All of lore.kernel.org
 help / color / mirror / Atom feed
From: Aleksandar Markovic <aleksandar.qemu.devel@gmail.com>
To: Huacai Chen <chenhuacai@gmail.com>
Cc: "Huacai Chen" <zltjiangshi@gmail.com>,
	"Thomas Huth" <thuth@redhat.com>,
	"Philippe Mathieu-Daudé" <f4bug@amsat.org>,
	"QEMU Developers" <qemu-devel@nongnu.org>,
	"Aleksandar Markovic" <aleksandar.m.mail@gmail.com>,
	"Aleksandar Rikalo" <aleksandar.rikalo@rt-rk.com>,
	"Aurelien Jarno" <aurelien@aurel32.net>
Subject: Re: [PATCH for-5.1 V4 1/4] hw/mips: Implement the kvm_type() hook in MachineClass
Date: Tue, 16 Jun 2020 23:17:31 +0200	[thread overview]
Message-ID: <CAHiYmc4kqAFE=DsR74W8Y6HPPXUCqopTt2D-OCxuN9jYQY8m_A@mail.gmail.com> (raw)
In-Reply-To: <CAAhV-H4oTR=LATQmZ0xNmnu+_O4XXa84SGzHSZ83HWq9VhKKDA@mail.gmail.com>

[-- Attachment #1: Type: text/plain, Size: 8415 bytes --]

уторак, 16. јун 2020., Huacai Chen <chenhuacai@gmail.com> је написао/ла:

> Hi, Thomas and Aleksandar,
>
> On Tue, Jun 16, 2020 at 3:45 AM Aleksandar Markovic
> <aleksandar.m.mail@gmail.com> wrote:
> >
> > On Mon, Jun 15, 2020 at 10:55 AM Thomas Huth <thuth@redhat.com> wrote:
> > >
> > > On 15/06/2020 02.52, Huacai Chen wrote:
> > > > Hi, Aleksandar,
> > > >
> > > > On Sun, Jun 14, 2020 at 4:07 PM Aleksandar Markovic
> > > > <aleksandar.qemu.devel@gmail.com> wrote:
> > > >>
> > > >>
> > > >>
> > > >> уто, 2. јун 2020. у 04:38 Huacai Chen <zltjiangshi@gmail.com> је
> написао/ла:
> > > >>>
> > > >>> MIPS has two types of KVM: TE & VZ, and TE is the default type.
> Now we
> > > >>> can't create a VZ guest in QEMU because it lacks the kvm_type()
> hook in
> > > >>> MachineClass. Besides, libvirt uses a null-machine to detect the
> kvm
> > > >>> capability, so by default it will return "KVM not supported" on a
> VZ
> > > >>> platform. Thus, null-machine also need the kvm_type() hook.
> > > >>>
> > > >>> Reviewed-by: Aleksandar Markovic <aleksandar.qemu.devel@gmail.com>
> > > >>> Signed-off-by: Huacai Chen <chenhc@lemote.com>
> > > >>> Co-developed-by: Jiaxun Yang <jiaxun.yang@flygoat.com>
> > > >>> ---
> > > >>
> > > >> Huacai,
> > > >>
> > > >> Please take a look at Peter's remarks at:
> > > >>
> > > >> https://lists.gnu.org/archive/html/qemu-devel/2020-06/msg01878.html
> > > >>
> > > >> ...and refactor this patch for v5. My general advice: The simpler,
> the batter.
> > > >>
> > > > OK, I will rework this patch.
> > >
> > >  Hi,
> > >
> > > is there maybe also a way to do this without moving null-machine.o from
> > > common-obj-y to obj-y, and to avoid the target-specific hacks in this
> > > file ? We just moved the null-machine from obj-y to common-obj-y two
> > > years ago (see commit 3858ff763985fb9e), since it's more desirable to
> > > have as much code in common-obj to save compilation time and maybe to
> be
> > > able to link a qemu with more than one target CPU in one binary one
> day...
> > >
> > > ppc64 has also more than one kvm_type (kvm-hv and kvm-pr), and
> > > apparently it also works without hacks to the null-machine code there
> > > ... so maybe you can peek into the ppc64 code to see how it is solved
> there?
> > >
> >
> > Hi, Huacai,
> >
> > I think the optimal outcome for this release of QEMU would be if you
> > drop support for VZ. I think your scenario could work without VZ,
> > couldn't it?
> >
> > This is a fairly complex thing, and, as you see, it is a little
> > intrusive, it could negatively impact other targets. With enough
> > development time, you can easily provide that support in 5.2. - but
> > now we are close to 5.1 softfreeze.
> >
> > Even without VZ support, I would consider your contribution the most
> > significant for MIPS target in last two years, at least - a giant step
> > ahead!
> There are two problems: 1, qemu need kvm_type() to create a normal KVM
> guest. 2, libvirt use the null-machine to detect the KVM capability.
> For the first problem, just provide a kvm_type() hook for MIPS is
> enough, which is similar to ppc64. For the second problem, I think
> libvirt is also unable to detect the KVM capabilty on ppc64, unless
> ppc64 do the same hack on null-machine.
>
> V5 of this patch will only provide kvm_type() for MIPS (not hack
> null-machine.c), but all of us should think how to solve the second
> problem.
>
>
I think this is a good approach, Huacai. Looking forward to v5.

May health be with you,

Aleksandar




> Huacai
> >
> > Best Regards,
> > Aleksandar
> >
> > >  Thomas
> > >
> > >
> > > >>>  hw/core/Makefile.objs  |  2 +-
> > > >>>  hw/core/null-machine.c |  4 ++++
> > > >>>  hw/mips/Makefile.objs  |  2 +-
> > > >>>  hw/mips/common.c       | 42 ++++++++++++++++++++++++++++++
> ++++++++++++
> > > >>>  include/hw/mips/mips.h |  3 +++
> > > >>>  5 files changed, 51 insertions(+), 2 deletions(-)
> > > >>>  create mode 100644 hw/mips/common.c
> > > >>>
> > > >>> diff --git a/hw/core/Makefile.objs b/hw/core/Makefile.objs
> > > >>> index 1d540ed..b5672f4 100644
> > > >>> --- a/hw/core/Makefile.objs
> > > >>> +++ b/hw/core/Makefile.objs
> > > >>> @@ -17,11 +17,11 @@ common-obj-$(CONFIG_SOFTMMU) +=
> vm-change-state-handler.o
> > > >>>  common-obj-$(CONFIG_SOFTMMU) += qdev-properties-system.o
> > > >>>  common-obj-$(CONFIG_SOFTMMU) += sysbus.o
> > > >>>  common-obj-$(CONFIG_SOFTMMU) += machine.o
> > > >>> -common-obj-$(CONFIG_SOFTMMU) += null-machine.o
> > > >>>  common-obj-$(CONFIG_SOFTMMU) += loader.o
> > > >>>  common-obj-$(CONFIG_SOFTMMU) += machine-hmp-cmds.o
> > > >>>  common-obj-$(CONFIG_SOFTMMU) += numa.o
> > > >>>  common-obj-$(CONFIG_SOFTMMU) += clock-vmstate.o
> > > >>> +obj-$(CONFIG_SOFTMMU) += null-machine.o
> > > >>>  obj-$(CONFIG_SOFTMMU) += machine-qmp-cmds.o
> > > >>>
> > > >>>  common-obj-$(CONFIG_EMPTY_SLOT) += empty_slot.o
> > > >>> diff --git a/hw/core/null-machine.c b/hw/core/null-machine.c
> > > >>> index cb47d9d..94a36f9 100644
> > > >>> --- a/hw/core/null-machine.c
> > > >>> +++ b/hw/core/null-machine.c
> > > >>> @@ -17,6 +17,7 @@
> > > >>>  #include "sysemu/sysemu.h"
> > > >>>  #include "exec/address-spaces.h"
> > > >>>  #include "hw/core/cpu.h"
> > > >>> +#include "hw/mips/mips.h"
> > > >>>
> > > >>>  static void machine_none_init(MachineState *mch)
> > > >>>  {
> > > >>> @@ -50,6 +51,9 @@ static void machine_none_machine_init(MachineClass
> *mc)
> > > >>>      mc->max_cpus = 1;
> > > >>>      mc->default_ram_size = 0;
> > > >>>      mc->default_ram_id = "ram";
> > > >>> +#ifdef TARGET_MIPS
> > > >>> +    mc->kvm_type = mips_kvm_type;
> > > >>> +#endif
> > > >>>  }
> > > >>>
> > > >>>  DEFINE_MACHINE("none", machine_none_machine_init)
> > > >>> diff --git a/hw/mips/Makefile.objs b/hw/mips/Makefile.objs
> > > >>> index 739e2b7..3b3e6ea 100644
> > > >>> --- a/hw/mips/Makefile.objs
> > > >>> +++ b/hw/mips/Makefile.objs
> > > >>> @@ -1,4 +1,4 @@
> > > >>> -obj-y += addr.o mips_int.o
> > > >>> +obj-y += addr.o common.o mips_int.o
> > > >>>  obj-$(CONFIG_R4K) += r4k.o
> > > >>>  obj-$(CONFIG_MALTA) += gt64xxx_pci.o malta.o
> > > >>>  obj-$(CONFIG_MIPSSIM) += mipssim.o
> > > >>> diff --git a/hw/mips/common.c b/hw/mips/common.c
> > > >>> new file mode 100644
> > > >>> index 0000000..4d8e141
> > > >>> --- /dev/null
> > > >>> +++ b/hw/mips/common.c
> > > >>> @@ -0,0 +1,42 @@
> > > >>> +/*
> > > >>> + * Common MIPS routines
> > > >>> + *
> > > >>> + * Copyright (c) 2020 Huacai Chen (chenhc@lemote.com)
> > > >>> + * This code is licensed under the GNU GPL v2.
> > > >>> + */
> > > >>> +
> > > >>> +#include <linux/kvm.h>
> > > >>> +#include "qemu/osdep.h"
> > > >>> +#include "qemu-common.h"
> > > >>> +#include "hw/boards.h"
> > > >>> +#include "hw/mips/mips.h"
> > > >>> +#include "sysemu/kvm_int.h"
> > > >>> +
> > > >>> +#ifndef CONFIG_KVM
> > > >>> +
> > > >>> +int mips_kvm_type(MachineState *machine, const char *vm_type)
> > > >>> +{
> > > >>> +    return 0;
> > > >>> +}
> > > >>> +
> > > >>> +#else
> > > >>> +
> > > >>> +int mips_kvm_type(MachineState *machine, const char *vm_type)
> > > >>> +{
> > > >>> +    int r;
> > > >>> +    KVMState *s = KVM_STATE(machine->accelerator);
> > > >>> +
> > > >>> +    r = kvm_check_extension(s, KVM_CAP_MIPS_VZ);
> > > >>> +    if (r > 0) {
> > > >>> +        return KVM_VM_MIPS_VZ;
> > > >>> +    }
> > > >>> +
> > > >>> +    r = kvm_check_extension(s, KVM_CAP_MIPS_TE);
> > > >>> +    if (r > 0) {
> > > >>> +        return KVM_VM_MIPS_TE;
> > > >>> +    }
> > > >>> +
> > > >>> +    return -1;
> > > >>> +}
> > > >>> +
> > > >>> +#endif
> > > >>> diff --git a/include/hw/mips/mips.h b/include/hw/mips/mips.h
> > > >>> index 0af4c3d..2ac0580 100644
> > > >>> --- a/include/hw/mips/mips.h
> > > >>> +++ b/include/hw/mips/mips.h
> > > >>> @@ -20,4 +20,7 @@ void rc4030_dma_write(void *dma, uint8_t *buf,
> int len);
> > > >>>
> > > >>>  DeviceState *rc4030_init(rc4030_dma **dmas, IOMMUMemoryRegion
> **dma_mr);
> > > >>>
> > > >>> +/* common.c */
> > > >>> +int mips_kvm_type(MachineState *machine, const char *vm_type);
> > > >>> +
> > > >>>  #endif
> > > >>> --
> > > >>> 2.7.0
> > > >>>
> > > >
> > >
> > >
>

[-- Attachment #2: Type: text/html, Size: 12325 bytes --]

  reply	other threads:[~2020-06-16 21:18 UTC|newest]

Thread overview: 41+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-06-02  2:39 [PATCH for-5.1 V4 0/7] mips: Add Loongson-3 machine support (with KVM) Huacai Chen
2020-06-02  2:39 ` [PATCH for-5.1 V4 1/4] hw/mips: Implement the kvm_type() hook in MachineClass Huacai Chen
2020-06-03 14:34   ` Aleksandar Markovic
2020-06-04  0:57     ` Huacai Chen
2020-06-04 10:04       ` Aleksandar Markovic
2020-06-14  8:07   ` Aleksandar Markovic
2020-06-15  0:52     ` Huacai Chen
2020-06-15  8:55       ` Thomas Huth
2020-06-15 19:44         ` Aleksandar Markovic
2020-06-16  6:11           ` Huacai Chen
2020-06-16 21:17             ` Aleksandar Markovic [this message]
2020-06-02  2:39 ` [PATCH for-5.1 V4 2/4] target/mips: Add Loongson-3 CPU definition Huacai Chen
2020-06-06  7:30   ` Aleksandar Markovic
2020-06-02  2:39 ` [PATCH for-5.1 V4 3/4] hw/mips: Add Loongson-3 machine support (with KVM) Huacai Chen
2020-06-06  7:32   ` Aleksandar Markovic
2020-06-06  8:01     ` Aleksandar Markovic
2020-06-07  1:12       ` chen huacai
2020-06-07 20:00         ` Aleksandar Markovic
2020-06-08  3:56           ` Huacai Chen
2020-06-11  5:58   ` Jiaxun Yang
2020-06-11  7:49     ` Huacai Chen
2020-06-11  8:12       ` Jiaxun Yang
2020-06-11  8:50         ` Aleksandar Markovic
2020-06-12  6:07           ` Huacai Chen
2020-06-14  7:51   ` Aleksandar Markovic
2020-06-15  0:55     ` Huacai Chen
2020-06-15  4:42       ` Aleksandar Markovic
2020-06-15  4:50       ` Aleksandar Markovic
2020-06-15  5:36         ` Huacai Chen
2020-06-15  5:58           ` Aleksandar Markovic
2020-06-15  6:04           ` Aleksandar Markovic
2020-06-15  6:29             ` Huacai Chen
2020-06-15  6:44               ` Aleksandar Markovic
2020-06-02  2:39 ` [PATCH for-5.1 V4 4/4] MAINTAINERS: Add myself as Loongson-3 maintainer Huacai Chen
2020-06-02  8:12   ` Philippe Mathieu-Daudé
2020-06-02 12:03     ` chen huacai
2020-06-05  8:38 ` [PATCH for-5.1 V4 0/7] mips: Add Loongson-3 machine support (with KVM) Aleksandar Markovic
2020-06-05  8:40   ` Aleksandar Markovic
2020-06-05  9:05   ` Jiaxun Yang
2020-06-05  9:21     ` Huacai Chen
2020-06-05  9:27     ` Aleksandar Markovic

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to='CAHiYmc4kqAFE=DsR74W8Y6HPPXUCqopTt2D-OCxuN9jYQY8m_A@mail.gmail.com' \
    --to=aleksandar.qemu.devel@gmail.com \
    --cc=aleksandar.m.mail@gmail.com \
    --cc=aleksandar.rikalo@rt-rk.com \
    --cc=aurelien@aurel32.net \
    --cc=chenhuacai@gmail.com \
    --cc=f4bug@amsat.org \
    --cc=qemu-devel@nongnu.org \
    --cc=thuth@redhat.com \
    --cc=zltjiangshi@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.