Messages in this thread | | | Date | Tue, 28 May 2002 23:17:55 +0200 | From | Pavel Machek <> | Subject | Re: [PATCH] swsusp in 2.4.19-pre8-ac5 |
| |
Hi!
> > > > INTERESTING(p); > > if (p->flags & PF_FROZEN) > > continue; > > - > > + if (p->state == TASK_STOPPED) { /* this task is a stopped but not frozen one */ > > + p->flags |= PF_IOTHREAD; > > + _printk("+"); > > + continue; > > + } > > /* FIXME: smp problem here: we may not access other process' flags > > without locking */ > > p->flags |= PF_FREEZE; > > > > Are you sure this is correct? What if someone wakes it just after you > > given it PF_IOTHREAD? > > Good point. I need to be more precise.
Yup.
> > > > What's the point of all those PRINTS -> __prints changes? I do not > > like additional abstractions on the top of printk(). Are they really > > neccessary? > > Actually I tried to make the process prettier using a dedicated console. > The PRINT are for debugging, _print for the dedicated console (can be > deactivated using SUSPEND_CONSOLE) and __print are always written > (errors messages). The PRINTS PRINTR macros were used to separate > suspend and resume machine. It's not necessary but isn't that nicer when > you suspend ?
Are not "Suspend : " and "Resume : " superfluous if you have dedicated console, anyway?
Why don't you use generic printk() for messages that are printed, always?
> > @@ -1207,11 +1263,12 @@ > > pagedir_order = get_bitmask_order(nr_pgdir_pages); > > > > error = -ENOMEM; > > + free_page((unsigned long)cur); > > pagedir_nosave = (suspend_pagedir_t *)__get_free_pages(GFP_ATOMIC, pagedir_order); > > if(!pagedir_nosave) > > goto resume_read_error; > > > > - PRINTR( "%sReading pagedir, ", name_resume ); > > + PRINTR( "Reading pagedir\n" ); > > > > /* We get pages in reverse order of saving! */ > > error=-EIO; > > > > Why freeing it? This system is going to be overwritten, anyway. > > I don't like the idea not to free the page as soon as we don't need it > any more. Maybe we'll have an error later and try to recover a normal > boot in a future version.
Okay, applied.
> What about the CONFIG_SMP restriction ? Is it still pertinent ?
Yes, I'm afraid. If someone wants to donate me SMP pentium, I might try to debug that ;-). Pavel -- Casualities in World Trade Center: ~3k dead inside the building, cryptography in U.S.A. and free speech in Czech Republic. - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |