Muli Ben-Yehuda's journal

I spent most of the weekend working on the x86-64 DMA mapping ops patch and being sick, not necessarily in this order. The latest version of the patch (-K3) is available here. In a few minutes, once the automated tests will finish running, I’ll email it to Andi Kleen and lkml and ask for inclusion. I hope it doesn’t break anyone’s machines… and if it does, I hope it at least breaks then in new and interesting ways!

On Sunday I had a long day at work, mostly talking to various people about various things. Hmm, that’s way too amorphous. Let’s see what I can elaborate on: we’re seeing a strange scheduling bug with the vanilla kernel on a 4 CPU machine (2 CPUs, each hyperthreaded). I had a meeting with my boss and got tentative travel approval to attend the Xen summit in Austin in January. At night I went for a sportive walk with my dad and prepared the slides for a talk I gave on Monday, “The Xen Hypervisor and Its IO Subsystem” to the storage research seminar at work.

On Monday I woke up early to finish the slides, and then drove to Tel-Aviv to give the talk. The trip took 2.5 hours (as opposed to an hour or so normally) because of massive traffic jams. Eventually I found the site, gave the talk, and then had lunch with the team and discussed a pretty cool project we’re considering for next year (more details later; for the time being suffice to say it involves Xen and storage). Then I drove back to the office, worked for a couple of hours (for values of work that mostly involve transatlantic phone calls), and headed to Haifux.

At Haifux Ami Chayun gave a summary of his visit to the Blackhat Briefings and DEFCON, concentrating on ciscogate and a few other interesting talks. Since I’m reasonably familiar with Michael Lynn’s work, I spent most of the talk trying to figure out how Xen’s save and restore mechanism works, and how to fix it, since it’s currently broken. And that’s about all the news that’s fit to print!

Looks like there’s going to be another Xen summit in Austin in January. I wonder who else is going?

Oh yeah, one thing I wanted to write about: we went to see a play yesterday, “Six Dance Lessons in Six Weeks” (Hebrew review). It was surprisingly good, albeit somewhat depressing for a comedy. A bitter-sweet comedy, if you will. Great acting by Natan Datner.

We had balcony seats, so Orna brought along her bird-watching, field-grade binoculars. They were so good it was like being right there on stage. We could see every nuance, every facial expression, every wrinkle. Sometimes, a little blurriness is a good thing.

I was feeling really lousy this morning, then felt great for a few hours, and now I’m feeling lousy again. I’m sure there’s something I could do to regulate these wild mood swings, but I have no idea what.

I was going to write a bunch of other things, but I no longer remember what they were. Such is life.

It looks like I’m headed to the US again in a couple of weeks. I’m starting to get used to this.

Every great journey begins with a small step. Mine actually begins with two: a morning workout at home yesterday, and then a long overdue visit to the gym. While suffer^H^H^H^H^H^Hexercising on the treadmill, I listened to this CD to get in the mood. It worked.

On Friday I switched our home connection from ADSL to cable, since we’re getting cable practically for free from work and ADSL is decidedly non-free. I’ve been resisting this change for a while, since I have somewhat of an emotional attachment to ADSL, but Orna has been pushing for it and I gave up. Cable works quite well so far.

Then I continued working on the DMA mapping patch and after testing it on a bunch of machines sent the first version (-C1) to lkml.

In the evening we caught Kate and Leopold on TV, which was surprisingly enjoyable.

On Saturday, I woke up early and submitted a follow up Xen patch to this bug. The Cambridge guys committed both my and Jeremy Katz’s patches, so the follow-up reverted mine (Jeremy’s was much prettier). Then we drove to Tel Aviv for lunch at Orna’s sister’s and her boyfriend’s apartment. Lunch was highly enjoyable and their apartment is gorgeous.

In the afternoon we visited omerm and shapirac, and had a horribly geeky discussion. Much fun was had. shapirac mentioned that my PNS posts used to motivate her to work out, so mayhap I’ll resuscitate them. $DEITY knows I need the exercise.

When we got home Orna cooked an amazing dinner, and then I watched “quality TV”, aka The Wire. While watching (perhaps during the detective wake?) I realized that there was a way to eradicate one of the warts of the dma-mapping-ops patch.

The way I structured dma_xxx() is that they call a dma_mapping function if one exists, and otherwise call the default gart_xxx() function. This was done in order to keep the fast path fast (avoid a function pointer call). Unfortunately, for the nommu case, where no gart exists, this requires us to link in gart_xxx(), even though it’s never called. The solution is obvious, even if it did take me a week or so to come up with it – provide empty definitions of gart_xxx() if CONFIG_GART_IOMMU is not defined. While implementing it, I also discovered a couple of buglets – nommu_map_sg() and nommu_unmap_sg() were not getting called, and I erroneously removed a couple of EXPORT_SYMBOLS(). Fixed, tested, and released swiotlb-dma-mapping-ops-D1.

And that’s about it. How was your weekend, gentle reader?

What I did on the holidays: [RFC PATCH] clean up x86_64 DMA mapping dispatching. Also, various Xen bits.

Holy batman, I actually agree with Joel Spolsky!