Drop the NetSocketListenState struct and add a listen_fd field
to NetSocketState. When a -netdev socket,listen= instance is created
there will be a NetSocketState with fd=-1 and a valid listen_fd. The
net_socket_accept() handler waits for listen_fd to become readable and
then accepts the connection. When this state transition happens, we no
longer monitor listen_fd for incoming connections...until the client
disconnects again.
Stefan Hajnoczi [Tue, 24 Jul 2012 15:35:15 +0000 (16:35 +0100)]
net: Rename qemu_del_vlan_client() to qemu_del_net_client()
Another step in moving the vlan feature out of net core. Users only
deal with NetClientState and therefore qemu_del_vlan_client() should be
named qemu_del_net_client().
Stefan Hajnoczi [Tue, 24 Jul 2012 15:35:14 +0000 (16:35 +0100)]
net: Rename vc local variables to nc
Now that VLANClientState has been renamed to NetClientState all 'vc'
local variables should be 'nc'. Much of the code already used 'nc' but
there are places where 'vc' needs to be renamed.
Stefan Hajnoczi [Tue, 24 Jul 2012 15:35:13 +0000 (16:35 +0100)]
net: Rename VLANClientState to NetClientState
The vlan feature is no longer part of net core. Rename VLANClientState
to NetClientState because net clients are not explicitly associated with
a vlan at all, instead they have a peer net client to which they are
connected.
This patch is a mechanical search-and-replace except for a few
whitespace fixups where changing VLANClientState to NetClientState
misaligned whitespace.
Stefan Hajnoczi [Tue, 24 Jul 2012 15:35:12 +0000 (16:35 +0100)]
net: Rename non_vlan_clients to net_clients
There is no longer a distinction between vlan clients and non-vlan
clients in the net core. The net core only knows about point-to-point
clients which are connected to a peer. It's time to rename the global
list of net clients since it no longer refers to vlans at all.
Instead of using VLANState use net/hub.h to support the vlan qdev
property. The vlan qdev property becomes an alias for the peer qdev
property but is represented as a VLAN ID number. When a VLAN ID is
selected the device will really peer with a hub port.
Stefan Hajnoczi [Tue, 24 Jul 2012 15:35:08 +0000 (16:35 +0100)]
net: Drop vlan argument to qemu_new_net_client()
Since hubs are now used to implement the 'vlan' feature and the vlan
argument is always NULL, remove the argument entirely and update all net
clients that use qemu_new_net_client().
Stefan Hajnoczi [Tue, 24 Jul 2012 15:35:07 +0000 (16:35 +0100)]
hub: Check that hubs are configured correctly
Checks can be performed to make sure that hubs have at least one NIC and
one host device, warning the user if this is not the case.
Configurations which do not meet this rule tend to be broken but just
emit a warning. This patch preserves compatibility with the checks
performed by net core on vlans.
Stefan Hajnoczi [Tue, 24 Jul 2012 15:35:05 +0000 (16:35 +0100)]
net: Use hubs for the vlan feature
Stop using the special-case vlan code in net.c. Instead use the hub net
client to implement the vlan feature. The next patch will remove vlan
code from net.c completely.
Stefan Hajnoczi [Tue, 24 Jul 2012 15:35:04 +0000 (16:35 +0100)]
net: Add a hub net client
The vlan feature can be implemented in terms of hubs. By introducing a
hub net client it becomes possible to remove the special case vlan code
from net.c and push the vlan feature out of generic networking code.
net: Add interface to bridge when SIOCBRADDIF isn't available
The bridge helper uses the SIOCBRADDIF ioctl to add an inteface to
a bridge. SIOCBRADDIF is not available on old Linux versions. This
patch adds support to use the SIOCDEVPRIVATE ioctl with BRCTL_ADD_IF
if SIOCBRADDIF is not available.
Peter Maydell [Wed, 18 Jul 2012 14:10:27 +0000 (15:10 +0100)]
configure: -I\$(SRC_PATH) goes in QEMU_INCLUDES not QEMU_CFLAGS
If the smartcard configure check passes, add '-I\$(SRC_PATH)/libcacard'
to QEMU_INCLUDES, not QEMU_CFLAGS. Otherwise the unexpanded SRC_PATH
will cause a warning in every following configure test.
Peter Maydell [Wed, 18 Jul 2012 14:10:22 +0000 (15:10 +0100)]
configure: -march=i486 belongs in QEMU_CFLAGS, not CFLAGS
The distinction between QEMU_CFLAGS and CFLAGS is that the
former is for flags without which QEMU can't compile, whereas
the latter is for flags like "-g -O2" which the user can
safely override. "-march=i486" is in the former category, and
so belongs in QEMU_CFLAGS.
Stefan Weil [Wed, 18 Jul 2012 14:10:21 +0000 (15:10 +0100)]
configure: Replace bash code by standard shell code
"+=" does not work with dash and other simple /bin/sh implementations.
The new code prepends the flag while the old code either did not work
(it continued after an error message which typically was not read) or
appended the flag. That difference should not matter here.
Stefan Weil [Wed, 18 Jul 2012 14:10:20 +0000 (15:10 +0100)]
configure: Fix build with capabilities
Since commit 417c9d72d48275d19c60861896efd4962d21aca2 all configure tests
normally run with -Werror. Some of these tests now fail because they
raised a compiler warning.
Peter Maydell [Wed, 18 Jul 2012 14:10:18 +0000 (15:10 +0100)]
configure: Don't run configure tests with -Werror enabled
Don't run configure tests with -Werror in the compiler flags. The idea
of -Werror is that it makes problems very obvious to developers, so
they get fixed quickly. However, when running configure tests, failures
due to -Werror are far from obvious -- they simply result in the test
quietly failing when it should have passed. Not using -Werror is in
line with recommended practice in the Autoconf world.
This commit is essentially backing out the changes in commit 417c9d72.
Instead we fix the problem that commit was trying to address in a
different way: we add -Werror only for the test of the nss headers,
with a comment that this is specifically intended to detect a bug
in some releases of nss.
We also have to clean up a bug in the smartcard test where it was
trying to include smartcard_cflags in the test compile flags: this
would always result in a failure with -Werror, because they include
an escaped "$(SRC_PATH)" which is only valid when used in the final
makefile.
Peter Maydell [Mon, 30 Jul 2012 15:13:07 +0000 (16:13 +0100)]
configure: Split valgrind test into pragma test and valgrind.h test
Split the configure test that checks for valgrind into two, one
part checking whether we have the gcc pragma to disable unused-but-set
variables, and the other part checking for the existence of valgrind.h.
The first of these has to be compiled with -Werror and the second
does not and shouldn't generate any warnings.
This (a) allows us to enable "make errors in configure tests be
build failures" and (b) enables use of valgrind on systems with
a gcc which doesn't know about -Wunused-but-set-varibale, like
Debian squeeze.
Anthony Liguori [Mon, 30 Jul 2012 14:59:23 +0000 (09:59 -0500)]
Merge remote-tracking branch 'bonzini/scsi-next' into staging
* bonzini/scsi-next: (32 commits)
virtio-scsi: enable MSI-X support
virtio-scsi: add ioeventfd support
virtio-scsi: report parameter change events
virtio-scsi: do not report dropped events after reset
virtio-scsi: Report missed events
virtio-scsi: Implement hotplug support for virtio-scsi
scsi: report parameter changes to HBA drivers
scsi-disk: report resized disk via sense codes
scsi: establish precedence levels for unit attention
scsi: introduce hotplug() and hot_unplug() interfaces for SCSI bus
scsi: add tracepoint for scsi_req_cancel
scsi-disk: removable hard disks support load/eject
scsi-disk: Fail medium writes with proper sense for readonly LUNs
scsi-disk: improve the lba-out-of-range tests for read/write/verify
scsi-disk: rd/wr/vr-protect !=0 is an error
scsi-disk: support toggling the write cache
scsi-disk: parse MODE SELECT commands and parameters
scsi-disk: fix changeable values for MODE_PAGE_R_W_ERROR
scsi-disk: adjust offsets in MODE SENSE by 2
scsi-disk: support emulated TO_DEV requests
...
Anthony Liguori [Mon, 30 Jul 2012 14:58:41 +0000 (09:58 -0500)]
Merge commit 'quintela/migration-next-v5' into staging
* commit '6c779f22a93cc6e4565b940ef616e3efc5b50ba5':
Change ram_save_block to return -1 if there are no more changes
ram: save_live_setup() we don't need to synchronize the dirty bitmap.
ram: iterate phase
ram: save_live_complete() only do one loop
ram: save_live_setup() don't need to sent pages
savevm: split save_live into stage2 and stage3
savevm: split save_live_setup from save_live_state
savevm: introduce is_active method
savevm: Refactor cancel operation in its own operation
savevm: remove SaveLiveStateHandler
savevm: remove SaveSetParamsHandler
savevm: Live migration handlers register the struct directly
savevm: Use a struct to pass all handlers
x86: Fixed incorrect segment base address addition in 64-bits mode
According to the Intel manual
"Intel® 64 and IA-32 Architectures Software Developer’s Manual
Volume 3", "3.4.4 Segment Loading Instructions in IA-32e Mode":
"When in compatibility mode, FS and GS overrides operate as defined by
32-bit mode behavior regardless of the value loaded into the upper 32
linear-address bits of the hidden descriptor register base field.
Compatibility mode ignores the upper 32 bits when calculating an effective address."
However, the code misses the 64-bit mode case, where an instruction with
address and segment size override would be translated incorrectly. For example,
inc dword ptr gs:260h[ebx*4] gets incorrectly translated to:
LINK qemu-ga
gcc: error: qga/../qapi-generated/qga-qapi-types.o: No such file or directory
gcc: error: qga/../qapi-generated/qga-qapi-visit.o: No such file or directory
gcc: error: qga/../qapi-generated/qga-qmp-marshal.o: No such file or directory
make: *** [qemu-ga] Error 1
Commit cdc976b040f2760103ce2fb49f4d504093a7163f changes the
dependencies of qemu-ga to depend "../qapi-generated/qga-qapi-types.o",
which will be expanded to "qga/../qapi-generated/qga-qapi-types.o" when
building qemu-ga.
In top-level Makefile, we defined a target "qapi-generated/qga-qapi-types.o"
which was not equal to "qga/../qapi-generated/qga-qapi-types" in the
Makefile world. So "No such file" error happened when qemu-ga was linking.
The easy approach to fix is to change the target name to
"qga/../qapi-generated/qga-qapi-types.o", but it is weird.
So, in order to solve it more graciously, I move those temporary
files(qga-qapi-*.{c,h}) qemu-ga depends on to qemu-ga/qapi-generated,
this makes dependencies more clearer.
Peter Maydell [Tue, 24 Jul 2012 14:10:27 +0000 (15:10 +0100)]
hw/escc: Drop duplicate definition of 'disabled' property
Drop a duplicate definition of the 'disabled' property from
the escc qdev property list: this redefinition is currently
effectively ignored but will become an error. (The duplication
was inadvertently introduced in 2009 in commit ec02f7dec2.)
Paolo Bonzini [Wed, 11 Jul 2012 14:40:21 +0000 (16:40 +0200)]
build: get dependency file directories from object file names
After commit dcff25f2cd8c11a9368cc2369aeb0319c32d9e26, Dependency file
are taken from the directories that have a Makefile.objs file. This is
not enough, since files can be included from other directories.
So, pick them from directories that have an object file in them.
Paolo Bonzini [Wed, 11 Jul 2012 14:40:20 +0000 (16:40 +0200)]
build: include qapi-generated/ files in qga/Makefile.objs
No reason to leave them out, and it will ensure that the dependencies
are picked up. Later we can perhaps move the files to another
directory to avoid ../ usage.
Peter Maydell [Mon, 9 Jul 2012 04:28:30 +0000 (04:28 +0000)]
vl.c: Don't print errno after failed qemu_chr_new()
The qemu_chr_new() function doesn't set errno on failure, so
don't print strerror(errno) on the error handling path when
dealing with the -serial, -parallel and -virtioconsole arguments.
This avoids nonsensical error messages like:
$ ./arm-softmmu/qemu-system-arm -serial wombat
qemu: could not open serial device 'wombat': Success
We also rephrase the message slightly to make it a little clearer
that we're expecting the name of a QEMU chr backend rather than
a host or guest serial/parallel/etc device.
For most Xtensa instructions, bit numbering is irrelevant; only the BBC
and BBS instructions assign bit numbers to values on which the processor
operates. The BBC/BBS instructions use big-endian bit ordering (0 is the
most-significant bit) on a big-endian processor configuration.
Paolo Bonzini [Fri, 27 Jul 2012 12:38:03 +0000 (14:38 +0200)]
virtio-scsi: enable MSI-X support
While virtio-scsi does support multiqueue, the default number of
interrupt vectors is not enough to actually enable usage of
multiple queues in the driver; this is because with only 2
vectors the driver will not be able to use a separate
interrupt for each request queue. Derive the desired number
of vectors from the number of request queues.
Paolo Bonzini [Fri, 27 Jul 2012 12:36:07 +0000 (14:36 +0200)]
virtio-scsi: add ioeventfd support
Probably due to bad merge months ago, virtio-scsi-pci did not have
ioeventfd support. Fix this and enable it by default, as is the
case for other virtio-pci devices.
Paolo Bonzini [Mon, 16 Jul 2012 12:18:58 +0000 (14:18 +0200)]
scsi: establish precedence levels for unit attention
When a device is resized, we will report a unit attention condition
for CAPACITY DATA HAS CHANGED. However, we should ensure that this
condition does not override a more important unit attention condition.
Cong Meng [Wed, 20 Jun 2012 06:47:10 +0000 (14:47 +0800)]
scsi: introduce hotplug() and hot_unplug() interfaces for SCSI bus
Add two interfaces hotplug() and hot_unplug() to scsi bus info.
The scsi bus can implement these two interfaces to signal the HBA driver
of guest kernel to add/remove the scsi device in question.
Paolo Bonzini [Mon, 16 Jul 2012 13:59:57 +0000 (15:59 +0200)]
scsi-disk: removable hard disks support load/eject
Support for the LOEJ bit of the START/STOP UNIT command right now is
limited to CD-ROMs. This is wrong, since removable hard disks (in the
real world: SD card readers) also support it in pretty much the same way.
Without the LOEJ bit, START/STOP UNIT does nothing for all devices.
Ronnie Sahlberg [Mon, 16 Jul 2012 06:53:28 +0000 (08:53 +0200)]
scsi-disk: Fail medium writes with proper sense for readonly LUNs
Add sense code for DATA_PROTECT/WRITE_PROTECTED and return this error
for any WRITE*/WRITE_VERIFY* calls if the device is readonly=on,
i.e. write-protected
Ronnie Sahlberg [Thu, 12 Jul 2012 06:52:47 +0000 (16:52 +1000)]
scsi-disk: improve the lba-out-of-range tests for read/write/verify
Improve the tests for the LBA to cover more cases.
For the 16 byte opcodes, the lba is a uint64, so we need to check is to
make sure that we do not wrap. For example if an opcode would specify
the LBA:0xffffffffffffffff and LEN:2 then lba+len would wrap to 1.
Also verify that ALL requested blocks are available, not just the first one.
Ronnie Sahlberg [Thu, 12 Jul 2012 00:19:38 +0000 (10:19 +1000)]
scsi-disk: rd/wr/vr-protect !=0 is an error
The QEMU SCSI emulation does not support protection information,
so any READ/WRITE/VERIFY commands that has the protect bits set to
non-zero should fail with ILLEGAL_REQUEST/INVALID_FIELD_IN_CDB
From SCSI SBC :
If the logical unit does not support protection information,
then the device server should terminate the command with CHECK CONDITION
status with the sense key set to ILLEGAL REQUEST and the additional sense
code set to INVALID FIELD IN CDB.
Signed-off-by: Ronnie Sahlberg <[email protected]>
[ Rebase after scsi_dma_reqops introduction - Paolo ] Signed-off-by: Paolo Bonzini <[email protected]>
Paolo Bonzini [Tue, 10 Jul 2012 13:02:55 +0000 (15:02 +0200)]
scsi-disk: parse MODE SELECT commands and parameters
This adds the bulk of the parsing code for MODE SELECT, including
breaking out changes to different mode pages, and checking that only
changeable values are modified.
In order to report errors correctly two passes are made through the
parameters; the first only looks for errors, the second actually
applies the changes to the mode page.
Paolo Bonzini [Thu, 12 Jul 2012 14:08:24 +0000 (16:08 +0200)]
scsi-disk: adjust offsets in MODE SENSE by 2
This will make offsets the same when implementing MODE SELECT. This is
because MODE SELECT has to deal with both 2-byte and 4-byte headers.
Unfortunately, this means that the offsets are now off by two compared
to the descriptions in the SCSI specs, which include the header.
Paolo Bonzini [Wed, 8 Feb 2012 09:32:55 +0000 (10:32 +0100)]
scsi-disk: support emulated TO_DEV requests
This adds the implementation of write_data for the emulated
command case. The first time through it asks for more data,
the second time it finishes the processing of the command.
MODE SELECT and MODE SELECT(10) can now be re-enabled, but they
will not do much.
Paolo Bonzini [Wed, 8 Feb 2012 09:26:21 +0000 (10:26 +0100)]
scsi-disk: separate read_data/write_data implementation for emulate_reqops
The previous patch only separated the send_command callback.
Use different implementations also for read_data and write_data.
The latter is still unreachable, so it aborts for now.
read_data passes the data buffer that was prepared and completes
the command.
Paolo Bonzini [Wed, 8 Feb 2012 08:53:42 +0000 (09:53 +0100)]
scsi-disk: split scsi-disk reqops
Only checks for present medium were still done in scsi_send_command
for emulated commands. So move those to scsi_disk_emulate_command
and return different SCSIReqOps depending on the kind of command.
Checks for present medium can be done unconditionally for the
scsi_disk_dma_reqops case.
Paolo Bonzini [Wed, 8 Feb 2012 08:43:52 +0000 (09:43 +0100)]
scsi-disk: move all non-DMA commands to scsi_disk_emulate_command
We want to use separate SCSIReqOps for emulated commands needing an
allocated buffer vs. those that are zerocopy when the HBA supports
S/G lists. Ensure that all of the former are in scsi_disk_emulate_command.
Commands that do not have any parameters are more similar to emulated
commands, so also move them, even if they do I/O.
Finally, MODE SELECT and MODE SELECT(10) are broken because we do not
yet support passing parameter data _to_ emulated commands, so disable
them.
Paolo Bonzini [Mon, 9 Jul 2012 10:11:04 +0000 (12:11 +0200)]
scsi-block: remove properties that are not relevant for passthrough
scsi-block is a passthrough device and does not allow customization
of vendor, product, removable, DPOFUA, block size or any other piece of
information. Thus, drop DEFINE_SCSI_DISK_PROPERTIES() from the
list of qdev properties.
Paolo Bonzini [Mon, 2 Jul 2012 15:03:55 +0000 (17:03 +0200)]
lsi: use qdev_reset_all
By first resetting the devices, lsi_soft_reset will find the queue
already cleared so there is no need to do that forcibly (which may also
leak SCSIRequests, and/or worse due to dangling references to the
lsi_request in the hba_private field).
Paolo Bonzini [Thu, 12 Jul 2012 13:02:29 +0000 (15:02 +0200)]
megasas: fix misuse of scsi_req_abort
scsi_req_abort is for terminating a command with a non-zero status.
The ABORT task management function is invoked by scsi_req_cancel.
In fact, ABORTED_COMMAND is a sense key, not a SAM status code.
Hannes Reinecke [Wed, 11 Jul 2012 11:35:16 +0000 (13:35 +0200)]
megasas: Replace trace_megasas_dcmd_dump_frame()
trace_megasas_dcmd_dump_frame() takes 9 arguments, which is
rather much. Plus the trace infrastructure doesn't support
it. As we can get the information via other means it's pointless
to have it in the driver, so rather use some proper trace
point here and remove the old one.
Anthony Liguori [Mon, 23 Jul 2012 18:15:34 +0000 (13:15 -0500)]
Merge remote-tracking branch 'stefanha/net' into staging
* stefanha/net:
remove unused QemuOpts parameter from net init functions
convert net_init_bridge() to NetClientOptions
convert net_init_tap() to NetClientOptions
convert net_init_vde() to NetClientOptions
convert net_init_socket() to NetClientOptions
convert net_init_slirp() to NetClientOptions
convert net_init_dump() to NetClientOptions
convert net_init_nic() to NetClientOptions
convert net_client_init() to OptsVisitor
hw, net: "net_client_type" -> "NetClientOptionsKind" (qapi-generated)
qapi schema: add Netdev types
qapi schema: remove trailing whitespace
qapi: introduce OptsVisitor
expose QemuOpt and QemuOpts struct definitions to interested parties
qapi: introduce "size" type
qapi: generate C types for fixed-width integers
qapi: add test case for deallocating traversal of incomplete structure
qapi: fix error propagation
MAINTAINERS: Replace net maintainer Mark McLoughlin with Stefan Hajnoczi
Anthony Liguori [Mon, 23 Jul 2012 18:15:34 +0000 (13:15 -0500)]
Merge remote-tracking branch 'stefanha/trivial-patches' into staging
* stefanha/trivial-patches:
Fix some more Qemus in documentation and help text
qdev: Fix Open Firmware comment
cpus.c: Make all_cpu_threads_idle() static
Use macro QEMU_PACKED for new packed structures
Recognize PCID feature
powerpc pci: fixed packing of ranges[]
Anthony Liguori [Mon, 23 Jul 2012 18:15:34 +0000 (13:15 -0500)]
Merge remote-tracking branch 'pmaydell/arm-devs.for-upstream' into staging
* pmaydell/arm-devs.for-upstream:
exynos4210: add Exynos4210 i2c implementation
hw/exynos4210_rtc.c: remove unnecessary code
hw/exynos4210_rtc.c: Fix calculating for value of year
hw/vexpress.c: Allow >4GB of RAM for Cortex-A15 daughterboard
hw/arm_boot.c: Support DTBs which use 64 bit addresses
device_tree: Add support for reading device tree properties
hw/arm_boot.c: Check for RAM sizes exceeding ATAGS capacity
hw/arm_boot.c: Consistently use ram_size from arm_boot_info struct
hw/arm_boot.c: Make ram_size a uint64_t
hw/pl011.c: Avoid crash on read when no chr backend present
Anthony Liguori [Mon, 23 Jul 2012 18:15:34 +0000 (13:15 -0500)]
Merge remote-tracking branch 'stefanha/tracing' into staging
* stefanha/tracing:
Update simpletrace.py for new log format
Simpletrace v2: Support multiple arguments, strings.
monitor: remove unused do_info_trace
trace: added ability to comment out events in the list
projects / qemu.git / log