From: Guus Sliepen <guus@warande3094.warande.uu.nl>
To: linux-kernel@vger.kernel.org
Cc: Maxim Krasnyansky <max_mk@yahoo.com>,
Alan Cox <alan@lxorguk.ukuu.org.uk>
Subject: [PATCH] [RFC] dev_alloc_name() requirement of %d and the 100 names limit
Date: Sun, 21 Oct 2001 18:09:29 +0200 [thread overview]
Message-ID: <20011021180929.A29733@sliepen.warande.net> (raw)
[-- Attachment #1.1: Type: text/plain, Size: 1179 bytes --]
Hello,
Somewhere between 2.4.10 and 2.4.12 the dev_alloc_name() function was
changed to explicitly check for a %d in the name of a to-be allocated
network interface. The idea is that this function returns a unique name
or bails out.
The recent change broke at least tinc (a userspace VPN daemon), because
it tries to alloc a tun/tap device and gives it a name, but does not require
that the %d must be present (the documentation of the tun/tap driver nor
the example code in Documentation/ make any mention of that). I don't
see why the %d would have to be mandatory, so the attached patch allows
device names without a %d to work as well.
A comment in that function says that if anyone would need more than 100
devices the algorithm should be fixed. The patch does this, allowing a
virtually unlimitted amount of names, in O(log(n)) time, n being the
number of already allocated names.
I stresstested the patched, I was able to allocate 1901 network devices
using "test%d" as the name. After that my system claimed it didn't have
any file descriptors left.
--
Met vriendelijke groet / with kind regards,
Guus Sliepen <guus@sliepen.warande.net>
[-- Attachment #1.2: patch_dev_alloc_name --]
[-- Type: text/plain, Size: 1829 bytes --]
--- dev.c.old Sun Oct 21 15:39:44 2001
+++ dev.c Sun Oct 21 17:45:11 2001
@@ -550,30 +550,65 @@
int dev_alloc_name(struct net_device *dev, const char *name)
{
- int i;
- char buf[32];
+ int i, j, k;
+ char buf[IFNAMSIZ];
char *p;
/*
* Verify the string as this thing may have come from
- * the user. There must be one "%d" and no other "%"
+ * the user. There may only be one "%d" and no other "%"
* characters.
*/
p = strchr(name, '%');
- if (!p || p[1] != 'd' || strchr(p+2, '%'))
+ if (!p)
+ {
+ snprintf(buf,sizeof(buf),name);
+ if (__dev_get_by_name(buf) == NULL) {
+ strcpy(dev->name, buf);
+ return 0;
+ }
+ else
+ return -ENFILE;
+ }
+
+ if (p[1] != 'd' || strchr(p+2, '%'))
return -EINVAL;
/*
- * If you need over 100 please also fix the algorithm...
+ * New algorithm that works in O(log(n)) worst case time, where
+ * n is the number of already allocated devices with the
+ * same base name.
+ *
+ * The algorithm works by trying a device number j between i and k,
+ * and changes i and k depending on whether it was allocated or not.
+ *
+ * i is the lower bound of already allocated devices,
+ * k is the upper bound of already allocated devices,
+ * k = 0 means k is infinite.
*/
- for (i = 0; i < 100; i++) {
- snprintf(buf,sizeof(buf),name,i);
+ i = j = k = 0;
+ for (;;)
+ {
+ snprintf(buf,sizeof(buf),name,j);
if (__dev_get_by_name(buf) == NULL) {
- strcpy(dev->name, buf);
- return i;
+ if (j == i || j == i + 1)
+ {
+ strcpy(dev->name, buf);
+ return j;
+ }
+ else
+ k = j;
}
+ else
+ i = j;
+
+ if (k)
+ j = (i + k) / 2;
+ else
+ j *= 2;
+ if (j == i)
+ j++;
}
- return -ENFILE; /* Over 100 of the things .. bail out! */
}
/**
[-- Attachment #2: Type: application/pgp-signature, Size: 232 bytes --]
reply other threads:[~2001-10-21 16:09 UTC|newest]
Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20011021180929.A29733@sliepen.warande.net \
--to=guus@warande3094.warande.uu.nl \
--cc=alan@lxorguk.ukuu.org.uk \
--cc=linux-kernel@vger.kernel.org \
--cc=max_mk@yahoo.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).