tag:blogger.com,1999:blog-2590786243514585895.post5574383097978416957..comments2022-03-28T13:41:03.207+01:00Comments on Camltastic!: virt-ps and the Red Hat SummitRichard Joneshttp://www.blogger.com/profile/08315526595922432607noreply@blogger.comBlogger2125tag:blogger.com,1999:blog-2590786243514585895.post-70717094730688316762008-06-17T17:07:00.000+01:002008-06-17T17:07:00.000+01:00Jeremy:Yes, and CONFIG_* makes a difference too. ...Jeremy:<BR/><BR/>Yes, and CONFIG_* makes a difference too. We still use heuristics to determine the precise location of pointers within the structure. The interesting pointers are the ones which point to (list_heads of) other task_structs, so one can make assertions about where pointers lie, test them, and backtrack if found not to be correct. If that gets too slow, then we have the full utsname so using tables for each vendor is a possibility too, automation being the key to make that feasible.Richard Joneshttps://www.blogger.com/profile/08315526595922432607noreply@blogger.comtag:blogger.com,1999:blog-2590786243514585895.post-82196729967026575002008-06-17T16:56:00.000+01:002008-06-17T16:56:00.000+01:00Of course, the kernel version isn't definitive giv...Of course, the kernel version isn't definitive given vendor patching... unless you're going to have a table with how various vendors patch things.Jeremyhttps://www.blogger.com/profile/06689752052867722832noreply@blogger.com