From: "Arun Kodilkar, Sairaj" <sarunkod@amd.com>
To: Alejandro Jimenez <alejandro.j.jimenez@oracle.com>,
<qemu-devel@nongnu.org>
Cc: <pbonzini@redhat.com>, <mst@redhat.com>, <mjt@tls.msk.ru>,
<marcel.apfelbaum@gmail.com>, <vasant.hegde@amd.com>,
<suravee.suthikulpanit@amd.com>, <santosh.shukla@amd.com>,
<Wei.Huang2@amd.com>, <joao.m.martins@oracle.com>,
<boris.ostrovsky@oracle.com>
Subject: Re: [PATCH 6/6] amd_iommu: Do not assume passthrough translation for devices with DTE[TV]=0
Date: Thu, 20 Mar 2025 10:41:48 +0530 [thread overview]
Message-ID: <48b257fe-f3e6-4882-a7ee-9a790b1eab3e@amd.com> (raw)
In-Reply-To: <20250311152446.45086-7-alejandro.j.jimenez@oracle.com>
On 3/11/2025 8:54 PM, Alejandro Jimenez wrote:
> The AMD I/O Virtualization Technology (IOMMU) Specification (see Table 8: V,
> TV, and GV Fields in Device Table Entry), specifies that a DTE with V=0,
> TV=1 does not contain a valid address translation information. If a request
> requires a table walk, the walk is terminated when this condition is
> encountered.
>
> Do not assume that addresses for a device with DTE[TV]=0 are passed through
> (i.e. not remapped) and instead terminate the page table walk early.
>
> Cc: qemu-stable@nongnu.org
> Fixes: d29a09ca6842 ("hw/i386: Introduce AMD IOMMU")
> Signed-off-by: Alejandro Jimenez <alejandro.j.jimenez@oracle.com>
> ---
> hw/i386/amd_iommu.c | 88 +++++++++++++++++++++++++--------------------
> 1 file changed, 49 insertions(+), 39 deletions(-)
>
> diff --git a/hw/i386/amd_iommu.c b/hw/i386/amd_iommu.c
> index cf00450ebe..31d5522a62 100644
> --- a/hw/i386/amd_iommu.c
> +++ b/hw/i386/amd_iommu.c
> @@ -932,51 +932,61 @@ static void amdvi_page_walk(AMDVIAddressSpace *as, uint64_t *dte,
> uint64_t pte = dte[0], pte_addr, page_mask;
>
> /* make sure the DTE has TV = 1 */
> - if (pte & AMDVI_DEV_TRANSLATION_VALID) {
> - level = get_pte_translation_mode(pte);
> - if (level >= 7) {
> - trace_amdvi_mode_invalid(level, addr);
> + if (!(pte & AMDVI_DEV_TRANSLATION_VALID)) {
> + /*
> + * A DTE with V=1, TV=0 does not have a valid Page Table Root Pointer.
> + * An IOMMU processing a request that requires a table walk terminates
> + * the walk when it encounters this condition. Do the same and return
> + * instead of assuming that the address is forwarded without translation
> + * i.e. the passthrough case, as it is done for the case where DTE[V]=0.
> + */
Hi Alejandro,
According to AMD IOMMU specs TABLE 44 (IO_PAGE_FAULT Event Types), IOMMU
reports IO_PAGE_FAULT event when TV bit is not set in the DTE.
Hence you should use amdvi_page_fault to report the IO_PAGE_FAULT
event before returning.
Regards
Sairaj Kodilkar
> + return;
> + }
> +
> + level = get_pte_translation_mode(pte);
> + if (level >= 7) {
> + trace_amdvi_mode_invalid(level, addr);
> + return;
> + }
> + if (level == 0) {
> + goto no_remap;
> + }
> +
> + /* we are at the leaf page table or page table encodes a huge page */
> + do {
> + pte_perms = amdvi_get_perms(pte);
> + present = pte & 1;
> + if (!present || perms != (perms & pte_perms)) {
> + amdvi_page_fault(as->iommu_state, as->devfn, addr, perms);
> + trace_amdvi_page_fault(addr);
> return;
> }
> - if (level == 0) {
> - goto no_remap;
> - }
>
> - /* we are at the leaf page table or page table encodes a huge page */
> - do {
> - pte_perms = amdvi_get_perms(pte);
> - present = pte & 1;
> - if (!present || perms != (perms & pte_perms)) {
> - amdvi_page_fault(as->iommu_state, as->devfn, addr, perms);
> - trace_amdvi_page_fault(addr);
> - return;
> - }
> -
> - /* go to the next lower level */
> - pte_addr = pte & AMDVI_DEV_PT_ROOT_MASK;
> - /* add offset and load pte */
> - pte_addr += ((addr >> (3 + 9 * level)) & 0x1FF) << 3;
> - pte = amdvi_get_pte_entry(as->iommu_state, pte_addr, as->devfn);
> - if (!pte) {
> - return;
> - }
> - oldlevel = level;
> - level = get_pte_translation_mode(pte);
> - } while (level > 0 && level < 7);
> -
> - if (level == 0x7) {
> - page_mask = pte_override_page_mask(pte);
> - } else {
> - page_mask = pte_get_page_mask(oldlevel);
> + /* go to the next lower level */
> + pte_addr = pte & AMDVI_DEV_PT_ROOT_MASK;
> + /* add offset and load pte */
> + pte_addr += ((addr >> (3 + 9 * level)) & 0x1FF) << 3;
> + pte = amdvi_get_pte_entry(as->iommu_state, pte_addr, as->devfn);
> + if (!pte) {
> + return;
> }
> + oldlevel = level;
> + level = get_pte_translation_mode(pte);
> + } while (level > 0 && level < 7);
>
> - /* get access permissions from pte */
> - ret->iova = addr & page_mask;
> - ret->translated_addr = (pte & AMDVI_DEV_PT_ROOT_MASK) & page_mask;
> - ret->addr_mask = ~page_mask;
> - ret->perm = amdvi_get_perms(pte);
> - return;
> + if (level == 0x7) {
> + page_mask = pte_override_page_mask(pte);
> + } else {
> + page_mask = pte_get_page_mask(oldlevel);
> }
> +
> + /* get access permissions from pte */
> + ret->iova = addr & page_mask;
> + ret->translated_addr = (pte & AMDVI_DEV_PT_ROOT_MASK) & page_mask;
> + ret->addr_mask = ~page_mask;
> + ret->perm = amdvi_get_perms(pte);
> + return;
> +
> no_remap:
> ret->iova = addr & AMDVI_PAGE_MASK_4K;
> ret->translated_addr = addr & AMDVI_PAGE_MASK_4K;
next prev parent reply other threads:[~2025-03-20 5:18 UTC|newest]
Thread overview: 22+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-03-11 15:24 [PATCH 0/6] amd_iommu: Fixes to align with AMDVi specification Alejandro Jimenez
2025-03-11 15:24 ` [PATCH 1/6] amd_iommu: Fix Miscellanous Information Register 0 offsets Alejandro Jimenez
2025-03-17 12:37 ` Vasant Hegde
2025-03-11 15:24 ` [PATCH 2/6] amd_iommu: Fix Device ID decoding for INVALIDATE_IOTLB_PAGES command Alejandro Jimenez
2025-03-17 12:40 ` Vasant Hegde
2025-03-11 15:24 ` [PATCH 3/6] amd_iommu: Update bitmasks representing DTE reserved fields Alejandro Jimenez
2025-03-12 4:12 ` Arun Kodilkar, Sairaj
2025-03-13 14:23 ` Alejandro Jimenez
2025-03-16 9:34 ` Arun Kodilkar, Sairaj
2025-03-17 12:36 ` Vasant Hegde
2025-03-17 12:34 ` Vasant Hegde
2025-03-11 15:24 ` [PATCH 4/6] amd_iommu: Fix masks for Device Table Address Register Alejandro Jimenez
2025-03-12 5:32 ` Arun Kodilkar, Sairaj
2025-03-17 15:07 ` Vasant Hegde
2025-03-11 15:24 ` [PATCH 5/6] amd_iommu: Fix the calculation for Device Table size Alejandro Jimenez
2025-03-17 13:00 ` Vasant Hegde
2025-03-11 15:24 ` [PATCH 6/6] amd_iommu: Do not assume passthrough translation for devices with DTE[TV]=0 Alejandro Jimenez
2025-03-19 6:06 ` Vasant Hegde
2025-03-19 14:10 ` Alejandro Jimenez
2025-03-20 5:11 ` Arun Kodilkar, Sairaj [this message]
2025-03-20 16:56 ` Alejandro Jimenez
2025-03-21 8:37 ` Arun Kodilkar, Sairaj
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=48b257fe-f3e6-4882-a7ee-9a790b1eab3e@amd.com \
--to=sarunkod@amd.com \
--cc=Wei.Huang2@amd.com \
--cc=alejandro.j.jimenez@oracle.com \
--cc=boris.ostrovsky@oracle.com \
--cc=joao.m.martins@oracle.com \
--cc=marcel.apfelbaum@gmail.com \
--cc=mjt@tls.msk.ru \
--cc=mst@redhat.com \
--cc=pbonzini@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=santosh.shukla@amd.com \
--cc=suravee.suthikulpanit@amd.com \
--cc=vasant.hegde@amd.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 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).