[haiku-development] [Announcement] R1A4.1? Follow up to R1A4 to correct some issues seen by end users.

  • From: Alexander von Gluck IV <kallisti5@xxxxxxxxxxx>
  • To: <haiku-development@xxxxxxxxxxxxx>
  • Date: Wed, 14 Nov 2012 07:23:44 -0600

Good morning!

So, there have been quite a few reported issues with the R1A4 release.
The actual "number" of issues is small, but quite a few people have seen them.

During my testing these issues were few and far between... we had several people testing the pre-R1A4 images and no one could reproduce what I was seeing.

Given all the eyes on R1A4 however, we have had several people report the issues
I initially saw.

The major issues seen are generally as follows:

 1) Networking deadlock.
    On boot of the CD, sometimes the Deskbar can freeze up.
Opening a terminal and running ifconfig will result it hanging forever
    #9128, #9148

Solution: Rene was able to nail this down in hrev44832.. a pretty simple fix
              for a major problem.

 2) PCI bus manager gets loaded twice.
I never saw this in my tests... however it seems to happen when the following
    conditions are met:
* PCI bus manager gets loaded twice (bad, but doesn't guarantee a crash)
       * A PCI device has MSI support and we use it. (crash)
    #8684, #9139, #9153

Solution: mmlr is working through a fix right now to see if he can solve it
              in the short term.

Given the potential scope of users that will see these issues, and the fact that we don't want to dilute the importance of this release... Ryan and myself both
agree that an updated R1A4 image would be wise.

Generally an updated R1A4 image (R1A4.1?) would replace the existing R1A4 image
on the website.

This will solve the following situations:

* User couldn't boot? Download the new image and give it a go... issues fixed. * User already installed? Download the new image and install it over your R1A4 installation (to "upgrade" it)... any potential for issues down the road solved.

The general timeline is as soon as #2 is solved. We want to get this out the door
sooner rather then later.

I do want to apologize for not seeing this before R1A4. I had the poor luck that none of my 5 test machines were effected by these bugs. (1 showed up briefly and quickly disappeared on one machine, 2 was occurring on all of my machines but didn't
result in a crash)

Feedback welcome.

 -- Alexander von Gluck

Other related posts: