Messages in this thread | | | Subject | Re: [PATCH 2/3] printk: Add %ptc to safely print a task's comm | From | Joe Perches <> | Date | Tue, 17 May 2011 14:52:30 -0700 |
| |
On Tue, 2011-05-17 at 23:42 +0200, Jiri Slaby wrote: > On 05/17/2011 10:47 PM, John Stultz wrote: > > Accessing task->comm requires proper locking. However in the past > > access to current->comm could be done without locking. This > > is no longer the case, so all comm access needs to be done > > while holding the comm_lock. > > +static noinline_for_stack > I still fail to see why this should be slowed down by noinlining it. > Care to explain?
Any vsprintf is slow.
> With my setup, the code below inlined will use 32 bytes of stack. The > same as %pK case. Uninlined it obviously eats "only" 8 bytes for IP.
The idea is to avoid excess stack consumption for things like:
struct va_format vaf;
const char *fmt = "some format with %ptc";
vaf.fmt = fmt; vaf.va = &va_list;
printk("some format with %pV\n", &vaf);
> > +char *task_comm_string(char *buf, char *end, void *addr, > > + struct printf_spec spec, const char *fmt) > > +{ > > + struct task_struct *tsk = addr; > > + char *ret; > > + unsigned long flags; > > + > > + spin_lock_irqsave(&tsk->comm_lock, flags); > > + ret = string(buf, end, tsk->comm, spec); > > + spin_unlock_irqrestore(&tsk->comm_lock, flags); > > + > > + return ret; > > +}
I think it was more of a problem when "4k stacks" was the default than today, but I think it is still "good form".
| |