All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Roger Pau Monné" <roger.pau@citrix.com>
To: Lan Tianyu <tianyu.lan@intel.com>
Cc: andrew.cooper3@citrix.com, kevin.tian@intel.com,
	Chao Gao <chao.gao@intel.com>,
	jbeulich@suse.com, xen-devel@lists.xen.org
Subject: Re: [PATCH V3 28/29] x86/vvtd: Add queued invalidation (QI) support
Date: Fri, 20 Oct 2017 12:20:06 +0100	[thread overview]
Message-ID: <20171020112006.77dj3vipjql7anpp@dhcp-3-128.uk.xensource.com> (raw)
In-Reply-To: <1506049330-11196-29-git-send-email-tianyu.lan@intel.com>

On Thu, Sep 21, 2017 at 11:02:09PM -0400, Lan Tianyu wrote:
> From: Chao Gao <chao.gao@intel.com>
> 
> Queued Invalidation Interface is an expanded invalidation interface with
> extended capabilities. Hardware implementations report support for queued
> invalidation interface through the Extended Capability Register. The queued
> invalidation interface uses an Invalidation Queue (IQ), which is a circular
> buffer in system memory. Software submits commands by writing Invalidation
> Descriptors to the IQ.
> 
> In this patch, a new function viommu_process_iq() is used for emulating how
> hardware handles invalidation requests through QI.
> 
> Signed-off-by: Chao Gao <chao.gao@intel.com>
> Signed-off-by: Lan Tianyu <tianyu.lan@intel.com>
> ---
>  xen/drivers/passthrough/vtd/iommu.h |  19 ++-
>  xen/drivers/passthrough/vtd/vvtd.c  | 232 ++++++++++++++++++++++++++++++++++++
>  2 files changed, 250 insertions(+), 1 deletion(-)
> 
> diff --git a/xen/drivers/passthrough/vtd/iommu.h b/xen/drivers/passthrough/vtd/iommu.h
> index c69cd21..c2b83f1 100644
> --- a/xen/drivers/passthrough/vtd/iommu.h
> +++ b/xen/drivers/passthrough/vtd/iommu.h
> @@ -177,6 +177,21 @@
>  #define DMA_IRTA_S(val)         (val & 0xf)
>  #define DMA_IRTA_SIZE(val)      (1UL << (DMA_IRTA_S(val) + 1))
>  
> +/* IQA_REG */
> +#define DMA_IQA_ADDR(val)       (val & ~0xfffULL)
> +#define DMA_IQA_QS(val)         (val & 0x7)
> +#define DMA_IQA_RSVD            0xff8ULL
> +
> +/* IECTL_REG */
> +#define DMA_IECTL_IM_SHIFT 31
> +#define DMA_IECTL_IM            (1 << DMA_IECTL_IM_SHIFT)

Isn't this undefined behavior? It should be 1u.

You should consider using u for all the defines added.

> +#define DMA_IECTL_IP_SHIFT 30
> +#define DMA_IECTL_IP            (1 << DMA_IECTL_IP_SHIFT)
> +
> +/* ICS_REG */
> +#define DMA_ICS_IWC_SHIFT       0
> +#define DMA_ICS_IWC             (1 << DMA_ICS_IWC_SHIFT)
> +
>  /* PMEN_REG */
>  #define DMA_PMEN_EPM    (((u32)1) << 31)
>  #define DMA_PMEN_PRS    (((u32)1) << 0)
> @@ -211,7 +226,8 @@
>  #define DMA_FSTS_PPF (1U << DMA_FSTS_PPF_SHIFT)
>  #define DMA_FSTS_AFO (1U << 2)
>  #define DMA_FSTS_APF (1U << 3)
> -#define DMA_FSTS_IQE (1U << 4)
> +#define DMA_FSTS_IQE_SHIFT 4
> +#define DMA_FSTS_IQE (1U << DMA_FSTS_IQE_SHIFT)
>  #define DMA_FSTS_ICE (1U << 5)
>  #define DMA_FSTS_ITE (1U << 6)
>  #define DMA_FSTS_PRO_SHIFT 7
> @@ -562,6 +578,7 @@ struct qinval_entry {
>  
>  /* Queue invalidation head/tail shift */
>  #define QINVAL_INDEX_SHIFT 4
> +#define QINVAL_INDEX_MASK  0x7fff0ULL
>  
>  #define qinval_present(v) ((v).lo & 1)
>  #define qinval_fault_disable(v) (((v).lo >> 1) & 1)
> diff --git a/xen/drivers/passthrough/vtd/vvtd.c b/xen/drivers/passthrough/vtd/vvtd.c
> index 55f7a46..668d0c9 100644
> --- a/xen/drivers/passthrough/vtd/vvtd.c
> +++ b/xen/drivers/passthrough/vtd/vvtd.c
> @@ -28,6 +28,7 @@
>  #include <asm/current.h>
>  #include <asm/event.h>
>  #include <asm/hvm/domain.h>
> +#include <asm/hvm/support.h>
>  #include <asm/io_apic.h>
>  #include <asm/page.h>
>  #include <asm/p2m.h>
> @@ -419,6 +420,177 @@ static int vvtd_record_fault(struct vvtd *vvtd,
>      return X86EMUL_OKAY;
>  }
>  
> +/*
> + * Process a invalidation descriptor. Currently, only two types descriptors,
> + * Interrupt Entry Cache Invalidation Descritor and Invalidation Wait
> + * Descriptor are handled.
> + * @vvtd: the virtual vtd instance
> + * @i: the index of the invalidation descriptor to be processed
> + *
> + * If success return 0, or return non-zero when failure.
> + */
> +static int process_iqe(struct vvtd *vvtd, int i)

unsigned int.

> +{
> +    uint64_t iqa;
> +    struct qinval_entry *qinval_page;
> +    int ret = 0;
> +
> +    iqa = vvtd_get_reg_quad(vvtd, DMAR_IQA_REG);
> +    qinval_page = map_guest_page(vvtd->domain, DMA_IQA_ADDR(iqa)>>PAGE_SHIFT);

PFN_DOWN instead of open coding the shift. Both can be initialized
at declaration. Also AFAICT iqa is only used once, so the local
variable is not needed.

> +    if ( IS_ERR(qinval_page) )
> +    {
> +        gdprintk(XENLOG_ERR, "Can't map guest IRT (rc %ld)",
> +                 PTR_ERR(qinval_page));
> +        return PTR_ERR(qinval_page);
> +    }
> +
> +    switch ( qinval_page[i].q.inv_wait_dsc.lo.type )
> +    {
> +    case TYPE_INVAL_WAIT:
> +        if ( qinval_page[i].q.inv_wait_dsc.lo.sw )
> +        {
> +            uint32_t data = qinval_page[i].q.inv_wait_dsc.lo.sdata;
> +            uint64_t addr = (qinval_page[i].q.inv_wait_dsc.hi.saddr << 2);

Unneeded parentheses.

> +
> +            ret = hvm_copy_to_guest_phys(addr, &data, sizeof(data), current);
> +            if ( ret )
> +                vvtd_info("Failed to write status address");

Don't you need to return or do something here? (like raise some kind
of error?)

> +        }
> +
> +        /*
> +         * The following code generates an invalidation completion event
> +         * indicating the invalidation wait descriptor completion. Note that
> +         * the following code fragment is not tested properly.
> +         */
> +        if ( qinval_page[i].q.inv_wait_dsc.lo.iflag )
> +        {
> +            uint32_t ie_data, ie_addr;

Missing newline, but...

> +            if ( !vvtd_test_and_set_bit(vvtd, DMAR_ICS_REG, DMA_ICS_IWC_SHIFT) )
> +            {
> +                vvtd_set_bit(vvtd, DMAR_IECTL_REG, DMA_IECTL_IP_SHIFT);
> +                if ( !vvtd_test_bit(vvtd, DMAR_IECTL_REG, DMA_IECTL_IM_SHIFT) )
> +                {
> +                    ie_data = vvtd_get_reg(vvtd, DMAR_IEDATA_REG);
> +                    ie_addr = vvtd_get_reg(vvtd, DMAR_IEADDR_REG);
> +                    vvtd_generate_interrupt(vvtd, ie_addr, ie_data);

...you don't seem two need the two local variables. They are used only
once.

> +                    vvtd_clear_bit(vvtd, DMAR_IECTL_REG, DMA_IECTL_IP_SHIFT);
> +                }
> +            }
> +        }
> +        break;
> +
> +    case TYPE_INVAL_IEC:
> +        /*
> +         * Currently, no cache is preserved in hypervisor. Only need to update
> +         * pIRTEs which are modified in binding process.
> +         */
> +        break;
> +
> +    default:
> +        goto error;

There's no reason to use a label that's only used for the default
case. Simply place the code in the error label here.

> +    }
> +
> +    unmap_guest_page((void*)qinval_page);
> +    return ret;
> +
> + error:
> +    unmap_guest_page((void*)qinval_page);
> +    gdprintk(XENLOG_ERR, "Internal error in Queue Invalidation.\n");
> +    domain_crash(vvtd->domain);

Do you really need to crash the domain in such case?

> +    return ret;
> +}
> +
> +/*
> + * Invalidate all the descriptors in Invalidation Queue.
> + */
> +static void vvtd_process_iq(struct vvtd *vvtd)
> +{
> +    uint64_t iqh, iqt, iqa, max_entry, i;
> +    int err = 0;
> +
> +    /*
> +     * No new descriptor is fetched from the Invalidation Queue until
> +     * software clears the IQE field in the Fault Status Register
> +     */
> +    if ( vvtd_test_bit(vvtd, DMAR_FSTS_REG, DMA_FSTS_IQE_SHIFT) )
> +        return;
> +
> +    iqh = vvtd_get_reg_quad(vvtd, DMAR_IQH_REG);
> +    iqt = vvtd_get_reg_quad(vvtd, DMAR_IQT_REG);
> +    iqa = vvtd_get_reg_quad(vvtd, DMAR_IQA_REG);
> +
> +    max_entry = 1 << (QINVAL_ENTRY_ORDER + DMA_IQA_QS(iqa));
> +    iqh = MASK_EXTR(iqh, QINVAL_INDEX_MASK);
> +    iqt = MASK_EXTR(iqt, QINVAL_INDEX_MASK);

This should be done above, when they are initialized.

> +
> +    ASSERT(iqt < max_entry);
> +    if ( iqh == iqt )
> +        return;
> +
> +    for ( i = iqh; i != iqt; i = (i + 1) % max_entry )
> +    {
> +        err = process_iqe(vvtd, i);

process_iqe takes an int parameter, and here you are feeding it a
uint64_t.

> +        if ( err )
> +            break;
> +    }
> +    vvtd_set_reg_quad(vvtd, DMAR_IQH_REG, i << QINVAL_INDEX_SHIFT);
> +
> +    /*
> +     * When IQE set, IQH references the desriptor associated with the error.
> +     */
> +    if ( err )
> +        vvtd_report_non_recoverable_fault(vvtd, DMA_FSTS_IQE_SHIFT);
> +}
> +
> +static int vvtd_write_iqt(struct vvtd *vvtd, unsigned long val)

Not sure there's much point in making this function return int, when
all the return values are X86EMUL_OKAY.

Sionce val here is a register AFAICT, I would rather prefer that you
use an explicit type size (uint64_t or uint32_t as fits).

> +{
> +    uint64_t max_entry, iqa = vvtd_get_reg_quad(vvtd, DMAR_IQA_REG);
> +
> +    if ( val & ~QINVAL_INDEX_MASK )
> +    {
> +        vvtd_info("Attempt to set reserved bits in Invalidation Queue Tail");
> +        return X86EMUL_OKAY;
> +    }
> +
> +    max_entry = 1 << (QINVAL_ENTRY_ORDER + DMA_IQA_QS(iqa));

1ull please.

> +    if ( MASK_EXTR(val, QINVAL_INDEX_MASK) >= max_entry )
> +    {
> +        vvtd_info("IQT: Value %lx exceeded supported max index.", val);
> +        return X86EMUL_OKAY;
> +    }
> +
> +    vvtd_set_reg_quad(vvtd, DMAR_IQT_REG, val);
> +    vvtd_process_iq(vvtd);
> +
> +    return X86EMUL_OKAY;
> +}
> +
> +static int vvtd_write_iqa(struct vvtd *vvtd, unsigned long val)

Same here, it seems like this function should return void, because the
current return value is meaningless, and same comment about 'val'
being uintXX_t.

> +{
> +    uint32_t cap = vvtd_get_reg(vvtd, DMAR_CAP_REG);
> +    unsigned int guest_max_addr_width = cap_mgaw(cap);
> +
> +    if ( val & (~((1ULL << guest_max_addr_width) - 1) | DMA_IQA_RSVD) )
> +    {
> +        vvtd_info("Attempt to set reserved bits in Invalidation Queue Address");
> +        return X86EMUL_OKAY;
> +    }
> +
> +    vvtd_set_reg_quad(vvtd, DMAR_IQA_REG, val);
> +    return X86EMUL_OKAY;
> +}
> +
> +static int vvtd_write_ics(struct vvtd *vvtd, uint32_t val)
> +{
> +    if ( val & DMA_ICS_IWC )
> +    {
> +        vvtd_clear_bit(vvtd, DMAR_ICS_REG, DMA_ICS_IWC_SHIFT);
> +        /*When IWC field is cleared, the IP field needs to be cleared */
             ^ missing space.

> +        vvtd_clear_bit(vvtd, DMAR_IECTL_REG, DMA_IECTL_IP_SHIFT);
> +    }
> +    return X86EMUL_OKAY;

This function wants to be void.

> +}
> +
>  static int vvtd_write_frcd3(struct vvtd *vvtd, uint32_t val)
>  {
>      /* Writing a 1 means clear fault */
> @@ -430,6 +602,30 @@ static int vvtd_write_frcd3(struct vvtd *vvtd, uint32_t val)
>      return X86EMUL_OKAY;
>  }
>  
> +static int vvtd_write_iectl(struct vvtd *vvtd, uint32_t val)

void please.

> +{
> +    /*
> +     * Only DMA_IECTL_IM bit is writable. Generate pending event when unmask.
> +     */

Single line comments use /* ... */

> +    if ( !(val & DMA_IECTL_IM) )
> +    {
> +        /* Clear IM and clear IP */
> +        vvtd_clear_bit(vvtd, DMAR_IECTL_REG, DMA_IECTL_IM_SHIFT);
> +        if ( vvtd_test_and_clear_bit(vvtd, DMAR_IECTL_REG, DMA_IECTL_IP_SHIFT) )
> +        {
> +            uint32_t ie_data, ie_addr;
> +
> +            ie_data = vvtd_get_reg(vvtd, DMAR_IEDATA_REG);
> +            ie_addr = vvtd_get_reg(vvtd, DMAR_IEADDR_REG);
> +            vvtd_generate_interrupt(vvtd, ie_addr, ie_data);

No need for ie_data and ie_addr.

> +        }
> +    }
> +    else
> +        vvtd_set_bit(vvtd, DMAR_IECTL_REG, DMA_IECTL_IM_SHIFT);
> +
> +    return X86EMUL_OKAY;
> +}
> +
>  static int vvtd_write_fectl(struct vvtd *vvtd, uint32_t val)
>  {
>      /*
> @@ -476,6 +672,10 @@ static int vvtd_write_fsts(struct vvtd *vvtd, uint32_t val)
>      if ( !((vvtd_get_reg(vvtd, DMAR_FSTS_REG) & DMA_FSTS_FAULTS)) )
>          vvtd_clear_bit(vvtd, DMAR_FECTL_REG, DMA_FECTL_IP_SHIFT);
>  
> +    /* Continue to deal invalidation when IQE is clear */
> +    if ( !vvtd_test_bit(vvtd, DMAR_FSTS_REG, DMA_FSTS_IQE_SHIFT) )
> +        vvtd_process_iq(vvtd);
> +
>      return X86EMUL_OKAY;
>  }
>  
> @@ -611,6 +811,32 @@ static int vvtd_write(struct vcpu *v, unsigned long addr,
>          case DMAR_FECTL_REG:
>              return vvtd_write_fectl(vvtd, val);
>  
> +        case DMAR_IECTL_REG:
> +            return vvtd_write_iectl(vvtd, val);
> +
> +        case DMAR_ICS_REG:
> +            return vvtd_write_ics(vvtd, val);
> +
> +        case DMAR_IQT_REG:
> +            return vvtd_write_iqt(vvtd, (uint32_t)val);
> +
> +        case DMAR_IQA_REG:
> +        {
> +            uint32_t iqa_hi;
> +
> +            iqa_hi = vvtd_get_reg(vvtd, DMAR_IQA_REG_HI);

Initialization at declaration time, but since it's used only once, I
would rather prefer that you don't use a local variable at all.

> +            return vvtd_write_iqa(vvtd,
> +                                 (uint32_t)val | ((uint64_t)iqa_hi << 32));
> +        }
> +
> +        case DMAR_IQA_REG_HI:
> +        {
> +            uint32_t iqa_lo;
> +
> +            iqa_lo = vvtd_get_reg(vvtd, DMAR_IQA_REG);
> +            return vvtd_write_iqa(vvtd, (val << 32) | iqa_lo);

Same comment as above regarding iqa_lo.

Thanks, Roger.

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel

  reply	other threads:[~2017-10-20 11:20 UTC|newest]

Thread overview: 108+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-09-22  3:01 [PATCH V3 00/29] Lan Tianyu
2017-09-22  3:01 ` [PATCH V3 1/29] Xen/doc: Add Xen virtual IOMMU doc Lan Tianyu
2017-10-18 13:26   ` Roger Pau Monné
2017-10-19  2:26     ` Lan Tianyu
2017-10-19  8:49       ` Roger Pau Monné
2017-10-19 11:28         ` Jan Beulich
2017-10-24  7:16           ` Lan Tianyu
2017-09-22  3:01 ` [PATCH V3 2/29] VIOMMU: Add vIOMMU helper functions to create, destroy vIOMMU instance Lan Tianyu
2017-10-18 14:05   ` Roger Pau Monné
2017-10-19  6:31     ` Lan Tianyu
2017-10-19  8:47       ` Roger Pau Monné
2017-10-25  1:43         ` Lan Tianyu
2017-10-30  1:41           ` Lan Tianyu
2017-10-30  9:54             ` Roger Pau Monné
2017-10-30  1:51     ` Lan Tianyu
2017-11-06  8:19       ` Jan Beulich
2017-09-22  3:01 ` [PATCH V3 3/29] DOMCTL: Introduce new DOMCTL commands for vIOMMU support Lan Tianyu
2017-10-18 14:18   ` Roger Pau Monné
2017-10-19  6:41     ` Lan Tianyu
2017-09-22  3:01 ` [PATCH V3 4/29] tools/libacpi: Add DMA remapping reporting (DMAR) ACPI table structures Lan Tianyu
2017-10-18 14:36   ` Roger Pau Monné
2017-10-19  6:46     ` Lan Tianyu
2017-09-22  3:01 ` [PATCH V3 5/29] tools/libacpi: Add new fields in acpi_config for DMAR table Lan Tianyu
2017-10-18 15:12   ` Roger Pau Monné
2017-10-19  8:09     ` Lan Tianyu
2017-10-19  8:40       ` Roger Pau Monné
2017-10-25  6:06         ` Lan Tianyu
2017-10-19 11:31       ` Jan Beulich
2017-09-22  3:01 ` [PATCH V3 6/29] tools/libxl: Add a user configurable parameter to control vIOMMU attributes Lan Tianyu
2017-10-19  9:49   ` Roger Pau Monné
2017-10-20  1:36     ` Chao Gao
2017-09-22  3:01 ` [PATCH V3 7/29] tools/libxl: build DMAR table for a guest with one virtual VTD Lan Tianyu
2017-10-19 10:00   ` Roger Pau Monné
2017-10-20  1:44     ` Chao Gao
2017-09-22  3:01 ` [PATCH V3 8/29] tools/libxl: create vIOMMU during domain construction Lan Tianyu
2017-10-19 10:13   ` Roger Pau Monné
2017-10-26 12:05     ` Wei Liu
2017-10-27  1:58       ` Lan Tianyu
2017-09-22  3:01 ` [PATCH V3 9/29] tools/libxc: Add viommu operations in libxc Lan Tianyu
2017-10-19 10:17   ` Roger Pau Monné
2017-09-22  3:01 ` [PATCH V3 10/29] vtd: add and align register definitions Lan Tianyu
2017-10-19 10:21   ` Roger Pau Monné
2017-10-20  1:47     ` Chao Gao
2017-09-22  3:01 ` [PATCH V3 11/29] x86/hvm: Introduce a emulated VTD for HVM Lan Tianyu
2017-10-19 11:20   ` Roger Pau Monné
2017-10-20  2:46     ` Chao Gao
2017-10-20  6:56       ` Jan Beulich
2017-10-20  6:12         ` Chao Gao
2017-10-20  8:37         ` Lan Tianyu
2017-09-22  3:01 ` [PATCH V3 12/29] x86/vvtd: Add MMIO handler for VVTD Lan Tianyu
2017-10-19 11:34   ` Roger Pau Monné
2017-10-20  2:58     ` Chao Gao
2017-10-20  9:51       ` Roger Pau Monné
2017-09-22  3:01 ` [PATCH V3 13/29] x86/vvtd: Set Interrupt Remapping Table Pointer through GCMD Lan Tianyu
2017-10-19 11:56   ` Roger Pau Monné
2017-10-20  4:08     ` Chao Gao
2017-10-20  6:57       ` Jan Beulich
2017-09-22  3:01 ` [PATCH V3 14/29] x86/vvtd: Enable Interrupt Remapping " Lan Tianyu
2017-10-19 13:42   ` Roger Pau Monné
2017-09-22  3:01 ` [PATCH V3 15/29] x86/vvtd: Process interrupt remapping request Lan Tianyu
2017-10-19 14:26   ` Roger Pau Monné
2017-10-20  5:16     ` Chao Gao
2017-10-20 10:01       ` Roger Pau Monné
2017-10-23  6:44         ` Chao Gao
2017-09-22  3:01 ` [PATCH V3 16/29] x86/vvtd: decode interrupt attribute from IRTE Lan Tianyu
2017-10-19 14:39   ` Roger Pau Monné
2017-10-20  5:22     ` Chao Gao
2017-09-22  3:01 ` [PATCH V3 17/29] x86/vvtd: add a helper function to decide the interrupt format Lan Tianyu
2017-10-19 14:43   ` Roger Pau Monné
2017-09-22  3:01 ` [PATCH V3 18/29] VIOMMU: Add irq request callback to deal with irq remapping Lan Tianyu
2017-10-19 15:00   ` Roger Pau Monné
2017-09-22  3:02 ` [PATCH V3 19/29] x86/vioapic: Hook interrupt delivery of vIOAPIC Lan Tianyu
2017-10-19 15:37   ` Roger Pau Monné
2017-09-22  3:02 ` [PATCH V3 20/29] VIOMMU: Add get irq info callback to convert irq remapping request Lan Tianyu
2017-10-19 15:42   ` Roger Pau Monné
2017-10-25  7:30     ` Lan Tianyu
2017-10-25  7:43       ` Roger Pau Monné
2017-10-25  7:38         ` Lan Tianyu
2017-09-22  3:02 ` [PATCH V3 21/29] VIOMMU: Introduce callback of checking irq remapping mode Lan Tianyu
2017-10-19 15:43   ` Roger Pau Monné
2017-09-22  3:02 ` [PATCH V3 22/29] x86/vioapic: extend vioapic_get_vector() to support remapping format RTE Lan Tianyu
2017-10-19 15:49   ` Roger Pau Monné
2017-10-19 15:56     ` Jan Beulich
2017-10-20  1:04       ` Chao Gao
2017-09-22  3:02 ` [PATCH V3 23/29] passthrough: move some fields of hvm_gmsi_info to a sub-structure Lan Tianyu
2017-09-22  3:02 ` [PATCH V3 24/29] tools/libxc: Add a new interface to bind remapping format msi with pirq Lan Tianyu
2017-10-19 16:03   ` Roger Pau Monné
2017-10-20  5:39     ` Chao Gao
2017-09-22  3:02 ` [PATCH V3 25/29] x86/vmsi: Hook delivering remapping format msi to guest Lan Tianyu
2017-10-19 16:07   ` Roger Pau Monné
2017-10-20  6:48     ` Jan Beulich
2017-09-22  3:02 ` [PATCH V3 26/29] x86/vvtd: Handle interrupt translation faults Lan Tianyu
2017-10-19 16:31   ` Roger Pau Monné
2017-10-20  5:54     ` Chao Gao
2017-10-20 10:08       ` Roger Pau Monné
2017-10-20 14:20         ` Jan Beulich
2017-09-22  3:02 ` [PATCH V3 27/29] x86/vvtd: Enable Queued Invalidation through GCMD Lan Tianyu
2017-10-20 10:30   ` Roger Pau Monné
2017-09-22  3:02 ` [PATCH V3 28/29] x86/vvtd: Add queued invalidation (QI) support Lan Tianyu
2017-10-20 11:20   ` Roger Pau Monné [this message]
2017-10-23  7:50     ` Chao Gao
2017-10-23  8:57       ` Roger Pau Monné
2017-10-23  8:52         ` Chao Gao
2017-10-23 23:26           ` Tian, Kevin
2017-09-22  3:02 ` [PATCH V3 29/29] x86/vvtd: save and restore emulated VT-d Lan Tianyu
2017-10-20 11:25   ` Roger Pau Monné
2017-10-20 11:36 ` [PATCH V3 00/29] Roger Pau Monné
2017-10-23  1:23   ` Lan Tianyu

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=20171020112006.77dj3vipjql7anpp@dhcp-3-128.uk.xensource.com \
    --to=roger.pau@citrix.com \
    --cc=andrew.cooper3@citrix.com \
    --cc=chao.gao@intel.com \
    --cc=jbeulich@suse.com \
    --cc=kevin.tian@intel.com \
    --cc=tianyu.lan@intel.com \
    --cc=xen-devel@lists.xen.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 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.