From: Sasha Levin <levinsasha928@gmail.com>
To: Matt Evans <matt@ozlabs.org>
Cc: kvm@vger.kernel.org, kvm-ppc@vger.kernel.org
Subject: Re: [PATCH 28/28] kvm tools: Create arch-specific kvm_cpu__emulate_io()
Date: Tue, 06 Dec 2011 10:54:47 +0200 [thread overview]
Message-ID: <1323161687.3882.16.camel@lappy> (raw)
In-Reply-To: <4EDD8F51.2020007@ozlabs.org>
Can we possibly do it by getting the generic code to call both
'kvm_cpu__arch_emulate_io' and 'kvm_cpu__arch_emulate_mmio', and have
the ppc code have an empty static for 'kvm_cpu__arch_emulate_io'?
On Tue, 2011-12-06 at 14:43 +1100, Matt Evans wrote:
> Different architectures will deal with MMIO exits differently. For example,
> KVM_EXIT_IO is x86-specific, and I/O cycles are often synthesisted by steering
> into windows in PCI bridges on other architectures.
>
> This patch moves the IO/MMIO exit code from the main runloop into x86/kvm-cpu.c
>
> Signed-off-by: Matt Evans <matt@ozlabs.org>
> ---
> tools/kvm/include/kvm/kvm-cpu.h | 1 +
> tools/kvm/kvm-cpu.c | 37 +++++--------------------------------
> tools/kvm/x86/kvm-cpu.c | 37 +++++++++++++++++++++++++++++++++++++
> 3 files changed, 43 insertions(+), 32 deletions(-)
>
> diff --git a/tools/kvm/include/kvm/kvm-cpu.h b/tools/kvm/include/kvm/kvm-cpu.h
> index 15618f1..6f38c0c 100644
> --- a/tools/kvm/include/kvm/kvm-cpu.h
> +++ b/tools/kvm/include/kvm/kvm-cpu.h
> @@ -13,6 +13,7 @@ void kvm_cpu__run(struct kvm_cpu *vcpu);
> void kvm_cpu__reboot(void);
> int kvm_cpu__start(struct kvm_cpu *cpu);
> bool kvm_cpu__handle_exit(struct kvm_cpu *vcpu);
> +bool kvm_cpu__emulate_io(struct kvm_cpu *cpu, struct kvm_run *kvm_run);
>
> int kvm_cpu__get_debug_fd(void);
> void kvm_cpu__set_debug_fd(int fd);
> diff --git a/tools/kvm/kvm-cpu.c b/tools/kvm/kvm-cpu.c
> index 884a89f..c9fbc81 100644
> --- a/tools/kvm/kvm-cpu.c
> +++ b/tools/kvm/kvm-cpu.c
> @@ -103,49 +103,22 @@ int kvm_cpu__start(struct kvm_cpu *cpu)
> kvm_cpu__show_registers(cpu);
> kvm_cpu__show_code(cpu);
> break;
> - case KVM_EXIT_IO: {
> - bool ret;
> -
> - ret = kvm__emulate_io(cpu->kvm,
> - cpu->kvm_run->io.port,
> - (u8 *)cpu->kvm_run +
> - cpu->kvm_run->io.data_offset,
> - cpu->kvm_run->io.direction,
> - cpu->kvm_run->io.size,
> - cpu->kvm_run->io.count);
> -
> - if (!ret)
> + case KVM_EXIT_IO:
> + case KVM_EXIT_MMIO:
> + if (!kvm_cpu__emulate_io(cpu, cpu->kvm_run))
> goto panic_kvm;
> break;
> - }
> - case KVM_EXIT_MMIO: {
> - bool ret;
> -
> - ret = kvm__emulate_mmio(cpu->kvm,
> - cpu->kvm_run->mmio.phys_addr,
> - cpu->kvm_run->mmio.data,
> - cpu->kvm_run->mmio.len,
> - cpu->kvm_run->mmio.is_write);
> -
> - if (!ret)
> - goto panic_kvm;
> - break;
> - }
> case KVM_EXIT_INTR:
> if (cpu->is_running)
> break;
> goto exit_kvm;
> case KVM_EXIT_SHUTDOWN:
> goto exit_kvm;
> - default: {
> - bool ret;
> -
> - ret = kvm_cpu__handle_exit(cpu);
> - if (!ret)
> + default:
> + if (!kvm_cpu__handle_exit(cpu))
> goto panic_kvm;
> break;
> }
> - }
> kvm_cpu__handle_coalesced_mmio(cpu);
> }
>
> diff --git a/tools/kvm/x86/kvm-cpu.c b/tools/kvm/x86/kvm-cpu.c
> index a0d10cc..665d742 100644
> --- a/tools/kvm/x86/kvm-cpu.c
> +++ b/tools/kvm/x86/kvm-cpu.c
> @@ -217,6 +217,43 @@ bool kvm_cpu__handle_exit(struct kvm_cpu *vcpu)
> return false;
> }
>
> +bool kvm_cpu__emulate_io(struct kvm_cpu *cpu, struct kvm_run *kvm_run)
> +{
> + bool ret;
> + switch (kvm_run->exit_reason) {
> + case KVM_EXIT_IO: {
> + ret = kvm__emulate_io(cpu->kvm,
> + cpu->kvm_run->io.port,
> + (u8 *)cpu->kvm_run +
> + cpu->kvm_run->io.data_offset,
> + cpu->kvm_run->io.direction,
> + cpu->kvm_run->io.size,
> + cpu->kvm_run->io.count);
> +
> + if (!ret)
> + goto panic_kvm;
> + break;
> + }
> + case KVM_EXIT_MMIO: {
> + ret = kvm__emulate_mmio(cpu->kvm,
> + cpu->kvm_run->mmio.phys_addr,
> + cpu->kvm_run->mmio.data,
> + cpu->kvm_run->mmio.len,
> + cpu->kvm_run->mmio.is_write);
> +
> + if (!ret)
> + goto panic_kvm;
> + break;
> + }
> + default:
> + pr_warning("Unknown exit reason %d in %s\n", kvm_run->exit_reason, __FUNCTION__);
> + return false;
> + }
> + return true;
> +panic_kvm:
> + return false;
> +}
> +
> static void print_dtable(const char *name, struct kvm_dtable *dtable)
> {
> dprintf(debug_fd, " %s %016llx %08hx\n",
> --
> To unsubscribe from this list: send the line "unsubscribe kvm" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
--
Sasha.
next prev parent reply other threads:[~2011-12-06 8:55 UTC|newest]
Thread overview: 105+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <cover.1323141075.git.matt@ozlabs.org>
2011-12-06 3:37 ` [PATCH 01/28] kvm tools: Split x86 arch-specific bits into x86/ Matt Evans
2011-12-06 8:07 ` Sasha Levin
2011-12-06 10:10 ` Pekka Enberg
2011-12-06 3:37 ` [PATCH 02/28] kvm tools: Only build/init i8042 on x86 Matt Evans
2011-12-06 8:11 ` Sasha Levin
2011-12-06 18:59 ` Scott Wood
2011-12-07 7:39 ` Matt Evans
2011-12-06 3:37 ` [PATCH 03/28] kvm tools: Add Makefile parameter for kernel include path Matt Evans
2011-12-06 3:38 ` [PATCH 04/28] kvm tools: Re-arrange Makefile to heed CFLAGS before checking for optional libs Matt Evans
2011-12-06 3:38 ` [PATCH 05/28] kvm tools: 64-bit tidy; use PRIx64 when printf'ing u64s and link appropriately Matt Evans
2011-12-06 8:13 ` Sasha Levin
2011-12-06 8:28 ` Ingo Molnar
2011-12-06 10:05 ` Paul Mackerras
2011-12-06 10:24 ` Ingo Molnar
2011-12-07 7:01 ` Matt Evans
2011-12-07 8:16 ` Ingo Molnar
2011-12-07 12:49 ` Paolo Bonzini
2011-12-07 17:21 ` Pekka Enberg
2011-12-07 17:14 ` Pekka Enberg
2011-12-08 3:14 ` Matt Evans
2011-12-08 4:49 ` Ingo Molnar
2011-12-08 4:56 ` Matt Evans
2011-12-08 5:47 ` Matt Evans
2011-12-08 5:49 ` Ingo Molnar
2011-12-08 6:17 ` Matt Evans
2011-12-06 3:39 ` [PATCH 06/28] kvm tools: Add arch-specific KVM_RUN exit handling via kvm_cpu__handle_exit() Matt Evans
2011-12-06 3:39 ` [PATCH 07/28] kvm tools: Move 'kvm__recommended_cpus' to arch-specific code Matt Evans
2011-12-06 8:20 ` Sasha Levin
2011-12-07 6:17 ` Matt Evans
2011-12-07 6:34 ` Sasha Levin
2011-12-07 7:19 ` Matt Evans
2011-12-07 7:24 ` Alexander Graf
2011-12-07 7:28 ` Matt Evans
2011-12-07 8:29 ` Sasha Levin
2011-12-07 14:11 ` Avi Kivity
2011-12-07 14:22 ` Sasha Levin
2011-12-07 14:25 ` Avi Kivity
2011-12-07 15:00 ` Alexander Graf
2011-12-07 15:25 ` Sasha Levin
2011-12-07 15:58 ` Alexander Graf
2011-12-20 15:23 ` Alexander Graf
2011-12-21 22:17 ` Paul Mackerras
2011-12-23 14:05 ` Alexander Graf
2011-12-20 15:20 ` Alexander Graf
2011-12-06 3:39 ` [PATCH 08/28] kvm tools: Fix KVM_RUN exit code check Matt Evans
2011-12-06 8:22 ` Sasha Levin
2011-12-07 0:32 ` Matt Evans
2011-12-07 6:44 ` Alexander Graf
2011-12-07 14:12 ` Avi Kivity
2011-12-07 15:01 ` Alexander Graf
2011-12-07 15:05 ` Avi Kivity
2011-12-08 3:03 ` Matt Evans
2011-12-08 5:31 ` Sasha Levin
2011-12-22 10:03 ` Avi Kivity
2011-12-22 10:18 ` Sasha Levin
2011-12-22 10:21 ` Avi Kivity
2011-12-22 10:34 ` Sasha Levin
2011-12-06 3:39 ` [PATCH 09/28] kvm tools: Add kvm__arch_periodic_poll() Matt Evans
2011-12-06 3:40 ` [PATCH 10/28] kvm tools: term.h needs to include stdbool.h Matt Evans
2011-12-06 3:40 ` [PATCH 11/28] kvm tools: kvm.c needs to include sys/stat.h for mkdir Matt Evans
2011-12-06 3:40 ` [PATCH 12/28] kvm tools: Move arch-specific cmdline init into kvm__arch_set_cmdline() Matt Evans
2011-12-06 3:40 ` [PATCH 13/28] kvm tools: Add CONSOLE_HV term type and allow it to be selected Matt Evans
2011-12-06 3:40 ` [PATCH 14/28] kvm tools: Fix term_getc(), term_getc_iov() endian bugs Matt Evans
2011-12-06 10:24 ` Pekka Enberg
2011-12-06 12:00 ` Asias He
2011-12-07 2:39 ` Matt Evans
2011-12-06 3:40 ` [PATCH 15/28] kvm tools: Allow initrd_check() to match a cpio Matt Evans
2011-12-06 10:26 ` Pekka Enberg
2011-12-06 3:41 ` [PATCH 16/28] kvm tools: Allow load_flat_binary() to load an initrd alongside Matt Evans
2011-12-06 10:29 ` Pekka Enberg
2011-12-06 12:04 ` Cyrill Gorcunov
2011-12-07 0:42 ` Matt Evans
2011-12-07 6:33 ` Cyrill Gorcunov
2011-12-06 3:41 ` [PATCH 17/28] kvm tools: Only call symbol__init() if we have BFD Matt Evans
2011-12-06 8:26 ` Sasha Levin
2011-12-07 3:03 ` Matt Evans
2011-12-06 3:41 ` [PATCH 18/28] kvm tools: Initialise PCI before devices start getting registered with PCI Matt Evans
2011-12-06 3:41 ` [PATCH 19/28] kvm tools: Perform CPU and firmware setup after devices are added Matt Evans
2011-12-06 3:41 ` [PATCH 20/28] kvm tools: Init IRQs after determining nrcpus Matt Evans
2011-12-06 3:41 ` [PATCH 21/28] kvm tools: Add --hugetlbfs option to specify memory path Matt Evans
2011-12-06 8:32 ` Sasha Levin
2011-12-07 0:35 ` Matt Evans
2011-12-07 6:01 ` Sasha Levin
2011-12-06 3:42 ` [PATCH 22/28] kvm tools: Move PCI_MAX_DEVICES to pci.h Matt Evans
2011-12-06 3:42 ` [PATCH 23/28] kvm tools: Endian-sanitise pci.h and PCI device setup Matt Evans
2011-12-06 10:25 ` Pekka Enberg
2011-12-06 10:28 ` Cyrill Gorcunov
2011-12-06 11:41 ` Pekka Enberg
2011-12-06 11:47 ` Cyrill Gorcunov
2011-12-06 11:58 ` Pekka Enberg
2011-12-06 12:10 ` Cyrill Gorcunov
2011-12-06 13:29 ` Pekka Enberg
2011-12-06 13:38 ` Cyrill Gorcunov
2011-12-07 2:58 ` Matt Evans
2011-12-06 3:42 ` [PATCH 24/28] kvm tools: Fix virtio-pci endian bug when reading VIRTIO_PCI_QUEUE_NUM Matt Evans
2011-12-06 10:26 ` Pekka Enberg
2011-12-06 11:28 ` Asias He
2011-12-06 11:39 ` Pekka Enberg
2011-12-06 13:15 ` Sasha Levin
2011-12-06 3:42 ` [PATCH 25/28] kvm tools: Correctly set virtio-pci bar_size and remove hardwired address Matt Evans
2011-12-06 3:42 ` [PATCH 26/28] kvm tools: Add pci__config_{rd,wr}(), pci__find_dev() and fix PCI config register addressing Matt Evans
2011-12-06 3:43 ` [PATCH 27/28] kvm tools: Arch-specific define for PCI MMIO allocation area Matt Evans
2011-12-06 3:43 ` [PATCH 28/28] kvm tools: Create arch-specific kvm_cpu__emulate_io() Matt Evans
2011-12-06 8:54 ` Sasha Levin [this message]
2011-12-07 6:36 ` Matt Evans
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=1323161687.3882.16.camel@lappy \
--to=levinsasha928@gmail.com \
--cc=kvm-ppc@vger.kernel.org \
--cc=kvm@vger.kernel.org \
--cc=matt@ozlabs.org \
/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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).