All of lore.kernel.org
 help / color / mirror / Atom feed
From: Jan Beulich <jbeulich@suse.com>
To: "xen-devel@lists.xenproject.org" <xen-devel@lists.xenproject.org>
Cc: "Andrew Cooper" <andrew.cooper3@citrix.com>,
	"Paul Durrant" <paul@xen.org>,
	"Roger Pau Monné" <roger.pau@citrix.com>
Subject: [PATCH v4 15/21] AMD/IOMMU: free all-empty page tables
Date: Mon, 25 Apr 2022 10:42:19 +0200	[thread overview]
Message-ID: <16f5b398-56f6-a70c-9ce0-0ad72eab5058@suse.com> (raw)
In-Reply-To: <b92e294e-7277-d977-bb96-7c28d60000c6@suse.com>

When a page table ends up with no present entries left, it can be
replaced by a non-present entry at the next higher level. The page table
itself can then be scheduled for freeing.

Note that while its output isn't used there yet,
pt_update_contig_markers() right away needs to be called in all places
where entries get updated, not just the one where entries get cleared.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
---
v4: Re-base over changes earlier in the series.
v3: Re-base over changes earlier in the series.
v2: New.

--- a/xen/drivers/passthrough/amd/iommu_map.c
+++ b/xen/drivers/passthrough/amd/iommu_map.c
@@ -21,6 +21,9 @@
 
 #include "iommu.h"
 
+#define CONTIG_MASK IOMMU_PTE_CONTIG_MASK
+#include <asm/pt-contig-markers.h>
+
 /* Given pfn and page table level, return pde index */
 static unsigned int pfn_to_pde_idx(unsigned long pfn, unsigned int level)
 {
@@ -33,16 +36,20 @@ static unsigned int pfn_to_pde_idx(unsig
 
 static union amd_iommu_pte clear_iommu_pte_present(unsigned long l1_mfn,
                                                    unsigned long dfn,
-                                                   unsigned int level)
+                                                   unsigned int level,
+                                                   bool *free)
 {
     union amd_iommu_pte *table, *pte, old;
+    unsigned int idx = pfn_to_pde_idx(dfn, level);
 
     table = map_domain_page(_mfn(l1_mfn));
-    pte = &table[pfn_to_pde_idx(dfn, level)];
+    pte = &table[idx];
     old = *pte;
 
     write_atomic(&pte->raw, 0);
 
+    *free = pt_update_contig_markers(&table->raw, idx, level, PTE_kind_null);
+
     unmap_domain_page(table);
 
     return old;
@@ -85,7 +92,11 @@ static union amd_iommu_pte set_iommu_pte
     if ( !old.pr || old.next_level ||
          old.mfn != next_mfn ||
          old.iw != iw || old.ir != ir )
+    {
         set_iommu_pde_present(pde, next_mfn, 0, iw, ir);
+        pt_update_contig_markers(&table->raw, pfn_to_pde_idx(dfn, level),
+                                 level, PTE_kind_leaf);
+    }
     else
         old.pr = false; /* signal "no change" to the caller */
 
@@ -322,6 +333,9 @@ static int iommu_pde_from_dfn(struct dom
             smp_wmb();
             set_iommu_pde_present(pde, next_table_mfn, next_level, true,
                                   true);
+            pt_update_contig_markers(&next_table_vaddr->raw,
+                                     pfn_to_pde_idx(dfn, level),
+                                     level, PTE_kind_table);
 
             *flush_flags |= IOMMU_FLUSHF_modified;
         }
@@ -347,6 +361,9 @@ static int iommu_pde_from_dfn(struct dom
                 next_table_mfn = mfn_x(page_to_mfn(table));
                 set_iommu_pde_present(pde, next_table_mfn, next_level, true,
                                       true);
+                pt_update_contig_markers(&next_table_vaddr->raw,
+                                         pfn_to_pde_idx(dfn, level),
+                                         level, PTE_kind_table);
             }
             else /* should never reach here */
             {
@@ -474,8 +491,24 @@ int cf_check amd_iommu_unmap_page(
 
     if ( pt_mfn )
     {
+        bool free;
+
         /* Mark PTE as 'page not present'. */
-        old = clear_iommu_pte_present(pt_mfn, dfn_x(dfn), level);
+        old = clear_iommu_pte_present(pt_mfn, dfn_x(dfn), level, &free);
+
+        while ( unlikely(free) && ++level < hd->arch.amd.paging_mode )
+        {
+            struct page_info *pg = mfn_to_page(_mfn(pt_mfn));
+
+            if ( iommu_pde_from_dfn(d, dfn_x(dfn), level, &pt_mfn,
+                                    flush_flags, false) )
+                BUG();
+            BUG_ON(!pt_mfn);
+
+            clear_iommu_pte_present(pt_mfn, dfn_x(dfn), level, &free);
+            *flush_flags |= IOMMU_FLUSHF_all;
+            iommu_queue_free_pgtable(hd, pg);
+        }
     }
 
     spin_unlock(&hd->arch.mapping_lock);



  parent reply	other threads:[~2022-04-25  8:52 UTC|newest]

Thread overview: 106+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-04-25  8:29 [PATCH v4 00/21] IOMMU: superpage support when not sharing pagetables Jan Beulich
2022-04-25  8:30 ` [PATCH v4 01/21] AMD/IOMMU: correct potentially-UB shifts Jan Beulich
2022-04-27 13:08   ` Andrew Cooper
2022-04-27 13:57     ` Jan Beulich
2022-05-03 10:10   ` Roger Pau Monné
2022-05-03 14:34     ` Jan Beulich
2022-04-25  8:32 ` [PATCH v4 02/21] IOMMU: simplify unmap-on-error in iommu_map() Jan Beulich
2022-04-27 13:16   ` Andrew Cooper
2022-04-27 14:05     ` Jan Beulich
2022-05-03 10:25   ` Roger Pau Monné
2022-05-03 14:37     ` Jan Beulich
2022-05-03 16:22       ` Roger Pau Monné
2022-04-25  8:32 ` [PATCH v4 03/21] IOMMU: add order parameter to ->{,un}map_page() hooks Jan Beulich
2022-04-25  8:33 ` [PATCH v4 04/21] IOMMU: have iommu_{,un}map() split requests into largest possible chunks Jan Beulich
2022-05-03 12:37   ` Roger Pau Monné
2022-05-03 14:44     ` Jan Beulich
2022-05-04 10:20       ` Roger Pau Monné
2022-04-25  8:34 ` [PATCH v4 05/21] IOMMU/x86: restrict IO-APIC mappings for PV Dom0 Jan Beulich
2022-05-03 13:00   ` Roger Pau Monné
2022-05-03 14:50     ` Jan Beulich
2022-05-04  9:32       ` Jan Beulich
2022-05-04 10:30         ` Roger Pau Monné
2022-05-04 10:51           ` Jan Beulich
2022-05-04 12:01             ` Roger Pau Monné
2022-05-04 12:12               ` Jan Beulich
2022-05-04 13:00                 ` Roger Pau Monné
2022-05-04 13:19                   ` Jan Beulich
2022-05-04 13:46                     ` Roger Pau Monné
2022-05-04 13:55                       ` Jan Beulich
2022-05-04 15:22                         ` Roger Pau Monné
2022-04-25  8:34 ` [PATCH v4 06/21] IOMMU/x86: perform PV Dom0 mappings in batches Jan Beulich
2022-05-03 14:49   ` Roger Pau Monné
2022-05-04  9:46     ` Jan Beulich
2022-05-04 11:20       ` Roger Pau Monné
2022-05-04 12:27         ` Jan Beulich
2022-05-04 13:55           ` Roger Pau Monné
2022-05-04 14:26             ` Jan Beulich
2022-04-25  8:35 ` [PATCH v4 07/21] IOMMU/x86: support freeing of pagetables Jan Beulich
2022-05-03 16:20   ` Roger Pau Monné
2022-05-04 13:07     ` Jan Beulich
2022-05-04 15:06       ` Roger Pau Monné
2022-05-05  8:20         ` Jan Beulich
2022-05-05  9:57           ` Roger Pau Monné
2022-04-25  8:36 ` [PATCH v4 08/21] AMD/IOMMU: walk trees upon page fault Jan Beulich
2022-05-04 15:57   ` Roger Pau Monné
2022-04-25  8:37 ` [PATCH v4 09/21] AMD/IOMMU: return old PTE from {set,clear}_iommu_pte_present() Jan Beulich
2022-04-25  8:38 ` [PATCH v4 10/21] AMD/IOMMU: allow use of superpage mappings Jan Beulich
2022-05-05 13:19   ` Roger Pau Monné
2022-05-05 14:34     ` Jan Beulich
2022-05-05 15:26       ` Roger Pau Monné
2022-04-25  8:38 ` [PATCH v4 11/21] VT-d: " Jan Beulich
2022-05-05 16:20   ` Roger Pau Monné
2022-05-06  6:13     ` Jan Beulich
2022-04-25  8:40 ` [PATCH v4 12/21] IOMMU: fold flush-all hook into "flush one" Jan Beulich
2022-05-06  8:38   ` Roger Pau Monné
2022-05-06  9:59     ` Jan Beulich
2022-04-25  8:40 ` [PATCH v4 13/21] IOMMU/x86: prefill newly allocate page tables Jan Beulich
2022-05-06 11:16   ` Roger Pau Monné
2022-05-19 12:12     ` Jan Beulich
2022-05-20 10:47       ` Roger Pau Monné
2022-05-20 11:11         ` Jan Beulich
2022-05-20 11:13           ` Jan Beulich
2022-05-20 12:22             ` Roger Pau Monné
2022-05-20 12:36               ` Jan Beulich
2022-05-20 14:28                 ` Roger Pau Monné
2022-05-20 14:38                   ` Roger Pau Monné
2022-05-23  6:49                     ` Jan Beulich
2022-05-23  9:10                       ` Roger Pau Monné
2022-05-23 10:52                         ` Jan Beulich
2022-04-25  8:41 ` [PATCH v4 14/21] x86: introduce helper for recording degree of contiguity in " Jan Beulich
2022-05-06 13:25   ` Roger Pau Monné
2022-05-18 10:06     ` Jan Beulich
2022-05-20 10:22       ` Roger Pau Monné
2022-05-20 10:59         ` Jan Beulich
2022-05-20 11:27           ` Roger Pau Monné
2022-04-25  8:42 ` Jan Beulich [this message]
2022-05-10 13:30   ` [PATCH v4 15/21] AMD/IOMMU: free all-empty " Roger Pau Monné
2022-05-18 10:18     ` Jan Beulich
2022-04-25  8:42 ` [PATCH v4 16/21] VT-d: " Jan Beulich
2022-04-27  4:09   ` Tian, Kevin
2022-05-10 14:30   ` Roger Pau Monné
2022-05-18 10:26     ` Jan Beulich
2022-05-20  0:38       ` Tian, Kevin
2022-05-20 11:13       ` Roger Pau Monné
2022-05-27  7:40         ` Jan Beulich
2022-05-27  7:53           ` Jan Beulich
2022-05-27  9:21             ` Roger Pau Monné
2022-04-25  8:43 ` [PATCH v4 17/21] AMD/IOMMU: replace all-contiguous page tables by superpage mappings Jan Beulich
2022-05-10 15:31   ` Roger Pau Monné
2022-05-18 10:40     ` Jan Beulich
2022-05-20 10:35       ` Roger Pau Monné
2022-04-25  8:43 ` [PATCH v4 18/21] VT-d: " Jan Beulich
2022-05-11 11:08   ` Roger Pau Monné
2022-05-18 10:44     ` Jan Beulich
2022-05-20 10:38       ` Roger Pau Monné
2022-04-25  8:44 ` [PATCH v4 19/21] IOMMU/x86: add perf counters for page table splitting / coalescing Jan Beulich
2022-05-11 13:48   ` Roger Pau Monné
2022-05-18 11:39     ` Jan Beulich
2022-05-20 10:41       ` Roger Pau Monné
2022-04-25  8:44 ` [PATCH v4 20/21] VT-d: fold iommu_flush_iotlb{,_pages}() Jan Beulich
2022-04-27  4:12   ` Tian, Kevin
2022-05-11 13:50   ` Roger Pau Monné
2022-04-25  8:45 ` [PATCH v4 21/21] VT-d: fold dma_pte_clear_one() into its only caller Jan Beulich
2022-04-27  4:13   ` Tian, Kevin
2022-05-11 13:57   ` Roger Pau Monné
2022-05-18 12:50 ` [PATCH v4 00/21] IOMMU: superpage support when not sharing pagetables Jan Beulich

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=16f5b398-56f6-a70c-9ce0-0ad72eab5058@suse.com \
    --to=jbeulich@suse.com \
    --cc=andrew.cooper3@citrix.com \
    --cc=paul@xen.org \
    --cc=roger.pau@citrix.com \
    --cc=xen-devel@lists.xenproject.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.