]> git.proxmox.com Git - qemu.git/log
qemu.git
13 years agolinux-user: Define AT_RANDOM to support target stack protection mechanism.
Laurent ALFONSI [Mon, 20 Jun 2011 06:43:13 +0000 (08:43 +0200)]
linux-user: Define AT_RANDOM to support target stack protection mechanism.

The dynamic linker from the GNU C library v2.10+ uses the ELF
auxiliary vector AT_RANDOM [1] as a pointer to 16 bytes with random
values to initialize the stack protection mechanism.  Technically the
emulated GNU dynamic linker crashes due to a NULL pointer
derefencement if it is built with stack protection enabled and if
AT_RANDOM is not defined by the QEMU ELF loader.

[1] This ELF auxiliary vector was introduced in Linux v2.6.29.

This patch can be tested with the code above:

    #include <elf.h>       /* Elf*_auxv_t, AT_RANDOM, */
    #include <stdio.h>     /* printf(3), */
    #include <stdlib.h>    /* exit(3), EXIT_*, */
    #include <stdint.h>    /* uint8_t, */
    #include <string.h>    /* memcpy(3), */

    #if defined(__LP64__) || defined(__ILP64__) || defined(__LLP64__)
    #    define Elf_auxv_t Elf64_auxv_t
    #else
    #    define Elf_auxv_t Elf32_auxv_t
    #endif

    main(int argc, char* argv[], char* envp[])
    {
        Elf_auxv_t *auxv;

        /* *envp = NULL marks end of envp. */
        while (*envp++ != NULL);

        /* auxv->a_type = AT_NULL marks the end of auxv. */
        for (auxv = (Elf_auxv_t *)envp; auxv->a_type != AT_NULL; auxv++) {
            if (auxv->a_type == AT_RANDOM) {
                int i;
                uint8_t rand_bytes[16];

                printf("AT_RANDOM is: 0x%x\n", auxv->a_un.a_val);
                memcpy(rand_bytes, (const uint8_t *)auxv->a_un.a_val, sizeof(rand_bytes));
                printf("it points to: ");
                for (i = 0; i < 16; i++) {
                    printf("0x%02x ", rand_bytes[i]);
                }
                printf("\n");
                exit(EXIT_SUCCESS);
            }
        }
        exit(EXIT_FAILURE);
    }

Changes introduced in v2 and v3:

    * Fix typos + thinko (AT_RANDOM is used for stack canary, not for
      ASLR)

    * AT_RANDOM points to 16 random bytes stored inside the user
      stack.

    * Add a small test program.

Signed-off-by: Cédric VINCENT <cedric.vincent@st.com>
Signed-off-by: Laurent ALFONSI <laurent.alfonsi@st.com>
Signed-off-by: Riku Voipio <riku.voipio@iki.fi>
13 years agolinux-user: add pselect6 syscall support
Mike Frysinger [Fri, 3 Jun 2011 21:01:49 +0000 (17:01 -0400)]
linux-user: add pselect6 syscall support

Some architectures (like Blackfin) only implement pselect6 (and skip
select/newselect).  So add support for it.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Riku Voipio <riku.voipio@iki.fi>
13 years agolinux-user: Fix the computation of the requested heap size
vincent [Tue, 14 Jun 2011 21:56:33 +0000 (21:56 +0000)]
linux-user: Fix the computation of the requested heap size

There were several remaining bugs in the previous implementation of
do_brk():

    1. the value of "new_alloc_size" was one page too large when the
       requested brk was aligned on a host page boundary.

    2. no new pages should be (re-)allocated when the requested brk is
       in the range of the pages that were already allocated
       previsouly (for the same purpose).  Technically these pages are
       never unmapped in the current implementation.

The problem/fix can be reproduced/validated with the test-suite above:

    #include <unistd.h>       /* syscall(2),      */
    #include <sys/syscall.h>  /* SYS_brk,         */
    #include <stdio.h>        /* puts(3),         */
    #include <stdlib.h>       /* exit(3), EXIT_*, */
    #include <stdint.h>       /* uint*_t,         */
    #include <sys/mman.h>     /* mmap(2), MAP_*,  */
    #include <string.h>       /* memset(3), */

    int main()
    {
        int exit_status = EXIT_SUCCESS;
        uint8_t *current_brk = 0;
        uint8_t *initial_brk;
        uint8_t *new_brk;
        uint8_t *old_brk;
        int failure = 0;
        int i;

        void test_brk(int increment, int expected_result) {
            new_brk = (uint8_t *)syscall(SYS_brk, current_brk + increment);
            if ((new_brk == current_brk) == expected_result)
                failure = 1;
            current_brk = (uint8_t *)syscall(SYS_brk, 0);
        }

        void test_result() {
            if (!failure)
                puts("OK");
            else {
                puts("failure");
                exit_status = EXIT_FAILURE;
            }
        }

        void test_title(const char *title) {
            failure = 0;
            printf("%-45s : ", title);
            fflush(stdout);
        }

        test_title("Initialization");
        test_brk(0, 1);
        initial_brk = current_brk;
        test_result();

        test_title("Don't overlap \"brk\" pages");
        test_brk(HOST_PAGE_SIZE, 1);
        test_brk(HOST_PAGE_SIZE, 1);
        test_result();

        /* Preparation for the test "Re-allocated heap is initialized".  */
        old_brk = current_brk - HOST_PAGE_SIZE;
        memset(old_brk, 0xFF, HOST_PAGE_SIZE);

        test_title("Don't allocate the same \"brk\" page twice");
        test_brk(-HOST_PAGE_SIZE, 1);
        test_brk(HOST_PAGE_SIZE, 1);
        test_result();

        test_title("Re-allocated \"brk\" pages are initialized");
        for (i = 0; i < HOST_PAGE_SIZE; i++) {
            if (old_brk[i] != 0) {
                printf("(index = %d, value = 0x%x) ", i, old_brk[i]);
                failure = 1;
                break;
            }
        }
        test_result();

        test_title("Don't allocate \"brk\" pages over \"mmap\" pages");
        new_brk = mmap(current_brk, HOST_PAGE_SIZE / 2, PROT_READ, MAP_PRIVATE | MAP_ANONYMOUS | MAP_FIXED, -1, 0);
        if (new_brk == (void *) -1)
            puts("unknown");
        else {
            test_brk(HOST_PAGE_SIZE, 0);
            test_result();
        }

        test_title("All \"brk\" pages are writable (please wait)");
        if (munmap(current_brk, HOST_PAGE_SIZE / 2) != 0)
            puts("unknown");
        else {
            while (current_brk - initial_brk < 2*1024*1024*1024UL) {
                old_brk = current_brk;

                test_brk(HOST_PAGE_SIZE, -1);
                if (old_brk == current_brk)
                    break;

                for (i = 0; i < HOST_PAGE_SIZE; i++)
                    old_brk[i] = 0xAA;
            }
            puts("OK");
        }

        test_title("Maximum size of the heap > 16MB");
        failure = (current_brk - initial_brk) < 16*1024*1024;
        test_result();

        exit(exit_status);
    }

Changes introduced in patch v2:

    * extend the "brk" test-suite embedded within the commit message;

    * heap contents have to be initialized to zero, this bug was
      exposed by "tst-calloc.c" from the GNU C library;

    * don't [try to] allocate a new host page if the new "brk" is
      equal to the latest allocated host page ("brk_page"); and

    * print some debug information when DEBUGF_BRK is defined.

Signed-off-by: Cédric VINCENT <cedric.vincent@st.com>
Reviewed-by: Christophe Guillon <christophe.guillon@st.com>
Cc: Riku Voipio <riku.voipio@iki.fi>
Signed-off-by: Riku Voipio <riku.voipio@iki.fi>
13 years agom68k-semi.c: Use correct check for failure of do_brk()
Peter Maydell [Mon, 18 Apr 2011 15:34:26 +0000 (16:34 +0100)]
m68k-semi.c: Use correct check for failure of do_brk()

In the m68k semihosting implementation of HOSTED_INIT_SIM, use the correct
check for whether do_brk() has failed -- it does not return -1 but the
previous value of the break limit.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@iki.fi>
13 years agoarm-semi.c: Use correct check for failure of do_brk()
Peter Maydell [Mon, 18 Apr 2011 15:34:25 +0000 (16:34 +0100)]
arm-semi.c: Use correct check for failure of do_brk()

In the ARM semihosting implementation of SYS_HEAPINFO, use the correct
check for whether do_brk() has failed -- it does not return -1 but the
previous value of the break limit.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@iki.fi>
13 years agolinux-user: Don't use MAP_FIXED in do_brk()
Peter Maydell [Mon, 18 Apr 2011 15:34:24 +0000 (16:34 +0100)]
linux-user: Don't use MAP_FIXED in do_brk()

Since mmap() with MAP_FIXED will map over the top of existing mappings,
it's a bad idea to use it to implement brk(), because brk() with a
large size is likely to overwrite important things like qemu itself
or the host libc. So we drop MAP_FIXED and handle "mapped but at
different address" as an error case instead.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@iki.fi>
13 years agolinux-user: Handle images where lowest vaddr is not page aligned
Peter Maydell [Tue, 17 May 2011 12:34:46 +0000 (13:34 +0100)]
linux-user: Handle images where lowest vaddr is not page aligned

Fix a bug in the linux-user ELF loader code where it was not correctly
handling images where the lowest vaddr to be loaded was not page aligned.
The problem was that the code to probe for a suitable guest base address
was changing the 'loaddr' variable (by rounding it to a page boundary),
which meant that the load bias would then be incorrectly calculated
unless loaddr happened to already be page-aligned.

Binaries generated by gcc with the default linker script do start with
a loadable segment at a page-aligned vaddr, so were unaffected. This
bug was noticed with a binary created by the Google Go toolchain for ARM.

We fix the bug by refactoring the "probe for guest base" code out into
its own self-contained function.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@iki.fi>
13 years agolinux-user: Fix the load of ELF files that have no "useful" symbol
Cédric VINCENT [Wed, 1 Jun 2011 12:36:38 +0000 (14:36 +0200)]
linux-user: Fix the load of ELF files that have no "useful" symbol

This patch fixes a "double free()" due to "realloc(syms, 0)" in the
loader when the ELF file has no "useful" symbol, as with the following
example (compiled with "sh4-linux-gcc -nostdlib"):

    .text
    .align 1
    .global _start
    _start:
        mov     #1, r3
        trapa   #40     // syscall(__NR_exit)
        nop

The bug appears when the log (option "-d") is enabled.

Signed-off-by: Cédric VINCENT <cedric.vincent@st.com>
Signed-off-by: Yves JANIN <yves.janin@st.com>
Signed-off-by: Riku Voipio <riku.voipio@iki.fi>
Reviewed-by: Richard Henderson <rth@twiddle.net>
13 years agoDon't translate pointer when in restore_sigcontext
Mike McCormack [Wed, 1 Jun 2011 06:14:37 +0000 (15:14 +0900)]
Don't translate pointer when in restore_sigcontext

Fixes crash in i386 when user emulation base address is non-zero.

21797 rt_sigreturn(8,1082124603,1,0,1082126048,1082126248)Exit reason and status: signal 11

Signed-off-by: Mike McCormack <mj.mccormack@samsung.com>
Signed-off-by: Riku Voipio <riku.voipio@iki.fi>
13 years agohw/9118.c: Implement active-low interrupt support
Peter Maydell [Thu, 26 May 2011 16:34:11 +0000 (17:34 +0100)]
hw/9118.c: Implement active-low interrupt support

The 9118 ethernet controller interrupt line is active low unless
the IRQ config register is programmed to set both the IRQ_POL
(polarity: active-high) and IRQ_TYPE (type: push-pull) bits:
implement support for inverting the irq output in other configurations.
This also requires that we support setting the bits in the first
place, and that we correctly preserve them across software reset.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Andrzej Zaborowski <andrew.zaborowski@intel.com>
13 years agoIntroduce format string for pid_t
Andreas Färber [Thu, 2 Jun 2011 17:58:06 +0000 (19:58 +0200)]
Introduce format string for pid_t

BeOS and Haiku on i386 use long for 32-bit types, including pid_t.
Using %d with pid_t therefore results in a warning.

Unfortunately POSIX:2008 does not define a PRId* string for pid_t.

In some places pid_t was previously casted to long and %ld hardcoded.
The predecessor of this patch added another upcast for the simpletrace
filename but was not applied to date.

Since new uses of pid_t with %d keep creeping in, let's instead define
an OS-dependent format string and use that consistently.

Cc: Stefan Hajnoczi <stefanha@gmail.com>
Cc: Blue Swirl <blauwirbel@gmail.com>
Cc: Ingo Weinhold <ingo_weinhold@gmx.de>
Cc: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Andreas Färber <andreas.faerber@web.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
13 years agosmc91c111: qdevify reset
Juha Riihimäki [Tue, 31 May 2011 16:48:23 +0000 (17:48 +0100)]
smc91c111: qdevify reset

Register the smc91c111 reset function as a qdev reset function.

Signed-off-by: Juha Riihimäki <juha.riihimaki@nokia.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
13 years agoMerge branch 'cocoa-for-upstream' of git://repo.or.cz/qemu/afaerber
Blue Swirl [Wed, 15 Jun 2011 18:31:56 +0000 (18:31 +0000)]
Merge branch 'cocoa-for-upstream' of git://repo.or.cz/qemu/afaerber

* 'cocoa-for-upstream' of git://repo.or.cz/qemu/afaerber:
  Darwin: Fix compilation warning regarding the deprecated daemon() function
  cocoa: Avoid warning related to multiple handleEvent: definitions
  cocoa: Revert dependency on VNC
  cocoa: Provide central qemu_main() prototype
  Fix libfdt warnings on Darwin
  configure: Fix check for fdatasync()
  Remove warning in printf due to type mismatch
  Cocoa: avoid displaying window when command-line contains '-h' or '-help'
  Fix compilation warning due to incorrectly specified type
  cocoa: do not create a spurious window for -version

13 years agoCommand line support for altering the log file location
Matthew Fernandez [Tue, 7 Jun 2011 16:32:40 +0000 (16:32 +0000)]
Command line support for altering the log file location

Add command line support for logging to a location other than /tmp/qemu.log.

With logging enabled (command line option -d), the log is written to
the hard-coded path /tmp/qemu.log. This patch adds support for writing
the log to a different location by passing the -D option.

Signed-off-by: Matthew Fernandez <matthew.fernandez@gmail.com>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
13 years agoMerge remote-tracking branch 'alon/pull-libcacard-1' into staging
Anthony Liguori [Wed, 15 Jun 2011 14:03:49 +0000 (09:03 -0500)]
Merge remote-tracking branch 'alon/pull-libcacard-1' into staging

13 years agoMerge remote-tracking branch 'qemu-kvm/uq/master' into staging
Anthony Liguori [Wed, 15 Jun 2011 14:03:33 +0000 (09:03 -0500)]
Merge remote-tracking branch 'qemu-kvm/uq/master' into staging

13 years agoMerge remote-tracking branch 'kraxel/usb.16' into staging
Anthony Liguori [Wed, 15 Jun 2011 14:03:22 +0000 (09:03 -0500)]
Merge remote-tracking branch 'kraxel/usb.16' into staging

13 years agoMerge remote-tracking branch 'kwolf/for-anthony' into staging
Anthony Liguori [Wed, 15 Jun 2011 14:02:18 +0000 (09:02 -0500)]
Merge remote-tracking branch 'kwolf/for-anthony' into staging

13 years agoAllow nested qemu_bh_poll() after BH deletion
Kevin Wolf [Tue, 7 Jun 2011 15:51:21 +0000 (17:51 +0200)]
Allow nested qemu_bh_poll() after BH deletion

Without this, qemu segfaults when a BH handler first deletes its BH and
then calls another function which involves a nested qemu_bh_poll() call.

This can be reproduced by generating an I/O error (e.g. with blkdebug) on
an IDE device and using rerror/werror=stop to stop the VM. When continuing
the VM, qemu segfaults.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
13 years agoide: Clear error_status after restarting flush
Kevin Wolf [Fri, 10 Jun 2011 14:32:13 +0000 (16:32 +0200)]
ide: Clear error_status after restarting flush

Clearing the error status flag was missing for restarting flushes. Now that the
error status is separate from the BM status register, we can simply set it to 0
after restarting the request. This ensures that we never forget to clear a bit.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
13 years agoide: add TRIM support
Christoph Hellwig [Thu, 19 May 2011 08:58:19 +0000 (10:58 +0200)]
ide: add TRIM support

Add support for TRIM sub function of the data set management command,
and wire it up to the qemu discard infrastructure.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
13 years agoide: allow other dma comands than read and write
Christoph Hellwig [Thu, 19 May 2011 08:58:09 +0000 (10:58 +0200)]
ide: allow other dma comands than read and write

Replace the is_read flag with a dma_cmd flag to allow the dma and
restart logic to handler other commands like TRIM.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
13 years agomake dma_bdrv_io available to drivers
Christoph Hellwig [Thu, 19 May 2011 08:57:59 +0000 (10:57 +0200)]
make dma_bdrv_io available to drivers

Make dma_bdrv_io available for drivers, and pass an explicit I/O function
instead of hardcoding bdrv_aio_readv/bdrv_aio_writev.  This is required
to implement non-READ/WRITE dma commands in the ide driver, e.g. the
upcoming TRIM support.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
13 years agoide: Add forgotten VMSTATE_END_OF_LIST in subsection
Kevin Wolf [Thu, 9 Jun 2011 12:15:18 +0000 (14:15 +0200)]
ide: Add forgotten VMSTATE_END_OF_LIST in subsection

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
13 years agoide: Fix ide_drive_pio_state_needed()
Kevin Wolf [Thu, 9 Jun 2011 12:12:57 +0000 (14:12 +0200)]
ide: Fix ide_drive_pio_state_needed()

When a failed PIO request caused the VM to stop, we still need to transfer the
PIO state even though DRQ=0 at this point.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
13 years agoide: Split error status from status register
Kevin Wolf [Thu, 26 May 2011 13:00:34 +0000 (15:00 +0200)]
ide: Split error status from status register

When adding the werror=stop mode, some flags were added to s->status
which are used to determine what kind of operation should be restarted
when the VM is continued.

Unfortunately, it turns out that s->status is in fact a device register
and as such is visible to the guest (some of the abused bits are even
writable for the guest).

For migration we keep on using the old VMState field (renamed to
migration_compat_status) if the status register doesn't use any of the
previously abused bits. If it does, we use a subsection with a clean copy of
the status register.

The error status is always sent in a subsection if there is any error. It can't
use the old field because errors happen even without PCI.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
13 years agoqcow2: Fix in-flight list after qcow2_cache_put failure
Kevin Wolf [Tue, 24 May 2011 14:40:02 +0000 (16:40 +0200)]
qcow2: Fix in-flight list after qcow2_cache_put failure

If qcow2_cache_put returns an error during cluster allocation and the
allocation fails, it must be removed from the list of in-flight allocations.
Otherwise we'd get a loop in the list when the ACB is used for the next
allocation.

Luckily, this qcow2_cache_put shouldn't fail anyway because the L2 table is
only read, so that qcow2_cache_put doesn't even involve I/O.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
13 years agoReplaced tabs with spaces in block.h and block_int.h
Devin Nakamura [Thu, 9 Jun 2011 05:06:43 +0000 (01:06 -0400)]
Replaced tabs with spaces in block.h and block_int.h

Signed-off-by: Devin Nakamura <devin122@gmail.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
13 years agovdi: Avoid direct AIO callback
Kevin Wolf [Tue, 7 Jun 2011 14:12:58 +0000 (16:12 +0200)]
vdi: Avoid direct AIO callback

bdrv_aio_* must not call the callback before returning to its caller. In vdi,
this could happen in some error cases. This starts the real requests processing
in a BH to avoid this situation.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
13 years agoqcow: Avoid direct AIO callback
Kevin Wolf [Tue, 7 Jun 2011 13:20:44 +0000 (15:20 +0200)]
qcow: Avoid direct AIO callback

bdrv_aio_* must not call the callback before returning to its caller. In qcow,
this could happen in some error cases. This starts the real requests processing
in a BH to avoid this situation.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
13 years agoqcow2: Avoid direct AIO callback
Kevin Wolf [Tue, 7 Jun 2011 13:04:32 +0000 (15:04 +0200)]
qcow2: Avoid direct AIO callback

bdrv_aio_* must not call the callback before returning to its caller. In qcow2,
this could happen in some error cases. This starts the real requests processing
in a BH to avoid this situation.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
13 years agokvm: Enable CPU SMEP feature
Yang, Wei Y [Mon, 30 May 2011 15:17:42 +0000 (23:17 +0800)]
kvm: Enable CPU SMEP feature

This patchset enables a new CPU feature SMEP (Supervisor Mode Execution
Protection) in QEMU-KVM. SMEP prevents kernel from executing code in application.
Updated Intel SDM describes this CPU feature. The document will be published soon.

SMEP is identified by CPUID leaf 7 EBX[7], which is 0 before. Get the right value by query KVM kernel module, so that guest can get SMEP through CPUID.

Signed-off-by: Yang, Wei <wei.y.yang@intel.com>
Singed-off-by: Shan, Haitao <haitao.shan@intel.com>
Singed-off-by: Li, Xin <xin.li@intel.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
13 years agokvm: Add CPUID support for VIA CPU
brillywu@viatech.com.cn [Wed, 1 Jun 2011 01:59:52 +0000 (09:59 +0800)]
kvm: Add CPUID support for VIA CPU

When KVM is running on VIA CPU with host cpu's model, the
feautures of VIA CPU will be passed into kvm guest by calling
the CPUID instruction for Centaur.

Signed-off-by: BrillyWu<brillywu@viatech.com.cn>
Signed-off-by: KaryJin<karyjin@viatech.com.cn>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
13 years agousb-ccid: Plug memory leak on qdev exit()
Markus Armbruster [Tue, 24 May 2011 16:09:10 +0000 (18:09 +0200)]
usb-ccid: Plug memory leak on qdev exit()

ccid_initfn() allocates CCIDBus dynamically, but there is no exit
callback to free it.

Fix by getting rid of the allocation.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
13 years agolibcacard: add libcacard.la target
Alon Levy [Sun, 15 May 2011 08:51:28 +0000 (11:51 +0300)]
libcacard: add libcacard.la target

No flag to configure is required. Instead, added a libcacard.la target that
is not built by default, only when requested explicitly via:

mkdir build
cd build
../configure
make libcacard.la
make install-libcacard

Uses libtool to do actual linking of object files and shared library, and
installing. Tested only under linux, but supposed to work on other systems as
well.

If libtool isn't found you get a message complaining about that, only at build
time (since it is not a default target I did not add a message at configure
time).

New build artifacts:
 .libs subdirectories (at <buildroot> and <buildroot>/libcacard)
 *.lo files (at same locations as the respective o files)

Added %.lo : %.c rule that uses libtool.
Updated clean rule to clean up those artifacts.
Added specific rule to call dtrace with libtool wrapper (note that because of
a current upstream dtrace bug fixed by systemtap b1568fd85 commit the -fPIC flag
isn't actually passed on. still current dtrace+libtool produced object links fine).
If libtool is missing any of the following targets will complain and exit 1:
 any subdir: *.lo
 root and libcacard: libcacard.la, libcacard-instsall

Tested to link and load with all tracing backends.

13 years agoconfigure: add libdir and --libdir
Alon Levy [Sun, 15 May 2011 09:08:59 +0000 (12:08 +0300)]
configure: add libdir and --libdir

13 years agousb-uhci: fix expire time initialization.
Gerd Hoffmann [Fri, 10 Jun 2011 12:38:08 +0000 (14:38 +0200)]
usb-uhci: fix expire time initialization.

expire_time must be initialited when the guest activates the
usb scheduler, not at device creation time.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agohw/usb-ohci.c: Implement remote wakeup
Peter Maydell [Tue, 7 Jun 2011 18:50:12 +0000 (19:50 +0100)]
hw/usb-ohci.c: Implement remote wakeup

Implement the wakeup callback in the OHCI USBPortOps, so that when
a downstream device wakes up it correctly causes the OHCI controller
to come out of suspend.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agohw/usb-ohci.c: Ignore writes to HcPeriodCurrentED register
Peter Maydell [Tue, 7 Jun 2011 19:02:29 +0000 (20:02 +0100)]
hw/usb-ohci.c: Ignore writes to HcPeriodCurrentED register

HcPeriodCurrentED is read-only, but Linux writes to it anyway; silently
ignore this rather than printing a warning message.

(Specifically, drivers/usb/host/ohci-hub.c:ohci_rh_resume() writes a
0, in at least kernels 2.6.25 through 2.6.39.)

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agousb: Use defines for serial bus release number register for EHCI
Brad Hards [Thu, 2 Jun 2011 01:18:48 +0000 (11:18 +1000)]
usb: Use defines for serial bus release number register for EHCI

Signed-off-by: Brad Hards <bradh@frogmouth.net>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agousb: Use defines for serial bus release number register for UHCI
Brad Hards [Thu, 2 Jun 2011 01:18:47 +0000 (11:18 +1000)]
usb: Use defines for serial bus release number register for UHCI

Signed-off-by: Brad Hards <bradh@frogmouth.net>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agousb: Add defines for USB Serial Bus Release Number register
Brad Hards [Thu, 2 Jun 2011 01:18:46 +0000 (11:18 +1000)]
usb: Add defines for USB Serial Bus Release Number register

Signed-off-by: Brad Hards <bradh@frogmouth.net>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agousb-bus: Don't detach non attached devices on device exit
Hans de Goede [Tue, 31 May 2011 09:35:29 +0000 (11:35 +0200)]
usb-bus: Don't detach non attached devices on device exit

This causes an "Error: tried to detach unattached usb device " to be printed,
this can happen when deleting ie a usb host qdev, which did not
get attached (because a device matching the filter never got plugged in).

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agousb-bus: Add knowledge of USB_SPEED_SUPER to usb_speed helper
Hans de Goede [Tue, 31 May 2011 09:35:27 +0000 (11:35 +0200)]
usb-bus: Add knowledge of USB_SPEED_SUPER to usb_speed helper

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agousb-linux: Enlarge buffer for descriptors to 8192 bytes
Hans de Goede [Tue, 31 May 2011 09:35:26 +0000 (11:35 +0200)]
usb-linux: Enlarge buffer for descriptors to 8192 bytes

1024 bytes is way to small, one hd UVC webcam I have over here has so
many resolutions its descriptors take op close to 4k. Hopefully 8k will
be enough for all devices.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agousb: don't call usb_host_device_open from vl.c
Gerd Hoffmann [Wed, 1 Jun 2011 12:41:59 +0000 (14:41 +0200)]
usb: don't call usb_host_device_open from vl.c

Not needed any more, usb-host is qdev-ified these days.
Well, at least the linux version ...

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agousb-linux: only cleanup in host_close when host_open was successful.
Gerd Hoffmann [Mon, 6 Jun 2011 07:45:20 +0000 (09:45 +0200)]
usb-linux: only cleanup in host_close when host_open was successful.

13 years agousb-linux: Don't try to open the same device twice
Hans de Goede [Tue, 31 May 2011 09:35:24 +0000 (11:35 +0200)]
usb-linux: Don't try to open the same device twice

If a user wants to redirect 2 identical usb sticks, in theory this is
possible by doing:
usb_add host:1234:5678
usb_add host:1234:5678

But this will lead to us trying to open the first stick twice, since we
don't break the loop after having found a match in our filter list, so the next'
filter list entry will result in us trying to open the same device again.

Fix this by adding the missing break.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agousb-linux: Ensure devep != 0
Hans de Goede [Tue, 31 May 2011 09:35:22 +0000 (11:35 +0200)]
usb-linux: Ensure devep != 0

So that we don't index endp_table with a negative index.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agousb-linux: Don't do perror when errno is not set
Hans de Goede [Tue, 31 May 2011 09:35:20 +0000 (11:35 +0200)]
usb-linux: Don't do perror when errno is not set

Note that "op" also is not set, so before this change these error paths
would feed NULL to perror.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agousb-linux: Teach about super speed
Hans de Goede [Tue, 31 May 2011 09:35:19 +0000 (11:35 +0200)]
usb-linux: Teach about super speed

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agousb-linux: Get speed from sysfs rather then from the connectinfo ioctl
Hans de Goede [Tue, 31 May 2011 09:35:18 +0000 (11:35 +0200)]
usb-linux: Get speed from sysfs rather then from the connectinfo ioctl

The connectinfo ioctl only differentiates between lo speed devices, and
all other speeds, where as we would like to know the real speed. The real
speed is available in sysfs so use that when available.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agousb: documentation update
Gerd Hoffmann [Tue, 31 May 2011 10:21:43 +0000 (12:21 +0200)]
usb: documentation update

Add some more informations to docs/usb2.txt about using usb2 (also usb1)
devices.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agousb-ehci: split trace calls to handle arg count limits
Gerd Hoffmann [Mon, 6 Jun 2011 10:31:34 +0000 (12:31 +0200)]
usb-ehci: split trace calls to handle arg count limits

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agousb-ehci: itd handling fixes.
Gerd Hoffmann [Mon, 30 May 2011 14:09:08 +0000 (16:09 +0200)]
usb-ehci: itd handling fixes.

This patch fixes a bunch of issues in the itd descriptor handling.
Most important fix is to handle transfers which cross page borders
correctly by looking up the address of the next page.  Luckily the
linux uses physically contigous memory so the data used to hits the
correct location even with this bug instead of corrupting guest
memory.  Also the transfer length updates for outgoing transfers wasn't
correct.

While being at it DPRINTFs have been replaced by tracepoints.

The isoch_pause logic has been disabled.  Not clear to me which propose
this serves and I think it is incorrect too as we just skip processing
itds.  Even when no xfer happens we have to clear the active bit.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agoThe USB tablet should not claim boot protocol support.
Kevin O'Connor [Sat, 28 May 2011 15:47:50 +0000 (11:47 -0400)]
The USB tablet should not claim boot protocol support.

The USB tablet advertises that it supports the "boot" protocol.
However, its reports aren't "boot" protocol compatible.  So, it
shouldn't claim that.

Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agoFix USB mouse Set_Protocol behavior
Kevin O'Connor [Sat, 28 May 2011 15:47:15 +0000 (11:47 -0400)]
Fix USB mouse Set_Protocol behavior

The QEMU USB mouse claims to support the "boot" protocol
(bInterfaceSubClass is 1).  However, the mouse rejects the
Set_Protocol command.

The qemu mouse does support the "boot" protocol specification, so a
simple fix is to enable the Set_Protocol request.

Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agousb-ehci: drop EXECUTING checks.
Gerd Hoffmann [Fri, 20 May 2011 20:08:44 +0000 (22:08 +0200)]
usb-ehci: drop EXECUTING checks.

The state machine doesn't stop in EXECUTING state any more when async
packets are in flight, so the checks are not needed any more and can
be dropped.

Also kick out the check for the frame timer.  As we don't stop & sleep
any more on async packets this is obsolete.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agousb: cancel async packets on unplug
Gerd Hoffmann [Mon, 23 May 2011 15:37:12 +0000 (17:37 +0200)]
usb: cancel async packets on unplug

This patch adds USBBusOps struct with (for now) only a single callback
which is called when a device is about to be destroyed.  The USB Host
adapters are implementing this callback and use it to cancel any async
requests which might be in flight before the device actually goes away.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agoehci: fix a number of unused-but-set-variable warnings (new with gcc-4.6)
Hans de Goede [Mon, 23 May 2011 12:49:43 +0000 (14:49 +0200)]
ehci: fix a number of unused-but-set-variable warnings (new with gcc-4.6)

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agousb-ehci: fix error handling.
Gerd Hoffmann [Mon, 23 May 2011 07:50:27 +0000 (09:50 +0200)]
usb-ehci: fix error handling.

Set the correct bits for nodev, stall and babble errors.
Raise errint irq.  Fix state transition from WRITEBACK
to the next state.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agousb-ehci: fix offset writeback in ehci_buffer_rw
Gerd Hoffmann [Fri, 20 May 2011 15:25:07 +0000 (17:25 +0200)]
usb-ehci: fix offset writeback in ehci_buffer_rw

Two bugs at once:

First the mask is backwards, so the it used to keeps the offset and
clears the page address, which is not what we need when we update the
offset.

Second the offset calculation is wrong in case head isn't page aligned.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agousb-ehci: multiqueue support
Gerd Hoffmann [Thu, 19 May 2011 15:56:19 +0000 (17:56 +0200)]
usb-ehci: multiqueue support

This patch adds support for keeping multiple queues going at the same
time.  One slow device will not affect other devices any more.

The patch adds code to manage EHCIQueue structs.  It also does a number
of changes to the state machine:

 * The state machine will never ever stop in EXECUTING any more.
   Instead it will continue with the next queue (aka HORIZONTALQH) when
   the usb device returns USB_RET_ASYNC.
 * The state machine will stop processing when it figures it walks in
   circles (easy to figure now that we have a EHCIQueue struct for each
   QH we've processed).  The bailout logic should not be needed any
   more.  For now it is still in, but will assert() in case it triggers.
 * The state machine will just skip queues with a async USBPacket in
   flight.
 * The state machine will resume processing as soon as the async
   USBPacket is finished.

The patch also takes care to flush the QH struct back to guest memory
when needed, so we don't get stale data when (re-)loading it from guest
memory in FETCHQH state.

It also makes the writeback code to not touch the first three dwords of
the QH struct as the EHCI must not write them.  This actually fixes a
bug where QH chaining changes (next ptr) by the linux ehci driver where
overwritten by the emulated EHCI.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agousb-ehci: add queue data struct
Gerd Hoffmann [Thu, 19 May 2011 08:49:03 +0000 (10:49 +0200)]
usb-ehci: add queue data struct

Add EHCIQueue struct, move the fields needed to track the queue state
into that struct.  Pass the new struct instead of ehci state down to
functions which handle the queue state.  Lot of variable references have
changed due to that without an actual functional change.

Replace fetch_addr with two variables, one for async and one for
periodic schedule.  Add functions to get and set the fetch address.

Use EHCIQueue->usb_status (old name: EHCIState->exec_status) directly in
ehci_execute_complete instead of passing around the status using a
parameters and the return value.

ehci_state_fetchqh returns a EHCIQueue struct now.

No change in behavior.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agousb-ehci: trace buffer copy
Gerd Hoffmann [Thu, 19 May 2011 07:24:00 +0000 (09:24 +0200)]
usb-ehci: trace buffer copy

Add a trace point for buffer copies and drop the DPRINTF's.

No change in behavior.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agousb-ehci: improve mmio tracing
Gerd Hoffmann [Thu, 19 May 2011 06:55:09 +0000 (08:55 +0200)]
usb-ehci: improve mmio tracing

Add a separate tracepoint to log how register values change in response
to a mmio write.  Especially useful for registers which have read-only
or clear-on-write bits in them.

No change in behavior.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agousb-ehci: trace port state
Gerd Hoffmann [Thu, 19 May 2011 06:46:53 +0000 (08:46 +0200)]
usb-ehci: trace port state

Trace usb port operations (attach, detach, reset),
drop a few obsolete DPRINTF's.

No change in behavior.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agousb-ehci: trace state machine changes
Gerd Hoffmann [Wed, 18 May 2011 12:23:35 +0000 (14:23 +0200)]
usb-ehci: trace state machine changes

Add functions to get and set the current state of the state machine,
add tracepoints there to trace state transitions.  Add support for
traceing the queue heads and transfer descriptors as we look at them.

Drop a few DPRINTFs and all DPRINTF_ST lines, they are obsolete now.

No change in behavior.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agousb-ehci: trace mmio and usbsts
Gerd Hoffmann [Wed, 18 May 2011 08:12:58 +0000 (10:12 +0200)]
usb-ehci: trace mmio and usbsts

This patch starts adding trace support to ehci.  It traces
updates of the status register (USBSTS), mmio access and
controller reset.

It also adds functions to set and clear status register bits
and puts them in use everywhere.

Some DPRINTF's are dropped in favor of the new tracepoints.

No change in behavior.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agousb-linux: catch ENODEV in more places.
Gerd Hoffmann [Tue, 24 May 2011 14:12:31 +0000 (16:12 +0200)]
usb-linux: catch ENODEV in more places.

Factor out disconnect code (called when a device disappears) to a
separate function.  Add a check for ENODEV errno to a few more places
to make sure we notice disconnects.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
13 years agoblock/rbd: Remove unused local variable
Stefan Weil [Fri, 10 Jun 2011 20:05:30 +0000 (22:05 +0200)]
block/rbd: Remove unused local variable

Variable 'snap' is assigned a value that is never used.
Remove snap and the related code.

Cc: Christian Brunner <chb@muc.de>
Cc: Josh Durgin <josh.durgin@dreamhost.com>
Cc: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Stefan Weil <weil@mail.berlios.de>
Reviewed-by: Josh Durgin <josh.durgin@dreamhost.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
13 years agoDarwin: Fix compilation warning regarding the deprecated daemon() function
Alexandre Raymond [Tue, 7 Jun 2011 03:34:10 +0000 (23:34 -0400)]
Darwin: Fix compilation warning regarding the deprecated daemon() function

Changes since v1: create a wrapper function named qemu_daemon() in oslib-posix.c
instead of putting the OS specific workaround in qemu-nbd.c directly.

On OSX >= 10.5, daemon() is deprecated, resulting in the following warning:
----8<----
qemu-nbd.c: In function ‘main’:
qemu-nbd.c:371: warning: ‘daemon’ is deprecated (declared at /usr/include/stdlib.h:289)
----8<----

The following trick, used in mDNSResponder, takes care of this warning:
http://www.opensource.apple.com/source/mDNSResponder/mDNSResponder-258.18/mDNSPosix/PosixDaemon.c

On OSX, it temporarily renames the daemon() function before including stdlib.h
and declares it manually as an extern function. This way, the compiler does not
see the declaration from stdlib.h and thus does not display the warning.

Signed-off-by: Alexandre Raymond <cerbere@gmail.com>
Cc: Blue Swirl <blauwirbel@gmail.com>
Signed-off-by: Andreas Färber <andreas.faerber@web.de>
13 years agococoa: Avoid warning related to multiple handleEvent: definitions
Andreas Färber [Thu, 9 Jun 2011 18:53:32 +0000 (20:53 +0200)]
cocoa: Avoid warning related to multiple handleEvent: definitions

Avoid compiler confusion as to which method signature to use for the
handleEvent: selector on OSX >= 10.6 by making the variable type-safe
as opposed to generic 'id' type.
Requires moving the variable definition to after the class definition.

----8<----
ui/cocoa.m: In function ‘cocoa_refresh’:
ui/cocoa.m:997: warning: multiple methods named ‘-handleEvent:’ found
/System/Library/Frameworks/AppKit.framework/Headers/NSTextInputContext.h:84: warning: using ‘-(BOOL)handleEvent:(NSEvent *)theEvent’
ui/cocoa.m:272: warning: also found ‘-(void)handleEvent:(NSEvent *)event’
----8<---

Reported-by: Alexandre Raymond <cerbere@gmail.com>
Signed-off-by: Andreas Färber <andreas.faerber@web.de>
Tested-by: Alexandre Raymond <cerbere@gmail.com>
13 years agococoa: Revert dependency on VNC
Andreas Färber [Thu, 2 Jun 2011 18:51:22 +0000 (20:51 +0200)]
cocoa: Revert dependency on VNC

In 821601ea5b02a68ada479731a4d3d07a9876632a (Make VNC support optional)
cocoa.o was moved from ui-obj-$(CONFIG_COCOA) to vnc-obj-$(CONFIG_COCOA),
adding a dependency on $(CONFIG_VNC). That must've been unintentional.

Cc: Jes Sorensen <Jes.Sorensen@redhat.com>
Cc: Anthony Liguori <aliguori@us.ibm.com>
Signed-off-by: Andreas Färber <andreas.faerber@web.de>
13 years agococoa: Provide central qemu_main() prototype
Andreas Färber [Sun, 29 May 2011 17:42:51 +0000 (19:42 +0200)]
cocoa: Provide central qemu_main() prototype

This fixes a missing prototype warning in vl.c and obsoletes
the prototype in cocoa.m. Adjust callers in cocoa.m to supply
third argument, which is currently only used on Linux/ppc.

The prototype is designed so that it could be shared with SDL
and other frontends, if desired.

Cc: Alexandre Raymond <cerbere@gmail.com>
Signed-off-by: Andreas Färber <andreas.faerber@web.de>
13 years agoFix libfdt warnings on Darwin
Andreas Färber [Sat, 28 May 2011 13:45:18 +0000 (15:45 +0200)]
Fix libfdt warnings on Darwin

Building with libfdt results in the following warnings on Mac OS X:

  CC    ppc-softmmu/device_tree.o
In file included from /Users/andreas/QEMU/latest64/include/libfdt.h:54,
                 from /Users/andreas/QEMU/qemu/device_tree.c:26:
/Users/andreas/QEMU/qemu/libfdt_env.h:25:20: warning: endian.h: No such file or directory
/Users/andreas/QEMU/qemu/libfdt_env.h:26:22: warning: byteswap.h: No such file or directory
/Users/andreas/QEMU/qemu/libfdt_env.h:28:5: warning: "__BYTE_ORDER" is not defined
/Users/andreas/QEMU/qemu/libfdt_env.h:28:21: warning: "__BIG_ENDIAN" is not defined

Since QEMU's copy of libfdt_env.h only uses bswap_32() and bswap_64(),
let QEMU's bswap.h take care of the headers and use its endianness define.

Cc: Hollis Blanchard <hollis@penguinppc.org>
Signed-off-by: Andreas Färber <andreas.faerber@web.de>
Acked-by: Alexander Graf <agraf@suse.de>
13 years agoconfigure: Fix check for fdatasync()
Alexandre Raymond [Sun, 29 May 2011 22:22:48 +0000 (18:22 -0400)]
configure: Fix check for fdatasync()

Under Darwin, a symbol exists for the fdatasync() function, so that our
link test succeeds. However _POSIX_SYNCHRONIZED_IO is set to '-1'.

According to POSIX:2008, a value of -1 means the feature is not supported.
A value of 0 means supported at compilation time, and a value greater 0
means supported at both compilation and run time.

Enable fdatasync() only if _POSIX_SYNCHRONIZED_IO is '>0'.

Signed-off-by: Alexandre Raymond <cerbere@gmail.com>
Signed-off-by: Andreas Färber <andreas.faerber@web.de>
13 years agoRemove warning in printf due to type mismatch
Alexandre Raymond [Sun, 29 May 2011 22:22:50 +0000 (18:22 -0400)]
Remove warning in printf due to type mismatch

----8<----
qemu/target-lm32/translate.c: In function ‘gen_intermediate_code_internal’:
qemu/target-lm32/translate.c:1135: warning: format ‘%zd’ expects type ‘signed size_t’, but argument 4 has type ‘int’
----8<----

Both gen_opc_ptr and gen_opc_buf are "uint16_t *". The difference between
pointers is a ptrdiff_t so printf needs '%td'.

Signed-off-by: Alexandre Raymond <cerbere@gmail.com>
Acked-by: Stefan Weil <weil@mail.berlios.de>
Signed-off-by: Andreas Färber <andreas.faerber@web.de>
13 years agoCocoa: avoid displaying window when command-line contains '-h' or '-help'
Alexandre Raymond [Sun, 29 May 2011 22:22:49 +0000 (18:22 -0400)]
Cocoa: avoid displaying window when command-line contains '-h' or '-help'

There was already a check in place to avoid displaying a window
in certain modes such as vnc, nographic or curses.

Add a check for '-h' and '-help' to avoid displaying a window for a split-
second before showing the usage information.

Signed-off-by: Alexandre Raymond <cerbere@gmail.com>
Signed-off-by: Andreas Färber <andreas.faerber@web.de>
13 years agoFix compilation warning due to incorrectly specified type
Alexandre Raymond [Fri, 27 May 2011 17:22:28 +0000 (13:22 -0400)]
Fix compilation warning due to incorrectly specified type

In audio/coreaudio.c, a variable named "str" was assigned "const char" values,
which resulted in the following warnings:

-----8<-----
audio/coreaudio.c: In function ‘coreaudio_logstatus’:
audio/coreaudio.c:59: warning: initialization discards qualifiers from pointer target type
audio/coreaudio.c:63: warning: assignment discards qualifiers from pointer target type
(...)
-----8<-----

Signed-off-by: Alexandre Raymond <cerbere@gmail.com>
Acked-by: Stefan Weil <weil@mail.berlios.de>
Signed-off-by: Andreas Färber <andreas.faerber@web.de>
13 years agococoa: do not create a spurious window for -version
Tristan Gingold [Tue, 15 Mar 2011 13:18:22 +0000 (14:18 +0100)]
cocoa: do not create a spurious window for -version

When invoked with -version, qemu will exit just after displaying the version,
so there is no need to create a window.
Also handles --XXX options.

Signed-off-by: Tristan Gingold <gingold@adacore.com>
Signed-off-by: Andreas Färber <andreas.faerber@web.de>
13 years agoconfigure: Detect and don't try to use older libcurl
Peter Maydell [Thu, 9 Jun 2011 21:54:29 +0000 (22:54 +0100)]
configure: Detect and don't try to use older libcurl

Older versions of libcurl don't have some of the features we try to
use, in particular curl_multi_setopt(). Check for this in the 'is
libcurl available?' configure test so we disable curl support if the
library is too old.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@gmail.com>
13 years agoChangelog: Add explanatory note that this file is no longer updated
Peter Maydell [Thu, 9 Jun 2011 10:43:12 +0000 (11:43 +0100)]
Changelog: Add explanatory note that this file is no longer updated

Add an explanatory note to the top of Changelog pointing at the
wiki and git history for changelogs for more recent releases.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@gmail.com>
13 years agodoc: Minor typo fix.
Brad Hards [Wed, 8 Jun 2011 21:50:43 +0000 (07:50 +1000)]
doc: Minor typo fix.

Thanks to agraf_, stefanha and Snader_LB for their IRC assistance.

Thanks to Markus Armbruster and Alexander Graf (again) for their
assistance with the second version of this patch. No patch is too
simple to test...

Signed-off-by: Brad Hards <bradh@frogmouth.net>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@gmail.com>
13 years agosigfd: use pthread_sigmask
Alexander Graf [Wed, 8 Jun 2011 22:55:37 +0000 (00:55 +0200)]
sigfd: use pthread_sigmask

Qemu uses signalfd to figure out, if a signal occured without the need
to actually receive the signal. Instead, it can read from the fd to receive
its news.

Now, we obviously don't always have signalfd around. Especially not on
non-Linux systems. So what we do there is that we create a new thread,
block that thread on all signals and simply call sigwait to wait for a
signal we're interested in to occur.

This all sounds great, but what we're really doing is:

    sigset_t all;

    sigfillset(&all);
    sigprocmask(SIG_BLOCK, &all, NULL);

which - on Darwin - blocks all signals on the current _process_, not only
on the current thread. To block signals on the thread, we can use
pthread_sigmask().

This patch does that, assuming that my above analysis is correct, and thus
renders Qemu useable on Darwin again.

Reported-by: Andreas Färber <andreas.faerber@web.de>
Acked-by: Paolo Bonizni <pbonzini@redhat.com>
CC: Jan Kiszka <jan.kiszka@siemens.com>
CC: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@gmail.com>
13 years agoMerge remote branch 'rth/axp-next' into alpha-merge
Edgar E. Iglesias [Fri, 10 Jun 2011 20:21:14 +0000 (22:21 +0200)]
Merge remote branch 'rth/axp-next' into alpha-merge

* rth/axp-next: (26 commits)
  target-alpha: Implement TLB flush primitives.
  target-alpha: Use a fixed frequency for the RPCC in system mode.
  target-alpha: Trap for unassigned and unaligned addresses.
  target-alpha: Remap PIO space for 43-bit KSEG for EV6.
  target-alpha: Implement cpu_alpha_handle_mmu_fault for system mode.
  target-alpha: Implement more CALL_PAL values inline.
  target-alpha: Disable interrupts properly.
  target-alpha: All ISA checks to use TB->FLAGS.
  target-alpha: Swap shadow registers moving to/from PALmode.
  target-alpha: Implement do_interrupt for system mode.
  target-alpha: Add IPRs to be used by the emulation PALcode.
  target-alpha: Use kernel mmu_idx for pal_mode.
  target-alpha: Add various symbolic constants.
  target-alpha: Use do_restore_state for arithmetic exceptions.
  target-alpha: Tidy up arithmetic exceptions.
  target-alpha: Tidy exception constants.
  target-alpha: Enable the alpha-softmmu target.
  target-alpha: Rationalize internal processor registers.
  target-alpha: Merge HW_REI and HW_RET implementations.
  target-alpha: Cleanup MMU modes.
  ...

13 years agofsdev: Fix archs that dont use fsdev
Edgar E. Iglesias [Fri, 10 Jun 2011 13:29:11 +0000 (15:29 +0200)]
fsdev: Fix archs that dont use fsdev

Fix provided by: "Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>

Signed-off-by: Edgar E. Iglesias <edgar.iglesias@gmail.com>
13 years agoMerge remote-tracking branch 'qmp/for-anthony' into staging
Anthony Liguori [Wed, 8 Jun 2011 17:16:24 +0000 (12:16 -0500)]
Merge remote-tracking branch 'qmp/for-anthony' into staging

13 years agoMerge remote-tracking branch 'jvrao/for-anthony' into staging
Anthony Liguori [Wed, 8 Jun 2011 17:15:43 +0000 (12:15 -0500)]
Merge remote-tracking branch 'jvrao/for-anthony' into staging

13 years agoMerge remote-tracking branch 'spice/spice.v37' into staging
Anthony Liguori [Wed, 8 Jun 2011 17:15:11 +0000 (12:15 -0500)]
Merge remote-tracking branch 'spice/spice.v37' into staging

Conflicts:
vl.c

13 years agoMerge remote-tracking branch 'stefanha/trivial-patches' into staging
Anthony Liguori [Wed, 8 Jun 2011 17:13:58 +0000 (12:13 -0500)]
Merge remote-tracking branch 'stefanha/trivial-patches' into staging

13 years agoqemu-img create: Fix displayed default cluster size
Kevin Wolf [Tue, 31 May 2011 13:01:46 +0000 (15:01 +0200)]
qemu-img create: Fix displayed default cluster size

When not specifying a cluster size on the command line, qemu-img printed
a cluster size of 0:

    Formatting '/tmp/test.qcow2', fmt=qcow2 size=67108864
    encryption=off cluster_size=0

This patch adds the default cluster size to the QEMUOptionParameter list, so
that it displays the default value that is used.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
13 years agobdrv_img_create: Fix segfault
Kevin Wolf [Wed, 1 Jun 2011 12:03:31 +0000 (14:03 +0200)]
bdrv_img_create: Fix segfault

Block drivers that don't support creating images don't have a size option. Fail
gracefully instead of segfaulting when trying to access the option's value.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
13 years agoqcow2: Fix memory leaks in error cases
Kevin Wolf [Wed, 1 Jun 2011 08:50:00 +0000 (10:50 +0200)]
qcow2: Fix memory leaks in error cases

This fixes memory leaks that may be caused by I/O errors during L1 table growth
(can happen during save_vm) and in qemu-img check.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
13 years agorbd: Add bdrv_truncate implementation
Josh Durgin [Thu, 26 May 2011 23:07:34 +0000 (16:07 -0700)]
rbd: Add bdrv_truncate implementation

Reviewed-by: Christian Brunner <chb@muc.de>
Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
13 years agorbd: check return values when scheduling aio
Josh Durgin [Thu, 26 May 2011 23:07:33 +0000 (16:07 -0700)]
rbd: check return values when scheduling aio

If scheduling fails, the number of outstanding I/Os must be correct,
or there will be a hang when waiting for everything to be flushed.

Reviewed-by: Christian Brunner <chb@muc.de>
Reported-by: Stefan Hajnoczi <stefanha@gmail.com>
Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
13 years agorbd: allow configuration of rados from the rbd filename
Josh Durgin [Thu, 26 May 2011 23:07:32 +0000 (16:07 -0700)]
rbd: allow configuration of rados from the rbd filename

The new format is rbd:pool/image[@snapshot][:option1=value1[:option2=value2...]]
Each option is used to configure rados, and may be any Ceph option, or "conf".
The "conf" option specifies a Ceph configuration file to read.

This allows rbd volumes from more than one Ceph cluster to be used by
specifying different monitor addresses, as well as having different
logging levels or locations for different volumes.

Reviewed-by: Christian Brunner <chb@muc.de>
Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
13 years agorbd: use the higher level librbd instead of just librados
Josh Durgin [Thu, 26 May 2011 23:07:31 +0000 (16:07 -0700)]
rbd: use the higher level librbd instead of just librados

librbd stacks on top of librados to provide access
to rbd images.

Using librbd simplifies the qemu code, and allows
qemu to use new versions of the rbd format
with few (if any) changes.

Reviewed-by: Christian Brunner <chb@muc.de>
Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
13 years agoblock/raw-posix: get right partition size
Christoph Egger [Mon, 23 May 2011 12:31:17 +0000 (14:31 +0200)]
block/raw-posix: get right partition size

use the correct way to get the size of a disk device or partition

From: Adam Hamsik <haad@netbsd.org>
Signed-off-by: Christoph Egger <Christoph.Egger@amd.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
13 years agoblock/raw-posix: use a character device if a block device is given
Christoph Egger [Tue, 24 May 2011 09:30:29 +0000 (11:30 +0200)]
block/raw-posix: use a character device if a block device is given

On NetBSD a userland process is better with the character device
interface. In addition, a block device can't be opened twice; if a Xen
backend opens it, qemu can't and vice-versa.

Signed-off-by: Christoph Egger <Christoph.Egger@amd.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>