Ian Jackson [Tue, 31 Mar 2009 15:52:24 +0000 (16:52 +0100)]
passthrough: Allow slots 1e and 1f to be used
This fixes an oversight in my recent patch "Allow any unused PCI device to
be used for pass-through" whereby attempts to use slots 1e and 1f for
pass-through would silently fail.
* This affects both static and manual selection of slots
* This affects both hot-plug and static pass-through
Ian Jackson [Tue, 31 Mar 2009 14:37:58 +0000 (15:37 +0100)]
Fixes from Christoph Egger
- xen-vl-exra.c: Only compile pci_emulation_add() w/ CONFIG_PASSTHROUGH
- Fix compiler confusion with tokens after #endif
- qemu-common.h: Fix merge botch with upstream
- qemu-xen.h: remove redundant declaration of pci_xen_platform_init
- xen_platform.c: take declaration of pci_xen_platform_init from
xen_platform.h
- xen_platform.c: Make platform_fixed_ioport_save and
platform_fixed_ioport_load static. They have no prototypes and aren't
used elsewhere
Signed-off-by: Christoph Egger <Christoph.Egger@amd.com>
Ian Jackson [Tue, 31 Mar 2009 10:45:38 +0000 (11:45 +0100)]
passthrough: Fix duplicated MAC addr when multi-port NIC assigned to HVM domain
This patch fixes duplicated MAC address when multi-port NIC is
assigned to HVM domain.
Currently multi-function device is shown as two single function
devices. e1000e driver and igb driver for linux add 1 to MAC address of
second port of multi-port NIC. But they don't add 1 to MAC address of
second port if NIC is single-function device. Because of this, the MAC
addresses of the first port and the second port are duplicated.
The patch make Header Type register passthrough-type. This means a
multi-function device is shown as two multi-function devices which
implement only one function. So e1000e driver and igb driver add 1 to
MAC address of second port.
When we use windows guest, the same issue occurs, the patch fix it.
Ian Jackson [Tue, 31 Mar 2009 10:40:30 +0000 (11:40 +0100)]
ioemu: allow xend to specify the slot for pass-through devices
Currently a slot may be specified for a hot-plug device,
but not for a pass-through device that is inserted at boot time.
This patch adds support for the latter.
The syntax is:
BUS:DEV.FUNC[@VSLOT]
e.g: 0000:00:1d:0@7
This may be important as recent changes that allow any free PCI
slot to be used for pass-through (and hotplug) may case pass-through
devices to be assigned in different locations to before. Amongst
other things, specifying the slot will allow users to move them
back, if there is a need.
There is also a xend portion of this patch, which will be posted separately.
Ian Jackson [Tue, 31 Mar 2009 10:38:55 +0000 (11:38 +0100)]
ioemu: Do slot parsing inside of next_bdf
Currently only hotplug provides vslot information from xend.
A subsequent patch will have xend provide this information
for boot-time inserted pass-through devices too.
With this in mind, this patch makes some infrastructure
to parse bdf + slot information.
Ian Jackson [Wed, 25 Mar 2009 11:38:29 +0000 (11:38 +0000)]
passthrough: fix buffer overflow of vslots
Assuming we assign n devices, strlen(direct_pci) can be 13n and the
length of the old 'vslots' is 13n/3 which is smaller than 5n+1 (1
slot_str takes 5 bytes). So we have to malloc a bigger buffer for
vslots.
Ian Jackson [Wed, 25 Mar 2009 11:35:42 +0000 (11:35 +0000)]
passthrough: Register dpci_infos.php_devs.pt_dev in a common location
Currently power_on_php_slot() sets dpci_infos.php_devs[i].pt_dev itself
after the call to register_real_device(). While pt_init relies on
power_on_php_slot() to set this pointer.
It seems sensible to use the same behaviour for both callers.
There are no other callers of register_real_device()
Ian Jackson [Tue, 24 Mar 2009 18:23:29 +0000 (18:23 +0000)]
passthrough: fix pt_chk_bar_overlap
This patch fixes pt_chk_bar_overlap.
Current pt_chk_bar_overlap does not distinguish memory resources and
io resources. They are placed in different address space. So
pt_chk_bar_overlap should distinguish them.
This patch fixes MSI/MSI-X capability structure virtualization code.
Currently, xen does not support multiple message (multiple vector).
So multiple message capable field should be emulated and fixed to 0
(single vector).
With the patch, my FC-HBA works when I assign it to guest domain where
windows 2008 runs.
In addition to this, initial values of emulated registers should be
the same with initial values defined in PCI spec. If initial values
are not defined, they should be 0. The emulated field mask and
read-only field mask are also fixed.
Ian Jackson [Tue, 24 Mar 2009 18:12:39 +0000 (18:12 +0000)]
cirrus: Do not clear vram area to 0xff when not yet allocated.
When restoring, the vram data pointer is not valid until (depending on
the version of the savefile) the guest has informed us what memory to
use. Thus vram_ptr may be NULL. In this patch we avoid attempting to
clear the video ram in this case.
Signed-off-by: Ian Jackson <Ian.Jackson@eu.citrix.com>
Ian Jackson [Tue, 24 Mar 2009 13:30:10 +0000 (13:30 +0000)]
ioemu: emulate No_Soft_Reset in PMCSR
When pci_power_mgmt=0, the No_Soft_Reset field in power
management control/status register of a PCI device needs to be
emulated and fixed to 1. This bit indicates that devices
transitioning from D3 to D0 because of PowerState commands do not
perform an internal reset.
Ian Jackson [Mon, 23 Mar 2009 17:15:12 +0000 (17:15 +0000)]
fix stubdomain after DisplayState changes
this patch fixes few stubdom issues arised after the recent DisplayState
changes (not yet pushed to qemu-xen-unstable).
The main problem is that we need to touch the newly allocated pages
before sharing them, because minios allocates on first write.
Ian Jackson [Mon, 23 Mar 2009 17:00:50 +0000 (17:00 +0000)]
Fix passthrough regression
pt_init() iterates through the PHP slots independantly of
the assignment that occurs inside __insert_to_pci_slot
which is called by register_real_device(). It assumes
that vslots are assigned in order sarting at PHP_SLOT_START.
This was valid before my change, although why it didn't take the
simpler option of just checking what value had been assigned to
pt_dev->dev.devfn in register_real_device() is a mystery to
me [Simon]. Its also a mystery to what valid circumstance could lead
to pt_init() using 0 (a.k.a. unknown?) as the vslot.
My patch made one the assumptions that pt_init() made about slot numbers
invalid. That is, they don't start at PHP_SLOT_START, they start
wherever there is a free device.
A simple solution seems to be to use the value assigned to
pt_dev->dev.devfn in register_real_device().
Ian Jackson [Mon, 23 Mar 2009 16:47:35 +0000 (16:47 +0000)]
Fixes for git ioemu tree from Christoph Egger
Make ioemu build w/o CONFIG_PASSTHROUGH
- Make ioemu build w/ TAPGIFNAME defined
- Make pci_emulation_add non-static as it is used in different files
- remove redundant variable re-declarations/re-definitions
- Make xenstore functions static. They have no prototypes and aren't used
elsewhere
- Initialize xenpv_machine correctly (xen_init_pv didn't match what
QEMUmachine expected)
- Misc compiler warning fixes
Signed-off-by: Christoph Egger <Christoph.Egger@amd.com>
Ian Jackson [Mon, 23 Mar 2009 16:35:27 +0000 (16:35 +0000)]
Correct compilation error in bdf_to_slow
This commit:
commit 3fcc8c62058a60a31d762da627d4325137eea813
make __insert_to_pci_slot idempotent Signed-off-by: Simon Horman <horms@verge.net.au>
has a trivial compilation error which I now fix.
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Ian Jackson [Mon, 23 Mar 2009 15:58:28 +0000 (15:58 +0000)]
More stubdom build fixes.
* Do not fail to build stubdom if the compiler fails to optimise out
never-called static functions containing undefined references.
* Properly disable stubdom ioemu docs build.
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Ian Jackson [Wed, 18 Mar 2009 16:32:53 +0000 (16:32 +0000)]
Post-merge compilation fixes for stubdom
Much of this is very ugly. Sadly that's a consequence of the stubdom
and upstream build systems and nothing much can be done about it.
However on the upside with this change we discontinue use of the
upstream `configure' script for stubdom (as it gives wrong answers),
in favour of a new xen-setup-stubdom script.
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Before pt_config_delete() ->
qemu_free_timer(ptdev->pm_state->pm_timer), we should invoke
qemu_del_timer(), otherwise, qemu_run_timers() would access a
qemu_free_timer()-ed timer. The below patch fixes the issue.
Ian Jackson [Mon, 16 Mar 2009 18:03:23 +0000 (18:03 +0000)]
passthrough: make management of PCI D-states by guest optional
Commit 8c771eb6294afc5b3754a9e3de51568d4e5986c2 enables the guest OS
to program D0-D3hot states of the assigned device, however,
D3hot state in some PCI devices causes the failure of domain
creation/destruction.
Ian Jackson [Mon, 16 Mar 2009 17:25:35 +0000 (17:25 +0000)]
xenfb shared buffer
After the recent DisplayState changes is now possible to share the xenfb
backend buffer with the display frontend (sdl, vnc), avoid a memcpy for
each screen update.
Ian Jackson [Mon, 16 Mar 2009 16:52:21 +0000 (16:52 +0000)]
qemu-dm: Provide bounce-buffer based cpu_physical_memory_map
The xc_map_foreign_batch version was slow (and also apparently buggy).
The bounce buffer version works nicely, although we have to
clone-and-hack a bunch of stuff from exec.c.
Signed-off-by: Ian Jackson <Ian.Jackson@eu.citrix.com>
Ian Jackson [Mon, 16 Mar 2009 16:50:49 +0000 (16:50 +0000)]
Fix option parsing; override upstream `-d' option for `domain'.
Pending change to xend to pass arguments differently, we need to
suppress the -d logfile parameter. Also we accidentally dropped
HAS_ARG from -vcpus during the merge.
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Ian Jackson [Tue, 10 Mar 2009 18:17:31 +0000 (18:17 +0000)]
introducing a DisplayAllocator interface
This patch introduces a new DisplayAllocator interface to allow
frontends to allocate the DisplaySurface for the graphic device to use.
At the moment it is used only by xenfbfront, to be able to page align
the displaysurface and also keep the same backing data across multiple
DisplaySurface resize.
Howeveri in the near future it is going to be a useful improvement for
SDL as well.
Ian Jackson [Tue, 10 Mar 2009 18:16:52 +0000 (18:16 +0000)]
fix xenfb frontend and backend
Due to the DisplayState changes the initialization order is changed
as well.
Now graphic devices are the first to be initialized and only after them
any display frontend is initialized.
Obviously this change has a direct consequence on the xenfb backend and
frontend and this patch takes care of adapting them to the new
initialization order.
Ian Jackson [Tue, 10 Mar 2009 18:13:07 +0000 (18:13 +0000)]
fix screendump (Stefano Stabellini)
this patch fixes the screendump functionality that was recently broken;
it must be applied *after* PATCH 5, 6 and 7 of the original displaystate
change patch series.
In fact the other patches make much easier to solve the screendump
problem because they make the console switching mechanism more robust.
This patch changes the graphical_console_init function to return an
allocated DisplayState instead of a QEMUConsole.
This patch contains just the graphical_console_init change and few other
modifications mainly in console.c and vl.c.
It was necessary to move the display frontends (e.g. sdl and vnc)
initialization after machine->init in vl.c.
This patch does *not* include any required changes to any device, these
changes come with the following patches.
Patch 6/7
This patch changes the QEMUMachine init functions not to take a
DisplayState as an argument because is not needed any more;
In few places the graphic hardware initialization function was called
only if DisplayState was not NULL, now they are always called.
Apart from these cases, the rest are all mechanical substitutions.
Patch 7/7
This patch updates the graphic device code to use the new
graphical_console_init function.
As for the previous patch, in few places graphical_console_init was called
only if DisplayState was not NULL, now it is always called.
Apart from these cases, the rest are all mechanical substitutions.
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com> Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
As adapted by Stefano to qemu-xen-unstable in his posting:
[Xen-devel] [PATCH 6 of 13] graphical_console_init change
which also contains several following fixes.
Ian Jackson [Tue, 10 Mar 2009 18:04:43 +0000 (18:04 +0000)]
exploiting the new interface in vnc.c (Stefano Stabellini)
This patch exploits the new DisplaySurface and PixelFormat structures in
vnc, making the code easier to read allowing further improvements.
Compared to the last version I fixed a bug that prevented the hextile
encoding from working properly.
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com> Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
As adapted by Stefano to qemu-xen-unstable in his posting:
[Xen-devel] [PATCH 5 of 13] exploiting the new interface in vnc.c
This patch changes the DisplayState interface adding support for
multiple frontends at the same time (sdl and vnc) and implements most
of the benefit of the shared_buf patch without the added complexity.
Currently DisplayState is managed by sdl (or vnc) and sdl (or vnc) is
also responsible for allocating the data and setting the depth.
Vga.c (or another backend) will do any necessary conversion.
The idea is to change it so that is vga.c (or another backend) together
with console.c that fully manage the DisplayState interface allocating
data and setting the depth (either 16 or 32 bit, if the guest uses a
different resolution or is in text mode, vga.c (or another backend) is
in charge of doing the conversion seamlessly).
The other idea is that DisplayState supports *multiple* frontends
like sdl and vnc; each of them can register some callbacks to be called
when a display event occurs.
The interesting changes are:
- the new structures and related functions in console.h and console.c
in particular the following functions are very helpful to manage a
DisplaySurface:
- console_select and qemu_console_resize in console.c
this two functions manage multiple consoles on a single host display
- moving code around in hw/vga.c
as for the shared_buf patch this is necessary to be able to handle a dynamic
DisplaySurface bpp
- changes to vga_draw_graphic in hw/vga.c
this is the place where the DisplaySurface buffer is shared with the
videoram, when possible;
Compared to the last version the only changes are:
- do not remove support to dpy_copy in cirrus_vga
- change the name of the displaysurface handling functions
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com> Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
As adapted by Stefano to qemu-xen-unstable in his posting:
[Xen-devel] [PATCH 4 of 13] DisplayState interface change
Ian Jackson [Tue, 10 Mar 2009 17:59:47 +0000 (17:59 +0000)]
remove bgr (Stefano Stabellini)
Do not handle bgr host displays in the backends.
Right now a bgr flag exists so that sdl can set it, if the SDL_Surface
is bgr.
Afterwards the graphic device (e.g. vga.c) does the needed conversion.
With this patch series is sdl that is responsible for rendering the format
provided by the graphic device that must provide a DisplaySurface
(ds->surface) in 16 or 32 bpp, rgb.
Afterwards sdl creates a SDL_Surface from the given DisplaySurface and
blits it into the main SDL_Surface using SDL_BlitSurface.
Everything is handled by sdl transparently, because SDL_BlitSurface is
perfectly capable of handling bgr displays by itself.
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com> Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Patch adapted for qemu-xen-unstable by Stefano Stabellini. Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com> Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Patch adapted for qemu-xen-unstable by Stefano Stabellini. Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Also additional fixes supplied by Stefano in his message
[Xen-devel] [PATCH 2 of 13] Introduce accessors for DisplayState Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Ian Jackson [Tue, 10 Mar 2009 17:56:42 +0000 (17:56 +0000)]
Implement "info chardev" command. (Gerd Hoffmann)
This patch makes qemu keep track of the character devices in use and
implements a "info chardev" monitor command to print a list.
qemu_chr_open() sticks the devices into a linked list now. It got a new
argument (label), so there is a name for each device. It also assigns a
filename to each character device. By default it just copyes the
filename passed in. Individual drivers can fill in something else
though. qemu_chr_open_pty() sets the filename to name of the pseudo tty
allocated.
Ian Jackson [Mon, 2 Mar 2009 11:21:51 +0000 (11:21 +0000)]
Override default cache mode for disk images to write-back
Upstream qemu changed the default cache mode to write-through (ie,
O_DSYNC) which is much slower. We do not need this as we have
explicit control of cacheing with the IDE cache control commands.
Original patch by Yang Zhang modified by Ian Jackson.
Signed-off-by: Yang Zhang <yang.zhang@intel.com> Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Ian Jackson [Mon, 2 Mar 2009 11:15:37 +0000 (11:15 +0000)]
wake up PCI passthru device on boot
Commit 8c771eb6294afc5b3754a9e3de51568d4e5986c2 enables a guest
OS to program D0-D3hot states, thus PCI passsthru devices transit
into D3hot state when the guest Windows shutdown.
That disables reboot from passthru devices because the device is
not active in the POST phase. Actually, the option ROM can't be read
from the guest bios.
Ian Jackson [Mon, 2 Mar 2009 11:13:44 +0000 (11:13 +0000)]
passthrough: MSI-X mask bit acceleration
Read MSI-X mask bit directly from the device, since buffered version
may not be up-to-date when MSI-X mask bit interception is working.
Also rebind every MSI-X vector on guest PCI BAR rebalancing so that
MSI-X mask bit intercept handler can get the correct gpa
[ Also, fix declaration of pt_msix_update_remap in pt-msi.h, which
was misspelled pt_msi_update_remap. -iwj ]
Signed-off-by: Qing He <qing.he@intel.com> Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
The anonymouys structure inside PHPSlots only has one element,
so things can be simplified by moving the element into PHPSlots
and removing the anonymous structure.
Ian Jackson [Mon, 23 Feb 2009 11:33:56 +0000 (11:33 +0000)]
Cleanup PCI passthrough code
The patch cleanups the code of PCI passthrough.
- Use LIST_FOREACH(). Don't use lh_first, le_next directly.
- Use pci_{read, write}_block instead of "switch(len) case [124]
pci_{read, write}_{byte, word, long}();".
- Eliminate duplicate codes using PT_MEARGE_VALUE macro.
- Define PCI_ERR_* macro in pass-through.h if libpci is old.
- Remove the unreasonable loop from pt_aer_reg_{save,restore}.
- Enable pt_aer_reg_{save,restore}, even if libpci is old.
- Fix ro_mask and remove unnecessary pt_xxx_reg_write functions.
- Add "Error:" or "Warning:" to messages.
- Remove verbose messages.
Trailing whitespace does not survive some people's mailers, with the
result that their patches do not apply. Therefore in this commit I'm
removing it from the pass-through code, which is currently entirely in
the Xen tree.
This change is the result of this rune:
perl -i~ -pe 's/[\t ]*$//' hw/pass-through.[ch] hw/pt-msi.[ch]
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Ian Jackson [Thu, 19 Feb 2009 17:55:54 +0000 (17:55 +0000)]
fix raw_aio_remove list walk bug
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Cherry picked and cross-ported by Ian Jackson
from Stefano's submission to qemu-devel, 18 Feb 2009 16:56:16 GMT.
Ian Jackson [Fri, 6 Feb 2009 16:16:06 +0000 (16:16 +0000)]
hw/pass-through.c: workaround for old libpci
Old versions of libpci (including the ones on the automatic tests
which control Xen staging propagation) do not define
PCI_LIB_VERSION or the PCI_ERR_{UNCOR_MASK,...} constants.
This means that change 8c771eb6294afc5b3754a9e3de51568d4e5986c2 breaks
the build. In this changeset I apply what is intended to be a
workaround for this problem but it may not be completely correct; this
is therefore perhaps an interim fix.
The potential problem is that the save/restore of some PCI passthrough
error handling registers (across suspend/resume) may not work properly
with old versions of libpci. However non-passthrough and non-suspect
use cases should now be fine.
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Ian Jackson [Thu, 5 Feb 2009 15:54:13 +0000 (15:54 +0000)]
Enable guest OS to program D0-D3hot states of an assigned device
This patch enables guest OS to program D0-D3hot states of assigned
device.
This patch is revised version based on the review comments.
- Use LIST_FOREACH(). Don't use lh_first, le_next directly.
- Use pci_{read, write}_block instead of "switch(len) case [124]
pci_{read, write}_{byte, word, long}():".
Ian Jackson [Thu, 5 Feb 2009 14:45:25 +0000 (14:45 +0000)]
fix raw_aio_read/write error handling
Currently when qemu_paio_read or qemu_paio_write return an error we call
qemu_aio_release without removing the request from the list.
I know that in the current implementation qemu_paio_write/read don't return
any error, but still the behavior is wrong, especially considering
that the implementation of these two functions is likely to change in is
the future.
This patch fixes the problem adding a raw_aio_remove function that
removes the callback from the queue and also calls qemu_aio_release.
raw_aio_remove is called by raw_aio_read, raw_aio_write and
raw_aio_cancel.
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com> Backported-by: Ian Jackson <ian.jackson@eu.citrix.com>
projects / qemu-xen-3.4-testing.git / log