From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753948Ab2GJVdv (ORCPT ); Tue, 10 Jul 2012 17:33:51 -0400 Received: from mail.linuxfoundation.org ([140.211.169.12]:46667 "EHLO mail.linuxfoundation.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751138Ab2GJVdu (ORCPT ); Tue, 10 Jul 2012 17:33:50 -0400 Date: Tue, 10 Jul 2012 14:33:48 -0700 From: Andrew Morton To: Octavian Purdila Cc: "H. Peter Anvin" , linux-kernel@vger.kernel.org, Ram Pai , Jesse Barnes Subject: Re: [PATCH] resource: make sure requested range intersects root range Message-Id: <20120710143348.d977da44.akpm@linux-foundation.org> In-Reply-To: <1341057657-7823-1-git-send-email-octavian.purdila@intel.com> References: <1341057657-7823-1-git-send-email-octavian.purdila@intel.com> X-Mailer: Sylpheed 3.0.2 (GTK+ 2.20.1; x86_64-pc-linux-gnu) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Sat, 30 Jun 2012 15:00:57 +0300 Octavian Purdila wrote: > When the requested and root ranges do not intersect the logic in > __reserve_region_with_split will cause an infinite recursion which > will overflow the stack as seen in the warning bellow. > > This particular stack overflow was caused by requesting the > (100000000-107ffffff) range while the root range was (0-ffffffff). In > this case __request_resource would return the whole root range as > conflict range (i.e. 0-ffffffff). Then, the logic in > __reserve_region_with_split would continue the recursion requesting > the new range as (conflict->end+1, end) which incidentally in this > case equals the originally requested range. > > This patch aborts looking for a usable range when the requested one is > completely outside the root range to avoid the infinite recursion, and > since this indicates a problem in the layers above, it also prints an > error message indicating the requested and root range in order to make > the problem more easily traceable. I think we should also emit a stack trace so the faulty caller can be pinpointed. > ... > > --- a/kernel/resource.c > +++ b/kernel/resource.c > @@ -789,7 +789,13 @@ void __init reserve_region_with_split(struct resource *root, > const char *name) > { > write_lock(&resource_lock); > - __reserve_region_with_split(root, start, end, name); > + if (start > root->end || end < root->start) > + pr_err("Requested range (0x%llx-0x%llx) not in root range (0x%llx-0x%llx)\n", > + (unsigned long long)start, (unsigned long long)end, > + (unsigned long long)root->start, > + (unsigned long long)root->end); > + else > + __reserve_region_with_split(root, start, end, name); > write_unlock(&resource_lock); > } The fancy way of doing that is if (!WARN(start > root->end || end < root->start), "Requested range (0x%llx-0x%llx) not in root range (0x%llx-0x%llx)\n", (unsigned long long)start, (unsigned long long)end, (unsigned long long)root->start, (unsigned long long)root->end) __reserve_region_with_split(root, start, end, name); but that's quite the eyesore. How about doing it the simple way? --- a/kernel/resource.c~resource-make-sure-requested-range-intersects-root-range-fix +++ a/kernel/resource.c @@ -792,13 +792,15 @@ void __init reserve_region_with_split(st const char *name) { write_lock(&resource_lock); - if (start > root->end || end < root->start) + if (start > root->end || end < root->start) { pr_err("Requested range (0x%llx-0x%llx) not in root range (0x%llx-0x%llx)\n", (unsigned long long)start, (unsigned long long)end, (unsigned long long)root->start, (unsigned long long)root->end); - else + dump_stack(); + } else { __reserve_region_with_split(root, start, end, name); + } write_unlock(&resource_lock); } _