Alex Bennée [Mon, 3 Jun 2019 15:14:24 +0000 (16:14 +0100)]
tests/tcg: better detect truncated reads
If we've truncated a wider read we can detect the condition earlier by
looking at the number of zeros we've read. So we don't trip up on
cases where we have written zeros to the start of the buffer we also
ensure we only start each offset read from the right address.
Alex Bennée [Mon, 3 Jun 2019 14:56:32 +0000 (15:56 +0100)]
cputlb: use uint64_t for interim values for unaligned load
When running on 32 bit TCG backends a wide unaligned load ends up
truncating data before returning to the guest. We specifically have
the return type as uint64_t to avoid any premature truncation so we
should use the same for the interim types.
Alex Bennée [Thu, 30 May 2019 14:35:14 +0000 (15:35 +0100)]
semihosting: split console_out into string and char versions
This is ostensibly to avoid the weirdness of len looking like it might
come from a guest and sometimes being used. While we are at it fix up
the error checking for the arm-linux-user implementation of the API
which got flagged up by Coverity (CID 1401700).
Gerd Hoffmann [Mon, 20 May 2019 12:47:03 +0000 (14:47 +0200)]
scripts: use git archive in archive-source
Use git archive to create tarballs of qemu and submodules instead of
cloning the repository and the submodules. This is a order of magnitude
faster because it doesn't fetch the submodules from the internet each
time the script runs.
`make vm-build-centos` run docker-based tests on CentOS. The
created containers should have network otherwise some tests
fail. Also fixed the BUILD_SCRIPT template to correctly
evaluate "V=1" for verbose output.
Fixed tests/vm/basevm.py to run with Python 3:
- hashlib.sha1() requires an binary encoded object.
- uses floor division ("//") (PEP 238).
- decode bytes to unicode when needed.
Changed the vm-test makefile to execute python scripts
with the interpreter configured on build. This allows to
run vm-test targets properly in Linux distros with Python 3
only support.
Alex Bennée [Wed, 29 May 2019 16:16:32 +0000 (17:16 +0100)]
qemu-io-cmds: use clock_gettime for benchmarking
The previous use of gettimeofday() ran into undefined behaviour when
we ended up doing a div 0 for a very short operation. This is because
gettimeofday only works at the microsecond level as well as being
prone to discontinuous jumps in system time. Using clock_gettime with
CLOCK_MONOTONIC gives greater precision and alleviates some of the
potential problems with time jumping around.
We could use CLOCK_MONOTONIC_RAW to avoid being tripped up by NTP and
adjtime but that is Linux specific so I decided it would do for now.
Peter Maydell [Wed, 12 Jun 2019 13:43:47 +0000 (14:43 +0100)]
Merge remote-tracking branch 'remotes/dgibson/tags/ppc-for-4.1-20190612' into staging
ppc patch queue 2019-06-12
Next pull request against qemu-4.1. The big thing here is adding
support for hot plug of P2P bridges, and PCI devices under P2P bridges
on the "pseries" machine (which doesn't use SHPC). Other than that
there's just a handful of fixes and small enhancements.
* remotes/dgibson/tags/ppc-for-4.1-20190612:
ppc/xive: Make XIVE generate the proper interrupt types
ppc/pnv: activate the "dumpdtb" option on the powernv machine
target/ppc: Use tcg_gen_gvec_bitsel
spapr: Allow hot plug/unplug of PCI bridges and devices under PCI bridges
spapr: Direct all PCI hotplug to host bridge, rather than P2P bridge
spapr: Don't use bus number for building DRC ids
spapr: Clean up DRC index construction
spapr: Clean up spapr_drc_populate_dt()
spapr: Clean up dt creation for PCI buses
spapr: Clean up device tree construction for PCI devices
spapr: Clean up device node name generation for PCI devices
target/ppc: Fix lxvw4x, lxvh8x and lxvb16x
spapr_pci: Improve error message
* remotes/armbru/tags/pull-misc-2019-06-11-v3:
MAINTAINERS: Polish headline decorations
MAINTAINERS: Improve section headlines
MAINTAINERS: Remove duplicate entries of [email protected]
Clean up a header guard symbols (again)
Supply missing header guards
Clean up a few header guard symbols
scripts/clean-header-guards: Fix handling of trailing comments
Normalize position of header guard
Include qemu-common.h exactly where needed
Include qemu/module.h where needed, drop it from qemu-common.h
qemu-common: Move qemu_isalnum() etc. to qemu/ctype.h
qemu-common: Move tcg_enabled() etc. to sysemu/tcg.h
the user is left to wonder *which* of our three "Overall" sections
applies: the one under "Guest CPU cores (TCG)", or the one under
"Guest CPU Cores (KVM)", or the one under "Usermode emulation".
Rename sections under
* "Guest CPU cores (TCG)" from "FOO" to "FOO TCG CPUs"
* "Guest CPU Cores (KVM)" from "FOO" to "FOO KVM CPUs"
* "Guest CPU Cores (Xen)" from "FOO" to "FOO Xen CPUs"
* "Architecture support" from "FOO" to "FOO general architecture
support"
* "Usermode Emulation" from "Overall" to "Overall usermode emulation"
* "Tiny Code Generator (TCG)" from "FOO target" to "FOO TCG target",
and from "Common code" to "Common TCG code"
The list is always selected by the 'All patches CC here' section.
Signed-off-by: Philippe Mathieu-Daudé <[email protected]>
[Conflicts resolved by redoing the patch] Signed-off-by: Markus Armbruster <[email protected]>
Commit d52c454aad "contrib: add vhost-user-gpu" and "c68082c43a
virtio-gpu: split virtio-gpu-pci & virtio-vga" created headers with
unusual header guard symbols. Clean them up
Commit 58ea30f5145 "Clean up header guards that don't match their file
name" messed up contrib/elf2dmp/qemu_elf.h and
tests/migration/migration-test.h.
It missed target/cris/opcode-cris.h and
tests/uefi-test-tools/UefiTestToolsPkg/Include/Guid/BiosTablesTest.h
due to the scripts/clean-header-guards.pl bug fixed in the previous
commit.
Commit a8b991b52dc "Clean up ill-advised or unusual header guards"
missed include/hw/xen/io/ring.h for the same reason.
Commit 3979fca4b69 "disas: Rename include/disas/bfd.h back to
include/disas/dis-asm.h" neglected to update the guard symbol for the
rename.
Commit a331c6d7741 "semihosting: implement a semihosting console"
created include/hw/semihosting/console.h with an ill-advised guard
symbol.
A few of our headers have some #include before the guard.
target/tilegx/spr_def_64.h has #ifndef __DOXYGEN__ outside the guard.
A few more have the #define elsewhere.
Change them to match the common idiom. For spr_def_64.h, that means
dropping #ifndef __DOXYGEN__. While there, rename guard symbols to
make scripts/clean-header-guards.pl happy.
Peter Maydell [Wed, 12 Jun 2019 10:49:08 +0000 (11:49 +0100)]
Merge remote-tracking branch 'remotes/ehabkost/tags/python-next-pull-request' into staging
Python queue, 2019-06-11
* New boot_linux_console test cases (Philippe Mathieu-Daudé)
* Make check-acceptance Travis job more verbose (Eduardo Habkost)
# gpg: Signature made Tue 11 Jun 2019 18:13:35 BST
# gpg: using RSA key 2807936F984DC5A6
# gpg: Good signature from "Eduardo Habkost <[email protected]>" [full]
# Primary key fingerprint: 5A32 2FD5 ABC4 D3DB ACCF D1AA 2807 936F 984D C5A6
* remotes/ehabkost/tags/python-next-pull-request:
travis: Make check-acceptance job more verbose
BootLinuxConsoleTest: Run kerneltests BusyBox on Malta
BootLinuxConsoleTest: Test nanoMIPS kernels on the I7200 CPU
BootLinuxConsoleTest: Test the SmartFusion2 board
BootLinuxConsoleTest: Do not log empty lines
tests/boot_linux_console: Let extract_from_deb handle various compressions
David Gibson [Fri, 5 Apr 2019 05:34:19 +0000 (16:34 +1100)]
spapr: Allow hot plug/unplug of PCI bridges and devices under PCI bridges
The pseries machine type already allows PCI hotplug and unplug via the
PAPR mechanism, but only on the root bus of each PHB. This patch extends
this to allow PCI to PCI bridges to be hotplugged, and devices to be
hotplugged or unplugged under P2P bridges.
For now we disallow hot unplugging P2P bridges. I tried doing that, but
haven't managed to get it working, I think due to some guest side problems
that need further investigation.
To do this we dynamically construct DRCs when bridges are hot (or cold)
added, which can in turn be used to hotplug devices under the bridge.
David Gibson [Fri, 15 Feb 2019 05:50:40 +0000 (16:50 +1100)]
spapr: Direct all PCI hotplug to host bridge, rather than P2P bridge
A P2P bridge will attempt to handle the hotplug with SHPC, which doesn't
work in the PAPR environment. Instead we want to direct all PCI hotplug
actions to the PAPR specific host bridge which will use the PAPR hotplug
mechanism.
David Gibson [Wed, 10 Apr 2019 01:49:28 +0000 (11:49 +1000)]
spapr: Don't use bus number for building DRC ids
DRC ids are more or less arbitrary, as long as they're consistent. For
PCI, we notionally build them from the phb's index along with PCI bus
number, slot and function number.
Using bus number is broken, however, because it can change if the guest
re-enumerates the PCI topology for whatever reason (e.g. due to hotplug
of a bridge, which we don't support yet but want to).
Fortunately, there's an alternative. Bridges are required to have a unique
non-zero "chassis number" that we can use instead. Adjust the code to
use that instead.
This looks like it would introduce a guest visible breaking change, but
in fact it does not because we don't yet ever use non-zero bus numbers.
Both chassis and bus number are always 0 for the root bus, so there's no
change for the existing cases.
David Gibson [Fri, 5 Apr 2019 04:51:00 +0000 (15:51 +1100)]
spapr: Clean up DRC index construction
spapr_pci.c currently has several confusingly similarly named functions for
various conversions between representations of DRCs. Make things clearer
by renaming things in a more consistent XXX_from_YYY() manner and remove
some called-only-once variants in favour of open coding.
While we're at it, move this code together in the file to avoid some extra
forward references, and split out construction and removal of DRCs for the
host bridge into helper functions.
David Gibson [Fri, 5 Apr 2019 02:31:48 +0000 (13:31 +1100)]
spapr: Clean up dt creation for PCI buses
Device nodes for PCI bridges (both host and P2P) describe both the bridge
device itself and the bus hanging off it, handling of this is a bit of a
mess.
spapr_dt_pci_device() has a few things it only adds for non-bridges, but
always adds #address-cells and #size-cells which should only appear for
bridges. But the walking down the subordinate PCI bus is done in one of
its callers spapr_populate_pci_devices_dt(). The PHB dt creation in
spapr_populate_pci_dt() open codes some similar logic to the bridge case.
This patch consolidates things in a bunch of ways:
* Bus specific dt info is now created in spapr_dt_pci_bus() used for both
P2P bridges and the host bridge. This includes walking subordinate
devices
* spapr_dt_pci_device() now calls spapr_dt_pci_bus() when called on a
P2P bridge
* We do detection of bridges with the is_bridge field of the device class,
rather than checking PCI config space directly, for consistency with
qemu's core PCI code.
* Several things are renamed for brevity and clarity
David Gibson [Fri, 22 Mar 2019 05:17:14 +0000 (16:17 +1100)]
spapr: Clean up device tree construction for PCI devices
spapr_create_pci_child_dt() is a trivial wrapper around
spapr_populate_pci_child_dt(), but is the latter's only caller. So fold
them together into spapr_dt_pci_device(), which closer matches our modern
naming convention.
While there, make a number of cleanups to the function itself. This is
mostly using more temporary locals to avoid awkwardly long lines, and in
some cases avoiding double reads of PCI config space variables.
David Gibson [Fri, 22 Mar 2019 05:13:09 +0000 (16:13 +1100)]
spapr: Clean up device node name generation for PCI devices
spapr_populate_pci_child_dt() adds a 'name' property to the device tree
node for PCI devices. This is never necessary for a flattened device tree,
it is implicit in the name added when the node is constructed. In fact
anything we do add to a 'name' property will be overwritten with something
derived from the structural name in the guest firmware (but in fact it is
exactly the same bytes).
So, remove that. In addition, pci_get_node_name() is very simple, so fold
it into its (also simple) sole caller spapr_create_pci_child_dt().
While we're there rename pci_find_device_name() to the shorter and more
accurate dt_name_from_class().
Greg Kurz [Wed, 29 May 2019 17:15:09 +0000 (19:15 +0200)]
spapr_pci: Improve error message
Every PHB must have a unique index. This is checked at realize but when
a duplicate index is detected, an error message mentioning BUIDs is
printed. This doesn't help much, especially since BUID is an internal
concept that is no longer exposed to the user.
Fix the message to mention the index property instead of BUID. As a bonus
print a list of indexes already in use.
qemu-common: Move tcg_enabled() etc. to sysemu/tcg.h
Other accelerators have their own headers: sysemu/hax.h, sysemu/hvf.h,
sysemu/kvm.h, sysemu/whpx.h. Only tcg_enabled() & friends sit in
qemu-common.h. This necessitates inclusion of qemu-common.h into
headers, which is against the rules spelled out in qemu-common.h's
file comment.
Move tcg_enabled() & friends into their own header sysemu/tcg.h, and
adjust #include directives.
BootLinuxConsoleTest: Run kerneltests BusyBox on Malta
This tests boots a Linux kernel on a Malta machine up to a
busybox shell on the serial console. Few commands are executed
before halting the machine (via reboot).
We use the initrd cpio image from the kerneltests project:
https://kerneltests.org/
If MIPS is a target being built, "make check-acceptance" will
automatically include this test by the use of the "arch:mips" tags.
Alternatively, this test can be run using:
$ avocado --show=console run -t arch:mips tests/acceptance/boot_linux_console.py
[...]
console: Boot successful.
[...]
console: / # uname -a
console: Linux buildroot 4.5.0-2-4kc-malta #1 Debian 4.5.5-1 (2016-05-29) mips GNU/Linux
console: / # reboot
console: / # reboot: Restarting system
tests/boot_linux_console: Let extract_from_deb handle various compressions
Debian binary package format supports various compressions.
Per man deb(5):
NAME
deb - Debian binary package format
FORMAT
...
The third, last required member is named data.tar. It contains the
filesystem as a tar archive, either not compressed (supported since
dpkg 1.10.24), or compressed with gzip (with .gz extension),
xz (with .xz extension, supported since dpkg 1.15.6),
bzip2 (with .bz2 extension, supported since dpkg 1.10.24) or
lzma (with .lzma extension, supported since dpkg 1.13.25).
List the archive files to have the 3rd name with the correct extension.
The function avocado.utils.archive.extract() will handle the different
compression format for us.
Peter Maydell [Tue, 11 Jun 2019 15:02:07 +0000 (16:02 +0100)]
Merge remote-tracking branch 'remotes/ehabkost/tags/x86-next-pull-request' into staging
x86 queue, 2019-06-11
* "unavailable-features" QOM property (Eduardo Habkost)
* Save EFER for 32-bit targets (Pavel Dovgalyuk)
# gpg: Signature made Tue 11 Jun 2019 14:41:45 BST
# gpg: using RSA key 2807936F984DC5A6
# gpg: Good signature from "Eduardo Habkost <[email protected]>" [full]
# Primary key fingerprint: 5A32 2FD5 ABC4 D3DB ACCF D1AA 2807 936F 984D C5A6
* remotes/ehabkost/tags/x86-next-pull-request:
i386: Save EFER for 32-bit targets
i386: "unavailable-features" QOM property
i386: x86_cpu_list_feature_names() function
Add a "unavailable-features" QOM property to X86CPU objects that
have the same semantics of "unavailable-features" on
query-cpu-definitions. The new property has the same goal of
"filtered-features", but is generic enough to let any kind of CPU
feature to be listed there without relying on low level details
like CPUID leaves or MSR numbers.
Extract feature name listing code from
x86_cpu_class_check_missing_features(). It will be reused to
return information about CPU filtered features at runtime.
Peter Maydell [Mon, 10 Jun 2019 15:09:19 +0000 (16:09 +0100)]
Merge remote-tracking branch 'remotes/rth/tags/pull-tcg-20190610' into staging
Move softmmu tlb into CPUNegativeOffsetState
# gpg: Signature made Mon 10 Jun 2019 15:07:55 BST
# gpg: using RSA key 7A481E78868B4DB6A85A05C064DF38E8AF7E215F
# gpg: issuer "[email protected]"
# gpg: Good signature from "Richard Henderson <[email protected]>" [full]
# Primary key fingerprint: 7A48 1E78 868B 4DB6 A85A 05C0 64DF 38E8 AF7E 215F
* remotes/rth/tags/pull-tcg-20190610: (39 commits)
tcg/arm: Remove mostly unreachable tlb special case
tcg/arm: Use LDRD to load tlb mask+table
tcg/aarch64: Use LDP to load tlb mask+table
cpu: Remove CPU_COMMON
cpu: Move the softmmu tlb to CPUNegativeOffsetState
cpu: Move icount_decr to CPUNegativeOffsetState
cpu: Introduce CPUNegativeOffsetState
cpu: Introduce cpu_set_cpustate_pointers
cpu: Move ENV_OFFSET to exec/gen-icount.h
target/xtensa: Use env_cpu, env_archcpu
target/unicore32: Use env_cpu, env_archcpu
target/tricore: Use env_cpu
target/tilegx: Use env_cpu
target/sparc: Use env_cpu, env_archcpu
target/sh4: Use env_cpu, env_archcpu
target/s390x: Use env_cpu, env_archcpu
target/riscv: Use env_cpu, env_archcpu
target/ppc: Use env_cpu, env_archcpu
target/openrisc: Use env_cpu, env_archcpu
target/nios2: Use env_cpu, env_archcpu
...
tcg/arm: Remove mostly unreachable tlb special case
There was nothing armv7 specific about the bic+cmp sequence, however
looking at the set of guests more closely shows that the 8-bit immediate
operand for the bic can only be satisfied with one guest in tree:
baseline m-profile -- 10-bit pages with aligned 4-byte memory ops.
Therefore it does not seem useful to keep this path.
for armv7 hosts. Rearranging the register allocation in
order to avoid overlap between the two ldrd pairs causes
the patch to be larger than it ordinarily would be.
cpu: Move the softmmu tlb to CPUNegativeOffsetState
We have for some time had code within the tcg backends to
handle large positive offsets from env. This move makes
sure that need not happen. Indeed, we are able to assert
at build time that simple offsets suffice for all hosts.
Amusingly, we had already ignored the comment to keep this value
at the end of CPUState. This restores the minimum negative offset
from TCG_AREG0 for code generation.
For the couple of uses within qom/cpu.c, without NEED_CPU_H, add
a pointer from the CPUState object to the IcountDecr object within
CPUNegativeOffsetState.
Cleanup in the boilerplate that each target must define.
Replace xtensa_env_get_cpu with env_archcpu. The combination
CPU(xtensa_env_get_cpu) should have used ENV_GET_CPU to begin;
use env_cpu now.
Move cpu_get_tb_cpu_state below the include of "exec/cpu-all.h"
so that the definition of env_cpu is available.
Cleanup in the boilerplate that each target must define.
Replace uc32_env_get_cpu with env_archcpu. The combination
CPU(uc32_env_get_cpu) should have used ENV_GET_CPU to begin;
use env_cpu now.
Cleanup in the boilerplate that each target must define.
Replace tricore_env_get_cpu with env_archcpu. The combination
CPU(tricore_env_get_cpu) should have used ENV_GET_CPU to begin;
use env_cpu now.
Cleanup in the boilerplate that each target must define.
Replace tilegx_env_get_cpu with env_archcpu. The combination
CPU(tilegx_env_get_cpu) should have used ENV_GET_CPU to begin;
use env_cpu now.
Cleanup in the boilerplate that each target must define.
Replace sparc_env_get_cpu with env_archcpu. The combination
CPU(sparc_env_get_cpu) should have used ENV_GET_CPU to begin;
use env_cpu now.
Cleanup in the boilerplate that each target must define.
Replace sh_env_get_cpu with env_archcpu. The combination
CPU(sh_env_get_cpu) should have used ENV_GET_CPU to begin;
use env_cpu now.
Cleanup in the boilerplate that each target must define.
Replace s390_env_get_cpu with env_archcpu. The combination
CPU(s390_env_get_cpu) should have used ENV_GET_CPU to begin;
use env_cpu now.
Cleanup in the boilerplate that each target must define.
Replace riscv_env_get_cpu with env_archcpu. The combination
CPU(riscv_env_get_cpu) should have used ENV_GET_CPU to begin;
use env_cpu now.
Cleanup in the boilerplate that each target must define.
Replace ppc_env_get_cpu with env_archcpu. The combination
CPU(ppc_env_get_cpu) should have used ENV_GET_CPU to begin;
use env_cpu now.
Cleanup in the boilerplate that each target must define.
Replace openrisc_env_get_cpu with env_archcpu. The combination
CPU(openrisc_env_get_cpu) should have used ENV_GET_CPU to begin;
use env_cpu now.
Cleanup in the boilerplate that each target must define.
Replace nios2_env_get_cpu with env_archcpu. The combination
CPU(nios2_env_get_cpu) should have used ENV_GET_CPU to begin;
use env_cpu now.
Cleanup in the boilerplate that each target must define.
Replace moxie_env_get_cpu with env_archcpu. The combination
CPU(moxie_env_get_cpu) should have used ENV_GET_CPU to begin;
use env_cpu now.
Cleanup in the boilerplate that each target must define.
Replace mips_env_get_cpu with env_archcpu. The combination
CPU(mips_env_get_cpu) should have used ENV_GET_CPU to begin;
use env_cpu now.
Cleanup in the boilerplate that each target must define.
Replace mb_env_get_cpu with env_archcpu. The combination
CPU(mb_env_get_cpu) should have used ENV_GET_CPU to begin;
use env_cpu now.
Move cpu_mmu_index below the include of "exec/cpu-all.h",
so that the definition of env_archcpu is available.
Cleanup in the boilerplate that each target must define.
Replace lm32_env_get_cpu with env_archcpu. The combination
CPU(lm32_env_get_cpu) should have used ENV_GET_CPU to begin;
use env_cpu now.
Cleanup in the boilerplate that each target must define.
Replace x86_env_get_cpu with env_archcpu. The combination
CPU(x86_env_get_cpu) should have used ENV_GET_CPU to begin;
use env_cpu now.
Cleanup in the boilerplate that each target must define.
Replace hppa_env_get_cpu with env_archcpu. The combination
CPU(hppa_env_get_cpu) should have used ENV_GET_CPU to begin;
use env_cpu now.
Cleanup in the boilerplate that each target must define.
Replace cris_env_get_cpu with env_archcpu. The combination
CPU(cris_env_get_cpu) should have used ENV_GET_CPU to begin;
use env_cpu now.
Cleanup in the boilerplate that each target must define.
Replace arm_env_get_cpu with env_archcpu. The combination
CPU(arm_env_get_cpu) should have used ENV_GET_CPU to begin;
use env_cpu now.
Cleanup in the boilerplate that each target must define.
Replace alpha_env_get_cpu with env_archcpu. The combination
CPU(alpha_env_get_cpu) should have used ENV_GET_CPU to begin;
use env_cpu now.
Move all softmmu tlb data into this structure. Arrange the
members so that we are able to place mask+table together and
at a smaller absolute offset from ENV.
For all targets, into this new file move TARGET_LONG_BITS,
TARGET_PAGE_BITS, TARGET_PHYS_ADDR_SPACE_BITS,
TARGET_VIRT_ADDR_SPACE_BITS, and NB_MMU_MODES.
Include this new file from exec/cpu-defs.h.
This now removes the somewhat odd requirement that target/arch/cpu.h
defines TARGET_LONG_BITS before including exec/cpu-defs.h, so push the
bulk of the includes within target/arch/cpu.h to the top.
The AccelType type was converted to AccelClass QOM
object on b14a0b7469f, and the original data type had
a field to store the option name which in turn was
used to search an accelerator. The lookup method
(accel_find) changed too, making the option field
unnecessary but it became AccelClass::opt_name despite
that. Therefore, and given that none accelerator
implementation sets AccelClass::opt_name, let's
remove this attribute.
Laurent Vivier [Wed, 29 May 2019 16:07:47 +0000 (18:07 +0200)]
numa: improve cpu hotplug error message with a wrong node-id
On pseries, core-ids are strongly binded to a node-id by the command
line option. If an user tries to add a CPU to the wrong node, he has
an error but it is not really helpful:
projects / qemu.git / log