<div dir="ltr">I'll speak up. I'd use it if it was stable (again).<div><br></div><div>-Mark</div><div><br></div></div><div class="gmail_extra"><br><div class="gmail_quote">On Wed, Jan 24, 2018 at 9:54 PM, Dave McGuire <span dir="ltr"><<a href="mailto:mcguire@neurotica.com" target="_blank">mcguire@neurotica.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><br>
Let me just say that I, for one, am very glad that you are working on<br>
this. DECnet under (modern) Linux is something that is very important<br>
to me. I know there are others who feel the same way; inexplicably they<br>
don't speak up in support of your efforts.<br>
<br>
Thank you for working on this. Please continue, if you are willing/able.<br>
<br>
-Dave<br>
<div class="HOEnZb"><div class="h5"><br>
On 01/24/2018 04:06 PM, Erik Olofsen wrote:<br>
> Hi Johnny,<br>
><br>
> You're right, and I will also get tired of it...!<br>
><br>
> However, being in the current routine of trying various kernels,<br>
> I've now arrived at the latest one 4.14.15, something I never<br>
> expected to be possible.<br>
><br>
> Different things go wrong with different kernels. With the latest<br>
> one, I get a lot of kernel messages about things the decnet module<br>
> is doing wrong, rather than mysterious failures, so there lies<br>
> some hope. Note: I get these messages while DECnet is working!<br>
><br>
> So I will continue just a little bit, and I'll be testing this latest<br>
> kernel. I hope you won't be getting too many strange packets from RULLFL.<br>
> Can I ask you after some days/weeks if you did?<br>
><br>
> With the changes in the Linux API, the decnet code has been updated,<br>
> even while it was orphaned, so that it compiles; but it hasn't been<br>
> tested (of course). I read somewhere that they're going to omit the<br>
> decnet source from the tree.<br>
><br>
> When I get tired of it, I'll document a bit how to get things working.<br>
><br>
> Erik<br>
><br>
><br>
> On Tue, Jan 23, 2018 at 11:34:35PM +0100, Johnny Billquist wrote:<br>
>> Erik, thanks for all the analyzing.<br>
>> This both (to me) shows the issues with a constantly changing internal API<br>
>> of the Linux kernel, and that maintaining DECnet for Linux is a constant<br>
>> effort and annoyance. If someone is willing to do it, great, but I suspect<br>
>> most people will tire after some time.<br>
>> Also, we still have the issues with DECnet under Linux in general, where it<br>
>> seems it don't work right at all in some situations.<br>
>><br>
>> Johnny<br>
>><br>
>> On 2018-01-23 13:34, Erik Olofsen wrote:<br>
>>> One more note:<br>
>>><br>
>>> In the last longterm 3.x.y kernel, 3.18.92, sock_alloc_send_skb()<br>
>>> sets the errcode to ENOBUFS even when the sk_buff is allocated<br>
>>> successfully. This causes sendmsg() to stop with "No buffer space<br>
>>> available". That is easy to fix in dn_alloc_send_pskb().<br>
>>><br>
>>><br>
>>> On Mon, Jan 15, 2018 at 09:43:53PM +0100, Erik Olofsen wrote:<br>
>>>> Some more notes on this:<br>
>>>><br>
>>>> The problem of accessing a Linux machine seems to be related to<br>
>>>> the standard mtu of 1500 of a network interface.<br>
>>>><br>
>>>> Reducing the maximum packet size to 576 helps;<br>
>>>><br>
>>>> <a href="https://sourceforge.net/p/linux-decnet/mailman/message/2395733/" rel="noreferrer" target="_blank">https://sourceforge.net/p/<wbr>linux-decnet/mailman/message/<wbr>2395733/</a><br>
>>>><br>
>>>> This can be done using ifconfig or using 576 instead of dst_mtu()<br>
>>>> in dn_current_mss() af_decnet.c of the kernel module source.<br>
>>>><br>
>>>> tcpdump shows that a large data packet is one byte too long.<br>
>>>> Perhaps this occurs because the pad byte in the header is not counted<br>
>>>> in dn_mss_from_pmtu() in af_decnet.c; with an additional mtu--<br>
>>>> this seems to be fixed.<br>
>>>><br>
>>>> Kernel warnings related to dst.h;<br>
>>>><br>
>>>> <a href="https://www.spinics.net/lists/netdev/msg225996.html" rel="noreferrer" target="_blank">https://www.spinics.net/lists/<wbr>netdev/msg225996.html</a><br>
>>>><br>
>>>> indeed disappear by using dst_metric_raw at two places.<br>
>>>><br>
>>>> The above was tested with kernel 3.10.17, and can be tested<br>
>>>> with <a href="http://mim.update.uu.se/hecnet" rel="noreferrer" target="_blank">http://mim.update.uu.se/hecnet</a> and nodename RULLFL.<br>
>>>><br>
>>>> dntype mim::nodenames.dat works very well on a real machine; why<br>
>>>> it doesn't work very well with the virtual machine is not really<br>
>>>> solved yet...<br>
>>>><br>
>>>><br>
>>>> On Wed, Jan 03, 2018 at 04:44:54PM +0100, Johnny Billquist wrote:<br>
>>>>> On 2018-01-03 16:02, Erik Olofsen wrote:<br>
>>>>>> On a virtual machine with lubuntu 12.04, kernel 3.2.0, I have<br>
>>>>>> DECnet for Linux V.2.5.68s working, with dnprogs version 2.65.<br>
>>>>>> It is node RULLFL.<br>
>>>>>><br>
>>>>>> The daemons and utilities use decnet.conf with node information,<br>
>>>>>> so something useful would be to create it from mim::nodenames.dat.<br>
>>>>>><br>
>>>>>> On a VAX, TYPEing it works well, but on the Linux machine, dntype<br>
>>>>>> hangs after giving parts of the file; dndir mim:: works well.<br>
>>>>>><br>
>>>>>> Does anyone have dntype mim::nodenames.dat working properly (with<br>
>>>>>> perhaps different versions of the above)?<br>
>>>>><br>
>>>>> I can't help, but I can report that trying to read a file from a Linux<br>
>>>>> machine on MIM also hangs after a while.<br>
>>>>> A big problem with the DECnet/Linux implementation is that the developers<br>
>>>>> only ever tested against VMS, and nothing else.<br>
>>>>><br>
>>>>> I've known for ages that it does not play well with RSX. But I do not want<br>
>>>>> to dig into the Linux implementation.<br>
>>>>><br>
>>>>> (The Linux DAP claims to be a VMS system, by the way, which is probably also<br>
>>>>> not a good idea. Better if it had claimed to be some Unix derivative.)<br>
>>>>><br>
>>>>> Johnny<br>
>>>>><br>
>>>>> --<br>
>>>>> Johnny Billquist || "I'm on a bus<br>
>>>>> || on a psychedelic trip<br>
>>>>> email: <a href="mailto:bqt@softjar.se">bqt@softjar.se</a> || Reading murder books<br>
>>>>> pdp is alive! || tryin' to stay hip" - B. Idol<br>
>><br>
>><br>
>> --<br>
>> Johnny Billquist || "I'm on a bus<br>
>> || on a psychedelic trip<br>
>> email: <a href="mailto:bqt@softjar.se">bqt@softjar.se</a> || Reading murder books<br>
>> pdp is alive! || tryin' to stay hip" - B. Idol<br>
<br>
<br>
</div></div><span class="HOEnZb"><font color="#888888">--<br>
Dave McGuire, AK4HZ<br>
New Kensington, PA<br>
</font></span></blockquote></div><br></div>