MCA recovery verify pfn_valid
PATCH 1/1 MCA recovery verify pfn_valid
Verify the pfn is valid before calling pfn_to_page().
---------------------------------------------------------------------
arch/ia64/kernel/mca_drv.c | 3 +++ 1 files changed, 3 insertions(+)
Index: linux-2.6.git#test/arch/ia64/kernel/mca_drv.c
=================================--- linux-2.6.git#test.orig/arch/ia64/kernel/mca_drv.c 2005-09-16 16:28:11.857773257 -0500
+++ linux-2.6.git#test/arch/ia64/kernel/mca_drv.c 2005-09-16 16:29:32.119181669 -0500
@@ -87,6 +87,9 @@ mca_page_isolate(unsigned long paddr)
if ( !ia64_phys_addr_valid(paddr) )
return ISOLATE_NG;
+ if ( !pfn_valid(paddr))
+ return ISOLATE_OK;
+
/* convert physical address to physical page number */ p = pfn_to_page(paddr>>PAGE_SHIFT);
Quoted 22 lines · Jack Steiner
Quoted 14 lines · Russ Anderson
Verify the pfn is valid before calling
pfn_to_page().Signed-off-by Russ Anderson ([email protected])---------------------------------------------------------------------
arch/ia64/kernel/mca_drv.c | 3 +++ 1 files changed, 3 insertions(+)Index: linux-2.6.git#test/arch/ia64/kernel/mca_drv.c
=================================> > --- linux-2.6.git#test.orig/arch/ia64/kernel/mca_drv.c 2005-09-16 16:28:11.857773257 -0500
+++ linux-2.6.git#test/arch/ia64/kernel/mca_drv.c 2005-09-16 16:29:32.119181669 -0500Review mca_page_isolate(unsigned long paddr) · line 91
Quoted 5 lines
(quoted diff) +2
-0@@ -87,6 +87,9 @@ mca_page_isolate(unsigned long paddr) if ( !ia64_phys_addr_valid(paddr) ) return ISOLATE_NG; + if ( !pfn_valid(paddr)) + return ISOLATE_OK;Should this be
ISOLATE_OKorISOLATE_NG?
Quoted 5 lines
If the page is from another partition, the page has not been isolated? I realize that the return status is not used for anything but printing a message, but it seems misleading to print that isolation was successful when nothing was actually done.
Perhaps another status should be returned.
In this case, is any message really needed? As you say, success is is misleading because nothing was done. Failure is not accurate because it didn't fail. So perhaps a third status that does not print any page isolation message is the best solution.
For reference, the code:
----------------------------------------------------------
voidmca_handler_bh(unsigned long paddr)
{
printk(KERN_DEBUG "OS_MCA: process [pid: %d](%s) encounters MCA.\n",
current->pid, current->comm);
spin_lock(&mca_bh_lock);
if (mca_page_isolate(paddr) = ISOLATE_OK) {
printk(KERN_DEBUG "Page isolation: ( %lx ) success.\n", paddr);
} else {
printk(KERN_DEBUG "Page isolation: ( %lx ) failure.\n", paddr);
}
spin_unlock(&mca_bh_lock);
/* This process is about to be killed itself */
do_exit(SIGKILL);
}Signature · 2 lines
Russ Anderson, OS RAS/Partitioning Project Lead SGI - Silicon Graphics Inc [email protected]
How about this?
Thanks,
H.Seto
Verify the pfn is valid before calling pfn_to_page(), and cut isolation message if nothing was done.
Quoted 8 lines · Russ Anderson
Quoted 2 lines · Jack Steiner
Perhaps another status should be returned.
In this case, is any message really needed? As you say, success is is misleading because nothing was done. Failure is not accurate because it didn't fail. So perhaps a third status that does not print any page isolation message is the best solution.
arch/ia64/kernel/mca_drv.c | 21 +++++++++++++++------ 1 files changed, 15 insertions(+), 6 deletions(-)
arch/ia64/kernel/mca_drv.c +15 -6
@@ -56,8 +56,9 @@ static struct page *page_isolate[MAX_PAGstatic int num_page_isolate =0 ;typedef enum { -ISOLATE_NG =0 , -ISOLATE_OK =1 +ISOLATE_NG , +ISOLATE_OK , +ISOLATE_NONE }isolate_status_t ;/*
@@ -74,7 +75,7 @@ static struct {* @paddr: poisoned memory location * * Return value: -* ISOLATE_OK / ISOLATE_NG +* one of isolate_status_t, ISOLATE_OK/NG/NONE. */ static isolate_status_t
@@ -85,7 +86,10 @@ mca_page_isolate(unsigned long paddr)/* whether physical address is valid or not */ if (!ia64_phys_addr_valid (paddr)) -return ISOLATE_NG ; +return ISOLATE_NONE ; + +if (!pfn_valid (paddr)) +return ISOLATE_NONE ;/* convert physical address to physical page number */ p =pfn_to_page (paddr>>PAGE_SHIFT );
@@ -122,10 +126,15 @@ mca_handler_bh(unsigned long paddr) current->pid, current->comm);spin_lock (&mca_bh_lock); -if (mca_page_isolate (paddr) ==ISOLATE_OK ) { +switch (mca_page_isolate (paddr)) { +case ISOLATE_OK :printk (KERN_DEBUG "Page isolation: ( %lx ) success.\n" , paddr); - }else { +break ; +case ISOLATE_NG :printk (KERN_DEBUG "Page isolation: ( %lx ) failure.\n" , paddr); +break ; +default : +break ; }spin_unlock (&mca_bh_lock);
Quoted 10 lines · Hidetoshi Seto
How about this?
Thanks,
H.SetoVerify the pfn is valid before calling
pfn_to_page(), and cut isolation message if nothing was done.Signed-off-by Hidetoshi Seto <[email protected]>
Looks good to me.
Signature · 2 lines
Russ Anderson, OS RAS/Partitioning Project Lead SGI - Silicon Graphics Inc [email protected]