qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: "Philippe Mathieu-Daudé" <f4bug@amsat.org>
To: BALATON Zoltan <balaton@eik.bme.hu>, qemu-devel@nongnu.org
Cc: Huacai Chen <chenhuacai@kernel.org>, John Snow <jsnow@redhat.com>,
	Gerd Hoffmann <kraxel@redhat.com>
Subject: Re: [PATCH] via-ide: Avoid expensive operations in irq handler
Date: Mon, 18 Oct 2021 11:42:42 +0200	[thread overview]
Message-ID: <549ece11-990f-a19b-5dfe-508e315a6163@amsat.org> (raw)
In-Reply-To: <20211018014059.13E65746353@zero.eik.bme.hu>

On 10/18/21 03:36, BALATON Zoltan wrote:
> Cache the pointer to PCI function 0 (ISA bridge, that this IDE device
> has to use for IRQs) in the PCIIDEState and pass that as the opaque
> data for the interrupt handler to eliminate both the need to look up
> function 0 at every interrupt and also a QOM type cast of the opaque
> pointer as that's also expensive (mainly due to qom-debug being
> enabled by default).
> 
> Suggested-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
> Signed-off-by: BALATON Zoltan <balaton@eik.bme.hu>
> ---
>  hw/ide/via.c         | 11 ++++++-----
>  include/hw/ide/pci.h |  1 +
>  2 files changed, 7 insertions(+), 5 deletions(-)
> 
> diff --git a/hw/ide/via.c b/hw/ide/via.c
> index 82def819c4..30566bc409 100644
> --- a/hw/ide/via.c
> +++ b/hw/ide/via.c
> @@ -104,15 +104,15 @@ static void bmdma_setup_bar(PCIIDEState *d)
>  
>  static void via_ide_set_irq(void *opaque, int n, int level)
>  {
> -    PCIDevice *d = PCI_DEVICE(opaque);
> +    PCIIDEState *d = opaque;
>  
>      if (level) {
> -        d->config[0x70 + n * 8] |= 0x80;
> +        d->parent_obj.config[0x70 + n * 8] |= 0x80;
>      } else {
> -        d->config[0x70 + n * 8] &= ~0x80;
> +        d->parent_obj.config[0x70 + n * 8] &= ~0x80;
>      }

Better not to access parent_obj, so I'd rather keep the previous
code as it. The rest is OK, thanks. If you don't want to respin
I can fix and take via mips-next.

>  
> -    via_isa_set_irq(pci_get_function_0(d), 14 + n, level);
> +    via_isa_set_irq(d->func0, 14 + n, level);
>  }
>  
>  static void via_ide_reset(DeviceState *dev)
> @@ -188,7 +188,8 @@ static void via_ide_realize(PCIDevice *dev, Error **errp)
>      bmdma_setup_bar(d);
>      pci_register_bar(dev, 4, PCI_BASE_ADDRESS_SPACE_IO, &d->bmdma_bar);
>  
> -    qdev_init_gpio_in(ds, via_ide_set_irq, 2);
> +    d->func0 = pci_get_function_0(dev);
> +    qdev_init_gpio_in_named_with_opaque(ds, via_ide_set_irq, d, NULL, 2);
>      for (i = 0; i < 2; i++) {
>          ide_bus_init(&d->bus[i], sizeof(d->bus[i]), ds, i, 2);
>          ide_init2(&d->bus[i], qdev_get_gpio_in(ds, i));
> diff --git a/include/hw/ide/pci.h b/include/hw/ide/pci.h
> index d8384e1c42..89d14abf95 100644
> --- a/include/hw/ide/pci.h
> +++ b/include/hw/ide/pci.h
> @@ -50,6 +50,7 @@ struct PCIIDEState {
>      IDEBus bus[2];
>      BMDMAState bmdma[2];
>      uint32_t secondary; /* used only for cmd646 */
> +    PCIDevice *func0; /* used only by IDE functions of superio chips */
>      MemoryRegion bmdma_bar;
>      MemoryRegion cmd_bar[2];
>      MemoryRegion data_bar[2];
> 


  reply	other threads:[~2021-10-18  9:43 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-10-18  1:36 [PATCH] via-ide: Avoid expensive operations in irq handler BALATON Zoltan
2021-10-18  9:42 ` Philippe Mathieu-Daudé [this message]
2021-10-18  9:51   ` BALATON Zoltan
2021-10-18 10:10     ` Philippe Mathieu-Daudé
2021-10-20 14:36       ` Eduardo Habkost
2021-10-20 14:58         ` BALATON Zoltan
2021-10-20 15:18           ` Eduardo Habkost
2021-10-20 16:32             ` John Snow
2021-10-20 22:58         ` BALATON Zoltan

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=549ece11-990f-a19b-5dfe-508e315a6163@amsat.org \
    --to=f4bug@amsat.org \
    --cc=balaton@eik.bme.hu \
    --cc=chenhuacai@kernel.org \
    --cc=jsnow@redhat.com \
    --cc=kraxel@redhat.com \
    --cc=qemu-devel@nongnu.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).