Graeme Gregory [Tue, 3 Nov 2015 13:49:42 +0000 (13:49 +0000)]
hw/arm/virt-acpi-build: _CCA attribute is compulsory
According to ACPI specification 6.2.17 _CCA (Cache Coherency Attribute)
this attribute is compulsory on ARM systems. Add this attribute to
the PCI host bridges as required.
Without this the kernel will produce the error
[Firmware Bug]: PCI device 0000:00:00.0 fail to setup DMA.
Peter Maydell [Tue, 3 Nov 2015 13:49:42 +0000 (13:49 +0000)]
target-arm: Bring AArch64 debug CPU display of PSTATE into line with AArch32
The AArch64 debug CPU display of PSTATE as "PSTATE=200003c5 (flags --C-)"
on the end of the same line as the last of the general purpose registers
is unnecessarily different from the AArch32 display of PSR as
"PSR=200001d3 --C- A svc32" on its own line. Update the AArch64
code to put PSTATE in its own line and in the same format, including
printing the exception level (mode).
Change armv7m_init to return the DeviceState* for the NVIC.
This allows access to all GPIO blocks, not just the IRQ inputs.
Move qdev_get_gpio_in() calls out of armv7m_init() into
board code for stellaris and stm32f205 boards.
Add a Linux-specific pre-boot routine that matches the device-
specific bootloaders behaviour. This is needed for modern Linux that
expects the ARM PLL in SLCR to be a more even value (not 26).
Add an API for boards to inject their own preboot software (or
firmware) sequence.
The software then returns to the bootloader via the link register. This
allows boards to do their own little bits of firmware setup without
needed to replace the bootloader completely (which is the requirement
for existing firmware support).
The blob is loaded by a callback if and only if doing a linux boot
(similar to the existing write_secondary support).
These comments start immediately after the current longest name in the
list. Tab them out to the next tab stop to give a little breathing room
and prepare for FIXUP_BOARD_SETUP which will require more indent.
Peter Maydell [Tue, 3 Nov 2015 10:20:04 +0000 (10:20 +0000)]
Merge remote-tracking branch 'remotes/kraxel/tags/pull-ui-20151103-1' into staging
ui: fixes for vnc, opengl and curses.
# gpg: Signature made Tue 03 Nov 2015 09:53:24 GMT using RSA key ID D3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <[email protected]>"
# gpg: aka "Gerd Hoffmann <[email protected]>"
# gpg: aka "Gerd Hoffmann (private) <[email protected]>"
* remotes/kraxel/tags/pull-ui-20151103-1:
vnc: fix bug: vnc server can't start when 'to' is specified
vnc: allow fall back to RAW encoding
ui/opengl: Reduce build required libraries for opengl
ui/curses: Fix pageup/pagedown on -curses
ui/curses: Support line graphics chars on -curses mode
ui/curses: Fix monitor color with -curses when 256 colors
Yang Hongyang [Tue, 27 Oct 2015 06:10:52 +0000 (14:10 +0800)]
vnc: fix bug: vnc server can't start when 'to' is specified
commit e0d03b8ceb52 converted VNC startup to use SocketAddress,
the interface socket_listen don't have a port_offset param, so
we need to add the port offset (5900) to both 'port' and 'to' opts.
currently only 'port' is added by offset.
This patch add the port offset to 'to' opts.
Peter Lieven [Thu, 27 Aug 2015 12:46:25 +0000 (14:46 +0200)]
vnc: allow fall back to RAW encoding
I have observed that depending on the contents and the encoding it happens
that sending data as RAW sometimes would take less space than the encoded data.
This is especially the case for small updates or areas with high color images.
If sending RAW encoded data is beneficial allow a fall back to RAW encoding
for the framebuffer update.
OGAWA Hirofumi [Mon, 26 Oct 2015 17:45:48 +0000 (02:45 +0900)]
ui/opengl: Reduce build required libraries for opengl
We now use epoxy to load opengl libraries. This means we don't need to
link opengl libraries directly if interfaces handled by epoxy. With
this, we just need epoxy headers and epoxy's *.so to build.
Tested with epoxy-1.3.1.
- sdl2/gtk/console egl stuff doesn't require other than epoxy
- milkymist-tmu2 glx stuff doesn't require other than epoxy
(lm32 test is limited, because can't find mmone-bios.bin, so just test
to load libGL with "./lm32-softmmu/qemu-system-lm32 -M milkymist,accel=qtest")
OGAWA Hirofumi [Mon, 19 Oct 2015 12:23:10 +0000 (21:23 +0900)]
ui/curses: Fix monitor color with -curses when 256 colors
If TERM=xterm-256color, COLOR_PAIRS==256 and monitor passes chtype
like 0x74xx. Then, the code uses uninitialized color pair. As result,
monitor uses black for both of fg and bg color, i.e. terminal is
filled by black.
To fix, this initialize above than 64 with default color (fg=white,bg=black).
FIXME: on 256 color, curses may be possible better vga color emulation.
Peter Maydell [Mon, 2 Nov 2015 11:11:39 +0000 (11:11 +0000)]
Merge remote-tracking branch 'remotes/armbru/tags/pull-qapi-2015-11-02' into staging
QAPI patches
# gpg: Signature made Mon 02 Nov 2015 09:07:23 GMT using RSA key ID EB918653
# gpg: Good signature from "Markus Armbruster <[email protected]>"
# gpg: aka "Markus Armbruster <[email protected]>"
* remotes/armbru/tags/pull-qapi-2015-11-02: (25 commits)
qapi: Simplify gen_struct_field()
qapi: Reserve 'u' member name
qapi: Finish converting to new qapi union layout
tpm: Convert to new qapi union layout
memory: Convert to new qapi union layout
input: Convert to new qapi union layout
char: Convert to new qapi union layout
net: Convert to new qapi union layout
sockets: Convert to new qapi union layout
block: Convert to new qapi union layout
tests: Convert to new qapi union layout
qapi-visit: Convert to new qapi union layout
qapi: Start converting to new qapi union layout
qapi-visit: Remove redundant functions for flat union base
qapi: Unbox base members
qapi: Prefer typesafe upcasts to qapi base classes
qapi-types: Refactor base fields output
qapi-visit: Split off visit_type_FOO_fields forward decl
vnc: Hoist allocation of VncBasicInfo to callers
qapi: Reserve 'q_*' and 'has_*' member names
...
Eric Blake [Mon, 26 Oct 2015 22:35:03 +0000 (16:35 -0600)]
qapi: Simplify gen_struct_field()
Rather than having all callers pass a name, type, and optional
flag, have them instead pass a QAPISchemaObjectTypeMember which
already has all that information.
Eric Blake [Mon, 26 Oct 2015 22:35:02 +0000 (16:35 -0600)]
qapi: Reserve 'u' member name
Now that we have separated union tag values from colliding with
non-variant C names, by naming the union 'u', we should reserve
this name for our use. Note that we want to forbid 'u' even in
a struct with no variants, because it is possible for a future
qemu release to extend QMP in a backwards-compatible manner while
converting from a struct to a flat union. Fortunately, no
existing clients were using this member name. If we ever find
the need for QMP to have a member 'u', we could at that time
relax things, perhaps by having c_name() munge the QMP member to
'q_u'.
Note that we cannot forbid 'u' everywhere (by adding the
rejection code to check_name()), because the existing QKeyCode
enum already uses it; therefore we only reserve it as a struct
type member name.
Eric Blake [Mon, 26 Oct 2015 22:35:01 +0000 (16:35 -0600)]
qapi: Finish converting to new qapi union layout
We have two issues with our qapi union layout:
1) Even though the QMP wire format spells the tag 'type', the
C code spells it 'kind', requiring some hacks in the generator.
2) The C struct uses an anonymous union, which places all tag
values in the same namespace as all non-variant members. This
leads to spurious collisions if a tag value matches a non-variant
member's name.
This patch is the back end for a series that converts to a
saner qapi union layout. Now that all clients have been
converted to use 'type' and 'obj->u.value', we can drop the
temporary parallel support for 'kind' and 'obj->value'.
this is the overall effect, when compared to the state before
this series of patches:
| struct Foo {
|- FooKind kind;
|- union { /* union tag is @kind */
|+ FooKind type;
|+ union { /* union tag is @type */
| void *data;
| int64_t a;
| bool b;
|- };
|+ } u;
| };
The testsuite still contains some examples of artificial restrictions
(see flat-union-clash-type.json, for example) that are no longer
technically necessary, now that there is no longer a collision between
enum tag values and non-variant member names; but fixing this will be
done in later patches, in part because some further changes are required
to keep QAPISchema*.check() from asserting. Also, a later patch will
add a reservation for the member name 'u' to avoid a collision between a
user's non-variant names and our internal choice of C union name.
Note, however, that we do not rename the generated enum, which
is still 'FooKind'. A further patch could generate implicit
enums as 'FooType', but while the generator already reserved
the '*Kind' namespace (commit 4dc2e69), there are already QMP
constructs with '*Type' naming, which means changing our
reservation namespace would have lots of churn to C code to
deal with a forced name change.
Eric Blake [Mon, 26 Oct 2015 22:35:00 +0000 (16:35 -0600)]
tpm: Convert to new qapi union layout
We have two issues with our qapi union layout:
1) Even though the QMP wire format spells the tag 'type', the
C code spells it 'kind', requiring some hacks in the generator.
2) The C struct uses an anonymous union, which places all tag
values in the same namespace as all non-variant members. This
leads to spurious collisions if a tag value matches a non-variant
member's name.
Make the conversion to the new layout for TPM-related code.
Eric Blake [Mon, 26 Oct 2015 22:34:59 +0000 (16:34 -0600)]
memory: Convert to new qapi union layout
We have two issues with our qapi union layout:
1) Even though the QMP wire format spells the tag 'type', the
C code spells it 'kind', requiring some hacks in the generator.
2) The C struct uses an anonymous union, which places all tag
values in the same namespace as all non-variant members. This
leads to spurious collisions if a tag value matches a non-variant
member's name.
Make the conversion to the new layout for memory-related code.
Eric Blake [Mon, 26 Oct 2015 22:34:58 +0000 (16:34 -0600)]
input: Convert to new qapi union layout
We have two issues with our qapi union layout:
1) Even though the QMP wire format spells the tag 'type', the
C code spells it 'kind', requiring some hacks in the generator.
2) The C struct uses an anonymous union, which places all tag
values in the same namespace as all non-variant members. This
leads to spurious collisions if a tag value matches a non-variant
member's name.
Make the conversion to the new layout for input-related code.
Eric Blake [Mon, 26 Oct 2015 22:34:57 +0000 (16:34 -0600)]
char: Convert to new qapi union layout
We have two issues with our qapi union layout:
1) Even though the QMP wire format spells the tag 'type', the
C code spells it 'kind', requiring some hacks in the generator.
2) The C struct uses an anonymous union, which places all tag
values in the same namespace as all non-variant members. This
leads to spurious collisions if a tag value matches a non-variant
member's name.
Make the conversion to the new layout for character-related
code.
Eric Blake [Mon, 26 Oct 2015 22:34:56 +0000 (16:34 -0600)]
net: Convert to new qapi union layout
We have two issues with our qapi union layout:
1) Even though the QMP wire format spells the tag 'type', the
C code spells it 'kind', requiring some hacks in the generator.
2) The C struct uses an anonymous union, which places all tag
values in the same namespace as all non-variant members. This
leads to spurious collisions if a tag value matches a non-variant
member's name.
Make the conversion to the new layout for net-related code.
Eric Blake [Mon, 26 Oct 2015 22:34:55 +0000 (16:34 -0600)]
sockets: Convert to new qapi union layout
We have two issues with our qapi union layout:
1) Even though the QMP wire format spells the tag 'type', the
C code spells it 'kind', requiring some hacks in the generator.
2) The C struct uses an anonymous union, which places all tag
values in the same namespace as all non-variant members. This
leads to spurious collisions if a tag value matches a non-variant
member's name.
Make the conversion to the new layout for socket-related code.
Eric Blake [Mon, 26 Oct 2015 22:34:54 +0000 (16:34 -0600)]
block: Convert to new qapi union layout
We have two issues with our qapi union layout:
1) Even though the QMP wire format spells the tag 'type', the
C code spells it 'kind', requiring some hacks in the generator.
2) The C struct uses an anonymous union, which places all tag
values in the same namespace as all non-variant members. This
leads to spurious collisions if a tag value matches a non-variant
member's name.
Make the conversion to the new layout for block-related code.
Eric Blake [Mon, 26 Oct 2015 22:34:53 +0000 (16:34 -0600)]
tests: Convert to new qapi union layout
We have two issues with our qapi union layout:
1) Even though the QMP wire format spells the tag 'type', the
C code spells it 'kind', requiring some hacks in the generator.
2) The C struct uses an anonymous union, which places all tag
values in the same namespace as all non-variant members. This
leads to spurious collisions if a tag value matches a non-variant
member's name.
Make the conversion to the new layout for testsuite code.
Eric Blake [Mon, 26 Oct 2015 22:34:52 +0000 (16:34 -0600)]
qapi-visit: Convert to new qapi union layout
We have two issues with our qapi union layout:
1) Even though the QMP wire format spells the tag 'type', the
C code spells it 'kind', requiring some hacks in the generator.
2) The C struct uses an anonymous union, which places all tag
values in the same namespace as all non-variant members. This
leads to spurious collisions if a tag value matches a non-variant
member's name.
Make the conversion to the new layout for qapi-visit.py.
Eric Blake [Mon, 26 Oct 2015 22:34:51 +0000 (16:34 -0600)]
qapi: Start converting to new qapi union layout
We have two issues with our qapi union layout:
1) Even though the QMP wire format spells the tag 'type', the
C code spells it 'kind', requiring some hacks in the generator.
2) The C struct uses an anonymous union, which places all tag
values in the same namespace as all non-variant members. This
leads to spurious collisions if a tag value matches a non-variant
member's name.
This patch is the front end for a series that converts to a
saner qapi union layout. By the end of the series, we will no
longer have the type/kind mismatch, and all tag values will be
under a named union, which requires clients to access
'obj->u.value' instead of 'obj->value'. But since the
conversion touches a number of files, it is easiest if we
temporarily support BOTH layouts simultaneously.
make the following changes in generated qapi-types.h:
| struct Foo {
|- FooKind kind;
|- union { /* union tag is @kind */
|+ union {
|+ FooKind kind;
|+ FooKind type;
|+ };
|+ union { /* union tag is @type */
| void *data;
| int64_t a;
| bool b;
|+ union { /* union tag is @type */
|+ void *data;
|+ int64_t a;
|+ bool b;
|+ } u;
| };
| };
Flat unions do not need the anonymous union for the tag member,
as we already fixed that to use the member name instead of 'kind'
back in commit 0f61af3e.
One additional change is needed in qapi.py: check_union() now
needs to check for collisions with 'type' in addition to those
with 'kind'.
Later, when the conversions are complete, we will remove the
duplication hacks, and also drop the check_union() restrictions.
Note, however, that we do not rename the generated enum, which
is still 'FooKind'. A further patch could generate implicit
enums as 'FooType', but while the generator already reserved
the '*Kind' namespace (commit 4dc2e69), there are already QMP
constructs with '*Type' naming, which means changing our
reservation namespace would have lots of churn to C code to
deal with a forced name change.
Eric Blake [Mon, 26 Oct 2015 22:34:50 +0000 (16:34 -0600)]
qapi-visit: Remove redundant functions for flat union base
The code for visiting the base class of a child struct created
visit_type_Base_fields() which covers all fields of Base; while
the code for visiting the base class of a flat union created
visit_type_Union_fields() covering all fields of the base
except the discriminator. But since the base class includes
the discriminator of a flat union, we can just visit the entire
base, without needing a separate visit of the discriminator.
Not only is consistently visiting all fields easier to
understand, it lets us share code.
The generated code in qapi-visit.c loses several now-unused
visit_type_UNION_fields(), along with changes like:
and forward declarations where needed. Note that the cast of obj
to BASE ** is necessary to call visit_type_BASE_fields() (and we
can't use our upcast wrappers, because those work on pointers while
we have a pointer-to-pointer).
Eric Blake [Mon, 26 Oct 2015 22:34:49 +0000 (16:34 -0600)]
qapi: Unbox base members
Rather than storing a base class as a pointer to a box, just
store the fields of that base class in the same order, so that
a child struct can be directly cast to its parent. This gives
less malloc overhead, less pointer dereferencing, and even less
generated code. Compare to the earlier commit 1e6c1616a "qapi:
Generate a nicer struct for flat unions" (although that patch
had fewer places to change, as less of qemu was directly using
qapi structs for flat unions). It also allows us to turn on
automatic type-safe wrappers for upcasting to the base class
of a struct.
Changes to the generated code look like this in qapi-types.h:
(the cast is necessary, since our upcast wrappers only deal with a
single pointer, not pointer-to-pointer); plus the wholesale
elimination of some now-unused visit_type_implicit_FOO() functions.
Without boxing, the corner case of one empty struct having
another empty struct as its base type now requires inserting a
dummy member (previously, the 'Base *base' member sufficed).
And now that we no longer consume a 'base' member in the generated
C struct, we can delete the former negative struct-base-clash-base
test.
Eric Blake [Mon, 26 Oct 2015 22:34:48 +0000 (16:34 -0600)]
qapi: Prefer typesafe upcasts to qapi base classes
A previous patch (commit 1e6c1616) made it possible to
directly cast from a qapi flat union type to its base type.
However, it requires the use of a C cast, which turns off
compiler type-safety checks. Fortunately, no such casts
exist, just yet.
Regardless, add inline type-safe wrappers named
qapi_FOO_base() for any union type FOO that has a base,
which can be used for a safer upcast, and enhance the
testsuite to cover the new functionality.
A future patch will extend the upcast support to structs,
where such conversions do exist already.
Note that C makes const-correct upcasts annoying because
it lacks overloads; these functions cast away const so that
they can accept user pointers whether const or not, and the
result in turn can be assigned to normal or const pointers.
Alternatively, this could have been done with macros, but
type-safe macros are hairy, and not worthwhile here.
This patch just adds upcasts. None of our code needed to
downcast from a base qapi class to a child. Also, in the
case of grandchildren (such as BlockdevOptionsQcow2), the
caller will need to call two functions to get to the inner
base (although it wouldn't be too hard to generate a
qapi_FOO_base_base() if desired). If a user changes qapi
to alter the base class hierarchy, such as going from
'A -> C' to 'A -> B -> C', it will change the type of
'qapi_C_base()', and the compiler will point out the places
that are affected by the new base.
One alternative was proposed, but was deemed too ugly to use
in practice: the generators could output redundant
information using anonymous types:
| struct Child {
| union {
| struct {
| Type1 parent_member1;
| Type2 parent_member2;
| };
| Parent base;
| };
| };
With that ugly proposal, for a given qapi type, obj->member
and obj->base.member would refer to the same storage; allowing
convenience in working with members without needing 'base.'
allowing typesafe upcast without needing a C cast by accessing
'&obj->base', and allowing downcasts from the parent back to
the child possible through container_of(obj, Child, base).
Eric Blake [Mon, 26 Oct 2015 22:34:46 +0000 (16:34 -0600)]
qapi-visit: Split off visit_type_FOO_fields forward decl
We generate a static visit_type_FOO_fields() for every type
FOO. However, sometimes we need a forward declaration. Split
the code to generate the forward declaration out of
gen_visit_implicit_struct() into a new gen_visit_fields_decl(),
and also prepare for a forward declaration to be emitted
during gen_visit_struct(), so that a future patch can switch
from using visit_type_FOO_implicit() to the simpler
visit_type_FOO_fields() as part of unboxing the base class
of a struct.
Eric Blake [Mon, 26 Oct 2015 22:34:45 +0000 (16:34 -0600)]
vnc: Hoist allocation of VncBasicInfo to callers
A future qapi patch will rework generated structs with a base
class to be unboxed. In preparation for that, change the code
that allocates then populates an info struct to instead merely
populate the fields of an info field passed in as a parameter
(renaming vnc_basic_info_get* to vnc_init_basic_info*). Add
rudimentary Error handling at the lowest levels for cases
where the old code returned NULL; but rather than plumb Error
all the way through the stack, the callers drop the error and
return NULL as before.
Eric Blake [Mon, 26 Oct 2015 22:34:44 +0000 (16:34 -0600)]
qapi: Reserve 'q_*' and 'has_*' member names
c_name() produces names starting with 'q_' when protecting a
dictionary member name that would fail to directly compile, but
in doing so can cause clashes with any member name already
beginning with 'q-' or 'q_'. Likewise, we create a C name 'has_'
for any optional member that can clash with any member name
beginning with 'has-' or 'has_'.
Technically, rather than blindly reserving the namespace,
we could try to complain about user names only when an actual
collision occurs, or even teach c_name() how to munge names
to avoid collisions. But it is not trivial, especially when
collisions can occur across multiple types (such as via
inheritance or flat unions). Besides, no existing .json
files are trying to use these names. So it's easier to just
outright forbid the potential for collision. We can always
relax things in the future if a real need arises for QMP to
express member names that have been forbidden here.
'has_' only has to be reserved for struct/union member names,
while 'q_' is reserved everywhere (matching the fact that
only members can be optional, while we use c_name() for munging
both members and entities). Note that we could relax 'q_'
restrictions on entities independently from member names; for
example, c_name('qmp_' + 'unix') would result in a different
function name than our current 'qmp_' + c_name('unix').
Update and add tests to cover the new error messages.
Eric Blake [Mon, 26 Oct 2015 22:34:43 +0000 (16:34 -0600)]
qapi: Reserve '*List' type names for list types
Type names ending in 'List' can clash with qapi list types in
generated C. We don't currently use such names. It is easier to
outlaw them now than to worry about how to resolve such a clash
in the future. For precedence, see commit 4dc2e69, which did the
same for names ending in 'Kind' versus implicit enum types for
qapi unions.
Eric Blake [Mon, 26 Oct 2015 22:34:42 +0000 (16:34 -0600)]
qapi: More robust conditions for when labels are needed
We were using regular expressions to see if ret included
any earlier text that emitted a 'goto out;' line, to decide
whether we needed to output an 'out:' label. But this is
fragile, if the ret text can possibly combine more than one
generated function body, where the first function used a
goto but the second does not. Change the code to just check
for the known conditions which cause an error check to be
needed. Besides, it's slightly more efficient to use plain
checks than regular expression searching.
Eric Blake [Mon, 26 Oct 2015 22:34:41 +0000 (16:34 -0600)]
qapi: More idiomatic string operations
Rather than slicing the end of a string, we can use python's
endswith(). And rather than creating a set of characters,
we can search for a character within a string.
Eric Blake [Mon, 26 Oct 2015 22:34:40 +0000 (16:34 -0600)]
tests/qapi-schema: Test for reserved names, empty struct
Add some testsuite coverage to ensure future patches are on
the right track:
Our current C representation of qapi arrays is done by appending
'List' to the element name; but we are not preventing the
creation of an object type with the same name. Add
reserved-type-list.json to test this. Then rename
enum-union-clash.json to reserved-type-kind.json to cover the
reservation that we DO detect, and shorten it to match the fact
that the name is reserved even if there is no clash.
We are failing to detect a collision between a dictionary member
and the implicit 'has_*' flag for another optional member. The
easiest fix would be for a future patch to reserve the entire
"has[-_]" namespace for member names (the collision is also
possible for branch names within flat unions, but only as long as
branch names can collide with (non-variant) members; however,
since future patches are about to remove that, it is not worth
testing here). Add reserved-member-has.json to test this.
A similar collision exists between a dictionary member where
c_name() munges what might otherwise be a reserved name to start
with 'q_', and another member explicitly starts with "q[-_]".
Again, the easiest solution for a future patch will be reserving
the entire namespace, but here for commands as well as members.
Add reserved-member-q.json and reserved-command-q.json to test
this; separate tests since arguably our munging of command 'unix'
to 'qmp_q_unix()' could be done without a q_, which is different
than the munging of a member 'unix' to 'foo.q_unix'.
Finally, our testsuite does not have any compilation coverage
of struct inheritance with empty qapi structs. Update
qapi-schema-test.json to test this.
Note that there is currently no technical reason to forbid type
name patterns from member names, or member name patterns from
types, since the two are not in the same namespace in C and
won't collide; but it's not worth adding positive tests of these
corner cases at this time, especially while there is other churn
pending in patches that rearrange which collisions actually
happen.
sockets: allow port to be NULL when listening on IP address
Switching "port" from mandatory to optional causes the QAPI
code generator to add a 'has_port' field to the InetSocketAddress
struct. No code that created InetSocketAddress objects was updated
to set 'has_port = true', which caused the non-NULL port strings
to be silently dropped when copying InetSocketAddress objects.
Peter Maydell [Fri, 30 Oct 2015 19:47:47 +0000 (19:47 +0000)]
Merge remote-tracking branch 'remotes/stefanha/tags/block-pull-request' into staging
# gpg: Signature made Thu 29 Oct 2015 18:09:16 GMT using RSA key ID 81AB73C8
# gpg: Good signature from "Stefan Hajnoczi <[email protected]>"
# gpg: aka "Stefan Hajnoczi <[email protected]>"
* remotes/stefanha/tags/block-pull-request:
block: Consider all child nodes in bdrv_requests_pending()
target-arm: xlnx-zynqmp: Add sdhci support.
sdhci: Split sdhci.h for public and internal device usage
sd.h: Move sd.h to include/hw/sd/
virtio: sync the dataplane vring state to the virtqueue before virtio_save
gdb command: qemu handlers
virtio-blk: switch off scsi-passthrough by default
ppc/spapr: add 2.4 compat props
s390x: include HW_COMPAT_* props
qemu-gdb: add $qemu_coroutine_sp and $qemu_coroutine_pc
qemu-gdb: extract parts of "qemu coroutine" implementation
qemu-gdb: allow using glibc_pointer_guard() on core dumps
Peter Maydell [Fri, 30 Oct 2015 16:30:25 +0000 (16:30 +0000)]
Merge remote-tracking branch 'remotes/lalrae/tags/mips-20151030' into staging
MIPS patches 2015-10-30
Changes:
* R6 CPU can be woken up by non-enabled interrupts
* PC fix in KVM
* Coprocessor 0 XContext calculation fix
* various MIPS R6 updates
# gpg: Signature made Fri 30 Oct 2015 14:51:56 GMT using RSA key ID 0B29DA6B
# gpg: Good signature from "Leon Alrae <[email protected]>"
* remotes/lalrae/tags/mips-20151030:
target-mips: fix updating XContext on mmu exception
target-mips: add SIGRIE instruction
target-mips: Set Config5.XNP for R6 cores
target-mips: add PC, XNP reg numbers to RDHWR
hw/mips_malta: Fix KVM PC initialisation
target-mips: Add enum for BREAK32
target-mips: update writing to CP0.Status.KX/SX/UX in MIPS Release R6
target-mips: implement the CPU wake-up on non-enabled interrupts in R6
target-mips: move the test for enabled interrupts to a separate function
Yongbok Kim [Thu, 29 Oct 2015 17:17:52 +0000 (17:17 +0000)]
target-mips: fix updating XContext on mmu exception
Correct updating XContext.Region field on mmu exceptions.
If Config3.CTXTC = 0 then the R field of XContext has to be updated
with the value of bits 63..62 of the virtual address upon a TLB
exception.
Also fixed the below line which overs 80 characters.
Yongbok Kim [Thu, 29 Oct 2015 15:18:38 +0000 (15:18 +0000)]
target-mips: add SIGRIE instruction
Add SIGRIE (Signal Reserved Instruction Exception) for both MIPS and
microMIPS.
The instruction allows to use the 16-bit code field for software use.
This instruction is introduced by and required as of Release 6.
Yongbok Kim [Thu, 29 Oct 2015 15:18:39 +0000 (15:18 +0000)]
target-mips: add PC, XNP reg numbers to RDHWR
Add Performance Counter (4) and XNP (5) register numbers to RDHWR.
Add check_hwrena() to simplify access control checkings.
Add RDHWR support to microMIPS R6.
James Hogan [Mon, 12 Oct 2015 16:54:39 +0000 (17:54 +0100)]
hw/mips_malta: Fix KVM PC initialisation
Commit 71c199c81d29 ("mips_malta: provide ememsize env variable to
kernels") changed the meaning of loaderparams.ram_size to be the whole
of RAM rather than just the low part below where the boot code is placed
for KVM, but it didn't update the PC initialisation for KVM to use
ram_low_size. Fix that now.
Peter Maydell [Fri, 30 Oct 2015 09:41:14 +0000 (09:41 +0000)]
Merge remote-tracking branch 'remotes/armbru/tags/pull-monitor-2015-10-30' into staging
QMP and QObject patches
# gpg: Signature made Fri 30 Oct 2015 08:06:26 GMT using RSA key ID EB918653
# gpg: Good signature from "Markus Armbruster <[email protected]>"
# gpg: aka "Markus Armbruster <[email protected]>"
* remotes/armbru/tags/pull-monitor-2015-10-30:
docs: Document QMP event rate limiting
monitor: Throttle event VSERPORT_CHANGE separately by "id"
monitor: Turn monitor_qapi_event_state[] into a hash table
glib: add compatibility interface for g_hash_table_add()
monitor: Split MonitorQAPIEventConf off MonitorQAPIEventState
monitor: Switch from timer_new() to timer_new_ns()
monitor: Simplify event throttling
monitor: Reduce casting of QAPI event QDict
qstring: Make conversion from QObject * accept null
qlist: Make conversion from QObject * accept null
qfloat qint: Make conversion from QObject * accept null
qdict: Make conversion from QObject * accept null
qbool: Make conversion from QObject * accept null
qobject: Drop QObject_HEAD
monitor: Throttle event VSERPORT_CHANGE separately by "id"
VSERPORT_CHANGE is emitted when the guest opens or closes a
virtio-serial port. The event's member "id" identifies the port.
When several events arrive quickly, throttling drops all but the last
of them. Because of that, a QMP client must assume that *any* port
may have changed state when it receives a VSERPORT_CHANGE event and
throttling may have happened.
Make the event more useful by throttling it for each port separately.
Sai Pavan Boddu [Thu, 8 Oct 2015 13:21:02 +0000 (18:51 +0530)]
sdhci: Split sdhci.h for public and internal device usage
Split sdhci.h into pubilc version (i.e include/hw/sd/sdhci.h) and
internal version (i.e hw/sd/sdhci-interna.h) based on register
declarations and object declaration.
Pavel Butsykin [Mon, 26 Oct 2015 11:42:57 +0000 (14:42 +0300)]
virtio: sync the dataplane vring state to the virtqueue before virtio_save
When creating snapshot with the dataplane enabled, the snapshot file gets
not the actual state of virtqueue, because the current state is stored in
VirtIOBlockDataPlane. Therefore, before saving snapshot need to sync
the dataplane vring state to the virtqueue. The dataplane will resume its
work at the next notify virtqueue.
When snapshot loads with loadvm we get a message:
VQ 0 size 0x80 Guest index 0x15f5 inconsistent with Host index 0x0:
delta 0x15f5
error while loading state for instance 0x0 of device
'0000:00:08.0/virtio-blk'
Error -1 while loading VM state
to reproduce the error I used the following hmp commands:
savevm snap1
loadvm snap1
That dumps an AioContext list (by default qemu_aio_context)
possibly including a backtrace for cases it knows about
(with the verbose option). Intended to help find why something
is hanging waiting for IO.
Use 'qemu handlers --verbose iohandler_ctx' to find out why
your incoming migration is stuck.
Cornelia Huck [Fri, 16 Oct 2015 10:25:54 +0000 (12:25 +0200)]
virtio-blk: switch off scsi-passthrough by default
Devices that are compliant with virtio-1 do not support scsi
passthrough any more (and it has not been a recommended setup
anyway for quite some time). To avoid having to switch it off
explicitly in newer qemus that turn on virtio-1 by default, let's
switch the default to scsi=false for 2.5.
Paolo Bonzini [Mon, 12 Oct 2015 08:02:52 +0000 (10:02 +0200)]
qemu-gdb: allow using glibc_pointer_guard() on core dumps
get_fs_base() cannot be run on a core dump, because it uses the arch_prctl
system call. The fs base is the value that is returned by pthread_self(),
and it would be nice to just glean it from the "info threads" output:
but unfortunately the gdb API does not provide that. Instead, we can
look for the "arg" argument of the start_thread function if glibc debug
information are available. If not, fall back to the old mechanism.
Leon Alrae [Mon, 14 Sep 2015 12:51:31 +0000 (13:51 +0100)]
target-mips: update writing to CP0.Status.KX/SX/UX in MIPS Release R6
Implement the relationship between CP0.Status.KX, SX and UX. It should not
be possible to set UX bit if SX is 0, the same applies for setting SX if
KX is 0.
Leon Alrae [Mon, 14 Sep 2015 12:58:24 +0000 (13:58 +0100)]
target-mips: implement the CPU wake-up on non-enabled interrupts in R6
In Release 6, the behaviour of WAIT has been modified to make it a
requirement that a processor that has disabled operation as a result of
executing a WAIT will resume operation on arrival of an interrupt even if
interrupts are not enabled.
QObject_HEAD is a macro expanding into the common part of structs that
are sub-types of QObject. It's always been just QObject base, and
unlikely to change. Drop the macro, because the code is clearer with
out it.
Peter Maydell [Thu, 29 Oct 2015 09:49:52 +0000 (09:49 +0000)]
Merge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging
virtio, pc, memory: fixes+features for 2.5
New features:
This enables hotplug for multifunction devices.
Patches are very small, so I think it's OK to merge
at this stage.
There's also some new infrastructure for vhost-user testing
not enabled yet so it's harmless to merge.
I've reverted the "gap between DIMMs" workaround, as it seems too risky, and
applied my own patch in virtio, but not in dataplane code. This means that
dataplane is broken for some complex DIMM configurations for now. Waiting for
Stefan to review the dataplane fix.
Signed-off-by: Michael S. Tsirkin <[email protected]>
# gpg: Signature made Thu 29 Oct 2015 09:36:16 GMT using RSA key ID D28D5469
# gpg: Good signature from "Michael S. Tsirkin <[email protected]>"
# gpg: aka "Michael S. Tsirkin <[email protected]>"
* remotes/mst/tags/for_upstream:
enable multi-function hot-add
remove function during multi-function hot-add
tests/vhost-user-bridge: add vhost-user bridge application
Revert "memhp: extend address auto assignment to support gaps"
Revert "pc: memhp: force gaps between DIMM's GPA"
virtio: drop virtqueue_map_sg
virtio-scsi: convert to virtqueue_map
virtio-serial: convert to virtio_map
virtio-blk: convert to virtqueue_map
virtio: switch to virtio_map
virtio: introduce virtio_map
mmap-alloc: fix error handling
pc: memhp: do not emit inserting event for coldplugged DIMMs
vhost-user-test: fix up rhel6 build
vhost-user: cleanup msg size math
vhost-user: cleanup struct size math
The test existing in QEMU for vhost-user feature is good for
testing the management protocol, but does not allow actual
traffic. This patch proposes Vhost-User Bridge application, which
can serve the QEMU community as a comprehensive test by running
real internet traffic by means of vhost-user interface.
Essentially the Vhost-User Bridge is a very basic vhost-user
backend for QEMU. It runs as a standalone user-level process.
For packet processing Vhost-User Bridge uses an additional QEMU
instance with a backend configured by "-net socket" as a shared
VLAN. This way another QEMU virtual machine can effectively
serve as a shared bus by means of UDP communication.
For a more simple setup, the another QEMU instance running the
SLiRP backend can be the same QEMU instance running vhost-user
client.
This Vhost-User Bridge implementation is very preliminary. It is
missing many features. I has been studying vhost-user protocol
internals, so I've written vhost-user-bridge bit by bit as I
progressed through the protocol. Most probably its internal
architecture will change significantly.
To run Vhost-User Bridge application:
1. Build vhost-user-bridge with a regular procedure. This will
create a vhost-user-bridge executable under tests directory:
$ configure; make tests/vhost-user-bridge
2. Ensure the machine has hugepages enabled in kernel with
command line like:
The above will run vhost-user server listening for connections
on UNIX domain socket /tmp/vubr.sock, and will try to connect
by UDP to VLAN bridge to localhost:5555, while listening on
localhost:4444
vhost-user-bridge was tested very lightly: it's able to bringup a
linux on client VM with the virtio-net driver, and execute transmits
and receives to the internet. I tested with "wget redhat.com",
"dig redhat.com".
PS. I've consulted DPDK's code for vhost-user during Vhost-User
Bridge implementation.
This also fixes a minor bug:
- virtqueue_map_sg(port->elem.out_sg, port->elem.out_addr,
- port->elem.out_num, 1);
is wrong: out_sg is not written so should not be marked dirty.
virtio_map_sg currently fails if one of the entries it's mapping is
contigious in GPA but not HVA address space. Introduce virtio_map which
handles this by splitting sg entries.
This new API generally turns out to be a good idea since it's harder to
misuse: at least in one case the existing one was used incorrectly.
This will still fail if there's no space left in the sg, but luckily max
queue size in use is currently 256, while max sg size is 1024, so we
should be OK even is all entries happen to cross a single DIMM boundary.
Won't work well with very small DIMM sizes, unfortunately:
e.g. this will fail with 4K DIMMs where a single
request might span a large number of DIMMs.
Let's hope these are uncommon - at least we are not breaking things.
Note: virtio-scsi calls virtio_map_sg on data loaded from network, and
validates input, asserting on failure. Copy the validating code here -
it will be dropped from virtio-scsi in a follow-up patch.
Igor Mammedov [Fri, 23 Oct 2015 12:55:26 +0000 (14:55 +0200)]
pc: memhp: do not emit inserting event for coldplugged DIMMs
currently acpi_memory_plug_cb() sets is_inserting for
cold- and hot-plugged DIMMs as result ASL MHPD.MSCN()
method issues device check even for every coldplugged
DIMM. There isn't much harm in it but if we try to
unplug such DIMM, OSPM will issue device check
intstead of device eject event. So OSPM won't eject
memory module as expected and it will try to eject it
only when another memory device is hot-(un)plugged.
As a fix do not set 'is_inserting' event and do not
issue SCI for cold-plugged DIMMs as they are
enumerated and activated by OSPM during guest's boot.
projects / qemu.git / log