4 ms·
My understanding is this is primarily about driver code. Is that correct? One way out for VMWare would be to import FreeBSD drivers, although that wouldn't do
by tedunangst 12y ago
My understanding is this is primarily about driver code. Is that correct?
One way out for VMWare would be to import FreeBSD drivers, although that wouldn't do the Linux community much good. From what I've read, VMWare does provide the source for the drivers, but not the core of their kernel, so Linux would still benefit from any improvements they make. If they switch to another set of drivers, that won't happen.
I can understand the principle of the position, but given that releasing the crown jewel core of the vmkernel as GPL seems unlikely, I'm not sure what actual benefits will be realized here. Something of a Pyrrhic victory.
- masamune__ 12y agoIt's not driver itself but clever module that VMware developed so they can use linux drivers. (In prev thread : https://news.ycombinator.com/item?id=9153716 https://news.ycombinator.com/item?id=9153716) But VMK API is not just wrapper for linux drivers. You can specifically write device driver for ESXI, like GPU : https://developercenter.vmware.com/web/dp/sdks https://developercenter.vmware.com/web/dp/sdks According to LWN's article they found linux kernel's source (radix tree, scheduler) https://lwn.net/SubscriberLink/635290/e501ce0264c182f4/ https://lwn.net/SubscriberLink/635290/e501ce0264c182f4/ But it's not exactly clear if those codes were actually built and shipped with product according to 'former employ' https://news.ycombinator.com/item?id=9155251 https://news.ycombinator.com/item?id=9155251 This is all cluster*uck of wild guessworks unless they elaborate more on VMware's response or VMware makes formal statements on their position. What a waste of time
- travem 12y ago> "According to LWN's article they found linux kernel's source (radix tree, scheduler)" [1]. Presumably when LWN article states "code can be downloaded from VMware's web site" they got it from the open source tab on VMware vSphere's download page [2]. [1] https://lwn.net/SubscriberLink/635290/e501ce0264c182f4/ https://lwn.net/SubscriberLink/635290/e501ce0264c182f4/ [2] https://my.vmware.com/web/vmware/info/slug/datacenter_cloud_infrastructure/vmware_vsphere/5_5#open_source https://my.vmware.com/web/vmware/info/slug/datacenter_cloud_...
- belorn 12y agoAs per the FAQ, this is much more than only drivers (http://sfconservancy.org/linux-compliance/linux-vs-vmkernel_en.png http://sfconservancy.org/linux-compliance/linux-vs-vmkernel_...). It both about the internal code that makes up the vmklinux kernel, and drivers. The radix tree implementation for example is unlikely provided through a driver api. The bigger question I think is how a non-technical judge will think when presented with a description of the vmkernel. Does it constitute one work, or is it a larger number of isolated small works that simply happens to work together in a independent fashion? Are the drivers in question separateable from the kernel and the finish "product" that wmware sells? Are wmware following what can be reasonable understood as the intended use of the licensed work?