Re: [Minios-devel] [PATCH v3 00/19] mini-os: support of auto-ballooning

[Wei Liu <wei.liu2@citrix.com>]

On Fri, Aug 12, 2016 at 12:42:31PM +0100, Wei Liu wrote:
> On Thu, Aug 11, 2016 at 11:18:03AM +0200, Juergen Gross wrote:
> > Support ballooning Mini-OS automatically up in case of memory shortage.
> > 
> > Do some cleanups, a small correction and add some basic features to
> > lay groundwork for support of ballooning in Mini-OS (patches 1-14).
> > 
> > The main visible change is the virtual memory layout: to be able to
> > add memory to the running Mini-OS we need to have some spare areas
> > especially after the 1:1 mapping of physical memory.
> > 
> > Then add the ballooning functionality: the p2m map must be expanded,
> > the page allocator's bitmap must  be expanded and we must get new
> > memory from the hypervisor.
> > 
> > In case of a detected memory shortage the domain will balloon up until
> > either enough memory is available or the upper limit has been reached.
> > 
> > Ballooning has been tested with a xenstore stubdom.
> > Regression tests have been done with:
> > - pure mini-os
> > - ioemu stubdom
> > - pvgrub 64 bit
> > 
> 
> Unfortunately vtmpmgr stubdom build failed. :-/
> 


And that's probably because vtpmmgr uses printk or whatever...

Here is a patch that seems to fix the build:

---8<---
From d72510368cdc3c73af3c8918a404a8137f40bd9c Mon Sep 17 00:00:00 2001
From: Wei Liu <wei.liu2@citrix.com>
Date: Fri, 12 Aug 2016 11:32:57 +0100
Subject: [PATCH] x86/arch_mm.h: move p2m_chk_pfn to x86/mm.c

Making that function inlined won't buy us much and is causing error in
vtpm manager stubdom build.

Signed-off-by: Wei Liu <wei.liu2@citrix.com>
---
 arch/x86/mm.c         |  9 +++++++++
 include/x86/arch_mm.h | 11 +++--------
 2 files changed, 12 insertions(+), 8 deletions(-)

diff --git a/arch/x86/mm.c b/arch/x86/mm.c
index 8fa3b4c..88a928d 100644
--- a/arch/x86/mm.c
+++ b/arch/x86/mm.c
@@ -608,6 +608,15 @@ static void clear_bootstrap(void)
         printk("Unable to unmap NULL page. rc=%d\n", rc);
 }
 
+void p2m_chk_pfn(unsigned long pfn)
+{
+    if ( (pfn >> L3_P2M_SHIFT) > 0 )
+    {
+        printk("Error: Too many pfns.\n");
+        do_exit();
+    }
+}
+
 void arch_init_p2m(unsigned long max_pfn)
 {
     unsigned long *l2_list = NULL, *l3_list;
diff --git a/include/x86/arch_mm.h b/include/x86/arch_mm.h
index e5d9c57..690a919 100644
--- a/include/x86/arch_mm.h
+++ b/include/x86/arch_mm.h
@@ -190,14 +190,9 @@ typedef unsigned long pgentry_t;
 #define L2_P2M_IDX(pfn) (((pfn) >> L1_P2M_SHIFT) & P2M_MASK)
 #define L3_P2M_IDX(pfn) (((pfn) >> L2_P2M_SHIFT) & P2M_MASK)
 #define INVALID_P2M_ENTRY (~0UL)
-static inline void p2m_chk_pfn(unsigned long pfn)
-{
-    if ( (pfn >> L3_P2M_SHIFT) > 0 )
-    {
-        printk("Error: Too many pfns.\n");
-        do_exit();
-    }
-}
+
+void p2m_chk_pfn(unsigned long pfn);
+
 static inline unsigned long p2m_pages(unsigned long pages)
 {
     return (pages + P2M_ENTRIES - 1) >> L1_P2M_SHIFT;
-- 
2.1.4


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel