lkml.org 
[lkml]   [2006]   [Oct]   [28]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [take21 1/4] kevent: Core files.
On Sat, Oct 28, 2006 at 12:28:12PM +0200, Eric Dumazet (dada1@cosmosbay.com) wrote:
> +/*
> + * Called under kevent_user->ready_lock, so updates are always protected.
> + */
> +int kevent_user_ring_add_event(struct kevent *k)
> +{
> + unsigned int pidx, off;
> + struct kevent_mring *ring, *copy_ring;
> +
> + ring = k->user->pring[0];
> +
> + if ((ring->kidx + 1 == ring->uidx) ||
> + ((ring->kidx + 1 == KEVENT_MAX_EVENTS) && ring->uidx
> == 0)) {
> + if (k->user->overflow_kevent == NULL)
> + k->user->overflow_kevent = k;
> + return -EAGAIN;
> + }
> +
>
>
> I really dont understand how you manage to queue multiple kevents in the
> 'overflow list'. You just queue one kevent at most. What am I missing ?

There is no overflow list - it is a pointer to the first kevent in the
ready queue, which was not put into ring buffer. It is an optimisation,
which allows to not search for that position each time new event should
be placed into the buffer, when it starts to have an empty slot.

>
> >+
> >+ for (i=0; i<KEVENT_MAX_PAGES; ++i) {
> >+ u->pring[i] = (struct kevent_mring
> >*)__get_free_page(GFP_KERNEL);
> >+ if (!u->pring[i])
> >+ break;
> >+ }
> >+
> >+ if (i != KEVENT_MAX_PAGES)
> >+ goto err_out_free;
>
> Why dont you use goto directly ?
>
> if (!u->pring[i])
> goto err_out_free;
>

I used a fallback mode here which allowed to use smaller number of pages
for kevent ring buffer, but then decided to drop it.
So it is possible to use goto directly.

> >+
> >+ u->pring[0]->uidx = u->pring[0]->kidx = 0;
> >+
> >+ return 0;
> >+
> >+err_out_free:
> >+ for (i=0; i<KEVENT_MAX_PAGES; ++i) {
> >+ if (!u->pring[i])
> >+ break;
> >+
> >+ free_page((unsigned long)u->pring[i]);
> >+ }
> >+ return k;
> >+}
> >+
>
>
>
>
> >+static int kevent_user_ctl_add(struct kevent_user *u, unsigned int num,
> >void __user *arg)
> >+{
> >+ int err, cerr = 0, knum = 0, rnum = 0, i;
> >+ void __user *orig = arg;
> >+ struct ukevent uk;
> >+
> >+ mutex_lock(&u->ctl_mutex);
> >+
> >+ err = -EINVAL;
> >+ if (num > KEVENT_MIN_BUFFS_ALLOC) {
> >+ struct ukevent *ukev;
> >+
> >+ ukev = kevent_get_user(num, arg);
> >+ if (ukev) {
> >+ for (i = 0; i < num; ++i) {
> >+ err = kevent_user_add_ukevent(&ukev[i], u);
> >+ if (err) {
> >+ kevent_stat_im(u);
> >+ if (i != rnum)
> >+ memcpy(&ukev[rnum],
> >&ukev[i], sizeof(struct ukevent));
> >+ rnum++;
> >+ } else
> >+ knum++;
>
>
> Why are you using/counting knum ?

It should go avay.

> >+ }
> >+ if (copy_to_user(orig, ukev, rnum*sizeof(struct
> >ukevent)))
> >+ cerr = -EFAULT;
> >+ kfree(ukev);
> >+ goto out_setup;
> >+ }
> >+ }
> >+
> >+ for (i = 0; i < num; ++i) {
> >+ if (copy_from_user(&uk, arg, sizeof(struct ukevent))) {
> >+ cerr = -EFAULT;
> >+ break;
> >+ }
> >+ arg += sizeof(struct ukevent);
> >+
> >+ err = kevent_user_add_ukevent(&uk, u);
> >+ if (err) {
> >+ kevent_stat_im(u);
> >+ if (copy_to_user(orig, &uk, sizeof(struct ukevent)))
> >{
> >+ cerr = -EFAULT;
> >+ break;
> >+ }
> >+ orig += sizeof(struct ukevent);
> >+ rnum++;
> >+ } else
> >+ knum++;
> >+ }
> >+
> >+out_setup:
> >+ if (cerr < 0) {
> >+ err = cerr;
> >+ goto out_remove;
> >+ }
> >+
> >+ err = rnum;
> >+out_remove:
> >+ mutex_unlock(&u->ctl_mutex);
> >+
> >+ return err;
> >+}
> -
> To unsubscribe from this list: send the line "unsubscribe netdev" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html

--
Evgeniy Polyakov
-
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/

\
 
 \ /
  Last update: 2006-10-28 12:57    [W:0.086 / U:0.568 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site