From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-13.6 required=3.0 tests=BAYES_00,DKIM_INVALID, DKIM_SIGNED,HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id ABE57C433E0 for ; Mon, 22 Feb 2021 18:48:33 +0000 (UTC) Received: from kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by mail.kernel.org (Postfix) with ESMTP id 0EAFA6148E for ; Mon, 22 Feb 2021 18:48:32 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 0EAFA6148E Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=linux.ibm.com Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=owner-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix) id 6847F8D0001; Mon, 22 Feb 2021 13:48:32 -0500 (EST) Received: by kanga.kvack.org (Postfix, from userid 40) id 60C846B0074; Mon, 22 Feb 2021 13:48:32 -0500 (EST) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id 460608D0001; Mon, 22 Feb 2021 13:48:32 -0500 (EST) X-Delivered-To: linux-mm@kvack.org Received: from forelay.hostedemail.com (smtprelay0081.hostedemail.com [216.40.44.81]) by kanga.kvack.org (Postfix) with ESMTP id 0F4216B0073 for ; Mon, 22 Feb 2021 13:48:32 -0500 (EST) Received: from smtpin27.hostedemail.com (10.5.19.251.rfc1918.com [10.5.19.251]) by forelay05.hostedemail.com (Postfix) with ESMTP id A651918011AB8 for ; Mon, 22 Feb 2021 18:48:31 +0000 (UTC) X-FDA: 77846789622.27.55FFE21 Received: from mx0a-001b2d01.pphosted.com (mx0a-001b2d01.pphosted.com [148.163.156.1]) by imf24.hostedemail.com (Postfix) with ESMTP id DE5E0A00085B for ; Mon, 22 Feb 2021 18:48:25 +0000 (UTC) Received: from pps.filterd (m0098396.ppops.net [127.0.0.1]) by mx0a-001b2d01.pphosted.com (8.16.0.43/8.16.0.43) with SMTP id 11MIgiTv158848; Mon, 22 Feb 2021 13:47:48 -0500 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=ibm.com; h=date : from : to : cc : subject : message-id : references : mime-version : content-type : content-transfer-encoding : in-reply-to; s=pp1; bh=7g8nMPK0BWDP02m4h8rVpPLI++nL8C+0dAkk5BW4xow=; b=IFr69xQANSi7a0VXN4CAuSGV0DIWl7KVH/9GpyBcx8tGQhaEhkIhP7lQMfGpYsuR9JmP NtnrPrOOMiOwwn+REYhlyoOy5M1aXHxCcKwLqP60oZIof5jSckoo+GBWPK7vzsE3EpNo 4NYUJpWD6PbrTaYf/dwJsbyhl+EWNDUyFifJ/nkGgYAWmzb5YzWEXZfMPVVV6liCa8jz cXrMG/R03sECr410VASMeMTd+md4KuV9ZVv99fDsPtS1QqBXIM5pOEbK2rvnfYk+806t CQ2JJeg08eOnGSmRTrZMW2KJne+DscipysdAQ5DL2Y162jepp9/4yQwTQkt62Rxye20t tg== Received: from pps.reinject (localhost [127.0.0.1]) by mx0a-001b2d01.pphosted.com with ESMTP id 36vhy78k8s-1 (version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384 bits=256 verify=NOT); Mon, 22 Feb 2021 13:47:47 -0500 Received: from m0098396.ppops.net (m0098396.ppops.net [127.0.0.1]) by pps.reinject (8.16.0.43/8.16.0.43) with SMTP id 11MIgvj2159799; Mon, 22 Feb 2021 13:47:33 -0500 Received: from ppma04ams.nl.ibm.com (63.31.33a9.ip4.static.sl-reverse.com [169.51.49.99]) by mx0a-001b2d01.pphosted.com with ESMTP id 36vhy78j7x-1 (version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384 bits=256 verify=NOT); Mon, 22 Feb 2021 13:47:32 -0500 Received: from pps.filterd (ppma04ams.nl.ibm.com [127.0.0.1]) by ppma04ams.nl.ibm.com (8.16.0.42/8.16.0.42) with SMTP id 11MIgCZM014626; Mon, 22 Feb 2021 18:45:52 GMT Received: from b06cxnps4074.portsmouth.uk.ibm.com (d06relay11.portsmouth.uk.ibm.com [9.149.109.196]) by ppma04ams.nl.ibm.com with ESMTP id 36tt289wmu-1 (version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384 bits=256 verify=NOT); Mon, 22 Feb 2021 18:45:51 +0000 Received: from d06av25.portsmouth.uk.ibm.com (d06av25.portsmouth.uk.ibm.com [9.149.105.61]) by b06cxnps4074.portsmouth.uk.ibm.com (8.14.9/8.14.9/NCO v10.0) with ESMTP id 11MIjnkg43712902 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-GCM-SHA384 bits=256 verify=OK); Mon, 22 Feb 2021 18:45:49 GMT Received: from d06av25.portsmouth.uk.ibm.com (unknown [127.0.0.1]) by IMSVA (Postfix) with ESMTP id AFA0911C05B; Mon, 22 Feb 2021 18:45:49 +0000 (GMT) Received: from d06av25.portsmouth.uk.ibm.com (unknown [127.0.0.1]) by IMSVA (Postfix) with ESMTP id BCDF211C04A; Mon, 22 Feb 2021 18:45:45 +0000 (GMT) Received: from linux.ibm.com (unknown [9.145.51.238]) by d06av25.portsmouth.uk.ibm.com (Postfix) with ESMTPS; Mon, 22 Feb 2021 18:45:45 +0000 (GMT) Date: Mon, 22 Feb 2021 20:45:43 +0200 From: Mike Rapoport To: Konrad Rzeszutek Wilk Cc: David Hildenbrand , George Kennedy , Andrey Konovalov , Andrew Morton , Catalin Marinas , Vincenzo Frascino , Dmitry Vyukov , Will Deacon , Andrey Ryabinin , Alexander Potapenko , Marco Elver , Peter Collingbourne , Evgenii Stepanov , Branislav Rankov , Kevin Brodsky , Christoph Hellwig , kasan-dev , Linux ARM , Linux Memory Management List , LKML , Dhaval Giani Subject: Re: [PATCH] mm, kasan: don't poison boot memory Message-ID: <20210222184543.GA1741768@linux.ibm.com> References: <797fae72-e3ea-c0b0-036a-9283fa7f2317@oracle.com> <1ac78f02-d0af-c3ff-cc5e-72d6b074fc43@redhat.com> <56c97056-6d8b-db0e-e303-421ee625abe3@redhat.com> <4c7351e2-e97c-e740-5800-ada5504588aa@redhat.com> <20210222174036.GA399355@fedora> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Disposition: inline In-Reply-To: <20210222174036.GA399355@fedora> X-TM-AS-GCONF: 00 X-Proofpoint-Virus-Version: vendor=fsecure engine=2.50.10434:6.0.369,18.0.761 definitions=2021-02-22_06:2021-02-22,2021-02-22 signatures=0 X-Proofpoint-Spam-Details: rule=outbound_notspam policy=outbound score=0 impostorscore=0 mlxscore=0 phishscore=0 mlxlogscore=999 spamscore=0 suspectscore=0 adultscore=0 clxscore=1011 priorityscore=1501 lowpriorityscore=0 malwarescore=0 bulkscore=0 classifier=spam adjust=0 reason=mlx scancount=1 engine=8.12.0-2009150000 definitions=main-2102220163 X-Stat-Signature: tpo163hy691j73tfqk347re9yoz415te X-Rspamd-Server: rspam05 X-Rspamd-Queue-Id: DE5E0A00085B Received-SPF: none (linux.ibm.com>: No applicable sender policy available) receiver=imf24; identity=mailfrom; envelope-from=""; helo=mx0a-001b2d01.pphosted.com; client-ip=148.163.156.1 X-HE-DKIM-Result: pass/pass X-HE-Tag: 1614019705-269143 Content-Transfer-Encoding: quoted-printable X-Bogosity: Ham, tests=bogofilter, spamicity=0.000000, version=1.2.4 Sender: owner-linux-mm@kvack.org Precedence: bulk X-Loop: owner-majordomo@kvack.org List-ID: On Mon, Feb 22, 2021 at 12:40:36PM -0500, Konrad Rzeszutek Wilk wrote: > On Mon, Feb 22, 2021 at 05:39:29PM +0100, David Hildenbrand wrote: > > On 22.02.21 17:13, David Hildenbrand wrote: > > > On 22.02.21 16:13, George Kennedy wrote: > > > >=20 > > > >=20 > > > > On 2/22/2021 4:52 AM, David Hildenbrand wrote: > > > > > On 20.02.21 00:04, George Kennedy wrote: > > > > > >=20 > > > > > >=20 > > > > > > On 2/19/2021 11:45 AM, George Kennedy wrote: > > > > > > >=20 > > > > > > >=20 > > > > > > > On 2/18/2021 7:09 PM, Andrey Konovalov wrote: > > > > > > > > On Fri, Feb 19, 2021 at 1:06 AM George Kennedy > > > > > > > > wrote: > > > > > > > > >=20 > > > > > > > > >=20 > > > > > > > > > On 2/18/2021 3:55 AM, David Hildenbrand wrote: > > > > > > > > > > On 17.02.21 21:56, Andrey Konovalov wrote: > > > > > > > > > > > During boot, all non-reserved memblock memory is ex= posed to the > > > > > > > > > > > buddy > > > > > > > > > > > allocator. Poisoning all that memory with KASAN len= gthens boot > > > > > > > > > > > time, > > > > > > > > > > > especially on systems with large amount of RAM. Thi= s patch makes > > > > > > > > > > > page_alloc to not call kasan_free_pages() on all ne= w memory. > > > > > > > > > > >=20 > > > > > > > > > > > __free_pages_core() is used when exposing fresh mem= ory during > > > > > > > > > > > system > > > > > > > > > > > boot and when onlining memory during hotplug. This = patch adds a new > > > > > > > > > > > FPI_SKIP_KASAN_POISON flag and passes it to __free_= pages_ok() > > > > > > > > > > > through > > > > > > > > > > > free_pages_prepare() from __free_pages_core(). > > > > > > > > > > >=20 > > > > > > > > > > > This has little impact on KASAN memory tracking. > > > > > > > > > > >=20 > > > > > > > > > > > Assuming that there are no references to newly expo= sed pages > > > > > > > > > > > before they > > > > > > > > > > > are ever allocated, there won't be any intended (bu= t buggy) > > > > > > > > > > > accesses to > > > > > > > > > > > that memory that KASAN would normally detect. > > > > > > > > > > >=20 > > > > > > > > > > > However, with this patch, KASAN stops detecting wil= d and large > > > > > > > > > > > out-of-bounds accesses that happen to land on a fre= sh memory page > > > > > > > > > > > that > > > > > > > > > > > was never allocated. This is taken as an acceptable= trade-off. > > > > > > > > > > >=20 > > > > > > > > > > > All memory allocated normally when the boot is over= keeps getting > > > > > > > > > > > poisoned as usual. > > > > > > > > > > >=20 > > > > > > > > > > > Signed-off-by: Andrey Konovalov > > > > > > > > > > > Change-Id: Iae6b1e4bb8216955ffc14af255a7eaaa6f35324= d > > > > > > > > > > Not sure this is the right thing to do, see > > > > > > > > > >=20 > > > > > > > > > > https://lkml.kernel.org/r/bcf8925d-0949-3fe1-baa8-cc5= 36c529860@oracle.com > > > > > > > > > >=20 > > > > > > > > > >=20 > > > > > > > > > >=20 > > > > > > > > > > Reversing the order in which memory gets allocated + = used during > > > > > > > > > > boot > > > > > > > > > > (in a patch by me) might have revealed an invalid mem= ory access > > > > > > > > > > during > > > > > > > > > > boot. > > > > > > > > > >=20 > > > > > > > > > > I suspect that that issue would no longer get detecte= d with your > > > > > > > > > > patch, as the invalid memory access would simply not = get detected. > > > > > > > > > > Now, I cannot prove that :) > > > > > > > > > Since David's patch we're having trouble with the iBFT = ACPI table, > > > > > > > > > which > > > > > > > > > is mapped in via kmap() - see acpi_map() in "drivers/ac= pi/osl.c". > > > > > > > > > KASAN > > > > > > > > > detects that it is being used after free when ibft_init= () accesses > > > > > > > > > the > > > > > > > > > iBFT table, but as of yet we can't find where it get's = freed (we've > > > > > > > > > instrumented calls to kunmap()). > > > > > > > > Maybe it doesn't get freed, but what you see is a wild or= a large > > > > > > > > out-of-bounds access. Since KASAN marks all memory as fre= ed during the > > > > > > > > memblock->page_alloc transition, such bugs can manifest a= s > > > > > > > > use-after-frees. > > > > > > >=20 > > > > > > > It gets freed and re-used. By the time the iBFT table is ac= cessed by > > > > > > > ibft_init() the page has been over-written. > > > > > > >=20 > > > > > > > Setting page flags like the following before the call to km= ap() > > > > > > > prevents the iBFT table page from being freed: > > > > > >=20 > > > > > > Cleaned up version: > > > > > >=20 > > > > > > diff --git a/drivers/acpi/osl.c b/drivers/acpi/osl.c > > > > > > index 0418feb..8f0a8e7 100644 > > > > > > --- a/drivers/acpi/osl.c > > > > > > +++ b/drivers/acpi/osl.c > > > > > > @@ -287,9 +287,12 @@ static void __iomem *acpi_map(acpi_physi= cal_address > > > > > > pg_off, unsigned long pg_sz) > > > > > >=20 > > > > > > =A0 =A0=A0=A0=A0 pfn =3D pg_off >> PAGE_SHIFT; > > > > > > =A0 =A0=A0=A0=A0 if (should_use_kmap(pfn)) { > > > > > > +=A0=A0=A0 =A0=A0=A0 struct page *page =3D pfn_to_page(pfn); > > > > > > + > > > > > > =A0 =A0=A0=A0=A0 =A0=A0=A0 if (pg_sz > PAGE_SIZE) > > > > > > =A0 =A0=A0=A0=A0 =A0=A0=A0 =A0=A0=A0 return NULL; > > > > > > -=A0=A0=A0 =A0=A0=A0 return (void __iomem __force *)kmap(pfn_= to_page(pfn)); > > > > > > +=A0=A0=A0 =A0=A0=A0 SetPageReserved(page); > > > > > > +=A0=A0=A0 =A0=A0=A0 return (void __iomem __force *)kmap(page= ); > > > > > > =A0 =A0=A0=A0=A0 } else > > > > > > =A0 =A0=A0=A0=A0 =A0=A0=A0 return acpi_os_ioremap(pg_off, p= g_sz); > > > > > > =A0 =A0} > > > > > > @@ -299,9 +302,12 @@ static void acpi_unmap(acpi_physical_add= ress > > > > > > pg_off, void __iomem *vaddr) > > > > > > =A0 =A0=A0=A0=A0 unsigned long pfn; > > > > > >=20 > > > > > > =A0 =A0=A0=A0=A0 pfn =3D pg_off >> PAGE_SHIFT; > > > > > > -=A0=A0=A0 if (should_use_kmap(pfn)) > > > > > > -=A0=A0=A0 =A0=A0=A0 kunmap(pfn_to_page(pfn)); > > > > > > -=A0=A0=A0 else > > > > > > +=A0=A0=A0 if (should_use_kmap(pfn)) { > > > > > > +=A0=A0=A0 =A0=A0=A0 struct page *page =3D pfn_to_page(pfn); > > > > > > + > > > > > > +=A0=A0=A0 =A0=A0=A0 ClearPageReserved(page); > > > > > > +=A0=A0=A0 =A0=A0=A0 kunmap(page); > > > > > > +=A0=A0=A0 } else > > > > > > =A0 =A0=A0=A0=A0 =A0=A0=A0 iounmap(vaddr); > > > > > > =A0 =A0} > > > > > >=20 > > > > > > David, the above works, but wondering why it is now necessary= . kunmap() > > > > > > is not hit. What other ways could a page mapped via kmap() be= unmapped? > > > > > >=20 > > > > >=20 > > > > > Let me look into the code ... I have little experience with ACP= I > > > > > details, so bear with me. > > > > >=20 > > > > > I assume that acpi_map()/acpi_unmap() map some firmware blob th= at is > > > > > provided via firmware/bios/... to us. > > > > >=20 > > > > > should_use_kmap() tells us whether > > > > > a) we have a "struct page" and should kmap() that one > > > > > b) we don't have a "struct page" and should ioremap. > > > > >=20 > > > > > As it is a blob, the firmware should always reserve that memory= region > > > > > via memblock (e.g., memblock_reserve()), such that we either > > > > > 1) don't create a memmap ("struct page") at all (-> case b) ) > > > > > 2) if we have to create e memmap, we mark the page PG_reserved = and > > > > > =A0=A0 *never* expose it to the buddy (-> case a) ) > > > > >=20 > > > > >=20 > > > > > Are you telling me that in this case we might have a memmap for= the HW > > > > > blob that is *not* PG_reserved? In that case it most probably g= ot > > > > > exposed to the buddy where it can happily get allocated/freed. > > > > >=20 > > > > > The latent BUG would be that that blob gets exposed to the syst= em like > > > > > ordinary RAM, and not reserved via memblock early during boot. > > > > > Assuming that blob has a low physical address, with my patch it= will > > > > > get allocated/used a lot earlier - which would mean we trigger = this > > > > > latent BUG now more easily. > > > > >=20 > > > > > There have been similar latent BUGs on ARM boards that my patch > > > > > discovered where special RAM regions did not get marked as rese= rved > > > > > via the device tree properly. > > > > >=20 > > > > > Now, this is just a wild guess :) Can you dump the page when ma= pping > > > > > (before PageReserved()) and when unmapping, to see what the sta= te of > > > > > that memmap is? > > > >=20 > > > > Thank you David for the explanation and your help on this, > > > >=20 > > > > dump_page() before PageReserved and before kmap() in the above pa= tch: > > > >=20 > > > > [=A0=A0=A0 1.116480] ACPI: Core revision 20201113 > > > > [=A0=A0=A0 1.117628] XXX acpi_map: about to call kmap()... > > > > [=A0=A0=A0 1.118561] page:ffffea0002f914c0 refcount:0 mapcount:0 > > > > mapping:0000000000000000 index:0x0 pfn:0xbe453 > > > > [=A0=A0=A0 1.120381] flags: 0xfffffc0000000() > > > > [=A0=A0=A0 1.121116] raw: 000fffffc0000000 ffffea0002f914c8 ffffe= a0002f914c8 > > > > 0000000000000000 > > > > [=A0=A0=A0 1.122638] raw: 0000000000000000 0000000000000000 00000= 000ffffffff > > > > 0000000000000000 > > > > [=A0=A0=A0 1.124146] page dumped because: acpi_map pre SetPageRes= erved > > > >=20 > > > > I also added dump_page() before unmapping, but it is not hit. The > > > > following for the same pfn now shows up I believe as a result of = setting > > > > PageReserved: > > > >=20 > > > > [=A0=A0 28.098208] BUG:Bad page state in process mo dprobe=A0 pfn= :be453 > > > > [=A0=A0 28.098394] page:ffffea0002f914c0 refcount:0 mapcount:0 > > > > mapping:0000000000000000 index:0x1 pfn:0xbe453 > > > > [=A0=A0 28.098394] flags: 0xfffffc0001000(reserved) > > > > [=A0=A0 28.098394] raw: 000fffffc0001000 dead000000000100 dead000= 000000122 > > > > 0000000000000000 > > > > [=A0=A0 28.098394] raw: 0000000000000001 0000000000000000 0000000= 0ffffffff > > > > 0000000000000000 > > > > [=A0=A0 28.098394] page dumped because: PAGE_FLAGS_CHECK_AT_PREP = flag(s) set > > > > [=A0=A0 28.098394] page_owner info is not present (never set?) > > > > [=A0=A0 28.098394] Modules linked in: > > > > [=A0=A0 28.098394] CPU: 2 PID: 204 Comm: modprobe Not tainted 5.1= 1.0-3dbd5e3 #66 > > > > [=A0=A0 28.098394] Hardware name: QEMU Standard PC (i440FX + PIIX= , 1996), > > > > BIOS 0.0.0 02/06/2015 > > > > [=A0=A0 28.098394] Call Trace: > > > > [=A0=A0 28.098394]=A0 dump_stack+0xdb/0x120 > > > > [=A0=A0 28.098394]=A0 bad_page.cold.108+0xc6/0xcb > > > > [=A0=A0 28.098394]=A0 check_new_page_bad+0x47/0xa0 > > > > [=A0=A0 28.098394]=A0 get_page_from_freelist+0x30cd/0x5730 > > > > [=A0=A0 28.098394]=A0 ? __isolate_free_page+0x4f0/0x4f0 > > > > [=A0=A0 28.098394]=A0 ? init_object+0x7e/0x90 > > > > [=A0=A0 28.098394]=A0 __alloc_pages_nodemask+0x2d8/0x650 > > > > [=A0=A0 28.098394]=A0 ? write_comp_data+0x2f/0x90 > > > > [=A0=A0 28.098394]=A0 ? __alloc_pages_slowpath.constprop.103+0x21= 10/0x2110 > > > > [=A0=A0 28.098394]=A0 ? __sanitizer_cov_trace_pc+0x21/0x50 > > > > [=A0=A0 28.098394]=A0 alloc_pages_vma+0xe2/0x560 > > > > [=A0=A0 28.098394]=A0 do_fault+0x194/0x12c0 > > > > [=A0=A0 28.098394]=A0 ? write_comp_data+0x2f/0x90 > > > > [=A0=A0 28.098394]=A0 __handle_mm_fault+0x1650/0x26c0 > > > > [=A0=A0 28.098394]=A0 ? copy_page_range+0x1350/0x1350 > > > > [=A0=A0 28.098394]=A0 ? write_comp_data+0x2f/0x90 > > > > [=A0=A0 28.098394]=A0 ? write_comp_data+0x2f/0x90 > > > > [=A0=A0 28.098394]=A0 handle_mm_fault+0x1f9/0x810 > > > > [=A0=A0 28.098394]=A0 ? write_comp_data+0x2f/0x90 > > > > [=A0=A0 28.098394]=A0 do_user_addr_fault+0x6f7/0xca0 > > > > [=A0=A0 28.098394]=A0 exc_page_fault+0xaf/0x1a0 > > > > [=A0=A0 28.098394]=A0 asm_exc_page_fault+0x1e/0x30 > > > > [=A0=A0 28.098394] RIP: 0010:__clear_user+0x30/0x60 > > >=20 > > > I think the PAGE_FLAGS_CHECK_AT_PREP check in this instance means t= hat > > > someone is trying to allocate that page with the PG_reserved bit se= t. > > > This means that the page actually was exposed to the buddy. > > >=20 > > > However, when you SetPageReserved(), I don't think that PG_buddy is= set > > > and the refcount is 0. That could indicate that the page is on the = buddy > > > PCP list. Could be that it is getting reused a couple of times. > > >=20 > > > The PFN 0xbe453 looks a little strange, though. Do we expect ACPI t= ables > > > close to 3 GiB ? No idea. Could it be that you are trying to map a = wrong > > > table? Just a guess. >=20 > Nah, ACPI MADT enumerates the table and that is the proper location of = it. > >=20 > > ... but I assume ibft_check_device() would bail out on an invalid che= cksum. > > So the question is, why is this page not properly marked as reserved > > already. >=20 > The ibft_check_device ends up being called as module way way after the > kernel has cleaned the memory. >=20 > The funny thing about iBFT is that (it is also mentioned in the spec) > that the table can resize in memory .. or in the ACPI regions (which ^ reside I presume? > have no E820_RAM and are considered "MMIO" regions). >=20 > Either place is fine, so it can be in either RAM or MMIO :-( I'd say that the tables in this case are in E820_RAM, because with MMIO w= e wouldn't get to kmap() at the first place. It can be easily confirmed by comparing the problematic address with /proc/iomem. Can't say I have a clue about what's going on there, but the theory that somehow iBFT table does not get PG_Reserved during boot makes sense. Do you see "iBFT found at 0x" early in the kernel log? I don't know if ACPI relocates the tables, but I could not find anywhere that it reserves the original ones. The memblock_reserve() in acpi_table_upgrade() is merely a part of open coded memblock allocation. --=20 Sincerely yours, Mike.