public inbox for goredo-devel@lists.stargrave.org
Atom feed
From: "Niklas Böhm" <mail@jnboehm•com>
To: goredo-devel@lists.cypherpunks.su
Subject: Re: Handling EINTR in unix.FcntlFlock
Date: Sat, 4 Jan 2025 15:04:41 +0100	[thread overview]
Message-ID: <5d578ee9-d843-43b8-bb4d-6a74d4c409fb@jnboehm.com> (raw)
In-Reply-To: <Z3ks7Qizpqjcp_Qe@stargrave.org>

Thank you for the quick response and release.  I have updated the 
package on AUR.

Now that I was looking at the implementation, would it maybe make sense 
to change the sleep time from 10ms to something a bit more randomized? 
So that two processes don't repeatedly run into the same deadlock error? 
  (And I think it also would not hurt for an NFS to have the file calls 
de-synchronized.)

Could take a uniform distribution between 0 and 20ms so that on average 
they will sleep for 10ms still.

On 1/4/25 1:43 PM, Sergey Matveev wrote:
> Greetings, Niklas!
> 
> *** Niklas Böhm [2025-01-04 10:59]:
>> [...] I have attached the small diff [...]
> 
> Thanks for finding the issue and fixing it! Released in 2.6.4.
> 


  reply	other threads:[~2025-01-04 14:05 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-01-04  9:59 Handling EINTR in unix.FcntlFlock Niklas Böhm
2025-01-04 12:43 ` Sergey Matveev
2025-01-04 14:04   ` Niklas Böhm [this message]
2025-01-07 11:05     ` Sergey Matveev