Paolo Bonzini [Wed, 22 Aug 2012 16:45:12 +0000 (18:45 +0200)]
nbd: do not leak nbd_trip coroutines when a connection is torn down
Because nbd_client_close removes the I/O handlers for the client
socket, there is no way that any suspended coroutines are restarted.
This will be a problem with the QEMU embedded NBD server, because
we will have a QMP command to forcibly close all connections with
the clients.
Instead, we can exploit the reference counting of NBDClients; shutdown the
client socket, which will make it readable and writeable. Also call the
close callback, which will release the user's reference. The coroutines
then will fail and exit cleanly, and release all remaining references,
until the last refcount finally triggers the closure of the client.
Add an explicit CPUCRISState parameter instead of relying on AREG0, and
use cpu_ld* in translation and interrupt handling. Remove AREG0 swapping
in tlb_fill(). Switch to AREG0 free mode
Natanael Copa [Wed, 12 Sep 2012 09:06:51 +0000 (09:06 +0000)]
configure: properly check if -lrt and -lm is needed
Fixes build against uClibc.
uClibc provides 2 versions of clock_gettime(), one with realtime
support and one without (this is so you can avoid linking in -lrt
unless actually needed). This means that the clock_gettime() don't
need -lrt. We still need it for timer_create() so we check for this
function in addition.
optimizer.c contains some cases were the break is appearing in both the
if and the else parts. Fix that by moving it to the outer part. Also
move some common code there.
tcg/optimize: swap brcond/setcond arguments when possible
brcond and setcond ops are not commutative, but it's easy to compute the
new condition after swapping the arguments. Try to always put the constant
argument in second position like for commutative ops, to help backends to
generate better code.
Now that there are two passes of optimization (optimize.c, liveness)
there is no point of outputing the statistics of the liveness part
only. Update the code to take into account both optimizations.
Gerd Hoffmann [Thu, 30 Aug 2012 11:05:10 +0000 (13:05 +0200)]
xhci: rework interrupt handling
Split xhci_irq_update into a function which handles intx updates
(including lowering the irq line once the guests acks the interrupt)
and one which is used for raising an irq only.
Gerd Hoffmann [Tue, 28 Aug 2012 11:38:01 +0000 (13:38 +0200)]
xhci: update port handling
This patch changes the way xhci ports are linked to USBPorts. The fixed
1:1 relationship between xhci ports and USBPorts is gone. Now each
USBPort represents a physical plug which has usually two xhci ports
assigned: one usb2 and ond usb3 port. usb devices show up at one or the
other, depending on whenever they support superspeed or not.
This patch also makes the number of usb2 and usb3 ports runtime
configurable by adding 'p2' and 'p3' properties. It is allowed to
have different numbers of usb2 and usb3 ports. Specifying p2=4,p3=2
will give you an xhci adapter which supports all speeds on physical
ports 1+2 and usb2 only on ports 3+4.
Gerd Hoffmann [Thu, 23 Aug 2012 11:26:25 +0000 (13:26 +0200)]
xhci: update register layout
Change the register layout to be a bit more sparse and also not depend
on the number of ports. Useful when for making the number of ports
runtime-configurable.
Gerd Hoffmann [Fri, 17 Aug 2012 09:04:36 +0000 (11:04 +0200)]
xhci: drop buffering
This patch splits the xhci_xfer_data function into three.
The xhci_xfer_data function used to do does two things:
(1) copy transfer data between guest memory and a temporary buffer.
(2) report transfer results to the guest using events.
Now we three functions to handle this:
(1) xhci_xfer_map creates a scatter list for the transfer and
uses that (instead of the temporary buffer) to build a
USBPacket.
(2) xhci_xfer_unmap undoes the mapping.
(3) xhci_xfer_report sends out events.
The patch also fixes reporting of transaction errors which must be
reported unconditinally, not only in case the guest asks for it
using the ISP flag.
Gerd Hoffmann [Fri, 17 Aug 2012 12:05:21 +0000 (14:05 +0200)]
xhci: rip out background transfer code
original xhci code (the one which used libusb directly) used to use
'background transfers' for iso streams. In upstream qemu the iso
stream buffering is handled by usb-host & usb-redir, so we will
never ever need this. It has been left in as reference, but is dead
code anyway. Rip it out.
Hans de Goede [Mon, 3 Sep 2012 10:04:49 +0000 (12:04 +0200)]
usb-redir: Ensure our peer has the necessary caps when redirecting to XHCI
In order for redirection to work properly when redirecting to an emulated
XHCI controller, the usb-redir-host must support both
usb_redir_cap_ep_info_max_packet_size and usb_redir_cap_64bits_ids,
reject any devices redirected to an XHCI controller when these are not
supported.
Hans de Goede [Tue, 28 Aug 2012 07:08:45 +0000 (09:08 +0200)]
usb-redir: Get rid of unused async-struct dev member
This is a preparation patch for completely getting rid of the async-packet
struct in usb-redir, instead relying on the (new) per ep queues in the
qemu usb core.
Hans de Goede [Tue, 28 Aug 2012 07:05:38 +0000 (09:05 +0200)]
usb-redir: Get rid of local shadow copy of packet headers
The shadow copy only serves as an extra check (besides the packet-id) to
ensure the packet we get back is a reply to the packet we think it is.
This check has never triggered in all the time usb-redir is in use now,
and since the verified data in the returned packet-header is not used
otherwise, removing the check does not open any possibilities for the
usbredirhost to confuse us.
This is a preparation patch for completely getting rid of the async-packet
struct in usb-redir, instead relying on the (new) per ep queues in the
qemu usb core.
Hans de Goede [Mon, 27 Aug 2012 14:33:08 +0000 (16:33 +0200)]
usb-redir: Get rid of async-struct get member
This is a preparation patch for completely getting rid of the async-packet
struct in usb-redir, instead relying on the (new) per ep queues in the
qemu usb core.
Hans de Goede [Thu, 23 Aug 2012 14:37:19 +0000 (16:37 +0200)]
usb-redir: Don't delay handling of open events to a bottom half
There is no need for this, and doing so means that a backend trying to
write immediately after an open event will see qemu_chr_be_can_write
returning 0, which not all backends handle well as there is no wakeup
mechanism to detect when the frontend does become writable.
Hans de Goede [Fri, 17 Aug 2012 15:27:08 +0000 (17:27 +0200)]
usb-redir: Never return USB_RET_NAK for async handled packets
USB_RET_NAK is not a valid response for async handled packets (and will
trigger an assert as such).
Also drop the warning when receiving a status of cancelled for packets not
cancelled by qemu itself, this can happen when a device gets unredirected
by the usbredir-host while transfers are pending.
Hans de Goede [Mon, 3 Sep 2012 10:17:48 +0000 (12:17 +0200)]
ehci: Correct a comment in fetchqtd packet processing
Since my previous comment said "Should never happen", I tried changing the
next line to an assert(0), which did not go well, which as the new comments
explains is logical if you think about it for a moment.
Hans de Goede [Mon, 3 Sep 2012 09:01:13 +0000 (11:01 +0200)]
ehci: Fix memory leak in handling of NAK-ed packets
Currently each time we try to execute a NAK-ed packet we redo
ehci_init_transfer, and usb_packet_map, re-allocing (without freeing) the
sg list every time.
This patch fixes this, it does this by introducing another async state, so
that we also properly cleanup a NAK-ed packet on cancel.
Gerd Hoffmann [Fri, 31 Aug 2012 08:44:21 +0000 (10:44 +0200)]
ehci: trace guest bugs
make qemu_queue_{cancel,reset} return the number of packets released,
so the caller can figure whenever there have been active packets even
though there shouldn't have been any. Add tracepoint to log this.
Hans de Goede [Thu, 30 Aug 2012 13:18:24 +0000 (15:18 +0200)]
ehci: Properly report completed but not yet processed packets to the guest
Reported packets which have completed before being cancelled as such to the
host. Note that the new code path this patch adds is untested since it I've
been unable to actually trigger the race which needs this code path.
Hans de Goede [Wed, 29 Aug 2012 08:37:37 +0000 (10:37 +0200)]
ehci: Validate qh is not changed unexpectedly by the guest
-combine the qh check with the check for devaddr changes
-also ensure that p gets set to NULL when the queue gets cancelled on
devaddr change, which was not done properly before this patch
Hans de Goede [Wed, 29 Aug 2012 08:12:52 +0000 (10:12 +0200)]
Revert "ehci: don't flush cache on doorbell rings."
This reverts commit 9bc3a3a216e2689bfcdd36c3e079333bbdbf3ba0, which got
added to fix an issue where the real, underlying cause was not stopping
the ep queue on an error.
Now that the underlying cause is fixed by the "usb: Halt ep queue and
cancel pending packets on a packet error" patch, the "don't flush" fix
is no longer needed.
Not only is it not needed, it causes us to see cancellations (unlinks)
done by the Linux EHCI driver too late, which in combination with the new
usb-core packet-id generation where qtd addresses are used as ids, causes
duplicate ids for in flight packets.
Hans de Goede [Mon, 3 Sep 2012 10:48:49 +0000 (12:48 +0200)]
usb-core: Allow the first packet of a pipelined ep to complete immediately
This can happen with usb-redir live-migration when the packet gets re-queued
after the migration and the original queuing from the migration source side
has already finished.
Hans de Goede [Fri, 17 Aug 2012 09:39:16 +0000 (11:39 +0200)]
usb: controllers do not need to check for babble themselves
If an (emulated) usb-device tries to write more data to a packet then
its iov len, this will trigger an assert in usb_packet_copy(), and if
a driver somehow circumvents that check and writes more data to the
iov then there is space, we have a much bigger problem then not correctly
reporting babble to the guest.
In practice babble will only happen with (real) redirected devices, and there
both the usb-host os and the qemu usb-device code already check for it.
Add ability to force enable/disable of tools build
The qemu-img, qemu-nbd and qemu-io tools are built conditionally
based on whether any softmmu target is enabled. These are useful
self-contained tools which can be used in many other scenarios.
Add new --enable-tools/--disable-tools args to configure to allow
the user to explicitly turn on / off their build. The default
behaviour is now to build these tools are all times, regardless
of whether any softmmu target is enabled
Allow passing of '--target-list=' to configure to request that
all targets are to be disabled. This allows for doing a very
fast tools-only build of things like qemu-img, qemu-io, qemu-nbd.
Paolo Bonzini [Thu, 2 Aug 2012 16:04:12 +0000 (18:04 +0200)]
RTC: Get and set time without going through s->current_tm
This patch makes rtc_set_time and rtc_set_cmos work without reading
s->current_tm. In the case of rtc_set_time I introduce a new
function that retrieves the time and stores into a given struct tm
(not hard-coded to s->current_tm). In the case of rtc_set_cmos, the
current time is similarly taken from a struct tm rather than
s->current_tm.
Paolo Bonzini [Thu, 2 Aug 2012 16:04:11 +0000 (18:04 +0200)]
RTC: Do not fire timer periodically to catch next alarm
This patch limits further the usage of a periodic timer. It computes the
time of the next alarm, and uses it to skip all intermediate occurrences
of the timer.
Base_rtc is the RTC value when the RTC was last updated.
Guest_time_now is the guest time when the access happens.
Guest_time_last_update was the guest time when the RTC was last updated.
Offset is used when divider reset happens or the set bit is toggled.
The timer is kept in order to signal interrupts, but it only needs to
run when either UF or AF is cleared. When the bits are both set, the
timer does not run.
UIP is now synthesized when reading register A. If the timer is not set,
or if there is more than one second before it (as is the case at the
end of this series), the leading edge of UIP is computed and the rising
edge occurs 220us later. If the update timer occurs within one second,
however, the rising edge of the AF and UF bits should coincide withe
the falling edge of UIP. We do not know exactly when this will happen
because there could be delays in the servicing of the timer. Hence, in
this case reading register A only computes for the rising edge of UIP,
and latches the bit until the timer is fired and clears it.
projects / qemu.git / log