VFIO_PCI_NUM_REGIONS and VFIO_PCI_NUM_IRQS should never have been
used in this manner as it locks a specific kernel implementation.
Future features may introduce new regions or interrupt entries
(VGA may add legacy ranges, AER might add an IRQ for error
signalling). Fix this before it gets us into trouble.
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Cc: qemu-stable@nongnu.org
Guests typically enable MSI-X with all of the vectors in the MSI-X
vector table masked. Only when the vector is enabled does the vector
get unmasked, resulting in a vector_use callback. These two points,
enable and unmask, correspond to pci_enable_msix() and request_irq()
for Linux guests. Some drivers rely on VF/PF or PF/fw communication
channels that expect the physical state of the device to match the
guest visible state of the device. They don't appreciate lazily
enabling MSI-X on the physical device.
To solve this, enable MSI-X with a single vector when the MSI-X
capability is enabled and immediate disable the vector. This leaves
the physical device in exactly the same state between host and guest.
Furthermore, the brief gap where we enable vector 0, it fires into
userspace, not KVM, so the guest doesn't get spurious interrupts.
Ideally we could call VFIO_DEVICE_SET_IRQS with the right parameters
to enable MSI-X with zero vectors, but this will currently return an
error as the Linux MSI-X interfaces do not allow it.
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Cc: qemu-stable@nongnu.org
Commit 667d22d1ae (qdev: move bus removal
to object_unparent) made the assumption that at unparenting time
parent_bus is not NULL. This assumption is unjustified since
object_unparent() may well be called directly after object_initialize(),
without any qdev_set_parent_bus().
This did not cause any issues yet because qdev_[try_]create() does call
qdev_set_parent_bus(), falling back to SysBus if unsupplied.
While at it, ensure that this new function uses the device_ prefix and
make the name more neutral in light of this semantic change.
Reported-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
Tested-by: Igor Mammedov <imammedo@redhat.com>
The code depends on some functions from qemu-option.o, so add
qemu-option.o to universal-obj-y to make sure it's included.
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
* kraxel/usb.75: (32 commits)
uhci: stop using portio lists
usbredir: Add support for buffered bulk input (v2)
exynos4210: Add EHCI support
usb/ehci: Add SysBus EHCI device for Exynos4210
usb/ehci: Move capsbase and opregbase into SysBus EHCI class
usb/ehci: Clean up SysBus and PCI EHCI split
xhci: call set-address with dummy usbpacket
usb-redir: Add debugging to bufpq save / restore
usbredir: Add usbredir_init_endpoints() helper
usbredir: Verify we have 32 bits bulk length cap when redirecting to xhci
usbredir: Add ep_stopped USBDevice method
usbredir: Add USBEP2I and I2USBEP helper macros
usbredir: Add an usbredir_stop_ep helper function
usb: Add an usb_device_ep_stopped USBDevice method
usb: Fix usb_ep_find_packet_by_id
hid: Change idle handling to use a timer
uhci: Maximize how many frames we catch up when behind
uhci: Limit amount of frames processed in one go
uhci: Add a QH_VALID define
uhci: Fix pending interrupts getting lost on migration
...
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Buffered bulk mode is intended for bulk *input* endpoints, where the data is
of a streaming nature (not part of a command-response protocol). These
endpoints' input buffer may overflow if data is not read quickly enough.
So in buffered bulk mode the usb-host takes care of the submitting and
re-submitting of bulk transfers.
Buffered bulk mode is necessary for reliable operation with the bulk in
endpoints of usb to serial convertors. Unfortunatelty buffered bulk input
mode will only work with certain devices, therefor this patch also adds a
usb-id table to enable it for devices which need it, while leaving the
bulk ep handling for other devices unmodified.
Note that the bumping of the required usbredir from 0.5.3 to 0.6 does
not mean that we will now need a newer usbredir release then qemu-1.3,
.pc files reporting 0.5.3 have only ever existed in usbredir builds directly
from git, so qemu-1.3 needs the 0.6 release too.
Changes in v2:
-Split of quirk handling into quirks.c
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Add EHCI USB host controller to exynos4210.
Signed-off-by: Liming Wang <walimisdev@gmail.com>
Signed-off-by: Andreas Färber <andreas.faerber@web.de>
Reviewed-by: Igor Mitsyanko <i.mitsyanko@samsung.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
It uses a different capsbase and opregbase than the Xilinx device.
Signed-off-by: Liming Wang <walimisdev@gmail.com>
Signed-off-by: Andreas Färber <andreas.faerber@web.de>
Cc: Igor Mitsyanko <i.mitsyanko@samsung.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
This allows specific derived models to use different values.
Signed-off-by: Andreas Färber <andreas.faerber@web.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
SysBus EHCI was introduced in a hurry before 1.3 Soft Freeze.
To use QOM casts in place of DO_UPCAST() / FROM_SYSBUS(), we need an
identifying type. Introduce generic abstract base types for PCI and
SysBus EHCI to allow multiple types to access the shared fields.
While at it, move the state structs being amended with macros to the
header file so that they can be embedded.
The VMSTATE_PCI_DEVICE() macro does not play nice with the QOM
parent_obj naming convention, so defer that cleanup.
Signed-off-by: Andreas Färber <andreas.faerber@web.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Due to the way devices are addressed with xhci (done by hardware, not
the guest os) there is no packet when invoking the set-address control
request. Create a dummy packet in that case to avoid null pointer
dereferences.
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
The xhci-hcd may submit bulk transfers > 65535 bytes even when not using
bulk-in pipeling, so usbredir can only be used in combination with an xhci
hcd if the client has the 32 bits bulk length capability.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
To ensure that interrupt receiving is properly stopped when the guest is
no longer interested in an interrupt endpoint.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Some usb devices (host or network redirection) can benefit from knowing when
the guest stops using an endpoint. Redirection may involve submitting packets
independently from the guest (in combination with a fifo buffer between the
redirection code and the guest), to ensure that buffers of the real usb device
are timely emptied. This is done for example for isoc traffic and for interrupt
input endpoints. But when the (re)submission of packets is done by the device
code, then how does it know when to stop this?
For isoc endpoints this is handled by detecting a set interface (change alt
setting) command, which works well for isoc endpoints. But for interrupt
endpoints currently the redirection code never stops receiving data from
the device, which is less then ideal.
However the controller emulation is aware when a guest looses interest, as
then the qh for the endpoint gets unlinked (ehci, ohci, uhci) or the endpoint
is explicitly stopped (xhci). This patch adds a new ep_stopped USBDevice
method and modifies the hcd code to call this on queue unlink / ep stop.
This makes it possible for the redirection code to properly stop receiving
interrupt input (*) data when the guest no longer has interest in it.
*) And in the future also buffered bulk input.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
usb_ep_find_packet_by_id mistakenly only checks the first packet and if that
is not a match, keeps trying the first packet! This patch fixes this.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
This leads to cleaner code in usb-hid, and removes up to a 1000 calls / sec to
qemu_get_clock_ns(vm_clock) if idle-time is set to its default value of 0.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
If somehow we've gotten behind a lot, simply skip ahead, like the ehci code
does.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Before this patch uhci would process an unlimited amount of frames when
behind on schedule, by setting the timer to a time already past, causing the
timer subsys to immediately recall the frame_timer function gain.
This would cause invalid cancellations of bulk queues when the catching up
processed more then 32 frames at a moment when the bulk qh was temporarily
unlinked (which the Linux uhci driver does).
This patch fixes this by processing maximum 16 frames in one go, and always
setting the timer one ms later, making the code behave more like the ehci
code.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Rather then using the magic 32 value in various places.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Re-arrange how we process frames / increase frnum / report pending interrupts,
to avoid a 1 ms delay in interrupt reporting to the guest. This increases
the packet throughput for cases where the guest submits a single packet,
then waits for its completion then re-submits from 500 pkts / sec to
1000 pkts / sec. This impacts for example the use of redirected / virtual
usb to serial convertors.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
ehci_raise_irq(s, USBSTS_PCD), gets applied immediately so there is no need
to call commit_irq after it.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
I tried lowering the time between raising an interrupt and rescanning the
async schedule to see if the guest has queued a new transfer before, but
that did not have any positive effect. I now believe the cause for this is
that lowering this time made it more likely to hit the 1 ms interrupt
threshold penalty for the next packet, as described in my
"ehci: Use uframe precision for interrupt threshold checking" commit.
Now that we do interrupt threshold handling with uframe precision, futher
lowering this time from .5 to .25 ms gives an extra 15% improvement in speed
(MB/s) reading from a simple USB-2.0 thumb-drive.
While at it also properly set the int_req_by_async flag for short packet
completions.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Before this patch, the following could happen:
1) Transfer completes, raises interrupt
2) .5 ms later we check if the guest has queued up any new transfers
3) We find and execute a new transfer
4) .2 ms later the new transfer completes
5) We re-run our frame_timer to write back the completion, but less then
1 ms has passed since our last run, so frindex is not changed, so the
interrupt threshold code delays the interrupt
6) 1 ms from the re-run our frame-timer runs again and finally delivers
the interrupt
This leads to unnecessary large delays of interrupts, this code fixes this
by changing frindex to uframe precision and using that for interrupt threshold
control, making the interrupt fire at step 5 for guest which have low interrupt
threshold settings (like Linux).
Note that the guest still sees the frindex move in steps of 8 for migration
compatibility.
This boosts Linux read speed of a simple cheap USB thumb drive by 6 %.
Changes in v2:
-Make the guest see frindex move in steps of 8 by modifying ehci_opreg_read,
rather then using a shadow variable
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
ehci_fill_queue assumes that there is a one on one relationship between an ep
and a qh, this patch adds a check to ensure this.
Note I don't expect this to ever trigger, this is just something I noticed
the guest might do while working on other stuff. The only way this check can
trigger is if a guest mixes in and out qtd-s in a single qh for a non
control ep.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Remove the short-circuiting of fetchqtd in fetchqh, so that the
qtd gets properly verified before completing the transaction.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
This is not allowed, except for clearing active on cancellation, so don't
warn when the new token does not have its active bit set.
This unifies the cancellation path for modified qtd-s, and prepares
ehci_verify_qtd to be used ad an extra check inside
ehci_writeback_async_complete_packet().
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Also drop the warning printf, which was there mainly because this was an
untested code path (as the previous bug fixes to it show), but that no
longer is the case now :)
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
A device reset does not affect the link state, only set_link does.
Signed-off-by: Amos Kong <akong@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Commit b9d03e352c added link
auto-negotiation emulation, it would always set link up by
callback function. Problem exists if original link status
was down, link status should not be changed in auto-negotiation.
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Amos Kong <akong@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Discard packets longer than 16384 when !SBP to match the hardware behavior.
Signed-off-by: Michael Contreras <michael@inetric.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
pc-testdev.c cannot be compiled with MinGW (and other non POSIX hosts):
CC i386-softmmu/hw/i386/../pc-testdev.o
qemu/hw/i386/../pc-testdev.c:38:22: warning: sys/mman.h: file not found
qemu/hw/i386/../pc-testdev.c: In function ‘test_flush_page’:
qemu/hw/i386/../pc-testdev.c:103: warning: implicit declaration of function ‘mprotect’
...
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
This typically reduces the size from 512 bytes to 128 bytes.
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
sys/mman.h is not needed (tested on Linux) and unavailable for MinGW,
so remove it.
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
pc_fw_add_pflash_drv() ignores qemu_find_file() failure, and happily
creates a drive without a medium.
When pc_system_flash_init() asks for its size, bdrv_getlength() fails
with -ENOMEDIUM, which isn't checked either. It fails relatively
cleanly only because -ENOMEDIUM isn't a multiple of 4096:
$ qemu-system-x86_64 -S -vnc :0 -bios nonexistant
qemu: PC system firmware (pflash) must be a multiple of 0x1000
[Exit 1 ]
Fix by handling the qemu_find_file() failure.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
