On 18.01.2019 10:03, David Holland wrote: > On Fri, Jan 18, 2019 at 09:50:38AM +0100, Martin Husemann wrote: > > Sorry, I completely fail to parse this - can you start from scratch and > > just describe the problem you think you are seeing? > > My understanding is: > > - If you map a file such that part of the map is past EOF, > accessing that part should cause SIGBUS. > - It does. > - However, you don't get a coredump because the coredump code fails > on this region. > - The proposed patch works around this by writing out a page of > zeros when copyin_proc() returns EINVAL, instead of failing the > dump. > > This also implies that you can't get a coredump from a process that > has a file mapped past EOF, regardless of what triggers the coredump. > Right. > ISTM that it would better to skip the page than write out zeros... > This behavior is also fine. Should it be - in this example - a zero-sized segment?
Attachment:
signature.asc
Description: OpenPGP digital signature