| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <47158540.7000709@oracle.com>
Date: Tue, 16 Oct 2007 20:45:04 -0700
From: Tina Yang <tina.yang@...cle.com>
To: Matt Mackall <mpm@...enic.com>
CC: netdev@...r.kernel.org
Subject: [patch] net: avoid race between netpoll and network fast path
The current netpoll design and implementation has serveral race issues with the
network fast path that panics/hangs the system or causes interface timeout/reset
but the fix is likely to have impact on the overall system performance and could
involve a large number of drivers. The proposal is to disable the problem code
for normal operations but only to enable it at the time of crash in case polling
is necessary. Tests that have been done included the bug fix verification
as well as regression check on the netlog results in various crash modes.
Signed-off-by: Tina Yang <tina.yang@...cle.com>
---
--- linux-2.6.23.1/include/linux/kernel.h.orig 2007-10-12 09:43:44.000000000 -0700
+++ linux-2.6.23.1/include/linux/kernel.h 2007-10-15 22:05:27.000000000 -0700
@@ -184,6 +184,8 @@ static inline void console_verbose(void)
console_loglevel = 15;
}
+extern int netpoll_enable;
+
extern void bust_spinlocks(int yes);
extern void wake_up_klogd(void);
extern int oops_in_progress; /* If set, an oops, panic(), BUG() or die() is in progress */
--- linux-2.6.23.1/net/core/netpoll.c.orig 2007-10-12 09:43:44.000000000 -0700
+++ linux-2.6.23.1/net/core/netpoll.c 2007-10-15 22:20:15.000000000 -0700
@@ -150,15 +150,19 @@ static void service_arp_queue(struct net
}
}
+int netpoll_enable;
+EXPORT_SYMBOL(netpoll_enable);
void netpoll_poll(struct netpoll *np)
{
if (!np->dev || !netif_running(np->dev) || !np->dev->poll_controller)
return;
/* Process pending work on NIC */
- np->dev->poll_controller(np->dev);
- if (np->dev->poll)
- poll_napi(np);
+ if (unlikely(netpoll_enable)) {
+ np->dev->poll_controller(np->dev);
+ if (np->dev->poll)
+ poll_napi(np);
+ }
service_arp_queue(np->dev->npinfo);
--- linux-2.6.23.1/kernel/panic.c.orig 2007-10-12 09:43:44.000000000 -0700
+++ linux-2.6.23.1/kernel/panic.c 2007-10-15 22:07:25.000000000 -0700
@@ -66,6 +66,7 @@ NORET_TYPE void panic(const char * fmt,
unsigned long caller = (unsigned long) __builtin_return_address(0);
#endif
+ netpoll_enable = 1;
/*
* It's possible to come here directly from a panic-assertion and not
* have preempt disabled. Some functions called from here want
--- linux-2.6.23.1/arch/x86_64/kernel/traps.c.orig 2007-10-12 09:43:44.000000000 -0700
+++ linux-2.6.23.1/arch/x86_64/kernel/traps.c 2007-10-15 22:06:29.000000000 -0700
@@ -522,6 +522,8 @@ void __kprobes __die(const char * str, s
#endif
printk("\n");
notify_die(DIE_OOPS, str, regs, err, current->thread.trap_no, SIGSEGV);
+ if (kexec_should_crash(current))
+ netpoll_enable = 1;
show_registers(regs);
add_taint(TAINT_DIE);
/* Executive summary in case the oops scrolled away */
--- linux-2.6.23.1/arch/i386/kernel/traps.c.orig 2007-10-12 09:43:44.000000000 -0700
+++ linux-2.6.23.1/arch/i386/kernel/traps.c 2007-10-15 22:06:14.000000000 -0700
@@ -428,6 +428,8 @@ void die(const char * str, struct pt_reg
if (notify_die(DIE_OOPS, str, regs, err,
current->thread.trap_no, SIGSEGV) !=
NOTIFY_STOP) {
+ if (kexec_should_crash(current))
+ netpoll_enable = 1;
show_registers(regs);
/* Executive summary in case the oops scrolled away */
esp = (unsigned long) (®s->esp);
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists