qemu.git
5 years agoMerge remote-tracking branch 'remotes/rth/tags/pull-tcg-20200317' into staging
Peter Maydell [Wed, 18 Mar 2020 22:15:59 +0000 (22:15 +0000)]
Merge remote-tracking branch 'remotes/rth/tags/pull-tcg-20200317' into staging

Fix tcg/i386 bug vs sari_vec.
Fix tcg-runtime-gvec.c vs i386 without avx.

# gpg: Signature made Tue 17 Mar 2020 18:58:14 GMT
# gpg:                using RSA key 7A481E78868B4DB6A85A05C064DF38E8AF7E215F
# gpg:                issuer "richard.henderson@linaro.org"
# gpg: Good signature from "Richard Henderson <richard.henderson@linaro.org>" [full]
# Primary key fingerprint: 7A48 1E78 868B 4DB6 A85A  05C0 64DF 38E8 AF7E 215F

* remotes/rth/tags/pull-tcg-20200317:
  tcg: Remove tcg-runtime-gvec.c DO_CMP0
  tcg: Tidy tcg-runtime-gvec.c DUP*
  tcg: Tidy tcg-runtime-gvec.c types
  tcg: Remove CONFIG_VECTOR16
  tcg/i386: Bound shift count expanding sari_vec

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
5 years agoMerge remote-tracking branch 'remotes/stsquad/tags/pull-testing-and-gdbstub-170320...
Peter Maydell [Wed, 18 Mar 2020 20:25:23 +0000 (20:25 +0000)]
Merge remote-tracking branch 'remotes/stsquad/tags/pull-testing-and-gdbstub-170320-1' into staging

Testing and gdbstub updates:

  - docker updates for VirGL
  - re-factor gdbstub for static GDBState
  - re-factor gdbstub for dynamic arrays
  - add SVE support to arm gdbstub
  - add some guest debug tests to check-tcg
  - add aarch64 userspace register tests
  - remove packet size limit to gdbstub
  - simplify gdbstub monitor code
  - report vContSupported in gdbstub to use proper single-step

# gpg: Signature made Tue 17 Mar 2020 17:47:46 GMT
# gpg:                using RSA key 6685AE99E75167BCAFC8DF35FBD0DB095A9E2A44
# gpg: Good signature from "Alex Bennée (Master Work Key) <alex.bennee@linaro.org>" [full]
# Primary key fingerprint: 6685 AE99 E751 67BC AFC8  DF35 FBD0 DB09 5A9E 2A44

* remotes/stsquad/tags/pull-testing-and-gdbstub-170320-1: (28 commits)
  gdbstub: Fix single-step issue by confirming 'vContSupported+' feature to gdb
  gdbstub: do not split gdb_monitor_write payload
  gdbstub: change GDBState.last_packet to GByteArray
  tests/tcg/aarch64: add test-sve-ioctl guest-debug test
  tests/tcg/aarch64: add SVE iotcl test
  tests/tcg/aarch64: add a gdbstub testcase for SVE registers
  tests/guest-debug: add a simple test runner
  configure: allow user to specify what gdb to use
  tests/tcg/aarch64: userspace system register test
  target/arm: don't bother with id_aa64pfr0_read for USER_ONLY
  target/arm: generate xml description of our SVE registers
  target/arm: default SVE length to 64 bytes for linux-user
  target/arm: explicitly encode regnum in our XML
  target/arm: prepare for multiple dynamic XMLs
  gdbstub: extend GByteArray to read register helpers
  target/i386: use gdb_get_reg helpers
  target/m68k: use gdb_get_reg helpers
  target/arm: use gdb_get_reg helpers
  gdbstub: add helper for 128 bit registers
  gdbstub: move mem_buf to GDBState and use GByteArray
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
5 years agoMerge remote-tracking branch 'remotes/armbru/tags/pull-error-2020-03-17' into staging
Peter Maydell [Wed, 18 Mar 2020 17:57:40 +0000 (17:57 +0000)]
Merge remote-tracking branch 'remotes/armbru/tags/pull-error-2020-03-17' into staging

Error reporting patches for 2020-03-17

# gpg: Signature made Tue 17 Mar 2020 16:30:49 GMT
# gpg:                using RSA key 354BC8B3D7EB2A6B68674E5F3870B400EB918653
# gpg:                issuer "armbru@redhat.com"
# gpg: Good signature from "Markus Armbruster <armbru@redhat.com>" [full]
# gpg:                 aka "Markus Armbruster <armbru@pond.sub.org>" [full]
# Primary key fingerprint: 354B C8B3 D7EB 2A6B 6867  4E5F 3870 B400 EB91 8653

* remotes/armbru/tags/pull-error-2020-03-17:
  hw/sd/ssi-sd: fix error handling in ssi_sd_realize
  xen-block: Use one Error * variable instead of two
  hw/misc/ivshmem: Use one Error * variable instead of two
  Use &error_abort instead of separate assert()

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
5 years agoMerge remote-tracking branch 'remotes/dgibson/tags/ppc-for-5.0-20200317' into staging
Peter Maydell [Wed, 18 Mar 2020 15:07:57 +0000 (15:07 +0000)]
Merge remote-tracking branch 'remotes/dgibson/tags/ppc-for-5.0-20200317' into staging

ppc patch queue 2020-03-17

Here's my final pull request for the qemu-5.0 soft freeze.  Sorry this
is just under the wire - I hit some last minute problems that took a
while to fix up and retest.

Highlights are:
 * Numerous fixes for the FWNMI feature
 * A handful of cleanups to the device tree construction code
 * Numerous fixes for the spapr-vscsi device
 * A number of fixes and cleanups for real mode (MMU off) softmmu
   handling
 * Fixes for handling of the PAPR RMA
 * Better handling of hotplug/unplug events during boot
 * Assorted other fixes

# gpg: Signature made Tue 17 Mar 2020 09:55:07 GMT
# gpg:                using RSA key 75F46586AE61A66CC44E87DC6C38CACA20D9B392
# gpg: Good signature from "David Gibson <david@gibson.dropbear.id.au>" [full]
# gpg:                 aka "David Gibson (Red Hat) <dgibson@redhat.com>" [full]
# gpg:                 aka "David Gibson (ozlabs.org) <dgibson@ozlabs.org>" [full]
# gpg:                 aka "David Gibson (kernel.org) <dwg@kernel.org>" [unknown]
# Primary key fingerprint: 75F4 6586 AE61 A66C C44E  87DC 6C38 CACA 20D9 B392

* remotes/dgibson/tags/ppc-for-5.0-20200317: (45 commits)
  pseries: Update SLOF firmware image
  ppc/spapr: Ignore common "ibm,nmi-interlock" Linux bug
  ppc/spapr: Implement FWNMI System Reset delivery
  target/ppc: allow ppc_cpu_do_system_reset to take an alternate vector
  ppc/spapr: Allow FWNMI on TCG
  ppc/spapr: Fix FWNMI machine check interrupt delivery
  ppc/spapr: Add FWNMI System Reset state
  ppc/spapr: Change FWNMI names
  ppc/spapr: Fix FWNMI machine check failure handling
  spapr: Rename DT functions to newer naming convention
  spapr: Move creation of ibm,architecture-vec-5 property
  spapr: Move creation of ibm,dynamic-reconfiguration-memory dt node
  spapr/rtas: Reserve space for RTAS blob and log
  pseries: Update SLOF firmware image
  ppc/spapr: Move GPRs setup to one place
  target/ppc: Fix rlwinm on ppc64
  spapr/xive: use SPAPR_IRQ_IPI to define IPI ranges exposed to the guest
  hw/scsi/spapr_vscsi: Convert debug fprintf() to trace event
  hw/scsi/spapr_vscsi: Prevent buffer overflow
  hw/scsi/spapr_vscsi: Do not mix SRP IU size with DMA buffer size
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
5 years agoMerge remote-tracking branch 'remotes/bonzini/tags/for-upstream' into staging
Peter Maydell [Tue, 17 Mar 2020 18:33:05 +0000 (18:33 +0000)]
Merge remote-tracking branch 'remotes/bonzini/tags/for-upstream' into staging

* Bugfixes all over the place
* get/set_uint cleanups (Felipe)
* Lock guard support (Stefan)
* MemoryRegion ownership cleanup (Philippe)
* AVX512 optimization for buffer_is_zero (Robert)

# gpg: Signature made Tue 17 Mar 2020 15:01:54 GMT
# gpg:                using RSA key BFFBD25F78C7AE83
# gpg: Good signature from "Paolo Bonzini <bonzini@gnu.org>" [full]
# gpg:                 aka "Paolo Bonzini <pbonzini@redhat.com>" [full]
# Primary key fingerprint: 46F5 9FBD 57D6 12E7 BFD4  E2F7 7E15 100C CD36 69B1
#      Subkey fingerprint: F133 3857 4B66 2389 866C  7682 BFFB D25F 78C7 AE83

* remotes/bonzini/tags/for-upstream: (62 commits)
  hw/arm: Let devices own the MemoryRegion they create
  hw/arm: Remove unnecessary memory_region_set_readonly() on ROM alias
  hw/ppc/ppc405: Use memory_region_init_rom() with read-only regions
  hw/arm/stm32: Use memory_region_init_rom() with read-only regions
  hw/char: Let devices own the MemoryRegion they create
  hw/riscv: Let devices own the MemoryRegion they create
  hw/dma: Let devices own the MemoryRegion they create
  hw/display: Let devices own the MemoryRegion they create
  hw/core: Let devices own the MemoryRegion they create
  scripts/cocci: Patch to let devices own their MemoryRegions
  scripts/cocci: Patch to remove unnecessary memory_region_set_readonly()
  scripts/cocci: Patch to detect potential use of memory_region_init_rom
  hw/sparc: Use memory_region_init_rom() with read-only regions
  hw/sh4: Use memory_region_init_rom() with read-only regions
  hw/riscv: Use memory_region_init_rom() with read-only regions
  hw/ppc: Use memory_region_init_rom() with read-only regions
  hw/pci-host: Use memory_region_init_rom() with read-only regions
  hw/net: Use memory_region_init_rom() with read-only regions
  hw/m68k: Use memory_region_init_rom() with read-only regions
  hw/display: Use memory_region_init_rom() with read-only regions
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
5 years agoMerge remote-tracking branch 'remotes/stefanha/tags/block-pull-request' into staging
Peter Maydell [Tue, 17 Mar 2020 17:41:48 +0000 (17:41 +0000)]
Merge remote-tracking branch 'remotes/stefanha/tags/block-pull-request' into staging

Pull request

# gpg: Signature made Tue 17 Mar 2020 15:18:23 GMT
# gpg:                using RSA key 8695A8BFD3F97CDAAC35775A9CA4ABB381AB73C8
# gpg: Good signature from "Stefan Hajnoczi <stefanha@redhat.com>" [full]
# gpg:                 aka "Stefan Hajnoczi <stefanha@gmail.com>" [full]
# Primary key fingerprint: 8695 A8BF D3F9 7CDA AC35  775A 9CA4 ABB3 81AB 73C8

* remotes/stefanha/tags/block-pull-request:
  block/io: fix bdrv_co_do_copy_on_readv

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
5 years agogdbstub: Fix single-step issue by confirming 'vContSupported+' feature to gdb
Changbin Du [Mon, 16 Mar 2020 17:21:55 +0000 (17:21 +0000)]
gdbstub: Fix single-step issue by confirming 'vContSupported+' feature to gdb

Recently when debugging an arm32 system on qemu, I found sometimes the
single-step command (stepi) is not working. This can be reproduced by
below steps:
 1) start qemu-system-arm -s -S .. and wait for gdb connection.
 2) start gdb and connect to qemu. In my case, gdb gets a wrong value
    (0x60) for PC, which is an another bug.
 3) After connected, type 'stepi' and expect it will stop at next ins.

But, it has never stopped. This because:
 1) We doesn't report ‘vContSupported’ feature to gdb explicitly and gdb
    think we do not support it. In this case, gdb use a software breakpoint
    to emulate single-step.
 2) Since gdb gets a wrong initial value of PC, then gdb inserts a
    breakpoint to wrong place (PC+4).

Not only for the arm target, Philippe has also encountered this on MIPS.
Probably gdb has different assumption for different architectures.

Since we do support ‘vContSupported’ query command, so let's tell gdb that
we support it.

Before this change, gdb send below 'Z0' packet to implement single-step:
gdb_handle_packet: Z0,4,4

After this change, gdb send "vCont;s.." which is expected:
gdb_handle_packet: vCont?
put_packet: vCont;c;C;s;S
gdb_handle_packet: vCont;s:p1.1;c:p1.-1

Signed-off-by: Changbin Du <changbin.du@gmail.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20200221002559.6768-1-changbin.du@gmail.com>
[AJB: fix for static gdbstub]
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Luc Michel <luc.michel@greensocs.com>
Message-Id: <20200316172155.971-29-alex.bennee@linaro.org>

5 years agogdbstub: do not split gdb_monitor_write payload
Damien Hedde [Mon, 16 Mar 2020 17:21:54 +0000 (17:21 +0000)]
gdbstub: do not split gdb_monitor_write payload

Since we can now send packets of arbitrary length:
simplify gdb_monitor_write() and send the whole payload
in one packet.

Suggested-by: Luc Michel <luc.michel@greensocs.com>
Signed-off-by: Damien Hedde <damien.hedde@greensocs.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20191211160514.58373-3-damien.hedde@greensocs.com>
Message-Id: <20200316172155.971-28-alex.bennee@linaro.org>

5 years agogdbstub: change GDBState.last_packet to GByteArray
Damien Hedde [Mon, 16 Mar 2020 17:21:53 +0000 (17:21 +0000)]
gdbstub: change GDBState.last_packet to GByteArray

Remove the packet size upper limit by using a GByteArray
instead of a statically allocated array for last_packet.
Thus we can now send big packets.

Also remove the last_packet_len field and use last_packet->len
instead.

Signed-off-by: Damien Hedde <damien.hedde@greensocs.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20191211160514.58373-2-damien.hedde@greensocs.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20200316172155.971-27-alex.bennee@linaro.org>

5 years agotests/tcg/aarch64: add test-sve-ioctl guest-debug test
Alex Bennée [Mon, 16 Mar 2020 17:21:52 +0000 (17:21 +0000)]
tests/tcg/aarch64: add test-sve-ioctl guest-debug test

This test exercises the gdbstub while runing the sve-iotcl test. I
haven't plubmed it into make system as we need a way of verifying if
gdb has the right support for SVE.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20200316172155.971-26-alex.bennee@linaro.org>

5 years agotests/tcg/aarch64: add SVE iotcl test
Alex Bennée [Mon, 16 Mar 2020 17:21:51 +0000 (17:21 +0000)]
tests/tcg/aarch64: add SVE iotcl test

This is a fairly bare-bones test of setting the various vector sizes
for SVE which will only fail if the PR_SVE_SET_VL can't reduce the
user-space vector length by powers of 2.

However we will also be able to use it in a future test which
exercises the GDB stub.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20200316172155.971-25-alex.bennee@linaro.org>

5 years agotests/tcg/aarch64: add a gdbstub testcase for SVE registers
Alex Bennée [Mon, 16 Mar 2020 17:21:50 +0000 (17:21 +0000)]
tests/tcg/aarch64: add a gdbstub testcase for SVE registers

A very simple test case which sets and reads SVE registers while
running a test case. We don't really need to compile a SVE binary for
this case but we will later so keep it simple for now.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20200316172155.971-24-alex.bennee@linaro.org>

5 years agotests/guest-debug: add a simple test runner
Alex Bennée [Mon, 16 Mar 2020 17:21:49 +0000 (17:21 +0000)]
tests/guest-debug: add a simple test runner

The test runners job is to start QEMU with guest debug enabled and
then spawn a gdb process running a test script that exercises the
functionality it wants to test.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20200316172155.971-23-alex.bennee@linaro.org>

5 years agoconfigure: allow user to specify what gdb to use
Alex Bennée [Mon, 16 Mar 2020 17:21:48 +0000 (17:21 +0000)]
configure: allow user to specify what gdb to use

This is useful, especially when testing relatively new gdbstub
features that might not be in distro packages yet.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20200316172155.971-22-alex.bennee@linaro.org>

5 years agotests/tcg/aarch64: userspace system register test
Alex Bennée [Mon, 16 Mar 2020 17:21:47 +0000 (17:21 +0000)]
tests/tcg/aarch64: userspace system register test

This tests a bunch of registers that the kernel allows userspace to
read including the CPUID registers. We need a SVE aware compiler as we
are testing the id_aa64zfr0_el1 register in the set.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20200316172155.971-21-alex.bennee@linaro.org>

5 years agotarget/arm: don't bother with id_aa64pfr0_read for USER_ONLY
Alex Bennée [Mon, 16 Mar 2020 17:21:46 +0000 (17:21 +0000)]
target/arm: don't bother with id_aa64pfr0_read for USER_ONLY

For system emulation we need to check the state of the GIC before we
report the value. However this isn't relevant to exporting of the
value to linux-user and indeed breaks the exported value as set by
modify_arm_cp_regs.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20200316172155.971-20-alex.bennee@linaro.org>

5 years agotarget/arm: generate xml description of our SVE registers
Alex Bennée [Mon, 16 Mar 2020 17:21:45 +0000 (17:21 +0000)]
target/arm: generate xml description of our SVE registers

We also expose a the helpers to read/write the the registers.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Acked-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20200316172155.971-19-alex.bennee@linaro.org>

5 years agotarget/arm: default SVE length to 64 bytes for linux-user
Alex Bennée [Mon, 16 Mar 2020 17:21:44 +0000 (17:21 +0000)]
target/arm: default SVE length to 64 bytes for linux-user

The Linux kernel chooses the default of 64 bytes for SVE registers on
the basis that it is the largest size on known hardware that won't
grow the signal frame. We still honour the sve-max-vq property and
userspace can expand the number of lanes by calling PR_SVE_SET_VL.

This should not make any difference to SVE enabled software as the SVE
is of course vector length agnostic.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20200316172155.971-18-alex.bennee@linaro.org>

5 years agotarget/arm: explicitly encode regnum in our XML
Alex Bennée [Mon, 16 Mar 2020 17:21:43 +0000 (17:21 +0000)]
target/arm: explicitly encode regnum in our XML

This is described as optional but I'm not convinced of the numbering
when multiple target fragments are sent.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20200316172155.971-17-alex.bennee@linaro.org>

5 years agotarget/arm: prepare for multiple dynamic XMLs
Alex Bennée [Mon, 16 Mar 2020 17:21:42 +0000 (17:21 +0000)]
target/arm: prepare for multiple dynamic XMLs

We will want to generate similar dynamic XML for gdbstub support of
SVE registers (the upstream doesn't use XML). To that end lightly
rename a few things to make the distinction.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Acked-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20200316172155.971-16-alex.bennee@linaro.org>

5 years agogdbstub: extend GByteArray to read register helpers
Alex Bennée [Mon, 16 Mar 2020 17:21:41 +0000 (17:21 +0000)]
gdbstub: extend GByteArray to read register helpers

Instead of passing a pointer to memory now just extend the GByteArray
to all the read register helpers. They can then safely append their
data through the normal way. We don't bother with this abstraction for
write registers as we have already ensured the buffer being copied
from is the correct size.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Acked-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Damien Hedde <damien.hedde@greensocs.com>
Message-Id: <20200316172155.971-15-alex.bennee@linaro.org>

5 years agotarget/i386: use gdb_get_reg helpers
Alex Bennée [Mon, 16 Mar 2020 17:21:40 +0000 (17:21 +0000)]
target/i386: use gdb_get_reg helpers

This is cleaner than poking memory directly and will make later
clean-ups easier.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20200316172155.971-14-alex.bennee@linaro.org>

5 years agotarget/m68k: use gdb_get_reg helpers
Alex Bennée [Mon, 16 Mar 2020 17:21:39 +0000 (17:21 +0000)]
target/m68k: use gdb_get_reg helpers

This is cleaner than poking memory directly and will make later
clean-ups easier.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20200316172155.971-13-alex.bennee@linaro.org>

5 years agotarget/arm: use gdb_get_reg helpers
Alex Bennée [Mon, 16 Mar 2020 17:21:38 +0000 (17:21 +0000)]
target/arm: use gdb_get_reg helpers

This is cleaner than poking memory directly and will make later
clean-ups easier.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20200316172155.971-12-alex.bennee@linaro.org>

5 years agogdbstub: add helper for 128 bit registers
Alex Bennée [Mon, 16 Mar 2020 17:21:37 +0000 (17:21 +0000)]
gdbstub: add helper for 128 bit registers

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20200316172155.971-11-alex.bennee@linaro.org>

5 years agogdbstub: move mem_buf to GDBState and use GByteArray
Alex Bennée [Mon, 16 Mar 2020 17:21:36 +0000 (17:21 +0000)]
gdbstub: move mem_buf to GDBState and use GByteArray

This is in preparation for further re-factoring of the register API
with the rest of the code. Theoretically the read register function
could overwrite the MAX_PACKET_LENGTH buffer although currently all
registers are well within the size range.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Damien Hedde <damien.hedde@greensocs.com>
Tested-by: Damien Hedde <damien.hedde@greensocs.com>
Message-Id: <20200316172155.971-10-alex.bennee@linaro.org>

5 years agogdbstub: move str_buf to GDBState and use GString
Alex Bennée [Mon, 16 Mar 2020 17:21:35 +0000 (17:21 +0000)]
gdbstub: move str_buf to GDBState and use GString

Rather than having a static buffer replace str_buf with a GString
which we know can grow on demand. Convert the internal functions to
take a GString instead of a char * and length.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Damien Hedde <damien.hedde@greensocs.com>
Tested-by: Damien Hedde <damien.hedde@greensocs.com>
Message-Id: <20200316172155.971-9-alex.bennee@linaro.org>

5 years agogdbstub: stop passing GDBState * around and use global
Alex Bennée [Mon, 16 Mar 2020 17:21:34 +0000 (17:21 +0000)]
gdbstub: stop passing GDBState * around and use global

We only have one GDBState which should be allocated at the time we
process any commands. This will make further clean-up a bit easier.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Damien Hedde <damien.hedde@greensocs.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20200316172155.971-8-alex.bennee@linaro.org>

5 years agogdbstub: make GDBState static and have common init function
Alex Bennée [Mon, 16 Mar 2020 17:21:33 +0000 (17:21 +0000)]
gdbstub: make GDBState static and have common init function

Instead of allocating make this entirely static. We shall reduce the
size of the structure in later commits and dynamically allocate parts
of it. We introduce an init and reset helper function to keep all the
manipulation in one place.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Damien Hedde <damien.hedde@greensocs.com>
Message-Id: <20200316172155.971-7-alex.bennee@linaro.org>

5 years agotravis.yml: Set G_MESSAGES_DEBUG do report GLib errors
Philippe Mathieu-Daudé [Mon, 16 Mar 2020 17:21:32 +0000 (17:21 +0000)]
travis.yml: Set G_MESSAGES_DEBUG do report GLib errors

Since commit f5852efa293 we can display GLib errors with the QEMU
error reporting API. Set it to the 'error' level, as this helps
understanding failures from QEMU calls to GLib on Travis-CI.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20200316101544.22361-1-philmd@redhat.com>
Message-Id: <20200316172155.971-6-alex.bennee@linaro.org>

5 years agotests/docker: Update VirGL to v0.8.0
Philippe Mathieu-Daudé [Mon, 16 Mar 2020 17:21:31 +0000 (17:21 +0000)]
tests/docker: Update VirGL to v0.8.0

Building the qemu:debian-amd64 fails when building VirGL:

  make[2]: Entering directory '/usr/src/virglrenderer/src/gallium/auxiliary'
    CC       cso_cache/cso_cache.lo
    CC       cso_cache/cso_hash.lo
    CC       os/os_misc.lo
    CC       util/u_debug.lo
    CC       util/u_debug_describe.lo
    CC       util/u_format.lo
    GEN      util/u_format_table.c
  Traceback (most recent call last):
    File "./util/u_format_table.py", line 168, in <module>
      main()
    File "./util/u_format_table.py", line 164, in main
      write_format_table(formats)
    File "./util/u_format_table.py", line 132, in write_format_table
      print("   %s,\t/* is_array */" % (bool_map(format.is_array()),))
    File "/usr/src/virglrenderer/src/gallium/auxiliary/util/u_format_parse.py", line 164, in is_array
      return self.array_element() != None
    File "/usr/src/virglrenderer/src/gallium/auxiliary/util/u_format_parse.py", line 73, in __eq__
      return self.type == other.type and self.norm == other.norm and self.pure == other.pure and self.size == other.size
  AttributeError: 'NoneType' object has no attribute 'type'
  make[2]: Leaving directory '/usr/src/virglrenderer/src/gallium/auxiliary'
  make[2]: *** [Makefile:906: util/u_format_table.c] Error 1
  make[1]: *** [Makefile:631: install-recursive] Error 1

VirGL commits a8962eda1..a613dcc82 fix this problem.
Update to VirGL 0.8.0 which contains them.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20200212202709.12665-4-philmd@redhat.com>
Message-Id: <20200316172155.971-5-alex.bennee@linaro.org>

5 years agotests/docker: Remove obsolete VirGL --with-glx configure option
Philippe Mathieu-Daudé [Mon, 16 Mar 2020 17:21:30 +0000 (17:21 +0000)]
tests/docker: Remove obsolete VirGL --with-glx configure option

The GLX configure option has been removed in 71c75f201d [*].
We missed that when updating to v0.7.0 in commit fab3220f97.

This silents:

  configure: creating ./config.status
  config.status: creating virglrenderer.pc
  ...
  configure: WARNING: unrecognized options: --with-glx

[*] https://gitlab.freedesktop.org/virgl/virglrenderer/commit/71c75f201d

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20200212202709.12665-3-philmd@redhat.com>
Message-Id: <20200316172155.971-4-alex.bennee@linaro.org>

5 years agotests/docker: Update VirGL git repository URL
Philippe Mathieu-Daudé [Mon, 16 Mar 2020 17:21:29 +0000 (17:21 +0000)]
tests/docker: Update VirGL git repository URL

freedesktop.org is moving to a GitLab instance,
use the new url.

- https://www.fooishbar.org/blog/gitlab-fdo-introduction/
- https://gitlab.freedesktop.org/freedesktop/freedesktop/-/wikis/home

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20200212202709.12665-2-philmd@redhat.com>
Message-Id: <20200316172155.971-3-alex.bennee@linaro.org>

5 years agotests/docker: Install tools to cross-debug and build Linux kernels
Philippe Mathieu-Daudé [Mon, 16 Mar 2020 17:21:28 +0000 (17:21 +0000)]
tests/docker: Install tools to cross-debug and build Linux kernels

We often run Linux kernels to test QEMU. We sometimes need
to build them manually to use non-default features. We only
miss the tiny 'bc' tool.

The ncurses library is helpful to run 'make menuconfig'.

Finally, gdb-multiarch allow us to debug a TCG guest when its
architecture is different than the host.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20200212202738.12986-1-philmd@redhat.com>
Message-Id: <20200316172155.971-2-alex.bennee@linaro.org>

5 years agohw/sd/ssi-sd: fix error handling in ssi_sd_realize
Vladimir Sementsov-Ogievskiy [Tue, 17 Mar 2020 12:57:41 +0000 (15:57 +0300)]
hw/sd/ssi-sd: fix error handling in ssi_sd_realize

It's wrong to use same err object as errp parameter for several
function calls without intermediate checking for error: we'll crash if
try to set err object twice. Fix that.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20200317125741.15301-1-vsementsov@virtuozzo.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
5 years agoxen-block: Use one Error * variable instead of two
Markus Armbruster [Fri, 13 Mar 2020 17:05:17 +0000 (18:05 +0100)]
xen-block: Use one Error * variable instead of two

While there, tidy up indentation, and add return just for consistency
and robustness.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20200313170517.22480-4-armbru@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
[The "while there" cleanups squashed in]

5 years agoMerge remote-tracking branch 'remotes/amarkovic/tags/mips-queue-mar-17-2020' into...
Peter Maydell [Tue, 17 Mar 2020 16:23:03 +0000 (16:23 +0000)]
Merge remote-tracking branch 'remotes/amarkovic/tags/mips-queue-mar-17-2020' into staging

MIPS queue for March 17th, 2020

# gpg: Signature made Tue 17 Mar 2020 12:53:50 GMT
# gpg:                using RSA key D4972A8967F75A65
# gpg: Good signature from "Aleksandar Markovic <amarkovic@wavecomp.com>" [full]
# Primary key fingerprint: 8526 FBF1 5DA3 811F 4A01  DD75 D497 2A89 67F7 5A65

* remotes/amarkovic/tags/mips-queue-mar-17-2020:
  MAINTAINERS: Add a file to the main MIPS section
  MAINTAINERS: Adjust maintainer's email
  MAINTAINERS: Adjust maintainer's status for some MIPS items

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
5 years agotcg: Remove tcg-runtime-gvec.c DO_CMP0
Richard Henderson [Mon, 17 Feb 2020 02:43:13 +0000 (18:43 -0800)]
tcg: Remove tcg-runtime-gvec.c DO_CMP0

Partial cleanup from the CONFIG_VECTOR16 removal.
Replace DO_CMP0 with its scalar expansion, a simple negation.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
5 years agotcg: Tidy tcg-runtime-gvec.c DUP*
Richard Henderson [Mon, 17 Feb 2020 02:41:23 +0000 (18:41 -0800)]
tcg: Tidy tcg-runtime-gvec.c DUP*

Partial cleanup from the CONFIG_VECTOR16 removal.
Replace the DUP* expansions with the scalar argument.

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
5 years agotcg: Tidy tcg-runtime-gvec.c types
Richard Henderson [Mon, 17 Feb 2020 02:35:26 +0000 (18:35 -0800)]
tcg: Tidy tcg-runtime-gvec.c types

Partial cleanup from the CONFIG_VECTOR16 removal.
Replace the vec* types with their scalar expansions.

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
5 years agotcg: Remove CONFIG_VECTOR16
Richard Henderson [Mon, 17 Feb 2020 01:43:11 +0000 (17:43 -0800)]
tcg: Remove CONFIG_VECTOR16

The comment in tcg-runtime-gvec.c about CONFIG_VECTOR16 says that
tcg-op-gvec.c has eliminated size 8 vectors, and only passes on
multiples of 16.  This may have been true of the first few operations,
but is not true of all operations.

In particular, multiply, shift by scalar, and compare of 8- and 16-bit
elements are not expanded inline if host vector operations are not
supported.

For an x86_64 host that does not support AVX, this means that we will
fall back to the helper, which will attempt to use SSE instructions,
which will SEGV on an invalid 8-byte aligned memory operation.

This patch simply removes the CONFIG_VECTOR16 code and configuration
without further simplification.

Buglink: https://bugs.launchpad.net/bugs/1863508
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
5 years agotcg/i386: Bound shift count expanding sari_vec
Richard Henderson [Wed, 11 Mar 2020 05:14:26 +0000 (22:14 -0700)]
tcg/i386: Bound shift count expanding sari_vec

A given RISU testcase for SVE can produce

tcg-op-vec.c:511: do_shifti: Assertion `i >= 0 && i < (8 << vece)' failed.

because expand_vec_sari gave a shift count of 32 to a MO_32
vector shift.

In 44f1441dbe1, we changed from direct expansion of vector opcodes
to re-use of the tcg expanders.  So while the comment correctly notes
that the hw will handle such a shift count, we now have to take our
own sanity checks into account.  Which is easy in this particular case.

Fixes: 44f1441dbe1
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
5 years agohw/misc/ivshmem: Use one Error * variable instead of two
Markus Armbruster [Fri, 13 Mar 2020 17:05:16 +0000 (18:05 +0100)]
hw/misc/ivshmem: Use one Error * variable instead of two

Commit fe44dc9180 "migration: disallow migrate_add_blocker during
migration" accidentally added a second Error * variable.  Use the
first one instead.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20200313170517.22480-3-armbru@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
5 years agoUse &error_abort instead of separate assert()
Markus Armbruster [Fri, 13 Mar 2020 17:05:15 +0000 (18:05 +0100)]
Use &error_abort instead of separate assert()

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20200313170517.22480-2-armbru@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Acked-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
[Unused Error *variable deleted]

5 years agoMerge remote-tracking branch 'remotes/pmaydell/tags/pull-target-arm-20200317' into...
Peter Maydell [Tue, 17 Mar 2020 14:44:49 +0000 (14:44 +0000)]
Merge remote-tracking branch 'remotes/pmaydell/tags/pull-target-arm-20200317' into staging

target-arm:
 * hw/arm/pxa2xx: Do not wire up OHCI for PXA255
 * aspeed/smc: Fix number of dummy cycles for FAST_READ_4 command
 * m25p80: Improve command handling for Jedec and unsupported commands
 * hw/net/imx_fec: write TGSR and TCSR3 in imx_enet_write()
 * hw/arm/fsl-imx6, imx6ul: Wire up USB controllers
 * hw/arm/fsl-imx6ul: Instantiate unimplemented pwm and can devices

# gpg: Signature made Tue 17 Mar 2020 11:40:01 GMT
# gpg:                using RSA key E1A5C593CD419DE28E8315CF3C2525ED14360CDE
# gpg:                issuer "peter.maydell@linaro.org"
# gpg: Good signature from "Peter Maydell <peter.maydell@linaro.org>" [ultimate]
# gpg:                 aka "Peter Maydell <pmaydell@gmail.com>" [ultimate]
# gpg:                 aka "Peter Maydell <pmaydell@chiark.greenend.org.uk>" [ultimate]
# Primary key fingerprint: E1A5 C593 CD41 9DE2 8E83  15CF 3C25 25ED 1436 0CDE

* remotes/pmaydell/tags/pull-target-arm-20200317:
  hw/arm/pxa2xx: Do not wire up OHCI for PXA255
  aspeed/smc: Fix number of dummy cycles for FAST_READ_4 command
  m25p80: Improve command handling for unsupported commands
  m25p80: Improve command handling for Jedec commands
  m25p80: Convert to support tracing
  hw/net/imx_fec: write TGSR and TCSR3 in imx_enet_write()
  hw/arm/fsl-imx6: Wire up USB controllers
  hw/arm/fsl-imx6ul: Wire up USB controllers
  hw/arm/fsl-imx6ul: Instantiate unimplemented pwm and can devices
  hw/arm/fsl-imx6ul: Fix USB interrupt numbers
  hw/usb: Add basic i.MX USB Phy support

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
5 years agohw/arm: Let devices own the MemoryRegion they create
Philippe Mathieu-Daudé [Mon, 24 Feb 2020 17:32:23 +0000 (18:32 +0100)]
hw/arm: Let devices own the MemoryRegion they create

Avoid orphan memory regions being added in the /unattached QOM
container.

This commit was produced with the Coccinelle script
scripts/coccinelle/memory-region-housekeeping.cocci.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
5 years agohw/arm: Remove unnecessary memory_region_set_readonly() on ROM alias
Philippe Mathieu-Daudé [Mon, 24 Feb 2020 15:59:18 +0000 (16:59 +0100)]
hw/arm: Remove unnecessary memory_region_set_readonly() on ROM alias

This commit was produced with the Coccinelle script
scripts/coccinelle/memory-region-housekeeping.cocci.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
5 years agohw/ppc/ppc405: Use memory_region_init_rom() with read-only regions
Philippe Mathieu-Daudé [Mon, 24 Feb 2020 16:04:51 +0000 (17:04 +0100)]
hw/ppc/ppc405: Use memory_region_init_rom() with read-only regions

The scripts/coccinelle/memory-region-housekeeping.cocci reported:
* TODO [[view:./hw/ppc/ppc405_boards.c::face=ovl-face1::linb=195::colb=8::cole=30][potential use of memory_region_init_rom*() in  ./hw/ppc/ppc405_boards.c::195]]
* TODO [[view:./hw/ppc/ppc405_boards.c::face=ovl-face1::linb=464::colb=8::cole=30][potential use of memory_region_init_rom*() in  ./hw/ppc/ppc405_boards.c::464]]

We can indeed replace the memory_region_init_ram() and
memory_region_set_readonly() calls by memory_region_init_rom().

Acked-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
5 years agohw/arm/stm32: Use memory_region_init_rom() with read-only regions
Philippe Mathieu-Daudé [Mon, 24 Feb 2020 15:58:53 +0000 (16:58 +0100)]
hw/arm/stm32: Use memory_region_init_rom() with read-only regions

The scripts/coccinelle/memory-region-housekeeping.cocci reported:
* TODO [[view:./hw/arm/stm32f205_soc.c::face=ovl-face1::linb=96::colb=4::cole=26][potential use of memory_region_init_rom*() in  ./hw/arm/stm32f205_soc.c::96]]
* TODO [[view:./hw/arm/stm32f405_soc.c::face=ovl-face1::linb=98::colb=4::cole=26][potential use of memory_region_init_rom*() in  ./hw/arm/stm32f405_soc.c::98]]

We can indeed replace the memory_region_init_ram() and
memory_region_set_readonly() calls by memory_region_init_rom().

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
5 years agohw/char: Let devices own the MemoryRegion they create
Philippe Mathieu-Daudé [Sat, 22 Feb 2020 17:11:33 +0000 (18:11 +0100)]
hw/char: Let devices own the MemoryRegion they create

Avoid orphan memory regions being added in the /unattached QOM
container.

This commit was produced with the Coccinelle script
scripts/coccinelle/memory-region-housekeeping.cocci.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
5 years agohw/riscv: Let devices own the MemoryRegion they create
Philippe Mathieu-Daudé [Sat, 22 Feb 2020 17:12:57 +0000 (18:12 +0100)]
hw/riscv: Let devices own the MemoryRegion they create

Avoid orphan memory regions being added in the /unattached QOM
container.

This commit was produced with the Coccinelle script
scripts/coccinelle/memory-region-housekeeping.cocci.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
5 years agohw/dma: Let devices own the MemoryRegion they create
Philippe Mathieu-Daudé [Sat, 22 Feb 2020 17:12:37 +0000 (18:12 +0100)]
hw/dma: Let devices own the MemoryRegion they create

Avoid orphan memory regions being added in the /unattached QOM
container.

This commit was produced with the Coccinelle script
scripts/coccinelle/memory-region-housekeeping.cocci.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
5 years agohw/display: Let devices own the MemoryRegion they create
Philippe Mathieu-Daudé [Sat, 22 Feb 2020 17:12:09 +0000 (18:12 +0100)]
hw/display: Let devices own the MemoryRegion they create

Avoid orphan memory regions being added in the /unattached QOM
container.

This commit was produced with the Coccinelle script
scripts/coccinelle/memory-region-housekeeping.cocci.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
5 years agohw/core: Let devices own the MemoryRegion they create
Philippe Mathieu-Daudé [Sat, 22 Feb 2020 17:11:51 +0000 (18:11 +0100)]
hw/core: Let devices own the MemoryRegion they create

Avoid orphan memory regions being added in the /unattached QOM
container.

This commit was produced with the Coccinelle script
scripts/coccinelle/memory-region-housekeeping.cocci.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
5 years agoscripts/cocci: Patch to let devices own their MemoryRegions
Philippe Mathieu-Daudé [Fri, 21 Feb 2020 15:05:08 +0000 (16:05 +0100)]
scripts/cocci: Patch to let devices own their MemoryRegions

When a device creates a MemoryRegion without setting its ownership,
the MemoryRegion is added to the machine "/unattached" container in
the QOM tree.

Example with the Samsung SMDKC210 board:

  $ arm-softmmu/qemu-system-arm -M smdkc210 -S -monitor stdio
  (qemu) info qom-tree
  /machine (smdkc210-machine)
    /unattached (container)
      /io[0] (qemu:memory-region)
      /exynos4210.dram0[0] (qemu:memory-region)
      /exynos4210.irom[0] (qemu:memory-region)
      /exynos4210.iram[0] (qemu:memory-region)
      /exynos4210.chipid[0] (qemu:memory-region)
      ...
      /device[26] (exynos4210.uart)
        /exynos4210.uart[0] (qemu:memory-region)
    /soc (exynos4210)
      ^
       \__ [*]

The irom/iram/chipid regions should go under 'soc' at [*].

Add a semantic patch to let the device own the memory region.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
5 years agoscripts/cocci: Patch to remove unnecessary memory_region_set_readonly()
Philippe Mathieu-Daudé [Mon, 24 Feb 2020 18:55:35 +0000 (19:55 +0100)]
scripts/cocci: Patch to remove unnecessary memory_region_set_readonly()

Add a semantic patch to remove memory_region_set_readonly() calls
on ROM memory regions.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
5 years agoscripts/cocci: Patch to detect potential use of memory_region_init_rom
Philippe Mathieu-Daudé [Mon, 24 Feb 2020 18:52:31 +0000 (19:52 +0100)]
scripts/cocci: Patch to detect potential use of memory_region_init_rom

Add a semantic patch to detect potential replacement of
memory_region_init_ram(readonly) by memory_region_init_rom().

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
5 years agohw/sparc: Use memory_region_init_rom() with read-only regions
Philippe Mathieu-Daudé [Mon, 24 Feb 2020 18:51:24 +0000 (19:51 +0100)]
hw/sparc: Use memory_region_init_rom() with read-only regions

This commit was produced with the Coccinelle script
scripts/coccinelle/memory-region-housekeeping.cocci.

Reviewed-by: KONRAD Frederic <frederic.konrad@adacore.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
5 years agohw/sh4: Use memory_region_init_rom() with read-only regions
Philippe Mathieu-Daudé [Mon, 24 Feb 2020 18:51:24 +0000 (19:51 +0100)]
hw/sh4: Use memory_region_init_rom() with read-only regions

This commit was produced with the Coccinelle script
scripts/coccinelle/memory-region-housekeeping.cocci.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
5 years agohw/riscv: Use memory_region_init_rom() with read-only regions
Philippe Mathieu-Daudé [Mon, 24 Feb 2020 18:51:24 +0000 (19:51 +0100)]
hw/riscv: Use memory_region_init_rom() with read-only regions

This commit was produced with the Coccinelle script
scripts/coccinelle/memory-region-housekeeping.cocci.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
5 years agohw/ppc: Use memory_region_init_rom() with read-only regions
Philippe Mathieu-Daudé [Mon, 24 Feb 2020 18:51:36 +0000 (19:51 +0100)]
hw/ppc: Use memory_region_init_rom() with read-only regions

This commit was produced with the Coccinelle script
scripts/coccinelle/memory-region-housekeeping.cocci.

Acked-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
5 years agohw/pci-host: Use memory_region_init_rom() with read-only regions
Philippe Mathieu-Daudé [Mon, 24 Feb 2020 18:51:23 +0000 (19:51 +0100)]
hw/pci-host: Use memory_region_init_rom() with read-only regions

This commit was produced with the Coccinelle script
scripts/coccinelle/memory-region-housekeeping.cocci.

Acked-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
5 years agohw/net: Use memory_region_init_rom() with read-only regions
Philippe Mathieu-Daudé [Mon, 24 Feb 2020 18:51:23 +0000 (19:51 +0100)]
hw/net: Use memory_region_init_rom() with read-only regions

This commit was produced with the Coccinelle script
scripts/coccinelle/memory-region-housekeeping.cocci.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
5 years agohw/m68k: Use memory_region_init_rom() with read-only regions
Philippe Mathieu-Daudé [Mon, 24 Feb 2020 18:51:23 +0000 (19:51 +0100)]
hw/m68k: Use memory_region_init_rom() with read-only regions

This commit was produced with the Coccinelle script
scripts/coccinelle/memory-region-housekeeping.cocci.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
5 years agohw/display: Use memory_region_init_rom() with read-only regions
Philippe Mathieu-Daudé [Mon, 24 Feb 2020 18:50:29 +0000 (19:50 +0100)]
hw/display: Use memory_region_init_rom() with read-only regions

This commit was produced with the Coccinelle script
scripts/coccinelle/memory-region-housekeeping.cocci.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
5 years agohw/arm: Use memory_region_init_rom() with read-only regions
Philippe Mathieu-Daudé [Mon, 24 Feb 2020 18:50:18 +0000 (19:50 +0100)]
hw/arm: Use memory_region_init_rom() with read-only regions

This commit was produced with the Coccinelle script
scripts/coccinelle/memory-region-housekeeping.cocci.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
5 years agoscripts/cocci: Patch to replace memory_region_init_{ram,readonly -> rom}
Philippe Mathieu-Daudé [Mon, 24 Feb 2020 18:23:40 +0000 (19:23 +0100)]
scripts/cocci: Patch to replace memory_region_init_{ram,readonly -> rom}

Add a semantic patch to replace memory_region_init_ram(readonly)
by memory_region_init_rom().

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
5 years agoscripts/cocci: Rename memory-region-{init-ram -> housekeeping}
Philippe Mathieu-Daudé [Mon, 24 Feb 2020 13:45:07 +0000 (14:45 +0100)]
scripts/cocci: Rename memory-region-{init-ram -> housekeeping}

As we are going to add various semantic changes related to the memory
region API, rename this script to be more generic.
Add a 'usage' header, and an entry in MAINTAINERS to avoid checkpatch
warning.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
5 years agomemory: Simplify memory_region_init_rom_nomigrate() to ease review
Philippe Mathieu-Daudé [Mon, 24 Feb 2020 09:58:17 +0000 (10:58 +0100)]
memory: Simplify memory_region_init_rom_nomigrate() to ease review

memory_region_init_rom_nomigrate() has the same content than
memory_region_init_ram_shared_nomigrate(), with setting the
readonly mode. The code is easier to review as creating a
readonly ram/shared/nomigrate region.

Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
5 years agomemory: Correctly return alias region type
Philippe Mathieu-Daudé [Mon, 24 Feb 2020 09:13:00 +0000 (10:13 +0100)]
memory: Correctly return alias region type

Since memory region aliases are neither rom nor ram, they are
described as i/o, which is often incorrect. Return instead the
type of the original region we are aliasing.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
5 years agolockable: add QemuRecMutex support
Stefan Hajnoczi [Mon, 16 Mar 2020 11:09:57 +0000 (11:09 +0000)]
lockable: add QemuRecMutex support

The polymorphic locking macros don't support QemuRecMutex yet.  Add it
so that lock guards can be used with QemuRecMutex.

Convert TCG plugins functions that benefit from these macros.  Manual
qemu_rec_mutex_lock/unlock() callers are left unmodified in cases where
clarity would not improve by switching to the macros.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
5 years agolockable: add lock guards
Stefan Hajnoczi [Mon, 16 Mar 2020 11:09:56 +0000 (11:09 +0000)]
lockable: add lock guards

This patch introduces two lock guard macros that automatically unlock a
lock object (QemuMutex and others):

  void f(void) {
      QEMU_LOCK_GUARD(&mutex);
      if (!may_fail()) {
          return; /* automatically unlocks mutex */
      }
      ...
  }

and:

  WITH_QEMU_LOCK_GUARD(&mutex) {
      if (!may_fail()) {
          return; /* automatically unlocks mutex */
      }
  }
  /* automatically unlocks mutex here */
  ...

Convert qemu-timer.c functions that benefit from these macros as an
example.  Manual qemu_mutex_lock/unlock() callers are left unmodified in
cases where clarity would not improve by switching to the macros.

Many other QemuMutex users remain in the codebase that might benefit
from lock guards.  Over time they can be converted, if that is
desirable.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
[Use QEMU_MAKE_LOCKABLE_NONNULL. - Paolo]
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
5 years agolockable: add QEMU_MAKE_LOCKABLE_NONNULL
Paolo Bonzini [Tue, 17 Mar 2020 14:17:20 +0000 (15:17 +0100)]
lockable: add QEMU_MAKE_LOCKABLE_NONNULL

This will be needed for lock guards, because if the lock is NULL the
dummy for loop of the lock guard never runs.  This can cause confusion
and dummy warnings in the compiler, but even if it did not, aborting
with a NULL pointer dereference is a less surprising behavior.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
5 years agoMerge remote-tracking branch 'remotes/kraxel/tags/usb-20200317-pull-request' into...
Peter Maydell [Tue, 17 Mar 2020 14:00:56 +0000 (14:00 +0000)]
Merge remote-tracking branch 'remotes/kraxel/tags/usb-20200317-pull-request' into staging

usb: bugfixes for usb-serial @ xhci.

# gpg: Signature made Tue 17 Mar 2020 09:50:04 GMT
# gpg:                using RSA key 4CB6D8EED3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <kraxel@redhat.com>" [full]
# gpg:                 aka "Gerd Hoffmann <gerd@kraxel.org>" [full]
# gpg:                 aka "Gerd Hoffmann (private) <kraxel@gmail.com>" [full]
# Primary key fingerprint: A032 8CFF B93A 17A7 9901  FE7D 4CB6 D8EE D3E8 7138

* remotes/kraxel/tags/usb-20200317-pull-request:
  usb-serial: Fix timeout closing the device
  usb-serial: Increase receive buffer to 496
  usb-serial: chunk data to wMaxPacketSize
  usb-serial: Move USB_TOKEN_IN into a helper function

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
5 years agoMAINTAINERS: Add a file to the main MIPS section
Aleksandar Markovic [Fri, 13 Mar 2020 01:00:52 +0000 (02:00 +0100)]
MAINTAINERS: Add a file to the main MIPS section

A recently added acceptance test is important not only for
Malta machine, but for overall MIPS target, since it tests
smp feature.

CC: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Aleksandar Markovic <amarkovic@wavecomp.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <1584061252-16635-4-git-send-email-aleksandar.markovic@rt-rk.com>

5 years agoMAINTAINERS: Adjust maintainer's email
Aleksandar Markovic [Fri, 13 Mar 2020 01:00:51 +0000 (02:00 +0100)]
MAINTAINERS: Adjust maintainer's email

For some longish time I've been using multiple email addresses
for mailing list communication, and would like to consolidate it
into a single email address that is the most convenient to me.
My other emails, from rt-rk.com and wavecomp.com domains remain
active and I will respond from them too, if needed, but I would
like to manage almost all communication using gmail.com account.

Signed-off-by: Aleksandar Markovic <amarkovic@wavecomp.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <1584061252-16635-3-git-send-email-aleksandar.markovic@rt-rk.com>

5 years agoMAINTAINERS: Adjust maintainer's status for some MIPS items
Aleksandar Markovic [Fri, 13 Mar 2020 01:00:50 +0000 (02:00 +0100)]
MAINTAINERS: Adjust maintainer's status for some MIPS items

Aurelien has been and will forever remain an idol in QEMU for
MIPS world. However, since he decided to move on to other projects,
acknowledge the reality, and formally releive him from maintainer's
duties for QEMU for MIPS items. Aurelien is though welcome to come
back at any time. Some empty spots caused by this are filled in by
Aleksandar.

CC: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Aleksandar Markovic <amarkovic@wavecomp.com>
Acked-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <1584061252-16635-2-git-send-email-aleksandar.markovic@rt-rk.com>

5 years agohw/arm/pxa2xx: Do not wire up OHCI for PXA255
Guenter Roeck [Fri, 13 Mar 2020 16:02:15 +0000 (09:02 -0700)]
hw/arm/pxa2xx: Do not wire up OHCI for PXA255

PXA255 does not support a USB OHCI controller, so don't wire it up.

Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Message-id: 20200313160215.28155-1-linux@roeck-us.net
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
5 years agoaspeed/smc: Fix number of dummy cycles for FAST_READ_4 command
Guenter Roeck [Mon, 16 Mar 2020 15:52:24 +0000 (15:52 +0000)]
aspeed/smc: Fix number of dummy cycles for FAST_READ_4 command

The Linux kernel recently started using FAST_READ_4 commands.
This results in flash read failures. At the same time, the m25p80
emulation is seen to read 8 more bytes than expected. Adjusting the
expected number of dummy cycles to match FAST_READ fixes the problem.

Fixes: f95c4bffdc4c ("aspeed/smc: snoop SPI transfers to fake dummy cycles")
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
5 years agom25p80: Improve command handling for unsupported commands
Guenter Roeck [Mon, 16 Mar 2020 15:52:24 +0000 (15:52 +0000)]
m25p80: Improve command handling for unsupported commands

Whenever an unsupported command is encountered, the current code
interprets each transferred byte as new command. Most of the time, those
'commands' are interpreted as new unknown commands. However, in rare
cases, it may be that for example address or length information
passed with the original command is by itself a valid command.
If that happens, the state machine may get completely confused and,
worst case, start writing data into the flash or even erase it.

To avoid the problem, transition into STATE_READING_DATA and keep
sending a value of 0 until the chip is deselected after encountering
an unsupported command.

Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
5 years agom25p80: Improve command handling for Jedec commands
Guenter Roeck [Mon, 16 Mar 2020 15:52:23 +0000 (15:52 +0000)]
m25p80: Improve command handling for Jedec commands

When requesting JEDEC data using the JEDEC_READ command, the Linux kernel
always requests 6 bytes. The current implementation only returns three
bytes, and interprets the remaining three bytes as new commands.
While this does not matter most of the time, it is at the very least
confusing. To avoid the problem, always report up to 6 bytes of JEDEC
data. Fill remaining data with 0.

Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
5 years agom25p80: Convert to support tracing
Guenter Roeck [Mon, 16 Mar 2020 15:52:23 +0000 (15:52 +0000)]
m25p80: Convert to support tracing

While at it, add some trace messages to help debug problems
seen when running the latest Linux kernel.

Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
5 years agohw/net/imx_fec: write TGSR and TCSR3 in imx_enet_write()
Chen Qun [Fri, 13 Mar 2020 12:32:42 +0000 (20:32 +0800)]
hw/net/imx_fec: write TGSR and TCSR3 in imx_enet_write()

The current code causes clang static code analyzer generate warning:
hw/net/imx_fec.c:858:9: warning: Value stored to 'value' is never read
        value = value & 0x0000000f;
        ^       ~~~~~~~~~~~~~~~~~~
hw/net/imx_fec.c:864:9: warning: Value stored to 'value' is never read
        value = value & 0x000000fd;
        ^       ~~~~~~~~~~~~~~~~~~

According to the definition of the function, the two “value” assignments
 should be written to registers.

Reported-by: Euler Robot <euler.robot@huawei.com>
Signed-off-by: Chen Qun <kuhn.chenqun@huawei.com>
Message-id: 20200313123242.13236-1-kuhn.chenqun@huawei.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
5 years agohw/arm/fsl-imx6: Wire up USB controllers
Guenter Roeck [Fri, 13 Mar 2020 01:45:51 +0000 (18:45 -0700)]
hw/arm/fsl-imx6: Wire up USB controllers

With this patch, the USB controllers on 'sabrelite' are detected
and can be used to boot the system.

Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Message-id: 20200313014551.12554-6-linux@roeck-us.net
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
5 years agohw/arm/fsl-imx6ul: Wire up USB controllers
Guenter Roeck [Fri, 13 Mar 2020 01:45:50 +0000 (18:45 -0700)]
hw/arm/fsl-imx6ul: Wire up USB controllers

IMX6UL USB controllers are quite similar to IMX7 USB controllers.
Wire them up the same way.

The only real difference is that wiring up phy devices is necessary
to avoid phy reset timeouts in the Linux kernel.

Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Message-id: 20200313014551.12554-5-linux@roeck-us.net
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
5 years agohw/arm/fsl-imx6ul: Instantiate unimplemented pwm and can devices
Guenter Roeck [Fri, 13 Mar 2020 01:45:49 +0000 (18:45 -0700)]
hw/arm/fsl-imx6ul: Instantiate unimplemented pwm and can devices

Recent Linux kernels (post v4.20) crash due to accesses to flexcan
and pwm controllers. Instantiate as unimplemented devices to work
around the problem.

Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Message-id: 20200313014551.12554-4-linux@roeck-us.net
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
5 years agohw/arm/fsl-imx6ul: Fix USB interrupt numbers
Guenter Roeck [Fri, 13 Mar 2020 01:45:48 +0000 (18:45 -0700)]
hw/arm/fsl-imx6ul: Fix USB interrupt numbers

USB1 and USB2 interrupt numbers were swapped. USB_PHY2 interrupt number
is 45. That didn't really matter up to now since the interrupts were not
used, but it needs to be fixed to be able to wire up the USB controllers.

Fixes: 31cbf933f0e ("i.MX6UL: Add i.MX6UL SOC")
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Message-id: 20200313014551.12554-3-linux@roeck-us.net
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
5 years agohw/usb: Add basic i.MX USB Phy support
Guenter Roeck [Fri, 13 Mar 2020 01:45:47 +0000 (18:45 -0700)]
hw/usb: Add basic i.MX USB Phy support

Add basic USB PHY support as implemented in i.MX23, i.MX28, i.MX6,
and i.MX7 SoCs.

The only support really needed - at least to boot Linux - is support
for soft reset, which needs to reset various registers to their initial
value. Otherwise, just record register values.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Message-id: 20200313014551.12554-2-linux@roeck-us.net
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
5 years agoMerge remote-tracking branch 'remotes/palmer/tags/riscv-for-master-5.0-sf5' into...
Peter Maydell [Tue, 17 Mar 2020 11:05:08 +0000 (11:05 +0000)]
Merge remote-tracking branch 'remotes/palmer/tags/riscv-for-master-5.0-sf5' into staging

RISC-V Patches for the 5.0 Soft Freeze, Part 5

This tag contains the last of the patches I'd like to target for the 5.0 soft
freeze.  At this point we're mostly collecting fixes, but there are a few new
features.  The changes include:

* An OpenSBI update, including the various bits necessary to put CI together
  and an image for the 32-bit sifive_u board.
* A fix that disallows TSR when outside of machine mode.
* A fix for VS-mode interrupt forwarding.

# gpg: Signature made Tue 17 Mar 2020 03:59:58 GMT
# gpg:                using RSA key 2B3C3747446843B24A943A7A2E1319F35FBB1889
# gpg:                issuer "palmer@dabbelt.com"
# gpg: Good signature from "Palmer Dabbelt <palmer@dabbelt.com>" [unknown]
# gpg:                 aka "Palmer Dabbelt <palmer@sifive.com>" [unknown]
# gpg:                 aka "Palmer Dabbelt <palmerdabbelt@google.com>" [unknown]
# gpg: WARNING: This key is not certified with a trusted signature!
# gpg:          There is no indication that the signature belongs to the owner.
# Primary key fingerprint: 00CE 76D1 8349 60DF CE88  6DF8 EF4C A150 2CCB AB41
#      Subkey fingerprint: 2B3C 3747 4468 43B2 4A94  3A7A 2E13 19F3 5FBB 1889

* remotes/palmer/tags/riscv-for-master-5.0-sf5:
  target/riscv: Fix VS mode interrupts forwarding.
  gitlab-ci.yml: Add jobs to build OpenSBI firmware binaries
  riscv: sifive_u: Update BIOS_FILENAME for 32-bit
  roms: opensbi: Add 32-bit firmware image for sifive_u machine
  roms: opensbi: Upgrade from v0.5 to v0.6
  target/riscv: Correctly implement TSR trap

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
5 years agousb-serial: Fix timeout closing the device
Jason Andryuk [Mon, 16 Mar 2020 17:46:10 +0000 (13:46 -0400)]
usb-serial: Fix timeout closing the device

Linux guests wait ~30 seconds when closing the emulated /dev/ttyUSB0.
During that time, the kernel driver is sending many control URBs
requesting GetModemStat (5).  Real hardware returns a status with
FTDI_THRE (Transmitter Holding Register) and FTDI_TEMT (Transmitter
Empty) set.  QEMU leaves them clear, and it seems Linux is waiting for
FTDI_TEMT to be set to indicate the tx queue is empty before closing.

Set the bits when responding to a GetModemStat query and avoid the
shutdown delay.

Signed-off-by: Jason Andryuk <jandryuk@gmail.com>
Reviewed-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Message-id: 20200316174610.115820-5-jandryuk@gmail.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
5 years agousb-serial: Increase receive buffer to 496
Jason Andryuk [Mon, 16 Mar 2020 17:46:09 +0000 (13:46 -0400)]
usb-serial: Increase receive buffer to 496

A FTDI USB adapter on an xHCI controller can send 512 byte USB packets.
These are 8 * ( 2 bytes header + 62 bytes data).  A 384 byte receive
buffer is insufficient to fill a 512 byte packet, so bump the receive
size to 496 ( 512 - 2 * 8 ).

Signed-off-by: Jason Andryuk <jandryuk@gmail.com>
Reviewed-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Message-id: 20200316174610.115820-4-jandryuk@gmail.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
5 years agousb-serial: chunk data to wMaxPacketSize
Jason Andryuk [Mon, 16 Mar 2020 17:46:08 +0000 (13:46 -0400)]
usb-serial: chunk data to wMaxPacketSize

usb-serial has issues with xHCI controllers where data is lost in the
VM.  Inspecting the URBs in the guest, EHCI starts every 64 byte boundary
(wMaxPacketSize) with a header.  EHCI hands packets into
usb_serial_token_in() with size 64, so these cannot cross the 64 byte
boundary.  The xHCI controller has packets of 512 bytes and the usb-serial
will just write through the 64 byte boundary.  In the guest, this means
data bytes are interpreted as header, so data bytes don't make it out
the serial interface.

Re-work usb_serial_token_in to chunk data into 64 byte units - 2 byte
header and 62 bytes data.  The Linux driver reads wMaxPacketSize to find
the chunk size, so we match that.

Real hardware was observed to pass in 512 byte URBs (496 bytes data +
8 * 2 byte headers).  Since usb-serial only buffers 384 bytes of data,
usb-serial will pass in 6 64 byte blocks and 1 12 byte partial block for
462 bytes max.

Signed-off-by: Jason Andryuk <jandryuk@gmail.com>
Message-id: 20200316174610.115820-3-jandryuk@gmail.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
5 years agousb-serial: Move USB_TOKEN_IN into a helper function
Jason Andryuk [Mon, 16 Mar 2020 17:46:07 +0000 (13:46 -0400)]
usb-serial: Move USB_TOKEN_IN into a helper function

We'll be adding a loop, so move the code into a helper function.  breaks
are replaced with returns.  While making this change, add braces to
single line if statements to comply with coding style and keep
checkpatch happy.

Signed-off-by: Jason Andryuk <jandryuk@gmail.com>
Message-id: 20200316174610.115820-2-jandryuk@gmail.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
5 years agopseries: Update SLOF firmware image
Alexey Kardashevskiy [Mon, 16 Mar 2020 23:27:15 +0000 (10:27 +1100)]
pseries: Update SLOF firmware image

The only change here is moving the decision about rtas-size
to QEMU.

Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
5 years agoppc/spapr: Ignore common "ibm,nmi-interlock" Linux bug
Nicholas Piggin [Mon, 16 Mar 2020 14:26:13 +0000 (00:26 +1000)]
ppc/spapr: Ignore common "ibm,nmi-interlock" Linux bug

Linux kernels call "ibm,nmi-interlock" in their system reset handlers
contrary to PAPR. Returning an error because the CPU does not hold the
interlock here causes Linux to print warning messages. PowerVM returns
success in this case, so do the same for now.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Message-Id: <20200316142613.121089-9-npiggin@gmail.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
5 years agoppc/spapr: Implement FWNMI System Reset delivery
Nicholas Piggin [Mon, 16 Mar 2020 14:26:12 +0000 (00:26 +1000)]
ppc/spapr: Implement FWNMI System Reset delivery

PAPR requires that if "ibm,nmi-register" succeeds, then the hypervisor
delivers all system reset and machine check exceptions to the registered
addresses.

System Resets are delivered with registers set to the architected state,
and with no interlock.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Message-Id: <20200316142613.121089-8-npiggin@gmail.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
5 years agotarget/ppc: allow ppc_cpu_do_system_reset to take an alternate vector
Nicholas Piggin [Mon, 16 Mar 2020 14:26:11 +0000 (00:26 +1000)]
target/ppc: allow ppc_cpu_do_system_reset to take an alternate vector

Provide for an alternate delivery location, -1 defaults to the
architected address.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Message-Id: <20200316142613.121089-7-npiggin@gmail.com>
Reviewed-by: Greg Kurz <groug@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
5 years agoppc/spapr: Allow FWNMI on TCG
Nicholas Piggin [Mon, 16 Mar 2020 14:26:10 +0000 (00:26 +1000)]
ppc/spapr: Allow FWNMI on TCG

There should no longer be a reason to prevent TCG providing FWNMI.
System Reset interrupts are generated to the guest with nmi monitor
command and H_SIGNAL_SYS_RESET. Machine Checks can not be injected
currently, but this could be implemented with the mce monitor cmd
similarly to i386.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Message-Id: <20200316142613.121089-6-npiggin@gmail.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Greg Kurz <groug@kaod.org>
[dwg: Re-enable FWNMI in qtests, since that now works]
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
5 years agoppc/spapr: Fix FWNMI machine check interrupt delivery
Nicholas Piggin [Mon, 16 Mar 2020 14:26:09 +0000 (00:26 +1000)]
ppc/spapr: Fix FWNMI machine check interrupt delivery

FWNMI machine check delivery misses a few things that will make it fail
with TCG at least (which we would like to allow in future to improve
testing).

It's not nice to scatter interrupt delivery logic around the tree, so
move it to excp_helper.c and share code where possible.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Message-Id: <20200316142613.121089-5-npiggin@gmail.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
5 years agoppc/spapr: Add FWNMI System Reset state
Nicholas Piggin [Mon, 16 Mar 2020 14:26:08 +0000 (00:26 +1000)]
ppc/spapr: Add FWNMI System Reset state

The FWNMI option must deliver system reset interrupts to their
registered address, and there are a few constraints on the handler
addresses specified in PAPR. Add the system reset address state and
checks.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Message-Id: <20200316142613.121089-4-npiggin@gmail.com>
Reviewed-by: Greg Kurz <groug@kaod.org>
Reviwed-by: Mahesh Salgaonkar <mahesh@linux.ibm.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>