| To: | shmulik.hen@xxxxxxxxx |
|---|---|
| Subject: | Re: BUG or not? GFP_KERNEL with interrupts disabled. |
| From: | "David S. Miller" <davem@xxxxxxxxxx> |
| Date: | Thu, 27 Mar 2003 05:43:57 -0800 (PST) |
| Cc: | dane@xxxxxxxxxx, bonding-devel@xxxxxxxxxxxxxxxxxxxxx, bonding-announce@xxxxxxxxxxxxxxxxxxxxx, netdev@xxxxxxxxxxx, linux-kernel@xxxxxxxxxxxxxxx, linux-net@xxxxxxxxxxxxxxx, torvalds@xxxxxxxxxxxxx, mingo@xxxxxxxxxx, kuznet@xxxxxxxxxxxxx |
| In-reply-to: | <Pine.LNX.4.44.0303271406230.7106-100000@jrslxjul4.npdj.intel.com> |
| References: | <E791C176A6139242A988ABA8B3D9B38A01085638@hasmsx403.iil.intel.com> <Pine.LNX.4.44.0303271406230.7106-100000@jrslxjul4.npdj.intel.com> |
| Sender: | netdev-bounce@xxxxxxxxxxx |
From: shmulik.hen@xxxxxxxxx Date: Thu, 27 Mar 2003 15:32:02 +0200 (IST) Further more, holding a lock_irq doesn't mean bottom halves are disabled too, it just means interrupts are disabled and no *new* softirq can be queued. Consider the following situation: I think local_bh_enable() should check irqs_disabled() and honour that. What you are showing here, that BH's can run via local_bh_enable() even when IRQs are disabled, is a BUG(). IRQ disabling is meant to be stronger than softint disabling. Ingo/Linus? |
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | Re: ?completeness of IPsec feature-set, bert hubert |
|---|---|
| Next by Date: | Re: BUG or not? GFP_KERNEL with interrupts disabled., David S. Miller |
| Previous by Thread: | Re: BUG or not? GFP_KERNEL with interrupts disabled., shmulik . hen |
| Next by Thread: | Re: BUG or not? GFP_KERNEL with interrupts disabled., Linus Torvalds |
| Indexes: | [Date] [Thread] [Top] [All Lists] |