| To: | David Stevens <dlstevens@xxxxxxxxxx> |
|---|---|
| Subject: | Re: IPV6 RFC3542 compliance [PATCH] |
| From: | David Stevens <dlstevens@xxxxxxxxxx> |
| Date: | Mon, 6 Jun 2005 23:35:46 -0700 |
| Cc: | davem@xxxxxxxxxxxxx, netdev@xxxxxxxxxxx |
| In-reply-to: | <OF9F62ECDB.67FFA191-ON88257019.001F5EA8-88257019.00234AB6@us.ibm.com> |
| Sender: | netdev-bounce@xxxxxxxxxxx |
PS -
I should've said clearly; with the patch I submitted as-is, all
old binaries should return EINVAL on the socket options
that have changed. That's because all of those (except the
new IPV6_TCLASS, which didn't exist before) have option
arguments greater than int-size.
Recompiling those programs will still result in the setsockopt()
returning EINVAL, until the source is fixed to change the
socket options to the IPV6_RECVx.
sendmsg() and recvmsg() processing in old binaries should
still work, as-is.
So, with that patch, programs using the old names will give
a strong indication of what needs fixing.
+-DLS
|
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | Re: IPV6 RFC3542 compliance [PATCH], YOSHIFUJI Hideaki / 吉藤英明 |
|---|---|
| Next by Date: | Re: IPV6 RFC3542 compliance [PATCH], David Stevens |
| Previous by Thread: | Re: IPV6 RFC3542 compliance [PATCH], David Stevens |
| Next by Thread: | [PATCH] tg3: Fix link failure in 5701, Michael Chan |
| Indexes: | [Date] [Thread] [Top] [All Lists] |