Skip to contentSkip to conversation

x86, pat: fix free_memtype() for legacy 1MB range

PATCH In review Suresh Siddha 2 participants 2 messages Jan 14, 2009

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

PATCH 1/1 x86, pat: fix free_memtype() for legacy 1MB range

commit 0e95c190cf1b4adbf1fc67decff3f2bb20158ba8 in tip tree is n't complete, as we left out adding similar logic to free_memtype().

Appended is the delta patch against tip.

thanks,
suresh
---

[patch] x86, pat: fix free_memtype() for legacy 1MB range

Starting this commit:

Quoted 5 lines
	commit 0e95c190cf1b4adbf1fc67decff3f2bb20158ba8
	Author: Suresh Siddha <[email protected]>
	Date:   Fri Jan 9 14:35:20 2009 -0800

	    x86, pat: fix reserve_memtype() for legacy 1MB range

Attribute requests with in this 1MB region are tracked in reserve_memtype() using linear linked list (and don't bother if the range is RAM or non-RAM or mixed).

free_memtype() also needs to follow similar logic.

---

arch/x86/mm/pat.c +13 -5
index ec8cd49..3be3990 100644
@@ -446,11 +446,19 @@ int free_memtype(u64 start, u64 end)
 	if (is_ISA_range(start, end - 1))
 		return 0;
 
-	is_range_ram = pagerange_is_ram(start, end);
-	if (is_range_ram == 1)
-		return free_ram_pages_type(start, end);
-	else if (is_range_ram < 0)
-		return -EINVAL;
+	/*
+	 * For legacy reasons, some parts of the physical address range in the
+	 * legacy 1MB region is treated as non-RAM (even when listed as RAM in
+	 * the e820 tables).  So we will track the memory attributes of this
+	 * legacy 1MB region using the linear memtype_list always.
+	 */
+	if (end >= ISA_END_ADDRESS) {
+		is_range_ram = pagerange_is_ram(start, end);
+		if (is_range_ram == 1)
+			return free_ram_pages_type(start, end);
+		else if (is_range_ram < 0)
+			return -EINVAL;
+	}
 
 	spin_lock(&memtype_lock);
 	list_for_each_entry(entry, &memtype_list, nd) {
#1 lore raw
Quoted 8 lines · Suresh Siddha
arch/x86/mm/pat.c +3 -5
-	is_range_ram = pagerange_is_ram(start, end);
-	if (is_range_ram == 1)
-		return free_ram_pages_type(start, end);
-	else if (is_range_ram < 0)
-		return -EINVAL;
+	/*
+	 * For legacy reasons, some parts of the physical address range in the
+	 * legacy 1MB region is treated as non-RAM (even when listed as RAM in

arch/x86/mm/pat.c int free_memtype(u64 start, u64 end) · line 461

Quoted 10 lines
arch/x86/mm/pat.c +10 -0
+	 * the e820 tables).  So we will track the memory attributes of this
+	 * legacy 1MB region using the linear memtype_list always.
+	 */
+	if (end >= ISA_END_ADDRESS) {
+		is_range_ram = pagerange_is_ram(start, end);
+		if (is_range_ram == 1)
+			return free_ram_pages_type(start, end);
+		else if (is_range_ram < 0)
+			return -EINVAL;
+	}

this logic is now repeated in an open-coded way in two places in pat.c. Could you please send another delta patch that consolidates them into a single helper function, which is used both in free_memtype() and reserve_memtype()?

Thanks,

Ingo