Tidy up the message printed when qemu exits due to a signal, so that
it's clearer where the message is coming from and that it's not just
stray debug output.
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>
Reviewed-by: NStefan Hajnoczi <stefanha@linux.vnet.ibm.com>
Acked-by: NGleb Natapov <gleb@redhat.com>
Signed-off-by: NAurelien Jarno <aurelien@aurel32.net>

The common checksum functions were moved to net/checksum.c in commit
7200ac3c but the original net-checksum.c
was never deleted from the source tree.  Remove it now since all users
of the checksum functions link against net/checksum.o and net-checksum.c
is not even compiled anymore.
Signed-off-by: NStefan Hajnoczi <stefanha@linux.vnet.ibm.com>
Signed-off-by: NAurelien Jarno <aurelien@aurel32.net>

Rx and Tx descriptors are 16 byte aligned, so the lower bits are
ignored by real hardware. In fact, they always read back as zero on real
hardware, but probably nobody relies on that.
Signed-off-by: NKevin Wolf <mail@kevin-wolf.de>
Signed-off-by: NAurelien Jarno <aurelien@aurel32.net>

When the VM goes into stop state while there is a gdb frontend attached,
it makes sense to inform gdb about this fact and at least a bit about
the stop reason. Basically, all stops are interesting except for the
temporary VMSTOP_SAVE/LOADVM.

The patch maps the relevant VMSTOP reasons on unique and more or less
associatable signals that gdb understands.
Signed-off-by: NJan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: NAurelien Jarno <aurelien@aurel32.net>

These errors were reported by cppcheck:

tests/cris/check_openpf1.c:30: error:
Mismatching allocation and deallocation: f

tests/cris/check_openpf2.c:13: error:
Mismatching allocation and deallocation: f

tests/cris/check_stat3.c:16: error:
Buffer overrun possible for long cmd-line args

tests/cris/check_stat4.c:18: error:
Buffer overrun possible for long cmd-line args

The first two are obvious coding errors (fopen needs fclose, not close).

The last two may seem less important (nobody will start test code
with an argument of more than 1022 characters which raises a buffer
overrun). Fixing them nevertheless helps with static code checks
like those done by cppcheck.
Signed-off-by: NStefan Weil <weil@mail.berlios.de>
Signed-off-by: NEdgar E. Iglesias <edgar.iglesias@gmail.com>

- the trigger mode is edge at first
- During initializatoin, the interrupt is raised as edge which is masked.
  The corresponding bit of irr is set.
- Then the mode is switched to level and it's unmasked.
- the bit of irr is set, so the interrupt is raised repeatedly by
  ioapic_service().
- OS considers that the irq line is broken and falls back to polling mode.

This patch fixes the issues.
After raising edige, clear the bit of irr.

> Bringing up interface eth0:
> Determining IP information for eth0...irq 18: nobody cared (try booting with the "irqpoll" option)
> Pid: 4126, comm: ip Not tainted 2.6.38-rc7 #1
> Call Trace:
>  <IRQ>  [<ffffffff8105b009>] ? __report_bad_irq+0x38/0x87
>  [<ffffffff8105b177>] ? note_interrupt+0x11f/0x188
>  [<ffffffff8105bacf>] ? handle_fasteoi_irq+0xa7/0xd1
>  [<ffffffff810046ff>] ? handle_irq+0x83/0x8c
>  [<ffffffff81003eb9>] ? do_IRQ+0x48/0xaf
>  [<ffffffff81300513>] ? ret_from_intr+0x0/0xe
>  [<ffffffff81031ab8>] ? __do_softirq+0x4f/0x114
>  [<ffffffff81002d6c>] ? call_softirq+0x1c/0x28
>  [<ffffffff81004647>] ? do_softirq+0x33/0x68
>  [<ffffffff810316fb>] ? irq_exit+0x36/0x38
>  [<ffffffff81015f2c>] ? smp_apic_timer_interrupt+0x88/0x96
>  [<ffffffff81002853>] ? apic_timer_interrupt+0x13/0x20
>  <EOI>  [<ffffffff810177ed>] ? __ioapic_set_affinity+0x68/0x7c
>  [<ffffffff813000f0>] ? _raw_spin_unlock_irqrestore+0x8/0xa
>  [<ffffffff8105a84f>] ? __setup_irq+0x224/0x2cb
>  [<ffffffff8120e3c5>] ? e1000_intr+0x0/0x103
>  [<ffffffff8105a9c7>] ? request_threaded_irq+0xd1/0x114
>  [<ffffffff8120e396>] ? e1000_request_irq+0x34/0x63
>  [<ffffffff8121237d>] ? e1000_open+0x81/0x11f
>  [<ffffffff8129097c>] ? call_netdevice_notifiers+0x45/0x4a
>  [<ffffffff81290d8d>] ? __dev_open+0x97/0xc4
>  [<ffffffff8128e9c5>] ? __dev_change_flags+0xb9/0x13d
>  [<ffffffff81290cc1>] ? dev_change_flags+0x1c/0x51
>  [<ffffffff812d0542>] ? devinet_ioctl+0x26e/0x594
>  [<ffffffff812d174c>] ? inet_ioctl+0x92/0xaa
>  [<ffffffff81281d75>] ? T.1003+0x13/0x32
>  [<ffffffff81282152>] ? sock_ioctl+0x1f2/0x1ff
>  [<ffffffff810ae2d3>] ? do_vfs_ioctl+0x498/0x4e7
>  [<ffffffff81281203>] ? sock_alloc_file+0xb3/0x115
>  [<ffffffff8109f79f>] ? fd_install+0x31/0x5d
>  [<ffffffff810ae364>] ? sys_ioctl+0x42/0x65
>  [<ffffffff81001f3b>] ? system_call_fastpath+0x16/0x1b
> handlers:
> [<ffffffff8120e3c5>] (e1000_intr+0x0/0x103)
> Disabling IRQ #18
Signed-off-by: NIsaku Yamahata <yamahata@valinux.co.jp>
Signed-off-by: NAurelien Jarno <aurelien@aurel32.net>

Replace endianess -> endianness.
Signed-off-by: NStefan Weil <weil@mail.berlios.de>
Signed-off-by: NAurelien Jarno <aurelien@aurel32.net>

Add a model of the ARM Versatile Express board (with A9MPx4
daughterboard).
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>
Signed-off-by: NAurelien Jarno <aurelien@aurel32.net>

Correct the argument and return types for the float<->int conversion helper
functions so that integer arguments and return values are declared as
uint32_t/uint64_t, not float32/float64. This allows us to remove the
hand-rolled functions which were doing bitwise copies between the types
via unions.
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>
Reviewed-by: NNathan Froyd <froydnj@codesourcery.com>
Signed-off-by: NAurelien Jarno <aurelien@aurel32.net>

Use the new softfloat min/max functions to implement the Neon VMAX
and VMIN instructions. This allows us to get the right behaviour
for NaN and negative zero.
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>
Signed-off-by: NAurelien Jarno <aurelien@aurel32.net>

Add min and max operations to softfloat. This allows us to implement
propagation of NaNs and handling of negative zero correctly (unlike
the approach of having target helper routines return one of the operands
based on the result of a comparison op).
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>
Signed-off-by: NAurelien Jarno <aurelien@aurel32.net>

Implement ABD by taking the absolute value of the difference
of the operands (as the ARM ARM specifies) rather than by
flipping the order of the operands to the subtract based
on the results of a comparison. The latter approch gives
the wrong answers for some edge cases like negative zero.
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>
Signed-off-by: NAurelien Jarno <aurelien@aurel32.net>

Implementing the floating-point versions of VCLE #0 and VCLT #0 by
doing a GT comparison and inverting the result gives the wrong
result if the input is a NaN. Implement as a GT comparison with the
operands swapped instead.
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>
Signed-off-by: NAurelien Jarno <aurelien@aurel32.net>

Fix the helper functions implementing the Neon floating point comparison
ops (VCGE, VCGT, VCEQ, VACGT, VACGE) to return the right answer when
one of the values being compared is a NaN.
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>
Signed-off-by: NAurelien Jarno <aurelien@aurel32.net>

Use the softfloat make_float32 and float32_val macros to convert between
softfloat's float32 type and raw uint32_t types, rather than private
conversion functions.
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>
Reviewed-by: NNathan Froyd <froydnj@codesourcery.com>
Signed-off-by: NAurelien Jarno <aurelien@aurel32.net>

Add documentation for the usb-ccid device and accompanying two card
devices, ccid-card-emulated and ccid-card-passthru.
Signed-off-by: NAlon Levy <alevy@redhat.com>

This devices uses libcacard (internal) to emulate a smartcard conforming
to the CAC standard. It attaches to the usb-ccid bus. Usage instructions
(example command lines) are in the following patch in docs/ccid.txt. It
uses libcacard which uses nss, so it can work with both hw cards and
certificates (files).
Signed-off-by: NAlon Levy <alevy@redhat.com>

---

changes from v20->v21: (Jes Sorenson review)
 * cosmetics
 * use qemu-thread and qemu_malloc/qemu_free

changes from v19->v20:
 * checkpatch.pl

changes from v18->v19:
 * add qdev.desc
 * backend: drop the enumeration property, back to using a string one.

changes from v16->v17:
 * use PROP_TYPE_ENUM for backend

changes from v15->v16:
 * fix error reporting in initfn
 * bump copyright year
 * update copyright license

changes from v1:
 * remove stale comments, use only c-style comments
 * bugfix, forgot to set recv_len
 * change reader name to 'Virtual Reader'

client to talk to ccid-card-passthru and use smartcard on client to
perform actual operations.

v23->v24 changes: (Jes Sorensen review 2)
 * use qemu_socket instead of socket
 * use fprintf(stderr,..) for errors
 * remove unneccessary includes since using qemu_common.h

libcacard emulates a Common Access Card (CAC) which is a standard
for smartcards. It is used by the emulated ccid card introduced in
a following patch. Docs are available in docs/libcacard.txt
Signed-off-by: NAlon Levy <alevy@redhat.com>

---

changes from v24->v25:
 * Fix out of tree builds.
 * Fix build with linux-user targets.

changes from v23->v24: (Jes Sorensen review 2)
 * Makefile.target: use obj-$(CONFIG_*) +=
 * remove unrequired includes, include qemu-common before qemu-thread
  * required adding #define NO_NSPR_10_SUPPORT (harmless)

changes from v22->v23:
 * configure fixes: (reported by Stefan Hajnoczi)
  * test a = b, not a == b (second isn't portable)
  * quote $source_path in case it contains spaces
   - this doesn't really help since there are many other places
     that need similar fixes, not introduced by this patch.

changes from v21->v22:
 * fix configure to not link libcacard if nss not found
    (reported by Stefan Hajnoczi)
 * fix vscclient linkage with simpletrace backend
    (reported by Stefan Hajnoczi)
 * card_7816.c: add missing break in ERROR_DATA_NOT_FOUND
    (reported by William van de Velde)

changes from v20->v21: (Jes Sorensen review)
 * use qemu infrastructure: qemu-thread, qemu-common (qemu_malloc
  and qemu_free), error_report
 * assert instead of ASSERT
 * cosmetic fixes
 * use strpbrk and isspace
 * add --disable-nss --enable-nss here, instead of in the final patch.
 * split vscclient, passthru and docs to following patches.

changes from v19->v20:
 * checkpatch.pl

changes from v15->v16:

Build:
 * don't erase self with distclean
 * fix make clean after make distclean
 * Makefile: make vscclient link quiet

Behavioral:
 * vcard_emul_nss: load coolkey in more situations
 * vscclient:
  * use hton,ntoh
  * send init on connect, only start vevent thread on response
  * read payload after header check, before type switch
  * remove Reconnect
  * update for vscard_common changes, empty Flush implementation

Style/Whitespace:
 * fix wrong variable usage
 * remove unused variable
 * use only C style comments
  * add copyright header
  * fix tabulation
Signed-off-by: NAlon Levy <alevy@redhat.com>

libcacard: fix out of tree builds

The passthru ccid card is a device sitting on the usb-ccid bus and
using a chardevice to communicate with a remote device using the
VSCard protocol defined in libcacard/vscard_common.h

Usage docs available in following patch in docs/ccid.txt
Signed-off-by: NAlon Levy <alevy@redhat.com>

---

Changes from v23->v24:
 * fixed double license line in header.

Changes from v20->v21: (Jes Sorensen review)
 * add reference to COPYING in header
 * long comment reformatting

Changes from v19->v20:
 * checkpatch.pl

Changes from v18->v19:
 * add qdev.desc
 * remove .qdev.unplug (no hot unplug support for ccid bus)

Changes from v16->v17:
 * fix wrong cast when receiving VSC_Error
 * ccid-card-passthru: force chardev user wakeup by sending Init
   see lengthy comment below.

Changes from v15->v16:

Behavioral changes:
 * return correct size
 * return error instead of assert if client sent too large ATR
 * don't assert if client sent too large a size, but add asserts for indices to buffer
 * reset vscard_in indices on chardev disconnect
 * handle init from client
 * error if no chardev supplied
 * use ntoh, hton
 * eradicate reader_id_t
 * remove Reconnect usage (removed from VSCARD protocol)
 * send VSC_SUCCESS on card insert/remove and reader add/remove

Style fixes:
 * width of line fix
 * update copyright
 * remove old TODO's
 * update file header comment
 * use macros for debug levels
 * c++ style comment replacement
 * update copyright license
 * fix ATR size comment
 * fix whitespace in struct def
 * fix DPRINTF prefix
 * line width fix

ccid-card-passthru: force chardev user wakeup by sending Init

The problem: how to wakeup the user of the smartcard when the smartcard
device is initialized?

Long term solution: have a callback interface. This was done via
the deprecated so called chardev ioctl interface.

Short term solution: do a write. Specifically we write an Init message.
And we change the client to send it's own Init message regardless of
receiving this one. Additional Init messages will be regarded as
acceptable, the first one received after connection establishment is
the determining one wrt capabilities.

---
Signed-off-by: NAlon Levy <alevy@redhat.com>

v20->v21 changes: (Jes Sorensen review)
 * license set to 2+
 * long comment fixes, remove empty line at eof.
 * add reference to COPYING

v19->v20 changes:
 * checkpatch.pl

v15->v16 changes:

Protocol change:
 * VSCMsgInit capabilities and magic
 * removed ReaderResponse, will use Error instead with code==VSC_SUCCESS.
 * adaded Flush and FlushComplete, remove Reconnect.
 * define VSCARD_MAGIC
 * added error code VSC_SUCCESS.

Fixes:
 * update VSCMsgInit comment
 * fix message type enum
 * remove underscore from wrapping define
 * update copyright
 * updated comments.
 * Header comment updated
 * remove C++ style comment
 * fix comment for VSCMsgError
 * give names to enums in typedefs

A CCID device is a smart card reader. It is a USB device, defined at [1].
This patch introduces the usb-ccid device that is a ccid bus. Next patches will
introduce two card types to use it, a passthru card and an emulated card.

 [1] http://www.usb.org/developers/devclass_docs/DWG_Smart-Card_CCID_Rev110.
Signed-off-by: NAlon Levy <alevy@redhat.com>

---

changes from v20->v21: (Jes Sorenson review)
 * cosmetic changes - fix multi line comments.
 * reorder fields in USBCCIDState
 * add reference to COPYING
 * add --enable-smartcard and --disable-smartcard here (moved
 from last patch)

changes from v19->v20:
 * checkpatch.pl

changes from v18->v19:
 * merged: ccid.h: add copyright, fix define and remove non C89 comments
 * add qdev.desc

changes from v15->v16:

Behavioral changes:
 * fix abort on client answer after card remove
 * enable migration
 * remove side affect code from asserts
 * return consistent self-powered state
 * mask out reserved bits in ccid_set_parameters
 * add missing abRFU in SetParameters (no affect on linux guest)

whitefixes / comments / consts defines:
 * remove stale comment
 * remove ccid_print_pending_answers if no DEBUG_CCID
 * replace printf's with DPRINTF, remove DEBUG_CCID, add verbosity defines
 * use error_report
 * update copyright (most of the code is not original)
 * reword known bug comment
 * add missing closing quote in comment
 * add missing whitespace on one line
 * s/CCID_SetParameter/CCID_SetParameters/
 * add comments
 * use define for max packet size

Comment for "return consistent self-powered state":

the Configuration Descriptor bmAttributes claims we are self powered,
but we were returning not self powered to USB_REQ_GET_STATUS control message.

In practice, this message is not sent by a linux 2.6.35.10-74.fc14.x86_64
guest (not tested on other guests), unless you issue lsusb -v as root (for
example).

qemu-thread.h relies on uint64_t being defined, but doesn't include
inttypes.h explicitly. This makes it easier to use it from vscclient (part
of libcacard).

Improve the warnings we give if the user specified a combination of -net
options which don't make much sense:
 * Don't warn about anything if the config is the implicit default
   "-net user -net nic" rather than one specified by the user (this will
   only kick in for boards with no NIC or if CONFIG_SLIRP is not set)
 * Diagnose the case where the user asked for NICs which the board
   didn't instantiate (for example where the user asked for two NICs
   but the board only supports one)
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>
Signed-off-by: NAurelien Jarno <aurelien@aurel32.net>

Correct the condition determining whether we instantiate the onboard
NIC or a PCI card NIC on VersatilePB and Realview boards. This was broken
in two ways:
 (1) if the user asked for two default NICs ("-net nic -net nic") we would
crash trying to strcmp() a NULL pointer
 (2) if the user asked for two NICs explicitly of the same model as the
onboard NIC (eg "-net nic,model=smc91c111 -net nic,model=smc91c111")
we would try to instantiate two onboard NICs at the same address.
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>
Signed-off-by: NAurelien Jarno <aurelien@aurel32.net>

The problem is with definitions in hw/pcnet.c such as:

  #define CSR_CRDA(S)      ((S)->csr[28] | ((S)->csr[29] << 16))

"(S)->csr[29]" is a uint16_t, but "(S)->csr[29] << 16" gets promoted to
int, so the overall CSR_CRDA(s) is a (signed) int rather than a uint32_t.

This then gets assigned to a uint64_t using

  target_phys_addr_t crda = CSR_CRDA(s);

so when (S)->csr[29] has the high bit set, we end up with
crda=0xffffffffxxxxxxxx.

From: Michael Brown <mcb30@ipxe.org>
Signed-off-by: NAlex Williamson <alex.williamson@redhat.com>
Reviewed-by: NStefan Hajnoczi <stefanha@linux.vnet.ibm.com>
Signed-off-by: NAurelien Jarno <aurelien@aurel32.net>

Move the allocation and freeing of the TCG temp used for the address for
Neon load/store instructions so that we don't allocate the temporary
until we've done enough decoding to know that the instruction is not
an UNDEF pattern; this avoids leaking the TCG temp in these cases.
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>
Signed-off-by: NAurelien Jarno <aurelien@aurel32.net>

Fix several bugs in VLD of single element to all lanes:

The "single element to all lanes" form of VLD1 differs from those for
VLD2, VLD3 and VLD4 in that bit 5 indicates whether the loaded element
should be written to one or two Dregs (rather than being a register
stride). Handle this by special-casing VLD1 rather than trying to
have one loop which deals with both VLD1 and 2/3/4.

Handle VLD4.32 with 16 byte alignment specified, rather than UNDEFfing.

UNDEF for the invalid size and alignment combinations.
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>
Signed-off-by: NAurelien Jarno <aurelien@aurel32.net>

Signed-off-by: NTristan Gingold <gingold@adacore.com>
Reviewed-by: NMarkus Armbruster <armbru@redhat.com>
Signed-off-by: NAurelien Jarno <aurelien@aurel32.net>

There are no { name, keysym } values in name2keysym[] array for Polish
national characters so "-k pl" option has no effect.
Signed-off-by: NLeszek Grzegorek <leszek.grzegorek@gmail.com>
Signed-off-by: NAurelien Jarno <aurelien@aurel32.net>

qemu now has generic bitmap functions,
so don't redefine them in sheepdog.c,
use common header instead.  A small cleanup.

Here's only one function which is actually
used in sheepdog and gets replaced with
a generic one (simplified):

- static inline int test_bit(int nr, const volatile unsigned long *addr)
+ static inline int test_bit(int nr, const unsigned long *addr)
 {
-  return ((1UL << (nr % BITS_PER_LONG))
            & ((unsigned long*)addr)[nr / BITS_PER_LONG])) != 0;
+  return 1UL & (addr[nr / BITS_PER_LONG] >> (nr & (BITS_PER_LONG-1)));
 }

The body is equivalent, but the argument is not: there's
"volatile" in there.  Why it is used for - I'm not sure.
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>
Acked-by: NMORITA Kazutaka <morita.kazutaka@lab.ntt.co.jp>
Signed-off-by: NAurelien Jarno <aurelien@aurel32.net>

If these messages are not handled correctly the guest driver may hang.

Always mandatory:
- ABORT
- BUS DEVICE RESET

Mandatory if tagged queuing is implemented (which disks usually do):
- ABORT TAG
- CLEAR QUEUE
Signed-off-by: NBernhard Kohl <bernhard.kohl@nsn.com>
Signed-off-by: NAurelien Jarno <aurelien@aurel32.net>

I enable acpi_piix4 debug, and got the following build errors:
# make
  CC    libhw64/acpi_piix4.o
cc1: warnings being treated as errors
/home/wency/source/qemu/hw/acpi_piix4.c: In function ‘pm_ioport_write’:
/home/wency/source/qemu/hw/acpi_piix4.c:193: error: format ‘%04x’ expects type ‘unsigned int’, but argument 2 has type ‘uint64_t’
/home/wency/source/qemu/hw/acpi_piix4.c:193: error: format ‘%04x’ expects type ‘unsigned int’, but argument 3 has type ‘uint64_t’
/home/wency/source/qemu/hw/acpi_piix4.c: In function ‘pm_ioport_read’:
/home/wency/source/qemu/hw/acpi_piix4.c:219: error: format ‘%04x’ expects type ‘unsigned int’, but argument 2 has type ‘uint64_t’
make[1]: *** [acpi_piix4.o] Error 1
make: *** [subdir-libhw64] Error 2
Signed-off-by: NWen Congyang <wency@cn.fujitsu.com>
Signed-off-by: NAurelien Jarno <aurelien@aurel32.net>

Load an optional QEMU icon file. If there is no icon file named
qemu.bmp in QEMU's default search path, QEMU will run with
the usual system default icon.

A matching icon file  will be loaded and used by X Windows managers
or MS Windows while a QEMU instance is running.

SDL requires icon files in 32x32x4 bmp format.

Cc: Anthony Liguori <aliguori@us.ibm.com>
Signed-off-by: NStefan Weil <weil@mail.berlios.de>
Signed-off-by: NAurelien Jarno <aurelien@aurel32.net>

Signed-off-by: NAurelien Jarno <aurelien@aurel32.net>

Currently, the emulated pSeries machine requires the use of the
-kernel parameter in order to explicitly load a guest kernel.  This
means booting from the virtual disk, cdrom or network is not possible.

This patch addresses this limitation by inserting a within-partition
firmware image (derived from the "SLOF" free Open Firmware project).
If -kernel is not specified, qemu will now load the SLOF image, which
has access to the qemu boot device list through the device tree, and
can boot from any of the usual virtual devices.

In order to support the new firmware, an extension to the emulated
machine/hypervisor is necessary.  Unlike Linux, which expects
multi-CPU entry to be handled kexec() style, the SLOF firmware expects
only one CPU to be active at entry, and to use a hypervisor RTAS
method to enable the other CPUs one by one.

This patch also implements this 'start-cpu' method, so that SLOF can
start the secondary CPUs and marshal them into the kexec() holding
pattern ready for entry into the guest OS.  Linux should, and in the
future might directly use the start-cpu method to enable initially
disabled CPUs, but for now it does require kexec() entry.
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: NPaul Mackerras <paulus@samba.org>
Signed-off-by: NDavid Gibson <dwg@au1.ibm.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>

Shared-processor partitions are those where a CPU is time-sliced between
partitions, rather than being permanently dedicated to a single
partition.  qemu emulated partitions, since they are just scheduled with
the qemu user process, behave mostly like shared processor partitions.

In order to better support shared processor partitions (splpar), PAPR
defines the "VPA" (Virtual Processor Area), a shared memory communication
channel between the hypervisor and partitions.  There are also two
additional shared memory communication areas for specialized purposes
associated with the VPA.

A VPA is not essential for operating an splpar, though it can be necessary
for obtaining accurate performance measurements in the presence of
runtime partition switching.

Most importantly, however, the VPA is a prerequisite for PAPR's H_CEDE,
hypercall, which allows a partition OS to give up it's shared processor
timeslices to other partitions when idle.

This patch implements the VPA and H_CEDE hypercalls in qemu.  We don't
implement any of the more advanced statistics which can be communicated
through the VPA.  However, this is enough to make normal pSeries kernels
do an effective power-save idle on an emulated pSeries, significantly
reducing the host load of a qemu emulated pSeries running an idle guest OS.
Signed-off-by: NDavid Gibson <dwg@au1.ibm.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>

Usually, PAPR virtual IO devices use a virtual IOMMU mechanism, TCEs,
to mediate all DMA transfers.  While this is necessary for some sorts of
operation, it can be complex to program and slow for others.

This patch implements a mechanism for bypassing TCE translation, treating
"IO" addresses as plain (guest) physical memory addresses.  This has two
main uses:
 * Simple, but 64-bit aware programs like firmwares can use the VIO devices
without the complexity of TCE setup.
 * The guest OS can optionally use the TCE bypass to improve performance in
suitable situations.

The mechanism used is a per-device flag which disables TCE translation.
The flag is toggled with some (hypervisor-implemented) RTAS methods.
Signed-off-by: NBen Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: NDavid Gibson <dwg@au1.ibm.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>