[PATCH] 5/6 i386 non-reversible-fs-gs

[zach@vmware.com]

Subtle fix:  load_TLS has been moved after saving %fs and %gs segments to
avoid creating non-reversible segments.  This could conceivably cause a bug
if the kernel ever needed to save and restore fs/gs from the NMI handler.
It currently does not, but this is the safest approach to avoiding fs/gs
corruption.  SMIs are safe, since SMI saves the descriptor hidden state.

Diffs against: patch-2.6.13-rc4 + cpu-inline-cleanup + dt-inline-cleanup

Signed-off-by: Zachary Amsden <zach@vmware.com>
Index: linux-2.6.13/arch/i386/kernel/process.c
===================================================================
--- linux-2.6.13.orig/arch/i386/kernel/process.c	2005-07-29 11:17:02.000000000 -0700
+++ linux-2.6.13/arch/i386/kernel/process.c	2005-07-29 11:50:19.000000000 -0700
@@ -678,21 +678,26 @@
 	__unlazy_fpu(prev_p);
 
 	/*
-	 * Reload esp0, LDT and the page table pointer:
+	 * Reload esp0.
 	 */
 	load_esp0(tss, next);
 
 	/*
-	 * Load the per-thread Thread-Local Storage descriptor.
+	 * Save away %fs and %gs. No need to save %es and %ds, as
+	 * those are always kernel segments while inside the kernel.
+	 * Doing this before setting the new TLS descriptors avoids
+	 * the situation where we temporarily have non-reloadable
+	 * segments in %fs and %gs.  This could be an issue if the
+	 * NMI handler ever used %fs or %gs (it does not today), or
+	 * if the kernel is running inside of a hypervisor layer.
 	 */
-	load_TLS(next, cpu);
+	savesegment(fs, prev->fs);
+	savesegment(gs, prev->gs);
 
 	/*
-	 * Save away %fs and %gs. No need to save %es and %ds, as
-	 * those are always kernel segments while inside the kernel.
+	 * Load the per-thread Thread-Local Storage descriptor.
 	 */
-	asm volatile("mov %%fs,%0":"=m" (prev->fs));
-	asm volatile("mov %%gs,%0":"=m" (prev->gs));
+	load_TLS(next, cpu);
 
 	/*
 	 * Restore %fs and %gs if needed.