Thomas Huth [Fri, 8 Feb 2019 05:39:48 +0000 (06:39 +0100)]
hw/arm: Express dependencies of collie with Kconfig
Add Kconfig dependencies for the Strongarm collie machine.
This patch is based on earlier work by Ákos Kovács (i.e.
his "hw/arm/Kconfig: Add ARM Kconfig" patch).
Thomas Huth [Thu, 7 Feb 2019 19:26:34 +0000 (20:26 +0100)]
hw/arm: Express dependencies of xilinx-zynq with Kconfig
Add Kconfig dependencies for the xilinx-zynq-a9 board.
This patch is based on earlier work by Ákos Kovács (i.e.
his "hw/arm/Kconfig: Add ARM Kconfig" patch).
Thomas Huth [Fri, 8 Feb 2019 11:35:26 +0000 (12:35 +0100)]
hw/arm: Express dependencies of the PXA2xx machines with Kconfig
Add Kconfig dependencies for the PXA2xx machines (akita, borzoi,
connex and verdex gumstix, tosa, mainstone, spitz, terrier and z2).
This patch is based on earlier work by Ákos Kovács (i.e.
his "hw/arm/Kconfig: Add ARM Kconfig" patch).
Thomas Huth [Thu, 7 Feb 2019 16:12:14 +0000 (17:12 +0100)]
hw/arm: Express dependencies of the OMAP machines with Kconfig
Add Kconfig dependencies for the OMAP machines (cheetah, n800, n810,
sx1 and sx1-v1).
This patch is slightly based on earlier work by Ákos Kovács (i.e.
his "hw/arm/Kconfig: Add ARM Kconfig" patch).
Thomas Huth [Thu, 7 Feb 2019 15:51:56 +0000 (16:51 +0100)]
hw/arm: Express dependencies of the fsl-imx31 machine with Kconfig
Add Kconfig dependencies for the fsl-imx31 / kzm machine.
This patch is slightly based on earlier work by Ákos Kovács (i.e.
his "hw/arm/Kconfig: Add ARM Kconfig" patch).
Thomas Huth [Thu, 7 Feb 2019 09:35:22 +0000 (10:35 +0100)]
hw/arm: Express dependencies of the highbank machines with Kconfig
Add Kconfig dependencies for the highbank machine (and the midway
machine).
This patch is slightly based on earlier work by Ákos Kovács (i.e.
his "hw/arm/Kconfig: Add ARM Kconfig" patch).
Thomas Huth [Thu, 7 Feb 2019 09:28:01 +0000 (10:28 +0100)]
hw/arm: Express dependencies of the exynos machines with Kconfig
Add Kconfig dependencies for the Exynos-related boards (nuri and
smdkc210).
This patch is slightly based on earlier work by Ákos Kovács (i.e.
his "hw/arm/Kconfig: Add ARM Kconfig" patch).
Thomas Huth [Fri, 22 Feb 2019 14:48:11 +0000 (15:48 +0100)]
hw/ide/ahci: Add a Kconfig switch for the AHCI-ICH9 device
Some of our machines (like the ARM cubieboard) use CONFIG_AHCI for an AHCI
sysbus device, but do not use CONFIG_PCI since they do not feature a PCI
bus. With CONFIG_AHCI but without CONFIG_PCI, currently linking fails:
../hw/ide/ich.o: In function `pci_ich9_ahci_realize':
hw/ide/ich.c:124: undefined reference to `pci_allocate_irq'
hw/ide/ich.c:126: undefined reference to `pci_register_bar'
hw/ide/ich.c:128: undefined reference to `pci_register_bar'
hw/ide/ich.c:131: undefined reference to `pci_add_capability'
hw/ide/ich.c:147: undefined reference to `msi_init'
../hw/ide/ich.o: In function `pci_ich9_uninit':
hw/ide/ich.c:158: undefined reference to `msi_uninit'
../hw/ide/ich.o:(.data.rel+0x50): undefined reference to `vmstate_pci_device'
We must only compile ich.c if CONFIG_PCI is available, too, so introduce a
new config switch for this device.
Thomas Huth [Fri, 10 May 2019 08:11:59 +0000 (10:11 +0200)]
Makefile: Fix inclusion of the config-devices.mak.d Kconfig dependencies
The Makefile tries to include device Kconfig dependencies via
-include $(SUBDIR_DEVICES_MAK_DEP)
and thus expects files that match *-softmmu/config-devices.mak.d ...
however, the minikconf script currently generates files a la
"*-softmmu-config.devices.mak.d" instead, so the dependency files
simply got ignored so far. For example, after a "touch hw/arm/Kconfig",
the arm-softmmu/config-devices.mak file is currently not re-generated.
Fix it by putting the dependency files in the *-softmmu folders now.
Thomas Huth [Tue, 19 Feb 2019 16:00:10 +0000 (17:00 +0100)]
hw/pci/pci-stub: Add msi_enabled() and msi_notify() to the pci stubs
Some machines have an AHCI adapter, but no PCI. To be able to
compile hw/ide/ahci.c without CONFIG_PCI, we still need the two
functions msi_enabled() and msi_notify() for linking.
This is required for the new Kconfig-like build system, if a user
wants to compile a QEMU binary with just one machine that has AHCI,
but no PCI, like the ARM "cubieboard" for example.
scripts/clean-header-guards.pl warns these headers use reserved
identifier _XTENSA_CORE_CONFIGURATION_H as header guard symbol. It
additionally warns the guard doesn't match the file name.
Reuse of the same guard symbol in multiple headers is okay as long as
they cannot be included together.
Since we can avoid guard symbol reuse easily, do so: use the guard
symbol scripts/clean-header-guards.pl picks, less the TARGET_ prefix.
linux-user/nios2 linux-user/riscv: Clean up header guards
Reuse of the same guard symbol in multiple headers is okay as long as
they cannot be included together. scripts/clean-header-guards.pl
can't tell, so it warns.
Since we can avoid guard symbol reuse easily, do so: use guard symbol
${target^^}_${fname^^} for linux-user/$target/$fname, just like we did
in commit a9c94277f0..3500385697.
We're in the process of spinning out slirp/. tests/fp/platform.h is
has to include qemu/osdep.h because tests/fp/berkeley-softfloat-3/ and
tests/fp/berkeley-testfloat-3/ don't. tests/uefi-test-tools/ is guest
software. The remaining reverts are the same as in commit b7d89466dde.
cputlb: Move TLB_RECHECK handling into load/store_helper
Having this in io_readx/io_writex meant that we forgot to
re-compute index after tlb_fill. It also means we can use
the normal aligned memory load path. It also fixes a bug
in that we had cached a use of index across a tlb_fill.
Alex Bennée [Fri, 15 Feb 2019 14:31:13 +0000 (14:31 +0000)]
accel/tcg: demacro cputlb
Instead of expanding a series of macros to generate the load/store
helpers we move stuff into common functions and rely on the compiler
to eliminate the dead code for each variant.
Most of the existing users would continue around a loop which
would fault the tlb entry in via a normal load/store.
But for AArch64 SVE we have an existing emulation bug wherein we
would mark the first element of a no-fault vector load as faulted
(within the FFR, not via exception) just because we did not have
its address in the TLB. Now we can properly only mark it as faulted
if there really is no valid, readable translation, while still not
raising an exception. (Note that beyond the first element of the
vector, the hardware may report a fault for any reason whatsoever;
with at least one element loaded, forward progress is guaranteed.)
We can now use the CPUClass hook instead of a named function.
Create a static tlb_fill function to avoid other changes within
cputlb.c. This also isolates the asserts within. Remove the
named tlb_fill function from all of the targets.
Note that env->pc is removed from the qemu_log as that value is garbage.
The PC isn't recovered until cpu_restore_state, called from
cpu_loop_exit_restore, called from riscv_raise_exception.
Note that env->active_tc.PC is removed from the qemu_log as that value
is garbage. The PC isn't recovered until cpu_restore_state, called from
cpu_loop_exit_restore, called from do_raise_exception_err.
target/mips: Tidy control flow in mips_cpu_handle_mmu_fault
Since the only non-negative TLBRET_* value is TLBRET_MATCH,
the subsequent test for ret < 0 is useless. Use early return
to allow subsequent blocks to be unindented.
target/mips: Pass a valid error to raise_mmu_exception for user-only
At present we give ret = 0, or TLBRET_MATCH. This gets matched
by the default case, which falls through to TLBRET_BADADDR.
However, it makes more sense to use a proper value. All of the
tlb-related exceptions are handled identically in cpu_loop.c,
so TLBRET_BADADDR is as good as any other. Retain it.
Peter Maydell [Fri, 10 May 2019 17:44:59 +0000 (18:44 +0100)]
Merge remote-tracking branch 'remotes/kevin/tags/for-upstream' into staging
Block layer patches:
- block: Remove bdrv_read() and bdrv_write()
- qemu-img: Allow rebase with no input base
- blockjob: Fix coroutine thread after AioContext change
- MAINTAINERS updates for pflash, curl and gluster
# gpg: Signature made Fri 10 May 2019 16:39:54 BST
# gpg: using RSA key 7F09B272C88F2FD6
# gpg: Good signature from "Kevin Wolf <[email protected]>" [full]
# Primary key fingerprint: DC3D EB15 9A9A F95D 3D74 56FE 7F09 B272 C88F 2FD6
* remotes/kevin/tags/for-upstream:
iotests: Add test for rebase without input base
qemu-img: Use zero writes after source backing EOF
qemu-img: Allow rebase with no input base
qcow2: Remove BDRVQcow2State.cluster_sectors
block: Remove bdrv_read() and bdrv_write()
vvfat: Replace bdrv_{read,write}() with bdrv_{pread,pwrite}()
vdi: Replace bdrv_{read,write}() with bdrv_{pread,pwrite}()
qcow2: Replace bdrv_write() with bdrv_pwrite()
qemu-img: Use IEC binary prefixes for size constants
test-block-iothread: Job coroutine thread after AioContext switch
blockjob: Fix coroutine thread after AioContext change
qemu-iotests: Fix cleanup for 192
MAINTAINERS: Add an entry for the Parallel NOR Flash devices
MAINTAINERS: Downgrade status of block sections without "M:" to "Odd Fixes"
block: remove bs from lists before closing
This hook will replace the (user-only mode specific) handle_mmu_fault
hook, and the (system mode specific) tlb_fill function.
The handle_mmu_fault hook was written as if there was a valid
way to recover from an mmu fault, and had 3 possible return states.
In reality, the only valid action is to raise an exception,
return to the main loop, and deliver the SIGSEGV to the guest.
Note that all of the current implementations of handle_mmu_fault
for guests which support linux-user do in fact only ever return 1,
which is the signal to return to the main loop.
Using the hook for system mode requires that all targets be converted,
so for now the hook is (optionally) used only from user-only mode.
Alberto Garcia [Wed, 1 May 2019 18:13:57 +0000 (21:13 +0300)]
vvfat: Replace bdrv_{read,write}() with bdrv_{pread,pwrite}()
There's only a couple of bdrv_read() and bdrv_write() calls left in
the vvfat code, and they can be trivially replaced with the byte-based
bdrv_pread() and bdrv_pwrite().
Alberto Garcia [Wed, 1 May 2019 18:13:56 +0000 (21:13 +0300)]
vdi: Replace bdrv_{read,write}() with bdrv_{pread,pwrite}()
There's only a couple of bdrv_read() and bdrv_write() calls left in
the vdi code, and they can be trivially replaced with the byte-based
bdrv_pread() and bdrv_pwrite().
Kevin Wolf [Fri, 3 May 2019 17:17:43 +0000 (19:17 +0200)]
blockjob: Fix coroutine thread after AioContext change
Commit 463e0be10 ('blockjob: add AioContext attached callback') tried to
make block jobs robust against AioContext changes of their main node,
but it never made sure that the job coroutine actually runs in the new
thread.
Instead of waking up the job coroutine in whatever thread it ran before,
let's always pass the AioContext where it should be running now.
Kevin Wolf [Wed, 8 May 2019 12:00:44 +0000 (14:00 +0200)]
qemu-iotests: Fix cleanup for 192
Test case 192 calls _launch_qemu, so it also needs to _cleanup_qemu when
it's done, otherwise the QMP FIFOs stay around in scratch/. It also
creates a temporary NBD socket that needs to be removed as well at the
end of the test case.
Thomas Huth [Mon, 6 May 2019 06:18:54 +0000 (08:18 +0200)]
MAINTAINERS: Downgrade status of block sections without "M:" to "Odd Fixes"
Fixes might still get picked up via the qemu-block mailing list,
so the status is not "Orphan" yet.
Also add the gluster mailing list as suggested by Niels here:
Peter Maydell [Fri, 10 May 2019 13:49:36 +0000 (14:49 +0100)]
Merge remote-tracking branch 'remotes/stefanha/tags/block-pull-request' into staging
Pull request
# gpg: Signature made Fri 10 May 2019 14:02:22 BST
# gpg: using RSA key 8695A8BFD3F97CDAAC35775A9CA4ABB381AB73C8
# gpg: Good signature from "Stefan Hajnoczi <[email protected]>" [full]
# gpg: aka "Stefan Hajnoczi <[email protected]>" [full]
# Primary key fingerprint: 8695 A8BF D3F9 7CDA AC35 775A 9CA4 ABB3 81AB 73C8
* remotes/stefanha/tags/block-pull-request:
docs: add Security chapter to the documentation
docs: add Secure Coding Practices to developer docs
aio-posix: ensure poll mode is left when aio_notify is called
block/io.c: fix for the allocation failure
block: Add coroutine_fn to bdrv_check_co_entry
util/readline: Add braces to fix checkpatch errors
util: readline: replace tab indent by four spaces to fix checkpatch errors
util/readline: add a space to fix errors by checkpatch tool
Add a new virtio-input device, which connects to a vhost-user
backend.
Instead of reading configuration directly from an input device /
evdev (like virtio-input-host), it reads it over vhost-user protocol
with {SET,GET}_CONFIG messages. The vhost-user-backend handles the
queues & events setup.
Create a vhost-user-backend object that holds a connection to a
vhost-user backend (or "slave" process) and can be referenced from
virtio devices that support it. See later patches for input & gpu
usage.
Note: a previous iteration of this object made it user-creatable, and
allowed managed sub-process spawning, but that has been dropped for
now.
/home/elmarco/src/qemu/contrib/libvhost-user/libvhost-user.c: In function ‘vu_set_mem_table_exec_postcopy’:
/home/elmarco/src/qemu/contrib/libvhost-user/libvhost-user.c:546:31: warning: taking address of packed member of ‘struct VhostUserMsg’ may result in an unaligned pointer value [-Waddress-of-packed-member]
546 | VhostUserMemory *memory = &vmsg->payload.memory;
| ^~~~~~~~~~~~~~~~~~~~~
/home/elmarco/src/qemu/contrib/libvhost-user/libvhost-user.c: In function ‘vu_set_mem_table_exec’:
/home/elmarco/src/qemu/contrib/libvhost-user/libvhost-user.c:688:31: warning: taking address of packed member of ‘struct VhostUserMsg’ may result in an unaligned pointer value [-Waddress-of-packed-member]
688 | VhostUserMemory *memory = &vmsg->payload.memory;
| ^~~~~~~~~~~~~~~~~~~~~
/home/elmarco/src/qemu/contrib/libvhost-user/libvhost-user.c: In function ‘vu_set_vring_addr_exec’:
/home/elmarco/src/qemu/contrib/libvhost-user/libvhost-user.c:817:36: warning: taking address of packed member of ‘struct VhostUserMsg’ may result in an unaligned pointer value [-Waddress-of-packed-member]
817 | struct vhost_vring_addr *vra = &vmsg->payload.addr;
| ^~~~~~~~~~~~~~~~~~~
Alex Bennée [Thu, 2 May 2019 09:27:28 +0000 (10:27 +0100)]
linux-user: fix GPROF build failure
When linux-user/exit was introduced we failed to move the gprof
include at the same time. The CI didn't notice because it only builds
system emulation. Fix it for those that still find gprof useful.
Alex Bennée [Thu, 2 May 2019 14:58:46 +0000 (15:58 +0100)]
linux-user: avoid treading on gprof's SIGPROF signals
The guest tends to get confused when it receives signals it doesn't
know about. Given the gprof magic has also set up it's own handler we
would do well to avoid stomping on it as well.
Some PT_LOAD segments may be completely zeroed out and their p_filesize
is zero, in that case the loader should just allocate a page that's at
least p_memsz bytes large (plus eventual alignment padding).
Calling zero_bss does this job for us, all we have to do is make sure we
don't try to mmap a zero-length page.
projects / qemu.git / log