Re: A little installing experience
Hasso Tepper wrote: > * System behaves well when installing, but after booting into installed > system from raid volume (pre-devfs system) there are constant intr 11 > livelock messages in the console and the system is under 50% > interrupt load. This is gone if I disconnect SATA DVD (not ahci(4)) > device. Why it isn't the while installing? I'm answering to it myself - because the cd is mounted at install time. If I mount it in the working system, interrupt storm stops as well. Isn't there something low level annoying (probing?) the device if it isn't mounted and it's not aligigning requests correctly (AFAIR SATA ATAPI required requests being multiple of 4 (?) bytes)? -- Hasso Tepper
Re: September release home stretch!
:The 2.3.2 packages I have built/am building now should work for 2.4; we :can start right off the bat with those, build new 2.4 packages, and shift :to pkgsrc-2009Q3 packages when they are available and built without :holding the release up, by shifting the links on avalon. : Ok, I like this plan. This means we can release mid-Sep (~2 weeks) as was originally planned. -Matt Matthew Dillon
Re: September release home stretch!
On Tue, September 1, 2009 11:54 pm, Matthew Dillon wrote: > * The actual release date this month is still a bit uncertain as we > talk with the pkgsrc meisters on which pkgsrc quarterly to > synchronize with. I was hoping for mid-September (2 weeks), but > it may wind up being more late-september (3.5 weeks). I don't know when the 2009Q3 release will happen - month end, I thought. That plus a week for building new packages would push us out too far. We've never defined any gating condition for what happens with pkgsrc - as in, is there a certain number or set of packages that must build for a release to be OK? (not necessarily a question to answer in this thread) The 2.3.2 packages I have built/am building now should work for 2.4; we can start right off the bat with those, build new 2.4 packages, and shift to pkgsrc-2009Q3 packages when they are available and built without holding the release up, by shifting the links on avalon.
Re: Fedora developer interested in DMA; query about build process
Oh, a tarball iso good enough, thanks. -- Michel On Sep 3, 2009 3:34 PM, "Simon 'corecode' Schubert" < corec...@fs.ei.tum.de> wrote: Michel Alexandre Salim wrote: > > On Tue, Sep 1, 2009 at 11:42 PM, Matthew > Dillon
Re: Fedora developer interested in DMA; query about build process
Michel Alexandre Salim wrote: On Tue, Sep 1, 2009 at 11:42 PM, Matthew Dillon wrote: I like the concept of throwing together an export hierarchy, but it might be too much maintainance to physically separate the git components out within the primary repo. Using git submodules, the maintenance required should be almost zero, I should think: http://progit.org/book/ch6-6.html The only change needed is at first clone, and I guess at every update: one would need to check the list of submodules currently available, and initialize any that has not been initialized yet. I don't want to use submodules, because they change the work flow in the main repo. My question to you is: would you like to see a broken out repo, or would a tarball good enough for you? cheers simon
A little installing experience
I got a shiny new HP server to test and it might happen that I can't get all issues with DragonFly on it solved before I'll loose it for testing, therefore will document problems found so far. * Network doesn't work. The bce(4) driver in our tree needs and update to get support for 5709 devices. * PS2 keyboard doesn't work at all in DragonFly, the USB one works only sometimes. There is two issues. At first PS2 isn't actually a real PS2, but emulated one. This commit in OpenBSD is relevant probably: http://anoncvs.estpak.ee/cgi-bin/cgit/openbsd-src/commit/?id=08a81789 The second issue is that there is always at least one USB keyboard and mouse in the system - virtual ones from ILO management console. Therefore there are problems choosing between these and real ones. We need keyboard muxing to solve this issue. * The system doesn't boot from ciss(4) raid volume. Mounting root is attempted before volume (da0) is really detected. It's a devfs issue, pre-devfs system boots fine. * System behaves well when installing, but after booting into installed system from raid volume (pre-devfs system) there are constant intr 11 livelock messages in the console and the system is under 50% interrupt load. This is gone if I disconnect SATA DVD (not ahci(4)) device. Why it isn't the while installing? There might be more of course, help is welcome ;). -- Hasso Tepper
Re: Fedora developer interested in DMA; query about build process
On Tue, Sep 1, 2009 at 11:42 PM, Matthew Dillon wrote: > I like the concept of throwing together an export hierarchy, > but it might be too much maintainance to physically separate > the git components out within the primary repo. > Using git submodules, the maintenance required should be almost zero, I should think: http://progit.org/book/ch6-6.html The only change needed is at first clone, and I guess at every update: one would need to check the list of submodules currently available, and initialize any that has not been initialized yet. Regards, -- Michel Salim
