[patch 1/1] openmosix/openmosix-fix-mig_send-x8664.patch

This patch fixes the buffer overrun in arch_mig_send_proc_context ... 
TLS were copied, but a little too much ;)

Task are migrating fine here on x86_64 now, I am doing more tests see if 
I can do everything I want with it, then there will still be the 
spin_lock thing to do ...

Matt, tell me if you still have your problem with syscalls, and can you 
give me a program that makes the bug happen, so I can check if it is 
found too on x86_64 (and so I can fix it too).


So, to sum up, my next tasks are :
- check it works *all* fine
- fix the spin_lock things
- start looking at exec on remote
- make it smp safe

tab told me online that having the exec on remote thing working will 
bring openmosix-2.6 to the beta state, so we keep on going !

Florian

BTW, of course this patch apply on top of the lasts I already sent. if 
anyone is lost, please email me in private and I'll send all patches to 
make sure.

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

_______________________________________________
openMosix-devel mailing list
openMosix-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
https://lists.sourceforge.net/lists/listinfo/openmosix-devel

Index: linux/hpc/arch-x86_64.c
===================================================================
--- linux.orig/hpc/arch-x86_64.c	2006-09-29 18:39:16.000000000 +0200
+++ linux/hpc/arch-x86_64.c	2006-09-29 18:39:22.000000000 +0200
@@ -103,7 +103,7 @@
 
 	memcpy(&m->regs, regs, sizeof(struct pt_regs));
 
-	for (i = 0; i < TLS_SIZE; i++)
+	for (i = 0; i < GDT_ENTRY_TLS_ENTRIES; i++)
 		m->arch.tls_array[i] = p->thread.tls_array[i];
 
 	m->arch.ds = p->thread.ds;