Skip to contentSkip to conversation

blocking w/ spinlock or interrupt's disabled

CHECKER Manfred Spraul 4 participants 4 messages Mar 18, 2001 – Mar 20, 2001

lore
Share & tools
lore
#0 lore raw
Quoted 9 lines
Quoted 3 lines

enclosed are 163 potential bugs in 2.4.1 where blocking functions are called with either interrupts disabled or a spin lock held. The checker works by:

Here's the file manifest. Apologies.

drivers/atm/idt77105.c
[...]
drivers/char/cyclades.c

Unortunately schedule() with disabled interrupts is a feature, it's needed for the old (deprecated and waiting for termination in 2.5) sleep_on() functions.

Is it difficult to split it into "interrupts disabled" and "spin lock held"?

--

Manfred

#1 lore raw
Quoted 2 lines

Is it difficult to split it into "interrupts disabled" and "spin lock held"?

Nope, since it's already done ;-) The suffix of each error message should say whether it's because you have a spinlock, ints disabled, or both:

2.4.1/drivers/atm/idt77105.c:153:fetch_stats: ERROR:BLOCK:151:153:
calling blocking fn '__constant_copy_to_user' w/ int's disabled
2.4.1/drivers/atm/iphase.c:2426:ia_led_timer: ERROR:BLOCK:2423:2426:
calling blocking fn 'ia_tx_poll' w/ spin lock held

Dawson

#2 lore raw
Quoted 4 lines · Dawson Engler
Quoted 3 lines

Is it difficult to split it into "interrupts disabled" and "spin lock held"?

Is it difficult to test for matching spinlock pairs such as spin_lock_irq/spin_unlock_irq. Sometimes a spin_lock_irq is followed by a spin_unlock and a separate interrupt re-enable. This sort of usage, while not strictly wrong, does make it hard to use the spin_lock/unlock macros to do preemption. This said, pairing information would be very helpful. Note, there are several flavors here, not just the one I cited.

George

#3 lore raw
In reply to Manfred Spraul

In article <001801c0af8e$bda30c10$5517fea9@local>,

Quoted 4 lines · Manfred Spraul

Unortunately schedule() with disabled interrupts is a feature, it's needed for the old (deprecated and waiting for termination in 2.5) sleep_on() functions.

Yes. But that should only cover "sleep_on()" and it's interruptible cousing "sleep_on_interruptible()". No other blocking call should have interrupts disabled, I would hope.

The special-case is a fairly specific "some old-style drivers avoid race conditions by having interrupts disabled over explicit conditional sleeps", not a generic "you may have interrupts disabled before blocking".

Linus