All of lore.kernel.org
 help / color / mirror / Atom feed
From: Paolo Bonzini <pbonzini@redhat.com>
To: qemu-devel@nongnu.org
Subject: [Qemu-devel] [PULL 02/15] target-i386: do not duplicate page protection checks
Date: Fri, 15 Jan 2016 17:04:17 +0100	[thread overview]
Message-ID: <1452873871-138914-3-git-send-email-pbonzini@redhat.com> (raw)
In-Reply-To: <1452873871-138914-1-git-send-email-pbonzini@redhat.com>

x86_cpu_handle_mmu_fault is currently checking twice for writability
and executability of pages; the first time to decide whether to
trigger a page fault, the second time to compute the "prot" argument
to tlb_set_page_with_attrs.

Reorganize code so that first "prot" is computed, then it is used
to check whether to raise a page fault, then finally PROT_WRITE is
removed if the D bit will have to be set.

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
 target-i386/helper.c | 65 +++++++++++++++++++---------------------------------
 1 file changed, 23 insertions(+), 42 deletions(-)

diff --git a/target-i386/helper.c b/target-i386/helper.c
index d18be95..bf58242 100644
--- a/target-i386/helper.c
+++ b/target-i386/helper.c
@@ -890,38 +890,30 @@ do_check_protect_pse36:
         goto do_fault_rsvd;
     }
     ptep ^= PG_NX_MASK;
-    if ((ptep & PG_NX_MASK) && is_write1 == 2) {
+
+    /* can the page can be put in the TLB?  prot will tell us */
+    if (is_user && !(ptep & PG_USER_MASK)) {
         goto do_fault_protect;
     }
-    switch (mmu_idx) {
-    case MMU_USER_IDX:
-        if (!(ptep & PG_USER_MASK)) {
-            goto do_fault_protect;
-        }
-        if (is_write && !(ptep & PG_RW_MASK)) {
-            goto do_fault_protect;
-        }
-        break;
 
-    case MMU_KSMAP_IDX:
-        if (is_write1 != 2 && (ptep & PG_USER_MASK)) {
-            goto do_fault_protect;
+    prot = 0;
+    if (mmu_idx != MMU_KSMAP_IDX || !(ptep & PG_USER_MASK)) {
+        prot |= PAGE_READ;
+        if ((ptep & PG_RW_MASK) || (!is_user && !(env->cr[0] & CR0_WP_MASK))) {
+            prot |= PAGE_WRITE;
         }
-        /* fall through */
-    case MMU_KNOSMAP_IDX:
-        if (is_write1 == 2 && (env->cr[4] & CR4_SMEP_MASK) &&
-            (ptep & PG_USER_MASK)) {
-            goto do_fault_protect;
-        }
-        if ((env->cr[0] & CR0_WP_MASK) &&
-            is_write && !(ptep & PG_RW_MASK)) {
-            goto do_fault_protect;
-        }
-        break;
+    }
+    if (!(ptep & PG_NX_MASK) &&
+        (mmu_idx == MMU_USER_IDX ||
+         !((env->cr[4] & CR4_SMEP_MASK) && (ptep & PG_USER_MASK)))) {
+        prot |= PAGE_EXEC;
+    }
 
-    default: /* cannot happen */
-        break;
+    if ((prot & (1 << is_write1)) == 0) {
+        goto do_fault_protect;
     }
+
+    /* yes, it can! */
     is_dirty = is_write && !(pte & PG_DIRTY_MASK);
     if (!(pte & PG_ACCESSED_MASK) || is_dirty) {
         pte |= PG_ACCESSED_MASK;
@@ -931,25 +923,13 @@ do_check_protect_pse36:
         x86_stl_phys_notdirty(cs, pte_addr, pte);
     }
 
-    /* the page can be put in the TLB */
-    prot = PAGE_READ;
-    if (!(ptep & PG_NX_MASK) &&
-        (mmu_idx == MMU_USER_IDX ||
-         !((env->cr[4] & CR4_SMEP_MASK) && (ptep & PG_USER_MASK)))) {
-        prot |= PAGE_EXEC;
-    }
-    if (pte & PG_DIRTY_MASK) {
+    if (!(pte & PG_DIRTY_MASK)) {
         /* only set write access if already dirty... otherwise wait
            for dirty access */
-        if (is_user) {
-            if (ptep & PG_RW_MASK)
-                prot |= PAGE_WRITE;
-        } else {
-            if (!(env->cr[0] & CR0_WP_MASK) ||
-                (ptep & PG_RW_MASK))
-                prot |= PAGE_WRITE;
-        }
+        assert(!is_write);
+        prot &= ~PROT_WRITE;
     }
+
  do_mapping:
     pte = pte & env->a20_mask;
 
@@ -962,6 +942,7 @@ do_check_protect_pse36:
     page_offset = vaddr & (page_size - 1);
     paddr = pte + page_offset;
 
+    assert(prot & (1 << is_write1));
     tlb_set_page_with_attrs(cs, vaddr, paddr, cpu_get_mem_attrs(env),
                             prot, mmu_idx, page_size);
     return 0;
-- 
1.8.3.1

  parent reply	other threads:[~2016-01-15 16:04 UTC|newest]

Thread overview: 30+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-01-15 16:04 [Qemu-devel] [PULL 00/15] NBD, chardev, SCSI patches for 2015-01-15 Paolo Bonzini
2016-01-15 16:04 ` [Qemu-devel] [PULL 01/15] scsi: revert change to scsi_req_cancel_async and add assertions Paolo Bonzini
2016-01-15 16:04 ` Paolo Bonzini [this message]
2016-01-15 16:04 ` [Qemu-devel] [PULL 03/15] i386: avoid null pointer dereference Paolo Bonzini
2016-01-15 16:53   ` Eric Blake
2016-01-15 17:09     ` Paolo Bonzini
2016-01-15 19:46       ` P J P
2016-01-15 19:48         ` Paolo Bonzini
2016-01-15 16:04 ` [Qemu-devel] [PULL] " Paolo Bonzini
2016-01-15 16:04 ` [Qemu-devel] [PULL 04/15] scsi: initialise info object with appropriate size Paolo Bonzini
2016-01-15 16:04 ` [Qemu-devel] [PULL 05/15] vmw_pvscsi: x-disable-pcie, x-old-pci-configuration back-compat props are 2.5 specific Paolo Bonzini
2016-01-15 16:04 ` [Qemu-devel] [PULL 06/15] qemu-char: delete send_all/recv_all helper methods Paolo Bonzini
2016-01-15 16:04 ` [Qemu-devel] [PULL 07/15] iscsi: send readcapacity10 when readcapacity16 failed Paolo Bonzini
2016-01-15 16:04 ` [Qemu-devel] [PULL 08/15] SCSI device: fix to incomplete QOMify Paolo Bonzini
2016-01-15 16:04 ` [Qemu-devel] [PULL 09/15] nbd: Always call "close_fn" in nbd_client_new Paolo Bonzini
2016-01-15 16:04 ` [Qemu-devel] [PULL 10/15] nbd: Split nbd.c Paolo Bonzini
2016-01-15 16:04 ` [Qemu-devel] [PULL 11/15] nbd-server: Coroutine based negotiation Paolo Bonzini
2016-01-15 16:04 ` [Qemu-devel] [PULL 12/15] nbd-server: do not check request length except for reads and writes Paolo Bonzini
2016-01-15 16:04 ` [Qemu-devel] [PULL 13/15] nbd-server: do not exit on failed memory allocation Paolo Bonzini
2016-01-15 16:04 ` [Qemu-devel] [PULL 14/15] qemu-char: add logfile facility to all chardev backends Paolo Bonzini
2016-01-21  6:16   ` Hervé Poussineau
2016-01-21  8:56     ` Daniel P. Berrange
2016-02-12 16:49       ` Markus Armbruster
2016-02-12 16:54         ` Daniel P. Berrange
2016-02-12 17:12           ` Markus Armbruster
2016-02-12 18:04             ` Daniel P. Berrange
2016-02-12 21:53               ` Daniel P. Berrange
2016-02-15  8:23                 ` Markus Armbruster
2016-01-15 16:04 ` [Qemu-devel] [PULL 15/15] qemu-char: do not leak QemuMutex when freeing a character device Paolo Bonzini
2016-01-15 17:42 ` [Qemu-devel] [PULL 00/15] NBD, chardev, SCSI patches for 2015-01-15 Peter Maydell

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=1452873871-138914-3-git-send-email-pbonzini@redhat.com \
    --to=pbonzini@redhat.com \
    --cc=qemu-devel@nongnu.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.