Comments on: JOS Virtual Memory (part 2) http://blog.cpu.eti.br/2008/05/05/jos-virtual-memory-part-2/ Tue, 06 Jul 2010 12:43:16 -0200 http://wordpress.org/?v=2.8.3 hourly 1 By: Nencia http://blog.cpu.eti.br/2008/05/05/jos-virtual-memory-part-2/comment-page-1/#comment-56474 Nencia Fri, 06 Feb 2009 13:03:55 +0000 http://blog.cpu.eti.br/?p=99#comment-56474 Nice work! I’ll have to do a cross post on this one ;)

]]> By: Luiz Capitulino http://blog.cpu.eti.br/2008/05/05/jos-virtual-memory-part-2/comment-page-1/#comment-49537 Luiz Capitulino Mon, 08 Sep 2008 14:33:22 +0000 http://blog.cpu.eti.br/?p=99#comment-49537 Yes, you are right. The text pages are not copied, they will be shared among descendants. This is, of course, a general rule. I mean, today's operating systems have so many features that sometimes they may not follow such a rule strictly. I think what confused you was this sentence: "the kernel marks all the parent’s pages as read-only and inserts them into child address space". Speaking about JOS, the pages which are marked read-only are the the ones which have write permission enabled. The others (eg. text) are already read-only, so they are directly inserted into the child's address space. I'm replying you here (instead of sending an email) because your question may be relevant to others. I'm glad to know that this post has been useful to someone. :) Yes, you are right. The text pages are not copied, they will be shared among descendants.

This is, of course, a general rule. I mean, today’s operating systems have so many features that sometimes they may not follow such a rule strictly.

I think what confused you was this sentence: “the kernel marks all the parent’s pages as read-only and inserts them into child address space”. Speaking about JOS, the pages which are marked read-only are the the ones which have write permission enabled. The others (eg. text) are already read-only, so they are directly inserted into the child’s address space.

I’m replying you here (instead of sending an email) because your question may be relevant to others.

I’m glad to know that this post has been useful to someone. :)

]]> By: vivek http://blog.cpu.eti.br/2008/05/05/jos-virtual-memory-part-2/comment-page-1/#comment-49460 vivek Sun, 07 Sep 2008 06:05:29 +0000 http://blog.cpu.eti.br/?p=99#comment-49460 thanks luiz, it was pretty informative. i was struggling to find this bit of information on internet. there are lots of references of this on the web, but none of them mention about the mmu and pagefaults. and the generally mess up the COW and read only markers! i have one more doubt. does the fork mark (cow and read only) pages which are non read-only (like data segment,stack) at the begining? because if the child is going to be the exact copy of the parent for its lifetime, there is no point in copying the parent's text segment. thank you please mail me thanks luiz,

it was pretty informative. i was struggling to find this bit of information on internet.
there are lots of references of this on the web, but none of them mention about the mmu and pagefaults. and the generally mess up the COW and read only markers!

i have one more doubt.

does the fork mark (cow and read only) pages which are non read-only (like data segment,stack) at the begining?

because if the child is going to be the exact copy of the parent for its lifetime, there is no point in copying the parent’s text segment.

thank you
please mail me

]]>