message-id: 9a13764115ba78688ba61b56526c6de65fc3ef42.1608142916.git.alistair.francis@wdc.com
Previously if an interrupt occured during the claim process (after the
interrupt is claimed but before it's completed) it would never be
cleared.
This patch ensures that we also clear the hidden_pending bits as well.
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Tested-by: Jackie Ke <jackieke724@hotmail.com>
Message-id: 4e9786084a86f220689123cc8a7837af8fa071cf.1607100423.git.alistair.francis@wdc.com
The TW and TSR fields should be bits 21 and 22 and not 30/29.
This was found while comparing QEMU behaviour against the sail formal
model (https://github.com/rems-project/sail-riscv/).
Signed-off-by: Alex Richardson <Alexander.Richardson@cl.cam.ac.uk>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 20201130170117.71281-1-Alexander.Richardson@cl.cam.ac.uk
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
We found that the hypervisor virtual-machine load and store instructions,
included HLVX/HLV/HSV, couldn't access guest userspace memory.
In the riscv-privileged spec, HLVX/HLV/HSV is defined as follow:
"As usual when V=1, two-stage address translation is applied, and
the HS-level sstatus.SUM is ignored."
But get_physical_address() doesn't ignore sstatus.SUM, when HLVX/HLV/HSV
accesses guest userspace memory. So this patch fixes it.
Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
Signed-off-by: Yipeng Yin <yinyipeng1@huawei.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 20201130012810.899-1-jiangyifei@huawei.com
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
