mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Rick Edgecombe <rick.p.edgecombe@intel.com>
To: bp@alien8.de, chao.gao@intel.com, dave.hansen@intel.com,
	isaku.yamahata@intel.com, kai.huang@intel.com, kas@kernel.org,
	kvm@vger.kernel.org, linux-coco@lists.linux.dev,
	linux-kernel@vger.kernel.org, mingo@redhat.com,
	pbonzini@redhat.com, seanjc@google.com, tglx@linutronix.de,
	vannapurve@google.com, x86@kernel.org, yan.y.zhao@intel.com,
	xiaoyao.li@intel.com, binbin.wu@intel.com
Cc: rick.p.edgecombe@intel.com,
	"Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Subject: [PATCH v4 04/16] x86/virt/tdx: Allocate page bitmap for Dynamic PAMT
Date: Thu, 20 Nov 2025 16:51:13 -0800	[thread overview]
Message-ID: <20251121005125.417831-5-rick.p.edgecombe@intel.com> (raw)
In-Reply-To: <20251121005125.417831-1-rick.p.edgecombe@intel.com>

From: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>

The Physical Address Metadata Table (PAMT) holds TDX metadata for physical
memory and must be allocated by the kernel during TDX module
initialization.

The exact size of the required PAMT memory is determined by the TDX module
and may vary between TDX module versions. Currently it is approximately
0.4% of the system memory. This is a significant commitment, especially if
it is not known upfront whether the machine will run any TDX guests.

For normal PAMT, each memory region that the TDX module might use (TDMR)
needs three separate PAMT allocations. One for each supported page size
(1GB, 2MB, 4KB).

At a high level, Dynamic PAMT still has the 1GB and 2MB levels allocated
on TDX module initialization, but the 4KB level allocated dynamically at
TD runtime. However, in the details, the TDX module still needs some per
4KB page data. The TDX module exposed how many bits per page need to be
allocated (currently it is 1). The bits-per-page value can then be used to
calculate the size to pass in place of the 4KB allocations in the TDMR,
which TDX specs call "PAMT_PAGE_BITMAP".

So in effect, Dynamic PAMT just needs a different (smaller) size
allocation for the 4KB level part of the allocation. Although it is
functionally something different, it is passed in the same way the 4KB page
size PAMT allocation is.

Begin to implement Dynamic PAMT in the kernel by reading the bits-per-page
needed for Dynamic PAMT. Calculate the size needed for the bitmap,
and use it instead of the 4KB size determined for normal PAMT, in the case
of Dynamic PAMT. In doing so, reduce the static allocations to
approximately 0.004%, a 100x improvement.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
[Enhanced log]
Signed-off-by: Rick Edgecombe <rick.p.edgecombe@intel.com>
---
v4:
 - Use PAGE_ALIGN (Binbin)
 - Add comment about why tdx_supports_dynamic_pamt() is checked in
   metadata reading.

v3:
 - Simplify pamt_4k_size calculation after addition of refactor patch
 - Write log
---
 arch/x86/include/asm/tdx.h                  |  5 +++++
 arch/x86/include/asm/tdx_global_metadata.h  |  1 +
 arch/x86/virt/vmx/tdx/tdx.c                 | 19 ++++++++++++++++++-
 arch/x86/virt/vmx/tdx/tdx_global_metadata.c |  7 +++++++
 4 files changed, 31 insertions(+), 1 deletion(-)

diff --git a/arch/x86/include/asm/tdx.h b/arch/x86/include/asm/tdx.h
index 99bbeed8fb28..cf51ccd16194 100644
--- a/arch/x86/include/asm/tdx.h
+++ b/arch/x86/include/asm/tdx.h
@@ -130,6 +130,11 @@ int tdx_enable(void);
 const char *tdx_dump_mce_info(struct mce *m);
 const struct tdx_sys_info *tdx_get_sysinfo(void);
 
+static inline bool tdx_supports_dynamic_pamt(const struct tdx_sys_info *sysinfo)
+{
+	return false; /* To be enabled when kernel is ready */
+}
+
 int tdx_guest_keyid_alloc(void);
 u32 tdx_get_nr_guest_keyids(void);
 void tdx_guest_keyid_free(unsigned int keyid);
diff --git a/arch/x86/include/asm/tdx_global_metadata.h b/arch/x86/include/asm/tdx_global_metadata.h
index 060a2ad744bf..5eb808b23997 100644
--- a/arch/x86/include/asm/tdx_global_metadata.h
+++ b/arch/x86/include/asm/tdx_global_metadata.h
@@ -15,6 +15,7 @@ struct tdx_sys_info_tdmr {
 	u16 pamt_4k_entry_size;
 	u16 pamt_2m_entry_size;
 	u16 pamt_1g_entry_size;
+	u8  pamt_page_bitmap_entry_bits;
 };
 
 struct tdx_sys_info_td_ctrl {
diff --git a/arch/x86/virt/vmx/tdx/tdx.c b/arch/x86/virt/vmx/tdx/tdx.c
index f2b16a91ba58..6c522db71265 100644
--- a/arch/x86/virt/vmx/tdx/tdx.c
+++ b/arch/x86/virt/vmx/tdx/tdx.c
@@ -440,6 +440,18 @@ static int fill_out_tdmrs(struct list_head *tmb_list,
 	return 0;
 }
 
+static unsigned long tdmr_get_pamt_bitmap_sz(struct tdmr_info *tdmr)
+{
+	unsigned long pamt_sz, nr_pamt_entries;
+	int bits_per_entry;
+
+	bits_per_entry = tdx_sysinfo.tdmr.pamt_page_bitmap_entry_bits;
+	nr_pamt_entries = tdmr->size >> PAGE_SHIFT;
+	pamt_sz = DIV_ROUND_UP(nr_pamt_entries * bits_per_entry, BITS_PER_BYTE);
+
+	return PAGE_ALIGN(pamt_sz);
+}
+
 /*
  * Calculate PAMT size given a TDMR and a page size.  The returned
  * PAMT size is always aligned up to 4K page boundary.
@@ -507,7 +519,12 @@ static int tdmr_set_up_pamt(struct tdmr_info *tdmr,
 	 * Calculate the PAMT size for each TDX supported page size
 	 * and the total PAMT size.
 	 */
-	tdmr->pamt_4k_size = tdmr_get_pamt_sz(tdmr, TDX_PS_4K);
+	if (tdx_supports_dynamic_pamt(&tdx_sysinfo)) {
+		/* With Dynamic PAMT, PAMT_4K is replaced with a bitmap */
+		tdmr->pamt_4k_size = tdmr_get_pamt_bitmap_sz(tdmr);
+	} else {
+		tdmr->pamt_4k_size = tdmr_get_pamt_sz(tdmr, TDX_PS_4K);
+	}
 	tdmr->pamt_2m_size = tdmr_get_pamt_sz(tdmr, TDX_PS_2M);
 	tdmr->pamt_1g_size = tdmr_get_pamt_sz(tdmr, TDX_PS_1G);
 	tdmr_pamt_size = tdmr->pamt_4k_size + tdmr->pamt_2m_size + tdmr->pamt_1g_size;
diff --git a/arch/x86/virt/vmx/tdx/tdx_global_metadata.c b/arch/x86/virt/vmx/tdx/tdx_global_metadata.c
index 13ad2663488b..00ab0e550636 100644
--- a/arch/x86/virt/vmx/tdx/tdx_global_metadata.c
+++ b/arch/x86/virt/vmx/tdx/tdx_global_metadata.c
@@ -33,6 +33,13 @@ static int get_tdx_sys_info_tdmr(struct tdx_sys_info_tdmr *sysinfo_tdmr)
 		sysinfo_tdmr->pamt_2m_entry_size = val;
 	if (!ret && !(ret = read_sys_metadata_field(0x9100000100000012, &val)))
 		sysinfo_tdmr->pamt_1g_entry_size = val;
+	/*
+	 * Don't fail here if tdx_supports_dynamic_pamt() isn't supported. The
+	 * TDX code can fallback to normal PAMT if it's not supported.
+	 */
+	if (!ret && tdx_supports_dynamic_pamt(&tdx_sysinfo) &&
+	    !(ret = read_sys_metadata_field(0x9100000100000013, &val)))
+		sysinfo_tdmr->pamt_page_bitmap_entry_bits = val;
 
 	return ret;
 }
-- 
2.51.2


  parent reply	other threads:[~2025-11-21  0:51 UTC|newest]

Thread overview: 106+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-11-21  0:51 [PATCH v4 00/16] TDX: Enable " Rick Edgecombe
2025-11-21  0:51 ` [PATCH v4 01/16] x86/tdx: Move all TDX error defines into <asm/shared/tdx_errno.h> Rick Edgecombe
2025-11-25 22:30   ` Huang, Kai
2025-11-25 22:44     ` Huang, Kai
2025-11-26 23:15       ` Edgecombe, Rick P
2025-11-26 23:14     ` Edgecombe, Rick P
2025-11-21  0:51 ` [PATCH v4 02/16] x86/tdx: Add helpers to check return status codes Rick Edgecombe
2025-11-24  8:56   ` Binbin Wu
2025-11-24 19:31     ` Edgecombe, Rick P
2025-11-25 23:07   ` Huang, Kai
2025-11-26 23:26     ` Edgecombe, Rick P
2025-11-21  0:51 ` [PATCH v4 03/16] x86/virt/tdx: Simplify tdmr_get_pamt_sz() Rick Edgecombe
2025-11-24  9:26   ` Binbin Wu
2025-11-24 19:47     ` Edgecombe, Rick P
2025-11-25  1:27       ` Binbin Wu
2025-11-21  0:51 ` Rick Edgecombe [this message]
2025-11-25  1:50   ` [PATCH v4 04/16] x86/virt/tdx: Allocate page bitmap for Dynamic PAMT Binbin Wu
2025-11-26 17:56     ` Edgecombe, Rick P
2025-12-24  9:10   ` Xu Yilun
2026-01-05 22:06     ` Edgecombe, Rick P
2026-01-06  4:01       ` Xu Yilun
2026-01-06 17:00         ` Edgecombe, Rick P
2026-01-07  6:01           ` Xu Yilun
2026-01-07 14:41             ` Edgecombe, Rick P
2026-01-08 12:53               ` Xu Yilun
2026-01-08 16:52                 ` Edgecombe, Rick P
2026-01-09  2:18                   ` Xu Yilun
2026-01-09 16:05                     ` Edgecombe, Rick P
2026-01-12  0:24                       ` Xu Yilun
2025-11-21  0:51 ` [PATCH v4 05/16] x86/virt/tdx: Allocate reference counters for PAMT memory Rick Edgecombe
2025-11-21  0:51 ` [PATCH v4 06/16] x86/virt/tdx: Improve PAMT refcounts allocation for sparse memory Rick Edgecombe
2025-11-25  3:15   ` Binbin Wu
2025-11-26 20:47     ` Edgecombe, Rick P
2025-11-27 15:57       ` Kiryl Shutsemau
2025-12-01 22:14         ` Edgecombe, Rick P
2025-11-26 14:45   ` Nikolay Borisov
2025-11-26 20:47     ` Edgecombe, Rick P
2025-11-27  7:36       ` Nikolay Borisov
2025-12-11  0:07         ` Edgecombe, Rick P
2025-11-27 16:04       ` Kiryl Shutsemau
2025-11-21  0:51 ` [PATCH v4 07/16] x86/virt/tdx: Add tdx_alloc/free_page() helpers Rick Edgecombe
2025-11-25  8:09   ` Binbin Wu
2025-11-26 22:28     ` Edgecombe, Rick P
2026-01-29  1:19       ` Sean Christopherson
2026-01-29 17:18         ` Edgecombe, Rick P
2026-01-29 19:09           ` Sean Christopherson
2026-01-29 19:12             ` Edgecombe, Rick P
2025-11-26  1:21   ` Huang, Kai
2025-11-26 22:28     ` Edgecombe, Rick P
2025-11-27 12:29   ` Nikolay Borisov
2025-12-01 22:31     ` Edgecombe, Rick P
2025-11-27 16:11   ` Nikolay Borisov
2025-12-01 22:39     ` Edgecombe, Rick P
2025-12-02  7:38       ` Nikolay Borisov
2025-12-02 20:02         ` Edgecombe, Rick P
2025-12-03 13:46           ` Kiryl Shutsemau
2025-12-03 13:48             ` Nikolay Borisov
2025-12-03 15:41               ` Dave Hansen
2025-12-03 18:15                 ` Edgecombe, Rick P
2025-12-03 18:21                   ` Dave Hansen
2025-12-03 19:59                     ` Edgecombe, Rick P
2025-12-03 20:13                       ` Dave Hansen
2025-12-03 21:39                         ` Edgecombe, Rick P
2025-12-03 21:40                           ` Dave Hansen
2025-12-08  9:15   ` Yan Zhao
2025-12-08 20:27     ` Edgecombe, Rick P
2026-01-16 23:17   ` Sean Christopherson
2026-01-16 23:25     ` Edgecombe, Rick P
2026-01-16 23:40     ` Dave Hansen
2025-11-21  0:51 ` [PATCH v4 08/16] x86/virt/tdx: Optimize " Rick Edgecombe
2025-11-21  0:51 ` [PATCH v4 09/16] KVM: TDX: Allocate PAMT memory for TD control structures Rick Edgecombe
2025-11-25  9:11   ` Binbin Wu
2025-11-21  0:51 ` [PATCH v4 10/16] KVM: TDX: Allocate PAMT memory for vCPU " Rick Edgecombe
2025-11-25  9:14   ` Binbin Wu
2025-11-21  0:51 ` [PATCH v4 11/16] KVM: TDX: Add x86 ops for external spt cache Rick Edgecombe
2026-01-17  0:53   ` Sean Christopherson
2026-01-19  2:31     ` Yan Zhao
2026-01-20  8:42       ` Huang, Kai
2026-01-20  9:18         ` Yan Zhao
2026-01-20 10:00           ` Huang, Kai
2026-01-20 19:53     ` Edgecombe, Rick P
2026-01-21 22:12       ` Sean Christopherson
2026-01-21 22:34         ` Edgecombe, Rick P
2025-11-21  0:51 ` [PATCH v4 12/16] x86/virt/tdx: Add helpers to allow for pre-allocating pages Rick Edgecombe
2025-11-26  3:40   ` Binbin Wu
2025-11-26  5:21     ` Binbin Wu
2025-11-26 22:33     ` Edgecombe, Rick P
2025-11-27  2:38       ` Binbin Wu
2026-01-20  7:10         ` Huang, Kai
2026-01-20  7:46           ` Yan Zhao
2026-01-20  8:01             ` Huang, Kai
2026-01-17  1:02   ` Sean Christopherson
2026-01-21  0:52   ` Sean Christopherson
2026-01-21  0:58     ` Edgecombe, Rick P
2025-11-21  0:51 ` [PATCH v4 13/16] KVM: TDX: Handle PAMT allocation in fault path Rick Edgecombe
2025-11-26  5:56   ` Binbin Wu
2025-11-26 22:33     ` Edgecombe, Rick P
2025-11-21  0:51 ` [PATCH v4 14/16] KVM: TDX: Reclaim PAMT memory Rick Edgecombe
2025-11-26  8:53   ` Binbin Wu
2025-11-26 22:58     ` Edgecombe, Rick P
2025-11-21  0:51 ` [PATCH v4 15/16] x86/virt/tdx: Enable Dynamic PAMT Rick Edgecombe
2025-11-21  0:51 ` [PATCH v4 16/16] Documentation/x86: Add documentation for TDX's " Rick Edgecombe
2025-11-26 10:33   ` Binbin Wu
2025-11-26 20:05     ` Edgecombe, Rick P
2025-11-24 20:18 ` [PATCH v4 00/16] TDX: Enable " Sagi Shahar
2025-11-25 20:19 ` Vishal Annapurve

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=20251121005125.417831-5-rick.p.edgecombe@intel.com \
    --to=rick.p.edgecombe@intel.com \
    --cc=binbin.wu@intel.com \
    --cc=bp@alien8.de \
    --cc=chao.gao@intel.com \
    --cc=dave.hansen@intel.com \
    --cc=isaku.yamahata@intel.com \
    --cc=kai.huang@intel.com \
    --cc=kas@kernel.org \
    --cc=kirill.shutemov@linux.intel.com \
    --cc=kvm@vger.kernel.org \
    --cc=linux-coco@lists.linux.dev \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=seanjc@google.com \
    --cc=tglx@linutronix.de \
    --cc=vannapurve@google.com \
    --cc=x86@kernel.org \
    --cc=xiaoyao.li@intel.com \
    --cc=yan.y.zhao@intel.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

Powered by JetHome