openkylin
/
qemu
mirror of https://gitee.com/openkylin/qemu.git
9
0
Fork 0
Code Issues Projects Releases Wiki Activity
69,510 Commits 10 Branches 345 Tags 565 MiB
Commit Graph

69510 Commits

Author SHA1 Message Date
Peter Maydell 4bade28288 RISC-V Patches for the 4.1 Soft Freeze, Part 1 
This tag contains a handful of patches that I'd like to target for 4.1:
 
 * An emulation for SiFive's GPIO device.
 * A fix to disallow sfence.vma from userspace.
 * Additional decodetree cleanups that should have no functional impact.
 * C extension emulation fidelity fixes that were noticed as part of that
   cleanup process.
 * A new "spike" target, along with the deprecation of a handful of old
   targets and CPUs.
 * Some initial infastructure related to the hypervisor extension.
 * An emulation fidelity fix that prevents prevents arbitrary bits in the
   SIP CSR from being set.
 * A small performance improvement that avoids excessive TLB flushing
   when the ASID does not change.
 
 This time I've used a new testing workflow: I've tested on both 32-bit
 and 64-bit builds of OpenEmbedded, via the default OpenSBI-based boot
 flow.
 -----BEGIN PGP SIGNATURE-----
 
 iQJHBAABCAAxFiEEAM520YNJYN/OiG3470yhUCzLq0EFAlzohuUTHHBhbG1lckBk
 YWJiZWx0LmNvbQAKCRDvTKFQLMurQfcKD/9xSyQld3g41e4qrcq/MHHQhIou0Mpj
 FdJ9jb5k8ndlhJix1RCJorXQ4ld7KPco5ywdQGebEzrSlsI0hanlWkrtBDibR9dU
 OoNjrBRqKAjTQOkIsmgQMW961n9HnfSZNlmm/L1n94TC/6W4/g+hlvDXQNJjyDKX
 u8QalbfNFlolSv11vOxSlVFTo8Cb6/++yBitvbLdNnquWcDgxid6FYWKMEmYB4jK
 pOTzlb0ZDdvS7rGLdMkTB2pTqrvEykP1uLvXpF7qOurGgoeHul2lqKiRuVcFFU46
 6IOsaiwzDtMoUHzuFZypEJfnm0uJ4FEGEK8ZQBXlQhl2c0MDLXzXhIekKZIs/BmG
 K1Anio30zgfJscA6Zv6ZPataeEbPQ/l0LeAgS/Q8n7zIN5oRi3D9kdNTlo1MoSpD
 RLXESf0uXaTxkOAiyQ2hDj/NuPqja4TFD6UKj91eoWIhcga8i4olqsvLmsTsxhCK
 aaaiLGpYnmm/yhmBjMOAeBnu03w0ApTFS2UoPoMvTnrhZI4OXhm9XABEASpzdh57
 5mJL/U8eJ098QSxBui2Sjcpn8mxfy1FHqAQl+urDCAb6/P/94+729LVzyuvEV08U
 IHYMFjO8v4e56Sf6eys6rcPYpPV/jrLnyL8uFODN1PsXY9+OyH/KLCGVb1vyAgsP
 nhE2uPKTnIrPqA==
 =7LvI
 -----END PGP SIGNATURE-----

Merge remote-tracking branch 'remotes/palmer/tags/riscv-for-master-4.1-sf0' into staging

RISC-V Patches for the 4.1 Soft Freeze, Part 1

This tag contains a handful of patches that I'd like to target for 4.1:

* An emulation for SiFive's GPIO device.
* A fix to disallow sfence.vma from userspace.
* Additional decodetree cleanups that should have no functional impact.
* C extension emulation fidelity fixes that were noticed as part of that
  cleanup process.
* A new "spike" target, along with the deprecation of a handful of old
  targets and CPUs.
* Some initial infastructure related to the hypervisor extension.
* An emulation fidelity fix that prevents prevents arbitrary bits in the
  SIP CSR from being set.
* A small performance improvement that avoids excessive TLB flushing
  when the ASID does not change.

This time I've used a new testing workflow: I've tested on both 32-bit
and 64-bit builds of OpenEmbedded, via the default OpenSBI-based boot
flow.

# gpg: Signature made Sat 25 May 2019 01:05:57 BST
# gpg:                using RSA key 00CE76D1834960DFCE886DF8EF4CA1502CCBAB41
# gpg:                issuer "palmer@dabbelt.com"
# gpg: Good signature from "Palmer Dabbelt <palmer@dabbelt.com>" [unknown]
# gpg:                 aka "Palmer Dabbelt <palmer@sifive.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

* remotes/palmer/tags/riscv-for-master-4.1-sf0: (29 commits)
  target/riscv: Only flush TLB if SATP.ASID changes
  target/riscv: More accurate handling of `sip` CSR
  target/riscv: Add checks for several RVC reserved operands
  target/riscv: Add the HGATP register masks
  target/riscv: Add the HSTATUS register masks
  target/riscv: Add Hypervisor CSR macros
  target/riscv: Allow setting mstatus virtulisation bits
  target/riscv: Add the MPV and MTL mstatus bits
  target/riscv: Improve the scause logic
  target/riscv: Trigger interrupt on MIP update asynchronously
  target/riscv: Mark privilege level 2 as reserved
  riscv: spike: Add a generic spike machine
  target/riscv: Deprecate the generic no MMU CPUs
  target/riscv: Add a base 32 and 64 bit CPU
  target/riscv: Create settable CPU properties
  riscv: virt: Allow specifying a CPU via commandline
  linux-user/riscv: Add the CPU type as a comment
  target/riscv: Remove unused include of riscv_htif.h for virt board riscv
  target/riscv: Remove spaces from register names
  target/riscv: Split gen_arith_imm into functional and temp
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
 2019-05-28 11:52:53 +01:00
Peter Maydell 2b01c1b382 Machine Core queue, 2019-05-24 
* Display more helpful message when an object type is missing
   (Philippe Mathieu-Daudé)
 * Use object_initialize_child for correct reference counting
   (Philippe Mathieu-Daudé)
 -----BEGIN PGP SIGNATURE-----
 
 iQIcBAABCAAGBQJc6DhqAAoJECgHk2+YTcWmzhkP/A89XUshKY+FUGrSaMJy+ibQ
 RqTvkQU3ShzejIfmtZKgsq0wxeSAIbXY0OYlDMXusMxcb1JQN4lhQQn2GYty0pVq
 sTUPFWrFuFaxrFJBcjD93TEUUuvKqQJbuAqCCDyhFhGr784mtrZ4gplmXLu1MMLH
 4UtLx8Wg8f3UFzj6IjznFDjs4eXiHS8AwhTLfy2fS7cHT5Hh+JqMorPSkOLNmON1
 L0PTgHix14nX0VwHg/EHmLHwUWXoYUg0VShoFpRHZQTo8MjeHpk3xBm8UnRKN03P
 bKqGipemx2pawQpH4F55EsWFAWgxzM9P/6j4cET7/exTi3hgGh3iE9V3CSGpvHe9
 yf9rgAUadGDUteTrZi6t+BTzgxDBPLbuZaiIHArA+uFzrRJC8EAr6iojV+PYU3MI
 wlhF6GsagvcHgRHuVlxAY92Pc/5YV4s6ySMh1GB0yZ4LwMT0UwP+08Il+ZCul+jp
 pY92bN83o+fgVIGf7gbabDDURcHhsSa7DRWMb9W0b8fT1gEXtjkOP6wiL7HkSDVi
 quSdBaC3bDLBVrwiaLVPm0hNgk/ZjN7VeCIYdapLXmLFAauYvP6BW+Py+5E/VM1n
 WYCTAF9un0n5NiV3ltDkgzb2Nbrc2tKmRyIXE1UcsSAqlrvt2nAQP3GhmURbuMux
 PpZHe1N+2S5yaSk4rhTo
 =m2EC
 -----END PGP SIGNATURE-----

Merge remote-tracking branch 'remotes/ehabkost/tags/machine-next-pull-request' into staging

Machine Core queue, 2019-05-24

* Display more helpful message when an object type is missing
  (Philippe Mathieu-Daudé)
* Use object_initialize_child for correct reference counting
  (Philippe Mathieu-Daudé)

# gpg: Signature made Fri 24 May 2019 19:31:06 BST
# gpg:                using RSA key 2807936F984DC5A6
# gpg: Good signature from "Eduardo Habkost <ehabkost@redhat.com>" [full]
# Primary key fingerprint: 5A32 2FD5 ABC4 D3DB ACCF  D1AA 2807 936F 984D C5A6

* remotes/ehabkost/tags/machine-next-pull-request:
  hw/intc/nvic: Use object_initialize_child for correct reference counting
  hw/arm/mps2: Use object_initialize_child for correct reference counting
  hw/microblaze/zynqmp: Use object_initialize_child for correct ref. counting
  hw/microblaze/zynqmp: Use object_initialize_child for correct ref. counting
  hw/microblaze/zynqmp: Let the SoC manage the IPI devices
  hw/microblaze/zynqmp: Move the IPI state into the PMUSoC state
  hw/mips: Use object_initialize_child for correct reference counting
  hw/mips: Use object_initialize() on MIPSCPSState
  hw/arm: Use object_initialize_child for correct reference counting
  hw/arm/aspeed: Use object_initialize_child for correct ref. counting
  hw/arm/bcm2835: Use object_initialize_child for correct ref. counting
  hw/arm/bcm2835: Use object_initialize() on PL011State
  hw/arm/bcm2835: Use TYPE_PL011 instead of hardcoded string
  hw/virtio: Use object_initialize_child for correct reference counting
  hw/misc/macio: Use object_initialize_child for correct ref. counting
  hw/ppc/pnv: Use object_initialize_child for correct reference counting
  qom/object: Display more helpful message when an object type is missing

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
 2019-05-28 10:50:09 +01:00
Jonathan Behrens 1e0d985fa9
target/riscv: Only flush TLB if SATP.ASID changes 
There is an analogous change for ARM here:
https://patchwork.kernel.org/patch/10649857

Signed-off-by: Jonathan Behrens <jonathan@fintelia.io>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 12:09:25 -07:00
Jonathan Behrens 087b051a51
target/riscv: More accurate handling of `sip` CSR 
According to the spec, "All bits besides SSIP, USIP, and UEIP in the sip
register are read-only." Further, if an interrupt is not delegated to mode x,
then "the corresponding bits in xip [...] should appear to be hardwired to
zero. This patch implements both of those requirements.

Signed-off-by: Jonathan Behrens <jonathan@fintelia.io>
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 12:09:25 -07:00
Richard Henderson 4cc16b3b92
target/riscv: Add checks for several RVC reserved operands 
C.ADDI16SP, C.LWSP, C.JR, C.ADDIW, C.LDSP all have reserved
operands that were not diagnosed.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 12:09:25 -07:00
Alistair Francis e06431108b
target/riscv: Add the HGATP register masks 
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 12:09:25 -07:00
Alistair Francis d28b15a4d3
target/riscv: Add the HSTATUS register masks 
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviwed-by: Palmer Dabbelt <palmer@sifive.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 12:09:24 -07:00
Alistair Francis 71f09a5bb4
target/riscv: Add Hypervisor CSR macros 
Add the 1.10.1 Hypervisor CSRs and remove the 1.9.1 spec versions.

Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 12:09:24 -07:00
Alistair Francis 1f0419cb04
target/riscv: Allow setting mstatus virtulisation bits 
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Revieweb-by: Palmer Dabbelt <palmer@sifive.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 12:09:24 -07:00
Alistair Francis 49aaa3e534
target/riscv: Add the MPV and MTL mstatus bits 
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 12:09:24 -07:00
Alistair Francis 16fdb8ff64
target/riscv: Improve the scause logic 
No functional change, just making the code easier to read.

Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 12:09:24 -07:00
Alistair Francis 0a01f2eecb
target/riscv: Trigger interrupt on MIP update asynchronously 
The requirement of holding the iothread_mutex is burdersome when
swapping the background and foreground registers in the Hypervisor
extension. To avoid the requrirement let's set the interrupt
asynchronously.

Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 12:09:24 -07:00
Alistair Francis 356d74192a
target/riscv: Mark privilege level 2 as reserved 
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 12:09:24 -07:00
Alistair Francis cd69e3a60b
riscv: spike: Add a generic spike machine 
Add a generic spike machine (not tied to a version) and deprecate the
spike mahines that are tied to a specific version. As we can now specify
the CPU via the command line we no londer need specific versions of the
spike machines.

Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Acked-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 12:09:24 -07:00
Alistair Francis d64db71cf7
target/riscv: Deprecate the generic no MMU CPUs 
These can now be specified via the command line so we no longer need
these.

Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 12:09:23 -07:00
Alistair Francis 8903bf6e6d
target/riscv: Add a base 32 and 64 bit CPU 
At the same time deprecate the ISA string CPUs.

It is dobtful anyone specifies the CPUs, but we are keeping them for the
Spike machine (which is about to be depreated) so we may as well just
mark them as deprecated.

Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 12:09:23 -07:00
Alistair Francis c4e95030ba
target/riscv: Create settable CPU properties 
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 12:09:23 -07:00
Alistair Francis ceb2ffd56e
riscv: virt: Allow specifying a CPU via commandline 
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 12:09:23 -07:00
Alistair Francis 77c62400e5
linux-user/riscv: Add the CPU type as a comment 
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 12:09:23 -07:00
Jonathan Behrens e5ef9566af
target/riscv: Remove unused include of riscv_htif.h for virt board riscv 
Signed-off-by: Jonathan Behrens <fintelia@gmail.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 12:09:23 -07:00
Richard Henderson 7f9188e210
target/riscv: Remove spaces from register names 
These extra spaces make the "-d op" dump look weird.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 12:09:23 -07:00
Richard Henderson 598aa1160c
target/riscv: Split gen_arith_imm into functional and temp 
The tcg_gen_fooi_tl functions have some immediate constant
folding built in, which match up with some of the riscv asm
builtin macros, like mv and not.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 12:09:23 -07:00
Richard Henderson 0e68e240a9
target/riscv: Split RVC32 and RVC64 insns into separate files 
This eliminates all functions in insn_trans/trans_rvc.inc.c,
so the entire file can be removed.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 12:09:22 -07:00
Richard Henderson c2cfb97c01
target/riscv: Use pattern groups in insn16.decode 
This eliminates about half of the complicated decode
bits within insn_trans/trans_rvc.inc.c.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 12:09:22 -07:00
Richard Henderson 6cafec92f1
target/riscv: Merge argument decode for RVC shifti 
Special handling for IMM==0 is the only difference between
RVC shifti and RVI shifti.  This can be handled with !function.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 12:09:22 -07:00
Richard Henderson e1d455dd91
target/riscv: Merge argument sets for insn32 and insn16 
In some cases this allows us to directly use the insn32
translator function.  In some cases we still need a shim.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 12:09:22 -07:00
Richard Henderson 8177025558
target/riscv: Use --static-decode for decodetree 
The generated functions are only used within translate.c
and do not need to be global, or declared.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 12:09:22 -07:00
Richard Henderson e761799796
target/riscv: Name the argument sets for all of insn32 formats 
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 12:09:22 -07:00
Fabien Chouteau 6e2716d8ca
RISC-V: fix single stepping over ret and other branching instructions 
This patch introduces wrappers around the tcg_gen_exit_tb() and
tcg_gen_lookup_and_goto_ptr() functions that handle single stepping,
i.e. call gen_exception_debug() when single stepping is enabled.

Theses functions are then used instead of the originals, bringing single
stepping handling in places where it was previously ignored such as jalr
and system branch instructions (ecall, mret, sret, etc.).

Signed-off-by: Fabien Chouteau <chouteau@adacore.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 12:09:22 -07:00
Jonathan Behrens b86f416763
target/riscv: Do not allow sfence.vma from user mode 
The 'sfence.vma' instruction is privileged, and should only ever be allowed
when executing in supervisor mode or higher.

Signed-off-by: Jonathan Behrens <fintelia@gmail.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 12:09:19 -07:00
Fabien Chouteau 30efbf330a
SiFive RISC-V GPIO Device 
QEMU model of the GPIO device on the SiFive E300 series SOCs.

The pins are not used by a board definition yet, however this
implementation can already be used to trigger GPIO interrupts from the
software by configuring a pin as both output and input.

Signed-off-by: Fabien Chouteau <chouteau@adacore.com>
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
 2019-05-24 11:58:30 -07:00
Philippe Mathieu-Daudé 23d1f360f3 hw/intc/nvic: Use object_initialize_child for correct reference counting 
As explained in commit aff39be0ed97:

  Both functions, object_initialize() and object_property_add_child()
  increase the reference counter of the new object, so one of the
  references has to be dropped afterwards to get the reference
  counting right. Otherwise the child object will not be properly
  cleaned up when the parent gets destroyed.
  Thus let's use now object_initialize_child() instead to get the
  reference counting here right.

This patch was generated using the following Coccinelle script:

 @use_sysbus_init_child_obj_missing_parent@
 expression child_ptr;
 expression child_type;
 expression child_size;
 @@
 -   object_initialize(child_ptr, child_size, child_type);
     ...
 -   qdev_set_parent_bus(DEVICE(child_ptr), sysbus_get_default());
     ...
 ?-  object_unref(OBJECT(child_ptr));
 +   sysbus_init_child_obj(OBJECT(PARENT_OBJ), "CHILD_NAME", child_ptr,
 +                         child_size, child_type);

We let NVIC adopt the SysTick timer.

While the object_initialize() function doesn't take an
'Error *errp' argument, the object_initialize_child() does.
Since this code is used when a machine is created (and is not
yet running), we deliberately choose to use the &error_abort
argument instead of ignoring errors if an object creation failed.
This choice also matches when using sysbus_init_child_obj(),
since its code is:

  void sysbus_init_child_obj(Object *parent,
                             const char *childname, void *child,
                             size_t childsize, const char *childtype)
  {
      object_initialize_child(parent, childname, child, childsize,
                              childtype, &error_abort, NULL);

      qdev_set_parent_bus(DEVICE(child), sysbus_get_default());
  }

Suggested-by: Eduardo Habkost <ehabkost@redhat.com>
Inspired-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20190507163416.24647-17-philmd@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
 2019-05-24 15:29:02 -03:00
Philippe Mathieu-Daudé f9e803218a hw/arm/mps2: Use object_initialize_child for correct reference counting 
As explained in commit aff39be0ed97:

  Both functions, object_initialize() and object_property_add_child()
  increase the reference counter of the new object, so one of the
  references has to be dropped afterwards to get the reference
  counting right. Otherwise the child object will not be properly
  cleaned up when the parent gets destroyed.
  Thus let's use now object_initialize_child() instead to get the
  reference counting here right.

This patch was generated using the following Coccinelle script:

 @use_sysbus_init_child_obj_missing_parent@
 expression child_ptr;
 expression child_type;
 expression child_size;
 @@
 -   object_initialize(child_ptr, child_size, child_type);
     ...
 -   qdev_set_parent_bus(DEVICE(child_ptr), sysbus_get_default());
     ...
 ?-  object_unref(OBJECT(child_ptr));
 +   sysbus_init_child_obj(OBJECT(PARENT_OBJ), "CHILD_NAME", child_ptr,
 +                         child_size, child_type);

We let the MPS2 boards adopt the cpu core, the FPGA and the SCC children.

While the object_initialize() function doesn't take an
'Error *errp' argument, the object_initialize_child() does.
Since this code is used when a machine is created (and is not
yet running), we deliberately choose to use the &error_abort
argument instead of ignoring errors if an object creation failed.
This choice also matches when using sysbus_init_child_obj(),
since its code is:

  void sysbus_init_child_obj(Object *parent,
                             const char *childname, void *child,
                             size_t childsize, const char *childtype)
  {
      object_initialize_child(parent, childname, child, childsize,
                              childtype, &error_abort, NULL);

      qdev_set_parent_bus(DEVICE(child), sysbus_get_default());
  }

Suggested-by: Eduardo Habkost <ehabkost@redhat.com>
Inspired-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20190507163416.24647-16-philmd@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
 2019-05-24 15:29:02 -03:00
Philippe Mathieu-Daudé 47865c3760 hw/microblaze/zynqmp: Use object_initialize_child for correct ref. counting 
As explained in commit aff39be0ed97:

  Both functions, object_initialize() and object_property_add_child()
  increase the reference counter of the new object, so one of the
  references has to be dropped afterwards to get the reference
  counting right. Otherwise the child object will not be properly
  cleaned up when the parent gets destroyed.
  Thus let's use now object_initialize_child() instead to get the
  reference counting here right.

This patch was generated using the following Coccinelle script
(with a bit of manual fix-up for overly long lines):

 @use_object_initialize_child@
 expression parent_obj;
 expression child_ptr;
 expression child_name;
 expression child_type;
 expression child_size;
 expression errp;
 @@
 (
 -   object_initialize(child_ptr, child_size, child_type);
 +   object_initialize_child(parent_obj, child_name,  child_ptr, child_size,
 +                           child_type, &error_abort, NULL);
     ... when != parent_obj
 -   object_property_add_child(parent_obj, child_name, OBJECT(child_ptr), NULL);
     ...
?-   object_unref(OBJECT(child_ptr));
 |
 -   object_initialize(child_ptr, child_size, child_type);
 +   object_initialize_child(parent_obj, child_name,  child_ptr, child_size,
 +                            child_type, errp, NULL);
     ... when != parent_obj
 -   object_property_add_child(parent_obj, child_name, OBJECT(child_ptr), errp);
     ...
?-   object_unref(OBJECT(child_ptr));
 )

While the object_initialize() function doesn't take an
'Error *errp' argument, the object_initialize_child() does.
Since this code is used when a machine is created (and is not
yet running), we deliberately choose to use the &error_abort
argument instead of ignoring errors if an object creation failed.

Suggested-by: Eduardo Habkost <ehabkost@redhat.com>
Inspired-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20190507163416.24647-15-philmd@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
 2019-05-24 15:29:02 -03:00
Philippe Mathieu-Daudé ff5d4dc998 hw/microblaze/zynqmp: Use object_initialize_child for correct ref. counting 
As explained in commit aff39be0ed97:

  Both functions, object_initialize() and object_property_add_child()
  increase the reference counter of the new object, so one of the
  references has to be dropped afterwards to get the reference
  counting right. Otherwise the child object will not be properly
  cleaned up when the parent gets destroyed.
  Thus let's use now object_initialize_child() instead to get the
  reference counting here right.

This patch was generated using the following Coccinelle script
(then manually modified to use numbered IPI name)

 @use_sysbus_init_child_obj_missing_parent@
 expression child_ptr;
 expression child_type;
 expression child_size;
 @@
 -   object_initialize(child_ptr, child_size, child_type);
     ...
 -   qdev_set_parent_bus(DEVICE(child_ptr), sysbus_get_default());
     ...
 ?-  object_unref(OBJECT(child_ptr));
 +   sysbus_init_child_obj(OBJECT(PARENT_OBJ), "CHILD_NAME", child_ptr,
 +                         child_size, child_type);

We let the SoC adopt the IPI children.

While the object_initialize() function doesn't take an
'Error *errp' argument, the object_initialize_child() does.
Since this code is used when a machine is created (and is not
yet running), we deliberately choose to use the &error_abort
argument instead of ignoring errors if an object creation failed.
This choice also matches when using sysbus_init_child_obj(),
since its code is:

  void sysbus_init_child_obj(Object *parent,
                             const char *childname, void *child,
                             size_t childsize, const char *childtype)
  {
      object_initialize_child(parent, childname, child, childsize,
                              childtype, &error_abort, NULL);

      qdev_set_parent_bus(DEVICE(child), sysbus_get_default());
  }

Suggested-by: Eduardo Habkost <ehabkost@redhat.com>
Inspired-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20190507163416.24647-14-philmd@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
 2019-05-24 15:29:02 -03:00
Philippe Mathieu-Daudé da4aeff9b3 hw/microblaze/zynqmp: Let the SoC manage the IPI devices 
The Inter Processor Interrupt is a block part of the SoC, not the
"machine" (See Zynq UltraScale+ Device TRM UG1085, "Platform
Management Unit", Power Domains and Islands).

Move the IPI management from the machine to the SoC.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20190507163416.24647-13-philmd@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
 2019-05-24 15:29:02 -03:00
Philippe Mathieu-Daudé a8ae92e0ee hw/microblaze/zynqmp: Move the IPI state into the PMUSoC state 
The Inter Processor Interrupt is a block part of the SoC, not the
"machine" (talking about machine is borderline with the PMU, since
it is embedded into the ZynqMP SoC, but currentl QEMU doesn't
support multi-arch cores).

Move the IPI state to the SoC state, this will simplify the review
of the next patch.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20190507163416.24647-12-philmd@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
 2019-05-24 15:29:02 -03:00
Philippe Mathieu-Daudé 4626548b02 hw/mips: Use object_initialize_child for correct reference counting 
As explained in commit aff39be0ed97:

  Both functions, object_initialize() and object_property_add_child()
  increase the reference counter of the new object, so one of the
  references has to be dropped afterwards to get the reference
  counting right. Otherwise the child object will not be properly
  cleaned up when the parent gets destroyed.
  Thus let's use now object_initialize_child() instead to get the
  reference counting here right.

This patch was generated using the following Coccinelle script:

 @use_sysbus_init_child_obj_missing_parent@
 expression child_ptr;
 expression child_type;
 expression child_size;
 @@
 -   object_initialize(child_ptr, child_size, child_type);
     ...
 -   qdev_set_parent_bus(DEVICE(child_ptr), sysbus_get_default());
     ...
 ?-  object_unref(OBJECT(child_ptr));
 +   sysbus_init_child_obj(OBJECT(PARENT_OBJ), "CHILD_NAME", child_ptr,
 +                         child_size, child_type);

We let the Malta/Boston machines adopt the CPS child, and similarly
the CPS adopts the ITU/CPC/GIC/GCR children.

While the object_initialize() function doesn't take an
'Error *errp' argument, the object_initialize_child() does.
Since this code is used when a machine is created (and is not
yet running), we deliberately choose to use the &error_abort
argument instead of ignoring errors if an object creation failed.
This choice also matches when using sysbus_init_child_obj(),
since its code is:

  void sysbus_init_child_obj(Object *parent,
                             const char *childname, void *child,
                             size_t childsize, const char *childtype)
  {
      object_initialize_child(parent, childname, child, childsize,
                              childtype, &error_abort, NULL);

      qdev_set_parent_bus(DEVICE(child), sysbus_get_default());
  }

Suggested-by: Eduardo Habkost <ehabkost@redhat.com>
Inspired-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20190507163416.24647-11-philmd@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
 2019-05-24 15:29:02 -03:00
Philippe Mathieu-Daudé 2d5fac809c hw/mips: Use object_initialize() on MIPSCPSState 
Initialize the MIPSCPSState with object_initialize() instead of
object_new(). This will allow us to add it as children of the
machine container.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20190507163416.24647-10-philmd@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
 2019-05-24 15:29:02 -03:00
Philippe Mathieu-Daudé d031379803 hw/arm: Use object_initialize_child for correct reference counting 
As explained in commit aff39be0ed97:

  Both functions, object_initialize() and object_property_add_child()
  increase the reference counter of the new object, so one of the
  references has to be dropped afterwards to get the reference
  counting right. Otherwise the child object will not be properly
  cleaned up when the parent gets destroyed.
  Thus let's use now object_initialize_child() instead to get the
  reference counting here right.

This patch was generated using the following Coccinelle script
(with a bit of manual fix-up for overly long lines):

 @use_object_initialize_child@
 expression parent_obj;
 expression child_ptr;
 expression child_name;
 expression child_type;
 expression child_size;
 expression errp;
 @@
 (
 -   object_initialize(child_ptr, child_size, child_type);
 +   object_initialize_child(parent_obj, child_name,  child_ptr, child_size,
 +                           child_type, &error_abort, NULL);
     ... when != parent_obj
 -   object_property_add_child(parent_obj, child_name, OBJECT(child_ptr), NULL);
     ...
 ?-  object_unref(OBJECT(child_ptr));
 |
 -   object_initialize(child_ptr, child_size, child_type);
 +   object_initialize_child(parent_obj, child_name,  child_ptr, child_size,
 +                            child_type, errp, NULL);
     ... when != parent_obj
 -   object_property_add_child(parent_obj, child_name, OBJECT(child_ptr), errp);
     ...
 ?-  object_unref(OBJECT(child_ptr));
 )

 @use_sysbus_init_child_obj@
 expression parent_obj;
 expression dev;
 expression child_ptr;
 expression child_name;
 expression child_type;
 expression child_size;
 expression errp;
 @@
 (
 -   object_initialize_child(parent_obj, child_name, child_ptr, child_size,
 -                           child_type, errp, NULL);
 +   sysbus_init_child_obj(parent_obj, child_name, child_ptr, child_size,
 +                         child_type);
     ...
 -   qdev_set_parent_bus(DEVICE(child_ptr), sysbus_get_default());
 |
 -   object_initialize_child(parent_obj, child_name, child_ptr, child_size,
 -                           child_type, errp, NULL);
 +   sysbus_init_child_obj(parent_obj, child_name, child_ptr, child_size,
 +                         child_type);
 -   dev = DEVICE(child_ptr);
 -   qdev_set_parent_bus(dev, sysbus_get_default());
 )

While the object_initialize() function doesn't take an
'Error *errp' argument, the object_initialize_child() does.
Since this code is used when a machine is created (and is not
yet running), we deliberately choose to use the &error_abort
argument instead of ignoring errors if an object creation failed.
This choice also matches when using sysbus_init_child_obj(),
since its code is:

  void sysbus_init_child_obj(Object *parent,
                             const char *childname, void *child,
                             size_t childsize, const char *childtype)
  {
      object_initialize_child(parent, childname, child, childsize,
                              childtype, &error_abort, NULL);

      qdev_set_parent_bus(DEVICE(child), sysbus_get_default());
  }

Suggested-by: Eduardo Habkost <ehabkost@redhat.com>
Inspired-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20190507163416.24647-9-philmd@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
 2019-05-24 15:29:02 -03:00
Philippe Mathieu-Daudé 1b0ad56727 hw/arm/aspeed: Use object_initialize_child for correct ref. counting 
As explained in commit aff39be0ed97:

  Both functions, object_initialize() and object_property_add_child()
  increase the reference counter of the new object, so one of the
  references has to be dropped afterwards to get the reference
  counting right. Otherwise the child object will not be properly
  cleaned up when the parent gets destroyed.
  Thus let's use now object_initialize_child() instead to get the
  reference counting here right.

This patch was generated using the following Coccinelle script
(with a bit of manual fix-up for overly long lines):

 @use_object_initialize_child@
 expression parent_obj;
 expression child_ptr;
 expression child_name;
 expression child_type;
 expression child_size;
 expression errp;
 @@
 (
 -   object_initialize(child_ptr, child_size, child_type);
 +   object_initialize_child(parent_obj, child_name,  child_ptr, child_size,
 +                           child_type, &error_abort, NULL);
     ... when != parent_obj
 -   object_property_add_child(parent_obj, child_name, OBJECT(child_ptr), NULL);
     ...
 ?-  object_unref(OBJECT(child_ptr));
 |
 -   object_initialize(child_ptr, child_size, child_type);
 +   object_initialize_child(parent_obj, child_name,  child_ptr, child_size,
 +                            child_type, errp, NULL);
     ... when != parent_obj
 -   object_property_add_child(parent_obj, child_name, OBJECT(child_ptr), errp);
     ...
 ?-  object_unref(OBJECT(child_ptr));
 )

 @use_sysbus_init_child_obj@
 expression parent_obj;
 expression dev;
 expression child_ptr;
 expression child_name;
 expression child_type;
 expression child_size;
 expression errp;
 @@
 (
 -   object_initialize_child(parent_obj, child_name, child_ptr, child_size,
 -                           child_type, errp, NULL);
 +   sysbus_init_child_obj(parent_obj, child_name, child_ptr, child_size,
 +                         child_type);
     ...
 -   qdev_set_parent_bus(DEVICE(child_ptr), sysbus_get_default());
 |
 -   object_initialize_child(parent_obj, child_name, child_ptr, child_size,
 -                           child_type, errp, NULL);
 +   sysbus_init_child_obj(parent_obj, child_name, child_ptr, child_size,
 +                         child_type);
 -   dev = DEVICE(child_ptr);
 -   qdev_set_parent_bus(dev, sysbus_get_default());
 )

While the object_initialize() function doesn't take an
'Error *errp' argument, the object_initialize_child() does.
Since this code is used when a machine is created (and is not
yet running), we deliberately choose to use the &error_abort
argument instead of ignoring errors if an object creation failed.
This choice also matches when using sysbus_init_child_obj(),
since its code is:

  void sysbus_init_child_obj(Object *parent,
                             const char *childname, void *child,
                             size_t childsize, const char *childtype)
  {
      object_initialize_child(parent, childname, child, childsize,
                              childtype, &error_abort, NULL);

      qdev_set_parent_bus(DEVICE(child), sysbus_get_default());
  }

Suggested-by: Eduardo Habkost <ehabkost@redhat.com>
Inspired-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Joel Stanley <joel@jms.id.au>
Message-Id: <20190507163416.24647-8-philmd@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
 2019-05-24 15:29:02 -03:00
Philippe Mathieu-Daudé 661488b94b hw/arm/bcm2835: Use object_initialize_child for correct ref. counting 
As explained in commit aff39be0ed97:

  Both functions, object_initialize() and object_property_add_child()
  increase the reference counter of the new object, so one of the
  references has to be dropped afterwards to get the reference
  counting right. Otherwise the child object will not be properly
  cleaned up when the parent gets destroyed.
  Thus let's use now object_initialize_child() instead to get the
  reference counting here right.

This patch was generated using the following Coccinelle script
(with a bit of manual fix-up for overly long lines):

 @use_object_initialize_child@
 expression parent_obj;
 expression child_ptr;
 expression child_name;
 expression child_type;
 expression child_size;
 expression errp;
 @@
 (
 -   object_initialize(child_ptr, child_size, child_type);
 +   object_initialize_child(parent_obj, child_name,  child_ptr, child_size,
 +                           child_type, &error_abort, NULL);
     ... when != parent_obj
 -   object_property_add_child(parent_obj, child_name, OBJECT(child_ptr), NULL);
     ...
 ?-  object_unref(OBJECT(child_ptr));
 |
 -   object_initialize(child_ptr, child_size, child_type);
 +   object_initialize_child(parent_obj, child_name,  child_ptr, child_size,
 +                            child_type, errp, NULL);
     ... when != parent_obj
 -   object_property_add_child(parent_obj, child_name, OBJECT(child_ptr), errp);
     ...
 ?-  object_unref(OBJECT(child_ptr));
 )

 @use_sysbus_init_child_obj@
 expression parent_obj;
 expression dev;
 expression child_ptr;
 expression child_name;
 expression child_type;
 expression child_size;
 expression errp;
 @@
 (
 -   object_initialize_child(parent_obj, child_name, child_ptr, child_size,
 -                           child_type, errp, NULL);
 +   sysbus_init_child_obj(parent_obj, child_name, child_ptr, child_size,
 +                         child_type);
     ...
 -   qdev_set_parent_bus(DEVICE(child_ptr), sysbus_get_default());
 |
 -   object_initialize_child(parent_obj, child_name, child_ptr, child_size,
 -                           child_type, errp, NULL);
 +   sysbus_init_child_obj(parent_obj, child_name, child_ptr, child_size,
 +                         child_type);
 -   dev = DEVICE(child_ptr);
 -   qdev_set_parent_bus(dev, sysbus_get_default());
 )

While the object_initialize() function doesn't take an
'Error *errp' argument, the object_initialize_child() does.
Since this code is used when a machine is created (and is not
yet running), we deliberately choose to use the &error_abort
argument instead of ignoring errors if an object creation failed.
This choice also matches when using sysbus_init_child_obj(),
since its code is:

  void sysbus_init_child_obj(Object *parent,
                             const char *childname, void *child,
                             size_t childsize, const char *childtype)
  {
      object_initialize_child(parent, childname, child, childsize,
                              childtype, &error_abort, NULL);

      qdev_set_parent_bus(DEVICE(child), sysbus_get_default());
  }

Suggested-by: Eduardo Habkost <ehabkost@redhat.com>
Inspired-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20190507163416.24647-7-philmd@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
 2019-05-24 15:29:02 -03:00
Philippe Mathieu-Daudé 948770b0a7 hw/arm/bcm2835: Use object_initialize() on PL011State 
To be coherent with the other peripherals contained in the
BCM2835PeripheralState structure, directly allocate the PL011State
(instead of using the pl011 uart as a pointer to a SysBusDevice).

Initialize the PL011State with object_initialize() instead of
object_new().

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20190507163416.24647-6-philmd@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
 2019-05-24 15:29:02 -03:00
Philippe Mathieu-Daudé 0a21950e43 hw/arm/bcm2835: Use TYPE_PL011 instead of hardcoded string 
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20190507163416.24647-5-philmd@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
 2019-05-24 15:29:02 -03:00
Philippe Mathieu-Daudé 3d2fc923ec hw/virtio: Use object_initialize_child for correct reference counting 
As explained in commit aff39be0ed97:

  Both functions, object_initialize() and object_property_add_child()
  increase the reference counter of the new object, so one of the
  references has to be dropped afterwards to get the reference
  counting right. Otherwise the child object will not be properly
  cleaned up when the parent gets destroyed.
  Thus let's use now object_initialize_child() instead to get the
  reference counting here right.

This patch was generated using the following Coccinelle script:

 @use_object_initialize_child@
 expression parent_obj;
 expression child_ptr;
 expression child_name;
 expression child_type;
 expression child_size;
 expression errp;
 @@
 (
 -   object_initialize(child_ptr, child_size, child_type);
 +   object_initialize_child(parent_obj, child_name,  child_ptr, child_size,
 +                           child_type, &error_abort, NULL);
     ... when != parent_obj
 -   object_property_add_child(parent_obj, child_name, OBJECT(child_ptr), NULL);
     ...
 ?-  object_unref(OBJECT(child_ptr));
 |
 -   object_initialize(child_ptr, child_size, child_type);
 +   object_initialize_child(parent_obj, child_name,  child_ptr, child_size,
 +                            child_type, errp, NULL);
     ... when != parent_obj
 -   object_property_add_child(parent_obj, child_name, OBJECT(child_ptr), errp);
     ...
 ?-  object_unref(OBJECT(child_ptr));
 )

While the object_initialize() function doesn't take an
'Error *errp' argument, the object_initialize_child() does.
Since this code is used when a machine is created (and is not
yet running), we deliberately choose to use the &error_abort
argument instead of ignoring errors if an object creation failed.

Suggested-by: Eduardo Habkost <ehabkost@redhat.com>
Inspired-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20190507163416.24647-4-philmd@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
 2019-05-24 15:29:02 -03:00
Philippe Mathieu-Daudé 954d97672f hw/misc/macio: Use object_initialize_child for correct ref. counting 
As explained in commit aff39be0ed97:

  Both functions, object_initialize() and object_property_add_child()
  increase the reference counter of the new object, so one of the
  references has to be dropped afterwards to get the reference
  counting right. Otherwise the child object will not be properly
  cleaned up when the parent gets destroyed.
  Thus let's use now object_initialize_child() instead to get the
  reference counting here right.

This patch was generated using the following Coccinelle script
(with a bit of manual fix-up for overly long lines):

 @use_object_initialize_child@
 expression parent_obj;
 expression child_ptr;
 expression child_name;
 expression child_type;
 expression child_size;
 expression errp;
 @@
 (
 -   object_initialize(child_ptr, child_size, child_type);
 +   object_initialize_child(parent_obj, child_name,  child_ptr, child_size,
 +                           child_type, &error_abort, NULL);
     ... when != parent_obj
 -   object_property_add_child(parent_obj, child_name, OBJECT(child_ptr), NULL);
     ...
 ?-  object_unref(OBJECT(child_ptr));
 |
 -   object_initialize(child_ptr, child_size, child_type);
 +   object_initialize_child(parent_obj, child_name,  child_ptr, child_size,
 +                            child_type, errp, NULL);
     ... when != parent_obj
 -   object_property_add_child(parent_obj, child_name, OBJECT(child_ptr), errp);
     ...
 ?-  object_unref(OBJECT(child_ptr));
 )

While the object_initialize() function doesn't take an
'Error *errp' argument, the object_initialize_child() does.
Since this code is used when a machine is created (and is not
yet running), we deliberately choose to use the &error_abort
argument instead of ignoring errors if an object creation failed.

Suggested-by: Eduardo Habkost <ehabkost@redhat.com>
Inspired-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20190507163416.24647-3-philmd@redhat.com>
Acked-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
 2019-05-24 15:29:02 -03:00
Philippe Mathieu-Daudé bc4c406c3e hw/ppc/pnv: Use object_initialize_child for correct reference counting 
As explained in commit aff39be0ed97:

  Both functions, object_initialize() and object_property_add_child()
  increase the reference counter of the new object, so one of the
  references has to be dropped afterwards to get the reference
  counting right. Otherwise the child object will not be properly
  cleaned up when the parent gets destroyed.
  Thus let's use now object_initialize_child() instead to get the
  reference counting here right.

This patch was generated using the following Coccinelle script
(with a bit of manual fix-up for overly long lines):

 @use_object_initialize_child@
 expression parent_obj;
 expression child_ptr;
 expression child_name;
 expression child_type;
 expression child_size;
 expression errp;
 @@
 (
 -   object_initialize(child_ptr, child_size, child_type);
 +   object_initialize_child(parent_obj, child_name,  child_ptr, child_size,
 +                           child_type, &error_abort, NULL);
     ... when != parent_obj
 -   object_property_add_child(parent_obj, child_name, OBJECT(child_ptr), NULL);
     ...
?-   object_unref(OBJECT(child_ptr));
 |
 -   object_initialize(child_ptr, child_size, child_type);
 +   object_initialize_child(parent_obj, child_name,  child_ptr, child_size,
 +                            child_type, errp, NULL);
     ... when != parent_obj
 -   object_property_add_child(parent_obj, child_name, OBJECT(child_ptr), errp);
     ...
?-   object_unref(OBJECT(child_ptr));
 )

While the object_initialize() function doesn't take an
'Error *errp' argument, the object_initialize_child() does.
Since this code is used when a machine is created (and is not
yet running), we deliberately choose to use the &error_abort
argument instead of ignoring errors if an object creation failed.

Suggested-by: Eduardo Habkost <ehabkost@redhat.com>
Inspired-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20190507163416.24647-2-philmd@redhat.com>
Acked-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
 2019-05-24 15:29:02 -03:00
Philippe Mathieu-Daudé e02bdf1cec qom/object: Display more helpful message when an object type is missing 
When writing a new board, adding device which uses other devices
(container) or simply refactoring, one can discover the hard way
his machine misses some devices. In the case of containers, the
error is not obvious:

  $ qemu-system-microblaze -M xlnx-zynqmp-pmu
  **
  ERROR:/source/qemu/qom/object.c:454:object_initialize_with_type: assertion failed: (type != NULL)
  Aborted (core dumped)

And we have to look at the coredump to figure the error:

  (gdb) bt
  #1  0x00007f84773cf895 in abort () at /lib64/libc.so.6
  #2  0x00007f847961fb53 in  () at /lib64/libglib-2.0.so.0
  #3  0x00007f847967a4de in g_assertion_message_expr () at /lib64/libglib-2.0.so.0
  #4  0x000055c4bcac6c11 in object_initialize_with_type (data=data@entry=0x55c4bdf239e0, size=size@entry=2464, type=<optimized out>) at /source/qemu/qom/object.c:454
  #5  0x000055c4bcac6e6d in object_initialize (data=data@entry=0x55c4bdf239e0, size=size@entry=2464, typename=typename@entry=0x55c4bcc7c643 "xlnx.zynqmp_ipi") at /source/qemu/qom/object.c:474
  #6  0x000055c4bc9ea474 in xlnx_zynqmp_pmu_init (machine=0x55c4bdd46000) at /source/qemu/hw/microblaze/xlnx-zynqmp-pmu.c:176
  #7  0x000055c4bca3b6cb in machine_run_board_init (machine=0x55c4bdd46000) at /source/qemu/hw/core/machine.c:1030
  #8  0x000055c4bc95f6d2 in main (argc=<optimized out>, argv=<optimized out>, envp=<optimized out>) at /source/qemu/vl.c:4479

Since the caller knows the type name requested, we can simply display it
to ease development.

With this patch applied we get:

  $ qemu-system-microblaze -M xlnx-zynqmp-pmu
  qemu-system-microblaze: missing object type 'xlnx.zynqmp_ipi'
  Aborted (core dumped)

Since the assert(type) check in object_initialize_with_type() is
now impossible, remove it.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20190427135642.16464-1-philmd@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
 2019-05-24 15:29:02 -03:00
Peter Maydell a7b21f6762 Add /proc/hardware and /proc/cpuinfo, 
update SIOCXXX ioctls,
 fix shmat emulation,
 add nanoseconds in stat,
 init field fp_abi on mips
 -----BEGIN PGP SIGNATURE-----
 
 iQIcBAABAgAGBQJc59R0AAoJEPMMOL0/L748px4QAJBgq468iQPge+6L2uwj4bX1
 xfclzs/HwNSZaxMj98HGHw+mMdMCZhcRMjZKBcF3bqMQ8vJEHEcHpZBVqGTxoQG9
 yDmMZzOSMPGOMjfYasKYaOLNXFexFigVI1tp1EHjuS/PEpaR/O3heFtp4ZqwBA/Y
 xIXbMVVod0my0uIBbs947CjPHFdFR7UrOvqQIUv0ptCukUsA1NNco4LxvYSur3Co
 YJgxXct9HIwueubwZKYVCxUI4SfDKuLD2aqkJGolhBSxiHMktzwR3kLdcosU0tnr
 b9ekBT6UdSVJZirLTHQOOAnU4Ph+bDrLZP3K+WaUdMndLIxWfgV6yaP0Fc5Sdmnv
 /NDctRthodYqqYXEUtSZrOAkQzVJsUoh6WHg8DKkNS22Q6Hzp6UkiAiKuKvhI/Hs
 V+5T3Wlkek8M6Zm3Hic7dCz9bATuaMrAb4j/2MI1VKbE1a2OCWruk+v7rAeMhi8d
 DOS94LJMmLHPUYviXoPyveEA60qm8jRIQAN2Yxuw3CLklDNmxY2UHY5bsfhKfkp5
 xTCyJalqUkYIMoj5j6Rtxjfo4SCg4lGyiu5/wKgD5Ry3YLr5NkTLsasutsqbvOt+
 D2JIJWny2JvCXMCkJLqffmlm96xglaPfsonrGlhufonuqw2nYyebddQLFo4rIz5b
 AYxBlo1cb5B77zpGFXC2
 =IEm6
 -----END PGP SIGNATURE-----

Merge remote-tracking branch 'remotes/vivier2/tags/linux-user-for-4.1-pull-request' into staging

Add /proc/hardware and /proc/cpuinfo,
update SIOCXXX ioctls,
fix shmat emulation,
add nanoseconds in stat,
init field fp_abi on mips

# gpg: Signature made Fri 24 May 2019 12:24:36 BST
# gpg:                using RSA key F30C38BD3F2FBE3C
# gpg: Good signature from "Laurent Vivier <lvivier@redhat.com>" [full]
# gpg:                 aka "Laurent Vivier <laurent@vivier.eu>" [full]
# gpg:                 aka "Laurent Vivier (Red Hat) <lvivier@redhat.com>" [full]
# Primary key fingerprint: CD2F 75DD C8E3 A4DC 2E4F  5173 F30C 38BD 3F2F BE3C

* remotes/vivier2/tags/linux-user-for-4.1-pull-request:
  linux-user: Pass through nanosecond timestamp components for stat syscalls
  linux-user: Align mmap_find_vma to host page size
  linux-user: Fix shmat emulation by honoring host SHMLBA
  linux-user: Sanitize interp_info and, for mips only, init field fp_abi
  linux-user: Add support for SIOC<G|S>IFPFLAGS ioctls for all targets
  linux-user: Add support for SIOCSPGRP ioctl for all targets
  linux-user: Fix support for SIOCATMARK and SIOCGPGRP ioctls for xtensa
  linux-user: add pseudo /proc/hardware for m68k
  linux-user: add pseudo /proc/cpuinfo for sparc

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
 2019-05-24 12:47:49 +01:00
Chen-Yu Tsai 5f992db605 linux-user: Pass through nanosecond timestamp components for stat syscalls 
Since Linux 2.6 the stat syscalls have mostly supported nanosecond
components for each of the file-related timestamps.

QEMU user mode emulation currently does not pass through the nanosecond
portion of the timestamp, even when the host system fills in the value.
This results in a mismatch when run on subsecond resolution filesystems
such as ext4 or XFS.

An example of this leading to inconsistency is cross-debootstraping a
full desktop root filesystem of Debian Buster. Recent versions of
fontconfig store the full timestamp (instead of just the second portion)
of the directory in its per-directory cache file, and checks this against
the directory to see if the cache is up-to-date. With QEMU user mode
emulation, the timestamp stored is incorrect, and upon booting the rootfs
natively, fontconfig discovers the mismatch, and proceeds to rebuild the
cache on the comparatively slow machine (low-power ARM vs x86). This
stalls the first attempt to open whatever application that incorporates
fontconfig.

This patch renames the "unused" padding trailing each timestamp element
to its nanosecond counterpart name if such an element exists in the
kernel sources for the given platform. Not all do. Then have the syscall
wrapper fill in the nanosecond portion if the host supports it, as
specified by the _POSIX_C_SOURCE and _XOPEN_SOURCE feature macros.

Recent versions of glibc only use stat64 and newfstatat syscalls on
32-bit and 64-bit platforms respectively. The changes in this patch
were tested by directly calling the stat, stat64 and newfstatat syscalls
directly, in addition to the glibc wrapper, on arm and aarch64 little
endian targets.

Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Chen-Yu Tsai <wens@csie.org>
Message-Id: <20190522162147.26303-1-wens@kernel.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
 2019-05-24 13:16:21 +02:00