From: Andrej Mihajlov <and@mullvad.net>
To: Houman <houmie@gmail.com>
Cc: WireGuard mailing list <wireguard@lists.zx2c4.com>
Subject: Re: Wireguard iOS crashes after upgrading to XCode 14
Date: Thu, 22 Sep 2022 11:44:31 +0200 [thread overview]
Message-ID: <6E42D7CE-E03C-4D87-B9CD-182567D9C087@mullvad.net> (raw)
In-Reply-To: <CABBZOs=9VVSDEGPw1eQWmusJotA5toLCn1-uNQeZ+moPthAOxg@mail.gmail.com>
Hi Houman,
I believe that the crash coming from withMemoryRebound is actually assertion (assert()). Very often assertions are stripped out from release builds, so I guess it could be the reason why it just worked in release builds.
Jason has already replied in regards of releasing an update.
Best regards,
Andrej
> On 22 Sep 2022, at 10:56, Houman <houmie@gmail.com> wrote:
>
> Hi Andrej,
>
> It works, well done!
>
> A strange thing though, before your patch I was still able to connect
> to the VPN server, if I changed the schema to Release instead of
> Debug. Now with your patch it also works under Debug schema, which is
> fantastic.
> What could be the technical reason that it still worked under Release?
>
> And what will happen now, are you able to actually get this patch
> released on the official repo? The repo hasn't been updated for a
> year. :-)
>
> Thanks,
> Houman
>
>
> On Thu, 22 Sept 2022 at 09:31, Andrej Mihajlov <and@mullvad.net> wrote:
>>
>> Hi,
>>
>> I think we have a bug. If I am right, basically in both IPv4 and IPv6 extensions, withMemoryRebound takes capacity which is actually a number of instances of a given type (sockaddr_ variant) and not the byte size of a struct.
>>
>> Could you please patch your WireGuardKit with the following commit and see if it helps?
>>
>> https://git.zx2c4.com/wireguard-apple/commit/?h=am/fix-addrinfo-crash
>>
>> Best regards,
>> Andrey Mikhaylov
>>
>>> On 13 Sep 2022, at 14:41, Houman <houmie@gmail.com> wrote:
>>>
>>> My existing Wireguard iOS implementation stopped working after
>>> upgrading to Xcode 14 today.
>>> When trying to connect to servers that support only IPv4, then it's
>>> fine. But if the server supports both IPv6 and IPv4 then the tunnel
>>> crashes:
>>>
>>> This IPv6 extension in
>>> wireguard-apple/Sources/WireGuardKit/IPAddress+AddrInfo.swift crashes
>>> with a Fatal Error at addrInfo.ai_addr.withMemoryRebound()
>>>
>>> The whole extension below:
>>>
>>> extension IPv6Address {
>>> init?(addrInfo: addrinfo) {
>>> guard addrInfo.ai_family == AF_INET6 else { return nil }
>>>
>>> let addressData = addrInfo.ai_addr.withMemoryRebound(to:
>>> sockaddr_in6.self, capacity: MemoryLayout<sockaddr_in6>.size) { ptr ->
>>> Data in
>>> return Data(bytes: &ptr.pointee.sin6_addr, count:
>>> MemoryLayout<in6_addr>.size)
>>> }
>>> self.init(addressData)
>>> }
>>> }
>>>
>>> Has anyone else experienced this problem?
>>>
>>> Thanks,
>>
next prev parent reply other threads:[~2022-09-22 9:55 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-09-13 12:41 Houman
2022-09-22 8:31 ` Andrej Mihajlov
2022-09-22 8:56 ` Houman
2022-09-22 9:38 ` Jason A. Donenfeld
2022-09-22 9:39 ` Houman
2022-12-05 17:15 ` Houman
2022-12-05 17:40 ` Jason A. Donenfeld
2022-09-22 9:44 ` Andrej Mihajlov [this message]
2022-09-22 9:46 ` Houman
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=6E42D7CE-E03C-4D87-B9CD-182567D9C087@mullvad.net \
--to=and@mullvad.net \
--cc=houmie@gmail.com \
--cc=wireguard@lists.zx2c4.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).