Skip to contentSkip to conversation

SGI UV: TLB shootdown using broadcast assist unit - v6 delta

PATCH Applied Cliff Wickman 2 participants 2 messages Jun 19, 2008 – Jun 20, 2008 x86

lore
Share & tools
lore
#0 lore raw

PATCH v6 1/1 SGI UV: TLB shootdown using broadcast assist unit - v6 delta

From: Cliff Wickman <[email protected]>

TLB shootdown for SGI UV.

This patch brings the code in Ingo's x86/x86/uv (June 19,2008) to v6

v1: 6/2 original
v2: 6/3 corrections/improvements per Ingo's review
v3: 6/4 split atomic operations off to a separate patch (Jeremy's review)
v4: 6/12 include <mach_apic.h> rather than <asm/mach-bigsmp/mach_apic.h>
(fixes a !SMP build problem)
fix the index on uv_table_bases[blade]
v5: 6/18 corrections/improvements per Ingo's second and third reviews
Ingo made coding improvements himself.
v6: 6/19 close the security hole in uv_ptc_proc_write())
and adds tlb_uv.o to the Makefile

Diffed against 2.6.26-rc6

---

Ingo,

(18Jun)

Quoted 7 lines

Found a potential security hole while doing that:
static ssize_t uv_ptc_proc_write(struct file *file, const char __user *user,
size_t count, loff_t *data)
if (copy_from_user(optstr, user, count))
return -EFAULT;

is count guaranteed to never be larger than 64?

is fixed below.

It adds tlb_uv.o to the Makefile.

And it's build-tested.

 arch/x86/kernel/Makefile |    2 +-
 arch/x86/kernel/tlb_uv.c |    2 ++
 2 files changed, 3 insertions(+), 1 deletion(-)
arch/x86/kernel/tlb_uv.c +2 -0
Index: linux/arch/x86/kernel/tlb_uv.c
===================================================================
@@ -492,6 +492,8 @@ static ssize_t uv_ptc_proc_write(struct 
 	long newmode;
 	char optstr[64];
 
+	if (count > 64)
+		return -EINVAL;
 	if (copy_from_user(optstr, user, count))
 		return -EFAULT;
 	optstr[count - 1] = '\0';
arch/x86/kernel/Makefile +1 -1
Index: linux/arch/x86/kernel/Makefile
===================================================================
@@ -93,7 +93,7 @@ obj-$(CONFIG_OLPC)		+= olpc.o
 ###
 # 64 bit specific files
 ifeq ($(CONFIG_X86_64),y)
-        obj-y				+= genapic_64.o genapic_flat_64.o genx2apic_phys.o genx2apic_uv_x.o
+        obj-y				+= genapic_64.o genapic_flat_64.o genx2apic_phys.o genx2apic_uv_x.o tlb_uv.o
         obj-$(CONFIG_X86_PM_TIMER)	+= pmtimer_64.o
         obj-$(CONFIG_AUDIT)		+= audit_64.o
 
Applied
#1 lore raw
Quoted 8 lines · Cliff Wickman
Quoted 7 lines

Found a potential security hole while doing that:
static ssize_t uv_ptc_proc_write(struct file *file, const char __user *user,
size_t count, loff_t *data)
if (copy_from_user(optstr, user, count))
return -EFAULT;

is count guaranteed to never be larger than 64?

is fixed below.

applied to tip/x86/uv, thanks Cliff.

Quoted 1 line

It adds tlb_uv.o to the Makefile.

hm, it was already there in tip/master, so i left this bit out.

arch/x86/kernel/tlb_uv.c static ssize_t uv_ptc_proc_write(struct · line 496

Quoted 4 lines
arch/x86/kernel/tlb_uv.c +2 -0
 	char optstr[64];
 
+	if (count > 64)
+		return -EINVAL;

also, this should not be in /proc, but be moved to debugfs. See arch/x86/mm/dump_pagetables.c about how to interface to debugfs.

And the "64" value should become something like:

	static const int max_option_len = 64;

or so.

Ingo