qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Aurelien Jarno <aurelien@aurel32.net>
To: qemu-devel@nongnu.org
Subject: [Qemu-devel] [5068] [sh4] MMU bug fix
Date: Fri, 22 Aug 2008 08:57:52 +0000	[thread overview]
Message-ID: <E1KWSSu-00057J-Hy@cvs.savannah.gnu.org> (raw)

Revision: 5068
          http://svn.sv.gnu.org/viewvc/?view=rev&root=qemu&revision=5068
Author:   aurel32
Date:     2008-08-22 08:57:52 +0000 (Fri, 22 Aug 2008)

Log Message:
-----------
[sh4] MMU bug fix

Some bugs on SH4 MMU are fixed.

- When a TLB entry is overwritten or invalidated, tlb_flush_page() should be
  invoked to invalidate old entry.
- When a ASID is changed, tlb_flush() should be invoke to invalidate entries
  which have old ASID.
- The check for shared bit in TLB entry causes multiple TLB hit exception.
  As SH3's MMU, shared bit is ignored.
- ASID is used when MMUCR's SV bit or SR's MD bit is zero.
  No need to check both bits are zero.

(Shin-ichiro KAWASAKI)

Modified Paths:
--------------
    trunk/hw/sh7750.c
    trunk/target-sh4/helper.c

Modified: trunk/hw/sh7750.c
===================================================================
--- trunk/hw/sh7750.c	2008-08-22 08:57:43 UTC (rev 5067)
+++ trunk/hw/sh7750.c	2008-08-22 08:57:52 UTC (rev 5068)
@@ -30,6 +30,7 @@
 #include "sh7750_regs.h"
 #include "sh7750_regnames.h"
 #include "sh_intc.h"
+#include "exec-all.h"
 #include "cpu.h"
 
 #define NB_DEVICES 4
@@ -356,6 +357,9 @@
 	s->cpu->mmucr = mem_value;
 	return;
     case SH7750_PTEH_A7:
+        /* If asid changes, clear all registered tlb entries. */
+	if ((s->cpu->pteh & 0xff) != (mem_value & 0xff))
+	    tlb_flush(s->cpu, 1);
 	s->cpu->pteh = mem_value;
 	return;
     case SH7750_PTEL_A7:

Modified: trunk/target-sh4/helper.c
===================================================================
--- trunk/target-sh4/helper.c	2008-08-22 08:57:43 UTC (rev 5067)
+++ trunk/target-sh4/helper.c	2008-08-22 08:57:52 UTC (rev 5068)
@@ -251,7 +251,7 @@
     for (i = 0; i < nbtlb; i++) {
 	if (!entries[i].v)
 	    continue;		/* Invalid entry */
-	if (use_asid && entries[i].asid != asid && !entries[i].sh)
+	if (use_asid && entries[i].asid != asid)
 	    continue;		/* Bad ASID */
 #if 0
 	switch (entries[i].sz) {
@@ -320,8 +320,14 @@
     else if (e == MMU_DTLB_MISS && update) {
 	e = find_tlb_entry(env, address, env->utlb, UTLB_SIZE, use_asid);
 	if (e >= 0) {
+	    tlb_t * ientry;
 	    n = itlb_replacement(env);
-	    env->itlb[n] = env->utlb[e];
+	    ientry = &env->itlb[n];
+	    if (ientry->v) {
+		if (!same_tlb_entry_exists(env->utlb, UTLB_SIZE, ientry))
+		    tlb_flush_page(env, ientry->vpn << 10);
+	    }
+	    *ientry = env->utlb[e];
 	    e = n;
 	} else if (e == MMU_DTLB_MISS)
 	    e = MMU_ITLB_MISS;
@@ -356,7 +362,7 @@
     int use_asid, is_code, n;
     tlb_t *matching = NULL;
 
-    use_asid = (env->mmucr & MMUCR_SV) == 0 && (env->sr & SR_MD) == 0;
+    use_asid = (env->mmucr & MMUCR_SV) == 0 || (env->sr & SR_MD) == 0;
     is_code = env->pc == address;	/* Hack */
 
     /* Use a hack to find if this is an instruction or data access */
@@ -540,6 +546,17 @@
     int n = cpu_mmucr_urc(env->mmucr);
     tlb_t * entry = &env->utlb[n];
 
+    if (entry->v) {
+        /* Overwriting valid entry in utlb. */
+        target_ulong address = entry->vpn << 10;
+	if (!same_tlb_entry_exists(env->itlb, ITLB_SIZE, entry)) {
+	    tlb_flush_page(env, address);
+	}
+    }
+
+    /* per utlb access cannot implemented. */
+    increment_urc(env);
+
     /* Take values into cpu status from registers. */
     entry->asid = (uint8_t)cpu_pteh_asid(env->pteh);
     entry->vpn  = cpu_pteh_vpn(env->pteh);

                 reply	other threads:[~2008-08-22  8:57 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

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=E1KWSSu-00057J-Hy@cvs.savannah.gnu.org \
    --to=aurelien@aurel32.net \
    --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 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).