Skip to contentSkip to conversation

powerpc/cell: make ptcal more reliable

PATCH In review [email protected] 3 participants 3 messages May 4, 2009 – May 5, 2009

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 powerpc/cell: make ptcal more reliable

This is for QS21. The following patch allocates pages only from the specified node, moves the ptcal area into the middle of the allocated page to avoid potential prefetch problems and prints the address of the ptcal area to facilitate diagnostics.

arch/powerpc/platforms/cell/ras.c +13 -3
Index: linux-2.6/arch/powerpc/platforms/cell/ras.c
===================================================================
@@ -122,12 +122,22 @@ static int __init cbe_ptcal_enable_on_no
 
 	area->nid = nid;
 	area->order = order;
-	area->pages = alloc_pages_node(area->nid, GFP_KERNEL, area->order);
+	area->pages = alloc_pages_node(area->nid, GFP_KERNEL | GFP_THISNODE, area->order);
 
-	if (!area->pages)
+	if (!area->pages) {
+		printk(KERN_INFO "%s: no page on node %d\n",
+			__FUNCTION__, area->nid);
 		goto out_free_area;
+	}
 
-	addr = __pa(page_address(area->pages));
+	/*
+	 * We move the ptcal area to the middle of the allocated
+	 * page, in order to avoid prefetches in memcpy and similar
+	 * functions stepping on it.
+	 */
+	addr = __pa(page_address(area->pages)) + (PAGE_SIZE >> 1);
+	printk(KERN_INFO "%s: enabling PTCAL on node %d address=0x%016lx PAGE_SIZE>>1=0x%016lx \n",
+			__FUNCTION__, area->nid, addr, PAGE_SIZE>>1);
 
 	ret = -EIO;
 	if (rtas_call(ptcal_start_tok, 3, 1, NULL, area->nid,

===================================================================

Best regards,

Gerhard Stenzel, Linux on Cell Development, LTC
-------------------------------------------------------------------------------------
IBM Deutschland Entwicklung GmbH
Vorsitzender des Aufsichtsrats: Martin Jetter | Geschaeftsfuehrung: Erich Baier Sitz der Gesellschaft: Boeblingen | Registergericht: Amtsgericht Stuttgart, HRB 243294

#1 lore raw

Hi Gerhard,

Quoted 4 lines

This is for QS21. The following patch allocates pages only from the specified node, moves the ptcal area into the middle of the allocated page to avoid potential prefetch problems and prints the address of the ptcal area to facilitate diagnostics.

You're seeing prefetches that cross a page boundary?

Quoted 3 lines
(quoted diff) +1 -1
-	area->pages = alloc_pages_node(area->nid, GFP_KERNEL, area->order);
+	area->pages = alloc_pages_node(area->nid, GFP_KERNEL |

GFP_THISNODE, area->order);

Best to keep this under 80 cols.

arch/powerpc/platforms/cell/ras.c static int __init cbe_ptcal_enable_on_no · line 131

Quoted 7 lines
arch/powerpc/platforms/cell/ras.c +4 -1
-	if (!area->pages)
+	if (!area->pages) {
+		printk(KERN_INFO "%s: no page on node %d\n",
+			__FUNCTION__, area->nid);
 		goto out_free_area;
+	}

That could probably be a KERN_ERR, as we don't have ptcal enabled on that node. Also, I believe __func__ is preferred over __FUNCTION__.

arch/powerpc/platforms/cell/ras.c static int __init cbe_ptcal_enable_on_no · line 138

Quoted 7 lines
arch/powerpc/platforms/cell/ras.c +6 -1
-	addr = __pa(page_address(area->pages));
+	/*
+	 * We move the ptcal area to the middle of the allocated
+	 * page, in order to avoid prefetches in memcpy and similar
+	 * functions stepping on it.
+	 */
+	addr = __pa(page_address(area->pages)) + (PAGE_SIZE >> 1);

Minor nitpick, but I think (PAGE_SIZE / 2) better illustrates that you're putting the addr in the middle of the page. But either should be fine.

arch/powerpc/platforms/cell/ras.c static int __init cbe_ptcal_enable_on_no · line 140

Quoted 3 lines
(quoted diff) +1 -0
+	printk(KERN_INFO "%s: enabling PTCAL on node %d address=0x%016lx

PAGE_SIZE>>1=0x%016lx \n",

arch/powerpc/platforms/cell/ras.c +1 -0
+			__FUNCTION__, area->nid, addr, PAGE_SIZE>>1);

80 cols again. Can we do this as a pr_debug?

Cheers,

Jeremy

#2 lore raw

Jeremy Kerr <[email protected]> wrote on 05/05/2009 02:57:18 AM:

Quoted 1 line

Hi Gerhard,

Jeremy, thanks for your comments

Quoted 9 lines
Quoted 5 lines

This is for QS21. The following patch allocates pages only from the specified node, moves the ptcal area into the middle of the allocated page to avoid potential prefetch problems and prints the address of the ptcal area to facilitate diagnostics.

You're seeing prefetches that cross a page boundary?

Quoted 1 line

- area->pages =3D alloc_pages_node(area->nid, GFP_KERNEL, area->o=

rder);

Quoted 4 lines
Quoted 2 lines

+ area->pages =3D alloc_pages_node(area->nid, GFP_KERNEL |
GFP_THISNODE, area->order);

Best to keep this under 80 cols.

ok

Quoted 11 lines

arch/powerpc/platforms/cell/ras.c static int __init cbe_ptcal_enable_on_no · line 131

Quoted 8 lines
arch/powerpc/platforms/cell/ras.c +4 -1
-   if (!area->pages)
+   if (!area->pages) {
+      printk(KERN_INFO "%s: no page on node %d\n",
+         __FUNCTION__, area->nid);
       goto out_free_area;
+   }

That could probably be a KERN_ERR, as we don't have ptcal enabled on that node. Also, I believe __func__ is preferred over __FUNCTION__.

Since this is the default for the kdump kernel in most cases, how about=

KERN_WARNING instead.:

Quoted 11 lines

arch/powerpc/platforms/cell/ras.c static int __init cbe_ptcal_enable_on_no

Quoted 8 lines
arch/powerpc/platforms/cell/ras.c +6 -1
-   addr =3D __pa(page_address(area->pages));
+   /*
+    * We move the ptcal area to the middle of the allocated
+    * page, in order to avoid prefetches in memcpy and similar
+    * functions stepping on it.
+    */
+   addr =3D __pa(page_address(area->pages)) + (PAGE_SIZE >> 1);

Minor nitpick, but I think (PAGE_SIZE / 2) better illustrates that you're putting the addr in the middle of the page. But either should =

be

Quoted 1 line

fine.

ok. I prefer to keep it like this. It should be clear from the comment.=

Quoted 2 lines
Quoted 2 lines

+ printk(KERN_INFO "%s: enabling PTCAL on node %d address=3D0x%01=

6lx

Quoted 4 lines
Quoted 2 lines

PAGE_SIZE>>1=3D0x%016lx \n",
+ __FUNCTION__, area->nid, addr, PAGE_SIZE>>1);

80 cols again. Can we do this as a pr_debug?

"PAGE_SIZE>>1" can be omitted anyway. Removing it brings it below 80 co= ls.
Regarding pr_debug, the message should appear only on QS21 where it can=
be
of interest. So, I would prefer to keep a printk.

Quoted 4 lines

Cheers,

Jeremy

Thanks again,

Best regards,

Gerhard Stenzel, Linux on Cell/Hybrid Technologies, LTC
-----------------------------------------------------------------------=
------------------------------------------------------------

IBM Deutschland Research & Development GmbH
Vorsitzender des Aufsichtsrats: Martin Jetter | Gesch=E4ftsf=FChrung: E= rich
Baier
Sitz der Gesellschaft: B=F6blingen | Registergericht: Amtsgericht Stutt= gart,
HRB 243294=