qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: "Michael S. Tsirkin" <mst@redhat.com>
To: Yuke Peng <pykfirst@gmail.com>
Cc: qemu-devel <qemu-devel@nongnu.org>, Peter Xu <peterx@redhat.com>,
	Jason Wang <jasowang@redhat.com>,
	Paolo Bonzini <pbonzini@redhat.com>,
	Richard Henderson <richard.henderson@linaro.org>,
	Eduardo Habkost <eduardo@habkost.net>,
	Marcel Apfelbaum <marcel.apfelbaum@gmail.com>
Subject: Re: [PATCH] hw/i386/intel_iommu: Block CFI when necessary
Date: Mon, 1 Jul 2024 16:51:07 -0400	[thread overview]
Message-ID: <20240701164648-mutt-send-email-mst@kernel.org> (raw)
In-Reply-To: <20240625112819.862282-1-pykfirst@gmail.com>

On Tue, Jun 25, 2024 at 07:28:19PM +0800, Yuke Peng wrote:
> According to Intel VT-d specification 5.1.4, CFI must be blocked when
> Extended Interrupt Mode is enabled or Compatibility format interrupts
> are disabled.
> 
> Signed-off-by: Yuke Peng <pykfirst@gmail.com>
> ---
>  hw/i386/intel_iommu.c         | 28 ++++++++++++++++++++++++++++
>  hw/i386/trace-events          |  1 +
>  include/hw/i386/intel_iommu.h |  1 +
>  3 files changed, 30 insertions(+)
> 
> diff --git a/hw/i386/intel_iommu.c b/hw/i386/intel_iommu.c
> index 5085a6fee3..dfa2f979e7 100644
> --- a/hw/i386/intel_iommu.c
> +++ b/hw/i386/intel_iommu.c
> @@ -2410,6 +2410,22 @@ static void vtd_handle_gcmd_ire(IntelIOMMUState *s, bool en)
>      }
>  }
>  
> +/* Handle Compatibility Format Interrupts Enable/Disable */
> +static void vtd_handle_gcmd_cfi(IntelIOMMUState *s, bool en)
> +{
> +    trace_vtd_cfi_enable(en);
> +
> +    if (en) {
> +        s->cfi_enabled = true;
> +        /* Ok - report back to driver */
> +        vtd_set_clear_mask_long(s, DMAR_GSTS_REG, 0, VTD_GSTS_CFIS);
> +    } else {
> +        s->cfi_enabled = false;
> +        /* Ok - report back to driver */
> +        vtd_set_clear_mask_long(s, DMAR_GSTS_REG, VTD_GSTS_CFIS, 0);
> +    }
> +}
> +
>  /* Handle write to Global Command Register */
>  static void vtd_handle_gcmd_write(IntelIOMMUState *s)
>  {
> @@ -2440,6 +2456,10 @@ static void vtd_handle_gcmd_write(IntelIOMMUState *s)
>          /* Interrupt remap enable/disable */
>          vtd_handle_gcmd_ire(s, val & VTD_GCMD_IRE);
>      }
> +    if (changed & VTD_GCMD_CFI) {
> +        /* Compatibility format interrupts enable/disable */
> +        vtd_handle_gcmd_cfi(s, val & VTD_GCMD_CFI);
> +    }
>  }
>  
>  /* Handle write to Context Command Register */
> @@ -3304,6 +3324,7 @@ static const VMStateDescription vtd_vmstate = {
>          VMSTATE_BOOL(dmar_enabled, IntelIOMMUState),
>          VMSTATE_BOOL(qi_enabled, IntelIOMMUState),
>          VMSTATE_BOOL(intr_enabled, IntelIOMMUState),
> +        VMSTATE_BOOL(cfi_enabled, IntelIOMMUState),
>          VMSTATE_BOOL(intr_eime, IntelIOMMUState),
>          VMSTATE_END_OF_LIST()
>      }


Looks like this will break migration compat.


The normal waywhould be that we use something like
VMSTATE_UINT32_TEST() or subsections for the new vmstate field.
sometimes it's possible to re-generate the field from
the registers, then there's no need to add a new field
to migration at all, instead calculate it in vtd_post_load.



> @@ -3525,6 +3546,12 @@ static int vtd_interrupt_remap_msi(IntelIOMMUState *iommu,
>  
>      /* This is compatible mode. */
>      if (addr.addr.int_mode != VTD_IR_INT_FORMAT_REMAP) {
> +        if (iommu->intr_eime || !iommu->cfi_enabled) {
> +            if (do_fault) {
> +                vtd_report_ir_fault(iommu, sid, VTD_FR_IR_REQ_COMPAT, 0);
> +            }
> +            return -EINVAL;
> +        }
>          memcpy(translated, origin, sizeof(*origin));
>          goto out;
>      }
> @@ -3950,6 +3977,7 @@ static void vtd_init(IntelIOMMUState *s)
>      s->root_scalable = false;
>      s->dmar_enabled = false;
>      s->intr_enabled = false;
> +    s->cfi_enabled = false;
>      s->iq_head = 0;
>      s->iq_tail = 0;
>      s->iq = 0;
> diff --git a/hw/i386/trace-events b/hw/i386/trace-events
> index 53c02d7ac8..ffd87db65f 100644
> --- a/hw/i386/trace-events
> +++ b/hw/i386/trace-events
> @@ -57,6 +57,7 @@ vtd_dmar_translate(uint8_t bus, uint8_t slot, uint8_t func, uint64_t iova, uint6
>  vtd_dmar_enable(bool en) "enable %d"
>  vtd_dmar_fault(uint16_t sid, int fault, uint64_t addr, bool is_write) "sid 0x%"PRIx16" fault %d addr 0x%"PRIx64" write %d"
>  vtd_ir_enable(bool en) "enable %d"
> +vtd_cfi_enable(bool en) "enable %d"
>  vtd_ir_irte_get(int index, uint64_t lo, uint64_t hi) "index %d low 0x%"PRIx64" high 0x%"PRIx64
>  vtd_ir_remap(int index, int tri, int vec, int deliver, uint32_t dest, int dest_mode) "index %d trigger %d vector %d deliver %d dest 0x%"PRIx32" mode %d"
>  vtd_ir_remap_type(const char *type) "%s"
> diff --git a/include/hw/i386/intel_iommu.h b/include/hw/i386/intel_iommu.h
> index 7fa0a695c8..38e20d0f2c 100644
> --- a/include/hw/i386/intel_iommu.h
> +++ b/include/hw/i386/intel_iommu.h
> @@ -294,6 +294,7 @@ struct IntelIOMMUState {
>  
>      /* interrupt remapping */
>      bool intr_enabled;              /* Whether guest enabled IR */
> +    bool cfi_enabled;               /* Whether CFI is enabled */
>      dma_addr_t intr_root;           /* Interrupt remapping table pointer */
>      uint32_t intr_size;             /* Number of IR table entries */
>      bool intr_eime;                 /* Extended interrupt mode enabled */
> -- 
> 2.34.1



      parent reply	other threads:[~2024-07-01 20:51 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-06-25 11:28 [PATCH] hw/i386/intel_iommu: Block CFI when necessary Yuke Peng
2024-06-26  6:16 ` cmd
2024-06-27  5:08   ` Yuke Peng
2024-06-27  5:15     ` cmd
2024-07-01 20:51 ` Michael S. Tsirkin [this message]

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=20240701164648-mutt-send-email-mst@kernel.org \
    --to=mst@redhat.com \
    --cc=eduardo@habkost.net \
    --cc=jasowang@redhat.com \
    --cc=marcel.apfelbaum@gmail.com \
    --cc=pbonzini@redhat.com \
    --cc=peterx@redhat.com \
    --cc=pykfirst@gmail.com \
    --cc=qemu-devel@nongnu.org \
    --cc=richard.henderson@linaro.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).