WARNING - OLD ARCHIVES

This is an archived copy of the Xen.org mailing list, which we have preserved to ensure that existing links to archives are not broken. The live archive, which contains the latest emails, can be found at http://lists.xen.org/
   
 
 
Xen 
 
Home Products Support Community News
 
   
 

xen-devel

[Xen-devel] [PATCH] Mini-os memory fixes

To: xen-devel <xen-devel@xxxxxxxxxxxxxxxxxxx>, Keir Fraser <Keir.Fraser@xxxxxxxxxxxx>
Subject: [Xen-devel] [PATCH] Mini-os memory fixes
From: Grzegorz Milos <gm281@xxxxxxxxx>
Date: Thu, 25 Aug 2005 17:19:47 +0100
Delivery-date: Fri, 26 Aug 2005 08:57:40 +0000
Envelope-to: www-data@xxxxxxxxxxxxxxxxxxx
List-help: <mailto:xen-devel-request@lists.xensource.com?subject=help>
List-id: Xen developer discussion <xen-devel.lists.xensource.com>
List-post: <mailto:xen-devel@lists.xensource.com>
List-subscribe: <http://lists.xensource.com/cgi-bin/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/cgi-bin/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=unsubscribe>
Sender: xen-devel-bounces@xxxxxxxxxxxxxxxxxxx
User-agent: Mozilla Thunderbird 1.0.2 (X11/20050317)
Memory management fixes. Page tables are created, buddy allocator now has ability to free memory (as well as allocate it).

Gregor.
# HG changeset patch
# User gmilos@xxxxxxxxxxxxxxxxxxxxx
# Node ID 922710ba08102ad00197bd705d9b63dd1f3e5d9a
# Parent  531ad4bde8f249c318588f41bdbb72d0b0e0ea9c
Mini-os memory manegement fixed (the pagetables spanning all the allocated 
memory are now constructed on mm init). Buddy allocator includes free_pages.

Signed-off-by: Grzegorz Milos <gm281@xxxxxxxxx>

diff -r 531ad4bde8f2 -r 922710ba0810 extras/mini-os/include/hypervisor.h
--- a/extras/mini-os/include/hypervisor.h       Mon Aug 22 10:21:18 2005
+++ b/extras/mini-os/include/hypervisor.h       Thu Aug 25 16:16:14 2005
@@ -80,16 +80,42 @@
 
 static __inline__ int HYPERVISOR_mmu_update(mmu_update_t *req, 
                                             int count, 
-                                            int *success_count)
-{
-    int ret;
-    __asm__ __volatile__ (
-        TRAP_INSTR
-        : "=a" (ret) : "0" (__HYPERVISOR_mmu_update), 
-        _a1 (req), _a2 (count), _a3 (success_count)  : "memory" );
-
-    return ret;
-}
+                                            int *success_count, 
+                                            domid_t domid)
+{
+    int ret;
+    unsigned long ign1, ign2, ign3, ign4;
+
+    __asm__ __volatile__ (
+        TRAP_INSTR
+        : "=a" (ret), "=b" (ign1), "=c" (ign2), "=d" (ign3), "=S" (ign4)
+        : "0" (__HYPERVISOR_mmu_update), "1" (req), "2" (count),
+          "3" (success_count), "4" (domid)
+        : "memory" );
+
+    return ret;
+}
+
+
+static __inline__ int HYPERVISOR_mmuext_op(struct mmuext_op *op, 
+                                           int count, 
+                                           int *success_count, 
+                                           domid_t domid)
+{
+    int ret;
+    unsigned long ign1, ign2, ign3, ign4;
+
+    __asm__ __volatile__ (
+        TRAP_INSTR
+        : "=a" (ret), "=b" (ign1), "=c" (ign2), "=d" (ign3), "=S" (ign4)
+        : "0" (__HYPERVISOR_mmuext_op), "1" (op), "2" (count),
+          "3" (success_count), "4" (domid)
+        : "memory" );
+
+    return ret;
+}
+
+
 
 static __inline__ int HYPERVISOR_set_gdt(unsigned long *frame_list, int 
entries)
 {
diff -r 531ad4bde8f2 -r 922710ba0810 extras/mini-os/include/mm.h
--- a/extras/mini-os/include/mm.h       Mon Aug 22 10:21:18 2005
+++ b/extras/mini-os/include/mm.h       Thu Aug 25 16:16:14 2005
@@ -43,13 +43,27 @@
 #define PADDR_MASK              ((1UL << PADDR_BITS)-1)
 #define VADDR_MASK              ((1UL << VADDR_BITS)-1)
 
-#define pte_to_mfn(_pte) (((_pte) & (PADDR_MASK&PAGE_MASK)) >> PAGE_SHIFT)
+#define pte_to_mfn(_pte) (((_pte) & (PADDR_MASK&PAGE_MASK)) >> 
L1_PAGETABLE_SHIFT)
+
+#endif
+
+
+
+#ifdef __i386__
+
+#define L1_PAGETABLE_SHIFT      12
+#define L2_PAGETABLE_SHIFT      22
+
+#define L1_PAGETABLE_ENTRIES    1024
+#define L2_PAGETABLE_ENTRIES    1024
+#endif
 
 /* Given a virtual address, get an entry offset into a page table. */
 #define l1_table_offset(_a) \
   (((_a) >> L1_PAGETABLE_SHIFT) & (L1_PAGETABLE_ENTRIES - 1))
 #define l2_table_offset(_a) \
   (((_a) >> L2_PAGETABLE_SHIFT) & (L2_PAGETABLE_ENTRIES - 1))
+#ifdef __x86_64__
 #define l3_table_offset(_a) \
   (((_a) >> L3_PAGETABLE_SHIFT) & (L3_PAGETABLE_ENTRIES - 1))
 #define l4_table_offset(_a) \
@@ -67,13 +81,16 @@
 #define _PAGE_PSE      0x080UL
 #define _PAGE_GLOBAL   0x100UL
 
-#define PAGE_SHIFT      12
-#define PAGE_SIZE       (1UL << PAGE_SHIFT)
+#define L1_PROT (_PAGE_PRESENT | _PAGE_RW | _PAGE_ACCESSED)
+#define L2_PROT (_PAGE_PRESENT | _PAGE_RW | _PAGE_ACCESSED | _PAGE_DIRTY | 
_PAGE_USER)
+
+#define PAGE_SIZE       (1UL << L1_PAGETABLE_SHIFT)
+#define PAGE_SHIFT      L1_PAGETABLE_SHIFT
 #define PAGE_MASK       (~(PAGE_SIZE-1))
 
-#define PFN_UP(x)      (((x) + PAGE_SIZE-1) >> PAGE_SHIFT)
-#define PFN_DOWN(x)    ((x) >> PAGE_SHIFT)
-#define PFN_PHYS(x)    ((x) << PAGE_SHIFT)
+#define PFN_UP(x)      (((x) + PAGE_SIZE-1) >> L1_PAGETABLE_SHIFT)
+#define PFN_DOWN(x)    ((x) >> L1_PAGETABLE_SHIFT)
+#define PFN_PHYS(x)    ((x) << L1_PAGETABLE_SHIFT)
 
 /* to align the pointer to the (next) page boundary */
 #define PAGE_ALIGN(addr)        (((addr)+PAGE_SIZE-1)&PAGE_MASK)
@@ -83,14 +100,14 @@
 #define mfn_to_pfn(_mfn) (machine_to_phys_mapping[(_mfn)])
 static __inline__ unsigned long phys_to_machine(unsigned long phys)
 {
-    unsigned long machine = pfn_to_mfn(phys >> PAGE_SHIFT);
-    machine = (machine << PAGE_SHIFT) | (phys & ~PAGE_MASK);
+    unsigned long machine = pfn_to_mfn(phys >> L1_PAGETABLE_SHIFT);
+    machine = (machine << L1_PAGETABLE_SHIFT) | (phys & ~PAGE_MASK);
     return machine;
 }
 static __inline__ unsigned long machine_to_phys(unsigned long machine)
 {
-    unsigned long phys = mfn_to_pfn(machine >> PAGE_SHIFT);
-    phys = (phys << PAGE_SHIFT) | (machine & ~PAGE_MASK);
+    unsigned long phys = mfn_to_pfn(machine >> L1_PAGETABLE_SHIFT);
+    phys = (phys << L1_PAGETABLE_SHIFT) | (machine & ~PAGE_MASK);
     return phys;
 }
 
@@ -105,7 +122,10 @@
 #define __va to_virt
 #define __pa to_phys
 
+#define virt_to_pfn(_virt)         (PFN_DOWN(to_phys(_virt)))
+
 void init_mm(void);
 unsigned long alloc_pages(int order);
+int is_mfn_mapped(unsigned long mfn);
 
 #endif /* _MM_H_ */
diff -r 531ad4bde8f2 -r 922710ba0810 extras/mini-os/kernel.c
--- a/extras/mini-os/kernel.c   Mon Aug 22 10:21:18 2005
+++ b/extras/mini-os/kernel.c   Thu Aug 25 16:16:14 2005
@@ -133,7 +133,7 @@
     for ( ; ; ) 
     {      
 //        HYPERVISOR_yield();
-        block(1);
+        block(100);
         i++;
     }
 }
diff -r 531ad4bde8f2 -r 922710ba0810 extras/mini-os/mm.c
--- a/extras/mini-os/mm.c       Mon Aug 22 10:21:18 2005
+++ b/extras/mini-os/mm.c       Thu Aug 25 16:16:14 2005
@@ -5,9 +5,9 @@
  *
  *        File: mm.c
  *      Author: Rolf Neugebauer (neugebar@xxxxxxxxxxxxx)
- *     Changes: 
+ *     Changes: Grzegorz Milos
  *              
- *        Date: Aug 2003
+ *        Date: Aug 2003, chages Aug 2005
  * 
  * Environment: Xen Minimal OS
  * Description: memory management related functions
@@ -41,86 +41,18 @@
 #include <types.h>
 #include <lib.h>
 
+
+#ifdef MM_DEBUG
+#define DEBUG(_f, _a...) \
+    printk("MINI_OS(file=mm.c, line=%d) " _f "\n", __LINE__, ## _a)
+#else
+#define DEBUG(_f, _a...)    ((void)0)
+#endif
+
 unsigned long *phys_to_machine_mapping;
 extern char *stack;
 extern char _text, _etext, _edata, _end;
 
-static void init_page_allocator(unsigned long min, unsigned long max);
-
-void init_mm(void)
-{
-
-    unsigned long start_pfn, max_pfn, max_free_pfn;
-
-    unsigned long *pgd = (unsigned long *)start_info.pt_base;
-
-    printk("MM: Init\n");
-
-    printk("  _text:        %p\n", &_text);
-    printk("  _etext:       %p\n", &_etext);
-    printk("  _edata:       %p\n", &_edata);
-    printk("  stack start:  %p\n", &stack);
-    printk("  _end:         %p\n", &_end);
-
-    /* set up minimal memory infos */
-    start_pfn = PFN_UP(to_phys(&_end));
-    max_pfn = start_info.nr_pages;
-
-    printk("  start_pfn:    %lx\n", start_pfn);
-    printk("  max_pfn:      %lx\n", max_pfn);
-
-    /*
-     * we know where free tables start (start_pfn) and how many we 
-     * have (max_pfn). 
-     * 
-     * Currently the hypervisor stores page tables it providesin the
-     * high region of the this memory range.
-     * 
-     * next we work out how far down this goes (max_free_pfn)
-     * 
-     * XXX this assumes the hypervisor provided page tables to be in
-     * the upper region of our initial memory. I don't know if this 
-     * is always true.
-     */
-
-    max_free_pfn = PFN_DOWN(to_phys(pgd));
-#ifdef __i386__
-    {
-        unsigned long *pgd = (unsigned long *)start_info.pt_base;
-        unsigned long  pte;
-        int i;
-        printk("  pgd(pa(pgd)): %lx(%lx)", (u_long)pgd, to_phys(pgd));
-
-        for ( i = 0; i < (HYPERVISOR_VIRT_START>>22); i++ )
-        {
-            unsigned long pgde = *pgd++;
-            if ( !(pgde & 1) ) continue;
-            pte = machine_to_phys(pgde & PAGE_MASK);
-            printk("  PT(%x): %lx(%lx)", i, (u_long)to_virt(pte), pte);
-            if (PFN_DOWN(pte) <= max_free_pfn) 
-                max_free_pfn = PFN_DOWN(pte);
-        }
-    }
-    max_free_pfn--;
-    printk("  max_free_pfn: %lx\n", max_free_pfn);
-
-    /*
-     * now we can initialise the page allocator
-     */
-    printk("MM: Initialise page allocator for %lx(%lx)-%lx(%lx)\n",
-           (u_long)to_virt(PFN_PHYS(start_pfn)), PFN_PHYS(start_pfn), 
-           (u_long)to_virt(PFN_PHYS(max_free_pfn)), PFN_PHYS(max_free_pfn));
-    init_page_allocator(PFN_PHYS(start_pfn), PFN_PHYS(max_free_pfn));   
-#endif
-
-
-    /* Now initialise the physical->machine mapping table. */
-
-
-    printk("MM: done\n");
-
-    
-}
 
 /*********************
  * ALLOCATION BITMAP
@@ -213,6 +145,59 @@
 
 #define round_pgdown(_p)  ((_p)&PAGE_MASK)
 #define round_pgup(_p)    (((_p)+(PAGE_SIZE-1))&PAGE_MASK)
+
+#ifdef MM_DEBUG
+/*
+ * Prints allocation[0/1] for @nr_pages, starting at @start
+ * address (virtual).
+ */
+static void print_allocation(void *start, int nr_pages)
+{
+    unsigned long pfn_start = virt_to_pfn(start);
+    int count;
+    for(count = 0; count < nr_pages; count++)
+        if(allocated_in_map(pfn_start + count)) printk("1");
+        else printk("0");
+        
+    printk("\n");        
+}
+
+/*
+ * Prints chunks (making them with letters) for @nr_pages starting
+ * at @start (virtual).
+ */
+static void print_chunks(void *start, int nr_pages)
+{
+    char chunks[1001], current='A';
+    int order, count;
+    chunk_head_t *head;
+    unsigned long pfn_start = virt_to_pfn(start);
+   
+    memset(chunks, (int)'_', 1000);
+    if(nr_pages > 1000) 
+    {
+        DEBUG("Can only pring 1000 pages. Increase buffer size.");
+    }
+    
+    for(order=0; order < FREELIST_SIZE; order++)
+    {
+        head = free_head[order];
+        while(!FREELIST_EMPTY(head))
+        {
+            for(count = 0; count < 1<< head->level; count++)
+            {
+                if(count + virt_to_pfn(head) - pfn_start < 1000)
+                    chunks[count + virt_to_pfn(head) - pfn_start] = current;
+            }
+            head = head->next;
+            current++;
+        }
+    }
+    chunks[nr_pages] = '\0';
+    printk("%s\n", chunks);
+}
+#endif
+
 
 
 /*
@@ -328,3 +313,198 @@
     return 0;
 }
 
+void free_pages(void *pointer, int order)
+{
+    chunk_head_t *freed_ch, *to_merge_ch;
+    chunk_tail_t *freed_ct;
+    unsigned long mask;
+    
+    /* First free the chunk */
+    map_free(virt_to_pfn(pointer), 1 << order);
+    
+    /* Create free chunk */
+    freed_ch = (chunk_head_t *)pointer;
+    freed_ct = (chunk_tail_t *)((char *)pointer + (1<<(order + PAGE_SHIFT)))-1;
+    
+    /* Now, possibly we can conseal chunks together */
+    while(order < FREELIST_SIZE)
+    {
+        mask = 1 << (order + PAGE_SHIFT);
+        if((unsigned long)freed_ch & mask) 
+        {
+            to_merge_ch = (chunk_head_t *)((char *)freed_ch - mask);
+            if(allocated_in_map(virt_to_pfn(to_merge_ch)) ||
+                    to_merge_ch->level != order)
+                break;
+            
+            /* Merge with predecessor */
+            freed_ch = to_merge_ch;   
+        }
+        else 
+        {
+            to_merge_ch = (chunk_head_t *)((char *)freed_ch + mask);
+            if(allocated_in_map(virt_to_pfn(to_merge_ch)) ||
+                    to_merge_ch->level != order)
+                break;
+            
+            /* Merge with successor */
+            freed_ct = (chunk_tail_t *)((char *)to_merge_ch + mask);
+        }
+        
+        /* We are commited to merging, unlink the chunk */
+        *(to_merge_ch->pprev) = to_merge_ch->next;
+        to_merge_ch->next->pprev = to_merge_ch->pprev;
+        
+        order++;
+    }
+
+    /* Link the new chunk */
+    freed_ch->level = order;
+    freed_ch->next  = free_head[order];
+    freed_ch->pprev = &free_head[order];
+    freed_ct->level = order;
+    
+    freed_ch->next->pprev = &freed_ch->next;
+    free_head[order] = freed_ch;   
+   
+}
+void build_pagetable(unsigned long *start_pfn, unsigned long *max_pfn)
+{
+    unsigned long pfn_to_map, pt_frame;
+    unsigned long mach_ptd, max_mach_ptd;
+    int count;
+    unsigned long mach_pte, virt_pte;
+    unsigned long *ptd = (unsigned long *)start_info.pt_base;
+    mmu_update_t mmu_updates[L1_PAGETABLE_ENTRIES + 1];
+    struct mmuext_op pin_request;
+    
+    /* Firstly work out what is the first pfn that is not yet in page tables
+       NB. Assuming that builder fills whole pt_frames (which it does at the
+       moment)
+     */  
+    pfn_to_map = (start_info.nr_pt_frames - 1) * L1_PAGETABLE_ENTRIES;
+    DEBUG("start_pfn=%ld, first pfn_to_map %ld, max_pfn=%ld", 
+            *start_pfn, pfn_to_map, *max_pfn);
+
+    /* Machine address of page table directory */
+    mach_ptd = phys_to_machine(to_phys(start_info.pt_base));
+    mach_ptd += sizeof(void *) * 
+        l2_table_offset((unsigned long)to_virt(PFN_PHYS(pfn_to_map)));
+  
+    max_mach_ptd = sizeof(void *) * 
+        l2_table_offset((unsigned long)to_virt(PFN_PHYS(*max_pfn)));
+    
+    /* Check that we are not trying to access Xen region */
+    if(max_mach_ptd > sizeof(void *) * l2_table_offset(HYPERVISOR_VIRT_START))
+    {
+        printk("WARNING: mini-os will not use all the memory supplied\n");
+        max_mach_ptd = sizeof(void *) * l2_table_offset(HYPERVISOR_VIRT_START);
+        *max_pfn = virt_to_pfn(HYPERVISOR_VIRT_START - PAGE_SIZE);
+    }
+    max_mach_ptd += phys_to_machine(to_phys(start_info.pt_base));
+    DEBUG("Max_mach_ptd 0x%lx", max_mach_ptd); 
+   
+    pt_frame = *start_pfn;
+    /* Should not happen - no empty, mapped pages */
+    if(pt_frame >= pfn_to_map)
+    {
+        printk("ERROR: Not even a single empty, mapped page\n");
+        *(int*)0=0;
+    }
+    
+    while(mach_ptd < max_mach_ptd)
+    {
+        /* Correct protection needs to be set for the new page table frame */
+        virt_pte = (unsigned long)to_virt(PFN_PHYS(pt_frame));
+        mach_pte = ptd[l2_table_offset(virt_pte)] & ~(PAGE_SIZE-1);
+        mach_pte += sizeof(void *) * l1_table_offset(virt_pte);
+        DEBUG("New page table page: pfn=0x%lx, mfn=0x%lx, virt_pte=0x%lx, "
+                "mach_pte=0x%lx", pt_frame, pfn_to_mfn(pt_frame), 
+                virt_pte, mach_pte);
+        
+        /* Update the entry */
+        mmu_updates[0].ptr = mach_pte;
+        mmu_updates[0].val = pfn_to_mfn(pt_frame) << PAGE_SHIFT | 
+                                                    (L1_PROT & ~_PAGE_RW);
+        if(HYPERVISOR_mmu_update(mmu_updates, 1, NULL, DOMID_SELF) < 0)
+        {
+            printk("PTE for new page table page could not be updated\n");
+            *(int*)0=0;
+        }
+        
+        /* Pin the page to provide correct protection */
+        pin_request.cmd = MMUEXT_PIN_L1_TABLE;
+        pin_request.mfn = pfn_to_mfn(pt_frame);
+        if(HYPERVISOR_mmuext_op(&pin_request, 1, NULL, DOMID_SELF) < 0)
+        {
+            printk("ERROR: pinning failed\n");
+            *(int*)0=0;
+        }
+        
+        /* Now fill the new page table page with entries.
+           Update the page directory as well. */
+        count = 0;
+        mmu_updates[count].ptr = mach_ptd;
+        mmu_updates[count].val = pfn_to_mfn(pt_frame) << PAGE_SHIFT |
+                                                         L2_PROT;
+        count++;
+        mach_ptd += sizeof(void *);
+        mach_pte = phys_to_machine(PFN_PHYS(pt_frame++));
+        
+        for(;count <= L1_PAGETABLE_ENTRIES && pfn_to_map <= *max_pfn; count++)
+        {
+            mmu_updates[count].ptr = mach_pte;
+            mmu_updates[count].val = 
+                pfn_to_mfn(pfn_to_map++) << PAGE_SHIFT | L1_PROT;
+            if(count == 1) DEBUG("mach_pte 0x%lx", mach_pte);
+            mach_pte += sizeof(void *);
+        }
+        if(HYPERVISOR_mmu_update(mmu_updates, count, NULL, DOMID_SELF) < 0) 
+        {            
+            printk("ERROR: mmu_update failed\n");
+            *(int*)0=0;
+        }
+        (*start_pfn)++;
+    }
+
+    *start_pfn = pt_frame;
+}
+
+void init_mm(void)
+{
+
+    unsigned long start_pfn, max_pfn;
+
+    printk("MM: Init\n");
+
+    printk("  _text:        %p\n", &_text);
+    printk("  _etext:       %p\n", &_etext);
+    printk("  _edata:       %p\n", &_edata);
+    printk("  stack start:  %p\n", &stack);
+    printk("  _end:         %p\n", &_end);
+
+    /* set up minimal memory infos */
+    phys_to_machine_mapping = (unsigned long *)start_info.mfn_list;
+   
+    /* First page follows page table pages and 3 more pages (store page etc) */
+    start_pfn = PFN_UP(__pa(start_info.pt_base)) + start_info.nr_pt_frames + 3;
+    max_pfn = start_info.nr_pages;
+
+    printk("  start_pfn:    %lx\n", start_pfn);
+    printk("  max_pfn:      %lx\n", max_pfn);
+
+
+    build_pagetable(&start_pfn, &max_pfn);
+    
+#ifdef __i386__
+    /*
+     * now we can initialise the page allocator
+     */
+    printk("MM: Initialise page allocator for %lx(%lx)-%lx(%lx)\n",
+           (u_long)to_virt(PFN_PHYS(start_pfn)), PFN_PHYS(start_pfn), 
+           (u_long)to_virt(PFN_PHYS(max_pfn)), PFN_PHYS(max_pfn));
+    init_page_allocator(PFN_PHYS(start_pfn), PFN_PHYS(max_pfn));   
+#endif
+
+    printk("MM: done\n");
+}
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel
<Prev in Thread] Current Thread [Next in Thread>
  • [Xen-devel] [PATCH] Mini-os memory fixes, Grzegorz Milos <=