On 02.02.2007 [10:32:40 -0600], Adam Litke wrote:
> On Fri, 2007-02-02 at 10:05 -0600, Steve Fox wrote:
> > On Fri, 2007-02-02 at 09:40 -0600, Adam Litke wrote:
> > 
> > > As of now, the extra_copy window abuts the filesz portion of the
> > > segment.  Therefore, we only really need to store extrasz.  The
> > > pagecache dropping code can then construct a for loop like the
> > > following:
> > > 
> > > for (p = vaddr; p < (vaddr + filesz + extrasz); p += HPAGE_SIZE) ...
> > 
> > While in practice that may occur frequently, possibly all the time, it
> > doesn't _have_ to. Remember that we are simply looking for the first
> > global symbol, which doesn't have to be immediately after filesz by
> > definition. We can theoretically end up with something like this:
> > 
> > |---------------|-----<--extra_copy-->----|
> >  filesz          memsz
> 
> Ok, fair enough.  I forgot that the extracopy code bumps moves start
> and end pointers for the copy window when examining the dynamic symbol
> table.  

Yep, this was the reason the code is a bit more complex than what you
had originally suggested. The current extracopy code definitely assumes
there can be a "gap" between the filesz and extracopy portions.

> ... Pay no mind to the muttering man in the corner ...  :)

Never do, except at pool halls.

-- 
Nishanth Aravamudan <[EMAIL PROTECTED]>
IBM Linux Technology Center

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier.
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
Libhugetlbfs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/libhugetlbfs-devel

Reply via email to