Search squid archive

Re: peek/splice working with lynx but not with firefox or chrome

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Amos Jeffries writes:

see Nathan Hoads thread just the other day about a setup same as yours
NOT working.

There are two patches that need applying. One already in the 3.5 series
snapshots to fix SNI on some traffic cases, one still in QA review for
adding an ACL "server_name" that can match SNI without the helper.

That's very useful for the SNI matching indeed. Thanks!

> Yes, I am, but since I'm only splicing the connection, the browser
> itself should be able to get the original certificate sent by the
> server, and handle it appropriately. Or am I mistaken there?

That is correct. But also they get it through the filter of your OpenSSL
version parsing and re-packing capabilities for the underlying TLS/SSL
protocol syntax.

Those errors hint at things like the SSLv2/SSLv3 syntax being offered
and rejected, ALPN being mangled, or some advanced timing-based feature
being screwed up by the peek operation.

It seems so.

>> > I'm using squid 3.5.2 built with openssl 0.9.8zc on Slackware 13.1.
>> > Traffic is redirected from port 443 top 3130 with iptables.
>>
>> ... and with an older version of OpenSSL missing many of the last few
>> years worth of TLS crypto features. IIRC the library releases are now up
>> to 1.1.* or something. Its best to keep that kind of thing operating the
>> latest versions.
>
> I know it missing the latest features, but security patches are
> backported. And I know it is old, but it's what I have to work with
> now.Do you think it might be the cause of the problem I'm having with
> peek/splice, or was it a general recommendation?

Its a potential source of problems. Chrome is very much on the front
line of the arms race attempting to stop things like SSL-Bump working.
Firefox implement their own crypto library which tracks the latest TLS
features at a similar speed of development.
OpenSSL will be perpetually behind both of them, but at least the latest
one(s) have better chances not to be advertising features they reject in
"considered harmful" grounds.

I'll have a go then at trying with a newer openssl and the patches from the thread you mentioned.

Thanks a lot so far,

Roel

_______________________________________________
squid-users mailing list
squid-users@xxxxxxxxxxxxxxxxxxxxx
http://lists.squid-cache.org/listinfo/squid-users





[Index of Archives]     [Linux Audio Users]     [Samba]     [Big List of Linux Books]     [Linux USB]     [Yosemite News]

  Powered by Linux