| Message ID | 20100422161224.465958039@suse.de (mailing list archive) |
|---|---|
| State | Accepted, archived |
| Delegated to: | Benjamin Herrenschmidt |
| Headers | show |
On Fri, 23 Apr 2010, npiggin@suse.de wrote: > As explained in commit 1c0fe6e3bd, we want to call the architecture independent > oom killer when getting an unexplained OOM from handle_mm_fault, rather than > simply killing current. > > Cc: linuxppc-dev@ozlabs.org > Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> > Cc: linux-arch@vger.kernel.org > Signed-off-by: Nick Piggin <npiggin@suse.de> > --- > Index: linux-2.6/arch/powerpc/mm/fault.c > =================================================================== > --- linux-2.6.orig/arch/powerpc/mm/fault.c > +++ linux-2.6/arch/powerpc/mm/fault.c > @@ -359,15 +359,10 @@ bad_area_nosemaphore: > */ > out_of_memory: > up_read(&mm->mmap_sem); > - if (is_global_init(current)) { > - yield(); > - down_read(&mm->mmap_sem); > - goto survive; > - } > - printk("VM: killing process %s\n", current->comm); > - if (user_mode(regs)) > - do_group_exit(SIGKILL); > - return SIGKILL; > + if (!user_mode(regs)) > + return SIGKILL; > + pagefault_out_of_memory(); > + return 0; Do we really want to return 0 and indicate that the fault was handled? It seems more consistent to do if (user_mode(regs)) pagefault_out_of_memory(); return SIGKILL;
On Thu, Apr 22, 2010 at 02:44:22PM -0700, David Rientjes wrote: > On Fri, 23 Apr 2010, npiggin@suse.de wrote: > > > As explained in commit 1c0fe6e3bd, we want to call the architecture independent > > oom killer when getting an unexplained OOM from handle_mm_fault, rather than > > simply killing current. > > > > Cc: linuxppc-dev@ozlabs.org > > Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> > > Cc: linux-arch@vger.kernel.org > > Signed-off-by: Nick Piggin <npiggin@suse.de> > > --- > > Index: linux-2.6/arch/powerpc/mm/fault.c > > =================================================================== > > --- linux-2.6.orig/arch/powerpc/mm/fault.c > > +++ linux-2.6/arch/powerpc/mm/fault.c > > @@ -359,15 +359,10 @@ bad_area_nosemaphore: > > */ > > out_of_memory: > > up_read(&mm->mmap_sem); > > - if (is_global_init(current)) { > > - yield(); > > - down_read(&mm->mmap_sem); > > - goto survive; > > - } > > - printk("VM: killing process %s\n", current->comm); > > - if (user_mode(regs)) > > - do_group_exit(SIGKILL); > > - return SIGKILL; > > + if (!user_mode(regs)) > > + return SIGKILL; > > + pagefault_out_of_memory(); > > + return 0; > > Do we really want to return 0 and indicate that the fault was handled? It > seems more consistent to do > > if (user_mode(regs)) > pagefault_out_of_memory(); > return SIGKILL; Well we don't necessarily want to kill current. pagefault_out_of_memory() could elect to send a SIGKILL if wants to, but if it decides current should not be terminated, then we have to just retry the page fault.
Index: linux-2.6/arch/powerpc/mm/fault.c =================================================================== --- linux-2.6.orig/arch/powerpc/mm/fault.c +++ linux-2.6/arch/powerpc/mm/fault.c @@ -359,15 +359,10 @@ bad_area_nosemaphore: */ out_of_memory: up_read(&mm->mmap_sem); - if (is_global_init(current)) { - yield(); - down_read(&mm->mmap_sem); - goto survive; - } - printk("VM: killing process %s\n", current->comm); - if (user_mode(regs)) - do_group_exit(SIGKILL); - return SIGKILL; + if (!user_mode(regs)) + return SIGKILL; + pagefault_out_of_memory(); + return 0; do_sigbus: up_read(&mm->mmap_sem);
As explained in commit 1c0fe6e3bd, we want to call the architecture independent oom killer when getting an unexplained OOM from handle_mm_fault, rather than simply killing current. Cc: linuxppc-dev@ozlabs.org Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: linux-arch@vger.kernel.org Signed-off-by: Nick Piggin <npiggin@suse.de> ---