| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <17362.1589022588@turing-police>
Date: Sat, 09 May 2020 07:09:48 -0400
From: "Valdis Klētnieks" <valdis.kletnieks@...edu>
To: Ben Skeggs <bskeggs@...hat.com>, David Airlie <airlied@...ux.ie>,
Christoph Hellwig <hch@....de>
Cc: dri-devel@...ts.freedesktop.org, nouveau@...ts.freedesktop.org,
linux-kernel@...r.kernel.org
Subject: linux-next 20200508 - build failure in kernel/resource.c w/ SPARSEMEM=n
So I did a 'make allmodconfig' and then a 'make' on an RPi4 ARM box, and it
decided that CONFIG_SPARSEMEM=n was OK (so an include of linux/mmzone.h doesn't
define some needed values).
The offending code in resource.c is wrapped in a #ifdef CONFIG_DEVICE_PRIVATE,
which throws a whinge during 'make menuconfig' or 'make allmodconfig':
WARNING: unmet direct dependencies detected for DEVICE_PRIVATE
Depends on [n]: ZONE_DEVICE [=n]
Selected by [m]:
- DRM_NOUVEAU_SVM [=y] && HAS_IOMEM [=y] && DRM_NOUVEAU [=m] && MMU [=y] && STAGING [=y]
after which I end up with CONFIG_DEVICE_PRIVATE=y in the .config file.
make menuconfig tells me:
Symbol: ZONE_DEVICE [=n]
Type : bool
Defined at mm/Kconfig:779
Prompt: Device memory (pmem, HMM, etc...) hotplug support
Depends on: MEMORY_HOTPLUG [=n] && MEMORY_HOTREMOVE [=n] && SPARSEMEM_VMEMMAP [=n] && ARCH_HAS_PTE_DEVMAP [=n]
Location:
(1) -> Memory Management options
Selects: XARRAY_MULTI [=n]
Pretty obviously a Kconfig whoops, but I have no idea what the proper Kconfig
fix is for this..
May be related to:
commit 0092908d16c604b8207c2141ec64b0fa4473bb03
Author: Christoph Hellwig <hch@....de>
Date: Wed Jun 26 14:27:06 2019 +0200
mm: factor out a devm_request_free_mem_region helper
which added the #ifdef CONFIG_DEVICE_PRIVATE code in question, except that's a
pretty old commit... The only thing I'm sure of is that DEVICE_PRIVATE=y and
SPARSEMEM=n blows up. :)
CC kernel/resource.o
In file included from ./include/linux/cache.h:5,
from ./include/linux/printk.h:9,
from ./include/linux/kernel.h:15,
from ./include/asm-generic/bug.h:19,
from ./arch/arm/include/asm/bug.h:60,
from ./include/linux/bug.h:5,
from ./include/linux/mmdebug.h:5,
from ./include/linux/gfp.h:5,
from ./include/linux/slab.h:15,
from kernel/resource.c:17:
kernel/resource.c: In function '__request_free_mem_region':
kernel/resource.c:1653:28: error: 'PA_SECTION_SHIFT' undeclared (first use in this function); did you mean 'SECTION_SHIFT'?
size = ALIGN(size, 1UL << PA_SECTION_SHIFT);
^~~~~~~~~~~~~~~~
./include/uapi/linux/kernel.h:11:47: note: in definition of macro '__ALIGN_KERNEL_MASK'
#define __ALIGN_KERNEL_MASK(x, mask) (((x) + (mask)) & ~(mask))
^~~~
./include/linux/kernel.h:33:22: note: in expansion of macro '__ALIGN_KERNEL'
#define ALIGN(x, a) __ALIGN_KERNEL((x), (a))
^~~~~~~~~~~~~~
kernel/resource.c:1653:9: note: in expansion of macro 'ALIGN'
size = ALIGN(size, 1UL << PA_SECTION_SHIFT);
^~~~~
kernel/resource.c:1653:28: note: each undeclared identifier is reported only once for each function it appears in
size = ALIGN(size, 1UL << PA_SECTION_SHIFT);
^~~~~~~~~~~~~~~~
./include/uapi/linux/kernel.h:11:47: note: in definition of macro '__ALIGN_KERNEL_MASK'
#define __ALIGN_KERNEL_MASK(x, mask) (((x) + (mask)) & ~(mask))
^~~~
./include/linux/kernel.h:33:22: note: in expansion of macro '__ALIGN_KERNEL'
#define ALIGN(x, a) __ALIGN_KERNEL((x), (a))
^~~~~~~~~~~~~~
kernel/resource.c:1653:9: note: in expansion of macro 'ALIGN'
size = ALIGN(size, 1UL << PA_SECTION_SHIFT);
^~~~~
In file included from ./include/asm-generic/bug.h:19,
from ./arch/arm/include/asm/bug.h:60,
from ./include/linux/bug.h:5,
from ./include/linux/mmdebug.h:5,
from ./include/linux/gfp.h:5,
from ./include/linux/slab.h:15,
from kernel/resource.c:17:
kernel/resource.c:1654:48: error: 'MAX_PHYSMEM_BITS' undeclared (first use in this function); did you mean 'MAX_UINSN_BYTES'?
end = min_t(unsigned long, base->end, (1UL << MAX_PHYSMEM_BITS) - 1);
^~~~~~~~~~~~~~~~
./include/linux/kernel.h:848:40: note: in definition of macro '__typecheck'
(!!(sizeof((typeof(x) *)1 == (typeof(y) *)1)))
^
./include/linux/kernel.h:872:24: note: in expansion of macro '__safe_cmp'
__builtin_choose_expr(__safe_cmp(x, y), \
^~~~~~~~~~
./include/linux/kernel.h:940:27: note: in expansion of macro '__careful_cmp'
#define min_t(type, x, y) __careful_cmp((type)(x), (type)(y), <)
^~~~~~~~~~~~~
kernel/resource.c:1654:8: note: in expansion of macro 'min_t'
end = min_t(unsigned long, base->end, (1UL << MAX_PHYSMEM_BITS) - 1);
^~~~~
./include/linux/kernel.h:872:2: error: first argument to '__builtin_choose_expr' not a constant
__builtin_choose_expr(__safe_cmp(x, y), \
^~~~~~~~~~~~~~~~~~~~~
./include/linux/kernel.h:940:27: note: in expansion of macro '__careful_cmp'
#define min_t(type, x, y) __careful_cmp((type)(x), (type)(y), <)
^~~~~~~~~~~~~
kernel/resource.c:1654:8: note: in expansion of macro 'min_t'
end = min_t(unsigned long, base->end, (1UL << MAX_PHYSMEM_BITS) - 1);
^~~~~
make[1]: *** [scripts/Makefile.build:273: kernel/resource.o] Error 1
make: *** [Makefile:1726: kernel] Error 2
Content of type "application/pgp-signature" skipped
Powered by blists - more mailing lists