Chuck Ebbert wrote: >On Wed, 10 Aug 2005 at 21:56:20 -0700, zach@xxxxxxxxxx wrote: > > > >>Patch-base: 2.6.13-rc5-mm1 >>Patch-keys: i386 desc cleanup >>Signed-off-by: Zachary Amsden <zach@xxxxxxxxxx> >>Index: linux-2.6.13/include/asm-i386/desc.h >>=================================================================== >>--- linux-2.6.13.orig/include/asm-i386/desc.h 2005-08-09 19:43:38.000000000 -0700 >>+++ linux-2.6.13/include/asm-i386/desc.h 2005-08-10 20:42:03.000000000 -0700 >>@@ -14,6 +14,28 @@ >> >> #include <asm/mmu.h> >> >>+#define desc_empty(desc) \ >>+ (!((desc)->a + (desc)->b)) >>+ >> >> > > I think that should be "|" instead of "+". > > I think so too. I merely moved the code here and didn't notice it in all this excitement. 0x00cf9a000xff306600 => Present CPL-0 32-bit code segment, base 0x0000ff30, limit 0xf6601 pages, for which desc_empty(desc) is true. Thankfully, this is not used as a security check, but it can falsely overwrite TLS segments with carefully chosen base / limits. I do not believe this is an issue in practice, but it is a kernel bug. Nice catch. Looks like it affects all 2.6.X kernels. Zach -------------- next part -------------- Chuck Ebbert noticed that the desc_empty macro is incorrect. Fix it. Signed-off-by: Zachary Amsden <zach@xxxxxxxxxx> Index: linux-2.6.13/include/asm-i386/desc.h =================================================================== --- linux-2.6.13.orig/include/asm-i386/desc.h 2005-08-15 11:23:32.000000000 -0700 +++ linux-2.6.13/include/asm-i386/desc.h 2005-08-16 10:49:03.000000000 -0700 @@ -18,7 +18,7 @@ #include <asm/mmu.h> #define desc_empty(desc) \ - (!((desc)->a + (desc)->b)) + (!((desc)->a | (desc)->b)) #define desc_equal(desc1, desc2) \ (((desc1)->a == (desc2)->a) && ((desc1)->b == (desc2)->b))