bitkeeper revision 1.153 (3e7a3a53LJF3EAuJITGpo9ybSgxf2w)

config.h, domain.c, dom_mem_ops.c, TODO:
  Fixed out of memory handling so we don't hang. Updated TODO file.

diff --git a/xen/TODO b/xen/TODO
index e81023e9952405f04fb27009b6d4b1fd02576ffc..41f0fca7d5fb2288347e6ffd16237122f055bc82 100644 (file)
--- a/xen/TODO
+++ b/xen/TODO
@@ -44,7 +44,15 @@ order to requests, just as we already do with block-device
 rings. We'll need to add an opaque identifier to ring entries,
 allowing matching of requests and responses, but that's about it.
 
-4. GDT AND LDT VIRTUALISATION 
+4. NETWORK CHECKSUM OFFLOAD 
+--------------------------- 
+All the NICs that we support can checksum packets on behalf of guest
+OSes. We need to add appropriate flags to and from each domain to
+indicate, on transmit, which packets need the checksum added and, on
+receive, which packets have been checked out as okay. We can steal
+Linux's interface, which is entirely sane given NIC limitations.
+
+5. GDT AND LDT VIRTUALISATION 
 ----------------------------- 
 We do not allow modification of the GDT, or any use of the LDT. This
 is necessary for support of unmodified applications (eg. Linux uses
@@ -55,14 +63,14 @@ I have some text on how to do this:
 /usr/groups/xeno/discussion-docs/memory_management/segment_tables.txt
 It's already half implemented, but the rest is still to do.
 
-5. DOMAIN 0 MANAGEMENT DAEMON
+6. DOMAIN 0 MANAGEMENT DAEMON
 -----------------------------
 A better control daemon is required for domain 0, which keeps proper
 track of machine resources and can make sensible policy choices. This
 may require support in Xen; for example, notifications (eg. DOMn is
 killed), and requests (eg. can DOMn allocate x frames of memory?).
 
-6. ACCURATE TIMERS AND WALL-CLOCK TIME
+7. ACCURATE TIMERS AND WALL-CLOCK TIME
 --------------------------------------
 Currently our long-term timebase free runs on CPU0, with no external
 calibration. We should run ntpd on domain 0 and allow this to warp
@@ -70,7 +78,7 @@ Xen's timebase. Once this is done, we can have a timebase per CPU and
 not worry about relative drift (since they'll all get sync'ed
 periodically by ntp).
 
-7. NEW DESIGN FEATURES
+8. NEW DESIGN FEATURES
 ----------------------
 This includes the last-chance page cache, and the unified buffer cache.
 

diff --git a/xen/common/dom_mem_ops.c b/xen/common/dom_mem_ops.c
index e919311c5c57b30f4eddf482a4d360e45ed8b444..07f7cf8f7d65d14674afb45da984b950e39e8b49 100644 (file)
--- a/xen/common/dom_mem_ops.c
+++ b/xen/common/dom_mem_ops.c
@@ -29,14 +29,24 @@ static long alloc_dom_mem(struct task_struct *p, balloon_def_op_t bop)
     unsigned long     i;
     unsigned long     flags;
 
-    /* POLICY DECISION: Each domain has a page limit. */
-    if( (p->tot_pages + bop.size) > p->max_pages )
+    /*
+     * POLICY DECISION: Each domain has a page limit.
+     * NB. The first part of test is because bop.size could be so big that
+     * tot_pages + bop.size overflows a u_long.
+     */
+    if( (bop.size > p->max_pages) ||
+        ((p->tot_pages + bop.size) > p->max_pages) )
         return -ENOMEM;
 
-    if ( free_pfns < bop.size ) 
-        return -ENOMEM;
-    
     spin_lock_irqsave(&free_list_lock, flags);
+
+    if ( free_pfns < (bop.size + (SLACK_DOMAIN_MEM_KILOBYTES << 
+                                  (PAGE_SHIFT-10))) ) 
+    {
+        spin_unlock_irqrestore(&free_list_lock, flags);
+        return -ENOMEM;
+    }
+
     spin_lock(&p->page_lock);
     
     temp = free_list.next;

diff --git a/xen/common/domain.c b/xen/common/domain.c
index 5fc4304c015a72a40bd9a9660a4c090c45bc1e57..d8ea1078a6274c7c9ed7af449e528aedca9d7eb2 100644 (file)
--- a/xen/common/domain.c
+++ b/xen/common/domain.c
@@ -161,8 +161,13 @@ unsigned int alloc_new_dom_mem(struct task_struct *p, unsigned int kbytes)
     spin_lock_irqsave(&free_list_lock, flags);
     
     /* is there enough mem to serve the request? */   
-    if ( req_pages > free_pfns ) return -1;
-    
+    if ( (req_pages + (SLACK_DOMAIN_MEM_KILOBYTES << (PAGE_SHIFT-10))) >
+         free_pfns )
+    {
+        spin_unlock_irqrestore(&free_list_lock, flags);
+        return -1;
+    }
+
     /* allocate pages and build a thread through frame_table */
     temp = free_list.next;
     for ( alloc_pfns = 0; alloc_pfns < req_pages; alloc_pfns++ )

diff --git a/xen/include/xeno/config.h b/xen/include/xeno/config.h
index 9dbceda88ea1bc91e8934f784c1c519cace6dbe2..0623d607db2895b955f6c6f2b58876f2daefc60d 100644 (file)
--- a/xen/include/xeno/config.h
+++ b/xen/include/xeno/config.h
@@ -85,6 +85,12 @@
 #define IOREMAP_VIRT_START    (MAPCACHE_VIRT_END)
 #define IOREMAP_VIRT_END      (IOREMAP_VIRT_START + (4*1024*1024))
 
+/*
+ * Amount of slack domain memory to leave in system, in megabytes.
+ * Prevents a hard out-of-memory crunch for thinsg like network receive.
+ */
+#define SLACK_DOMAIN_MEM_KILOBYTES 1024
+
 /* Linkage for x86 */
 #define FASTCALL(x)     x __attribute__((regparm(3)))
 #define asmlinkage        __attribute__((regparm(0)))