Patch series "kdump: crashkernel reservation from CMA", v5.
This series implements a way to reserve additional crash kernel memory
using CMA.
Currently, all the memory for the crash kernel is not usable by the 1st
(production) kernel. It is also unmapped so that it can't be corrupted by
the fault that will eventually trigger the crash. This makes sense for
the memory actually used by the kexec-loaded crash kernel image and initrd
and the data prepared during the load (vmcoreinfo, ...). However, the
reserved space needs to be much larger than that to provide enough
run-time memory for the crash kernel and the kdump userspace. Estimating
the amount of memory to reserve is difficult. Being too careful makes
kdump likely to end in OOM, being too generous takes even more memory from
the production system. Also, the reservation only allows reserving a
single contiguous block (or two with the "low" suffix). I've seen systems
where this fails because the physical memory is fragmented.
By reserving additional crashkernel memory from CMA, the main crashkernel
reservation can be just large enough to fit the kernel and initrd image,
minimizing the memory taken away from the production system. Most of the
run-time memory for the crash kernel will be memory previously available
to userspace in the production system. As this memory is no longer
wasted, the reservation can be done with a generous margin, making kdump
more reliable. Kernel memory that we need to preserve for dumping is
normally not allocated from CMA, unless it is explicitly allocated as
movable. Currently this is only the case for memory ballooning and zswap.
Such movable memory will be missing from the vmcore. User data is
typically not dumped by makedumpfile. When dumping of user data is
intended this new CMA reservation cannot be used.
There are five patches in this series:
The first adds a new ",cma" suffix to the recenly introduced generic
crashkernel parsing code. parse_crashkernel() takes one more argument to
store the cma reservation size.
The second patch implements reserve_crashkernel_cma() which performs the
reservation. If the requested size is not available in a single range,
multiple smaller ranges will be reserved.
The third patch updates Documentation/, explicitly mentioning the
potential DMA corruption of the CMA-reserved memory.
The fourth patch adds a short delay before booting the kdump kernel,
allowing pending DMA transfers to finish.
The fifth patch enables the functionality for x86 as a proof of
concept. There are just three things every arch needs to do:
- call reserve_crashkernel_cma()
- include the CMA-reserved ranges in the physical memory map
- exclude the CMA-reserved ranges from the memory available
through /proc/vmcore by excluding them from the vmcoreinfo
PT_LOAD ranges.
Adding other architectures is easy and I can do that as soon as this
series is merged.
With this series applied, specifying
crashkernel=100M craskhernel=1G,cma
on the command line will make a standard crashkernel reservation
of 100M, where kexec will load the kernel and initrd.
An additional 1G will be reserved from CMA, still usable by the production
system. The crash kernel will have 1.1G memory available. The 100M can
be reliably predicted based on the size of the kernel and initrd.
The new cma suffix is completely optional. When no
crashkernel=size,cma is specified, everything works as before.
This patch (of 5):
Add a new cma_size parameter to parse_crashkernel(). When not NULL, call
__parse_crashkernel to parse the CMA reservation size from
"crashkernel=size,cma" and store it in cma_size.
Set cma_size to NULL in all calls to parse_crashkernel().
Link: https://lkml.kernel.org/r/aEqnxxfLZMllMC8I@dwarf.suse.cz
Link: https://lkml.kernel.org/r/aEqoQckgoTQNULnh@dwarf.suse.cz
Signed-off-by: Jiri Bohac <jbohac@suse.cz>
Cc: Baoquan He <bhe@redhat.com>
Cc: Dave Young <dyoung@redhat.com>
Cc: Donald Dutile <ddutile@redhat.com>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Philipp Rudo <prudo@redhat.com>
Cc: Pingfan Liu <piliu@redhat.com>
Cc: Tao Liu <ltao@redhat.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: David Hildenbrand <david@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Convert arm32 to use the arch_cpu_is_hotpluggable() helper rather than
arch_register_cpu().
Signed-off-by: Jinjie Ruan <ruanjinjie@huawei.com>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Currently, almost all architectures have switched to GENERIC_CPU_DEVICES,
except for arm32. Also switch over to GENERIC_CPU_DEVICES, and provide an
arch_register_cpu() that populates the hotpluggable flag for arm32.
The struct cpu in struct cpuinfo_arm is never used directly, remove
it to use the one GENERIC_CPU_DEVICES provides.
This also has the effect of moving the registration of CPUs from subsys to
driver core initialisation, prior to any initcalls running.
Signed-off-by: Jinjie Ruan <ruanjinjie@huawei.com>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Pull tty and serial updates from Greg KH:
"Here is the big set of tty/serial driver changes for 6.7-rc1. Included
in here are:
- console/vgacon cleanups and removals from Arnd
- tty core and n_tty cleanups from Jiri
- lots of 8250 driver updates and cleanups
- sc16is7xx serial driver updates
- dt binding updates
- first set of port lock wrapers from Thomas for the printk fixes
coming in future releases
- other small serial and tty core cleanups and updates
All of these have been in linux-next for a while with no reported
issues"
* tag 'tty-6.7-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty: (193 commits)
serdev: Replace custom code with device_match_acpi_handle()
serdev: Simplify devm_serdev_device_open() function
serdev: Make use of device_set_node()
tty: n_gsm: add copyright Siemens Mobility GmbH
tty: n_gsm: fix race condition in status line change on dead connections
serial: core: Fix runtime PM handling for pending tx
vgacon: fix mips/sibyte build regression
dt-bindings: serial: drop unsupported samsung bindings
tty: serial: samsung: drop earlycon support for unsupported platforms
tty: 8250: Add note for PX-835
tty: 8250: Fix IS-200 PCI ID comment
tty: 8250: Add Brainboxes Oxford Semiconductor-based quirks
tty: 8250: Add support for Intashield IX cards
tty: 8250: Add support for additional Brainboxes PX cards
tty: 8250: Fix up PX-803/PX-857
tty: 8250: Fix port count of PX-257
tty: 8250: Add support for Intashield IS-100
tty: 8250: Add support for Brainboxes UP cards
tty: 8250: Add support for additional Brainboxes UC cards
tty: 8250: Remove UC-257 and UC-431
...
Pull non-MM updates from Andrew Morton:
"As usual, lots of singleton and doubleton patches all over the tree
and there's little I can say which isn't in the individual changelogs.
The lengthier patch series are
- 'kdump: use generic functions to simplify crashkernel reservation
in arch', from Baoquan He. This is mainly cleanups and
consolidation of the 'crashkernel=' kernel parameter handling
- After much discussion, David Laight's 'minmax: Relax type checks in
min() and max()' is here. Hopefully reduces some typecasting and
the use of min_t() and max_t()
- A group of patches from Oleg Nesterov which clean up and slightly
fix our handling of reads from /proc/PID/task/... and which remove
task_struct.thread_group"
* tag 'mm-nonmm-stable-2023-11-02-14-08' of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm: (64 commits)
scripts/gdb/vmalloc: disable on no-MMU
scripts/gdb: fix usage of MOD_TEXT not defined when CONFIG_MODULES=n
.mailmap: add address mapping for Tomeu Vizoso
mailmap: update email address for Claudiu Beznea
tools/testing/selftests/mm/run_vmtests.sh: lower the ptrace permissions
.mailmap: map Benjamin Poirier's address
scripts/gdb: add lx_current support for riscv
ocfs2: fix a spelling typo in comment
proc: test ProtectionKey in proc-empty-vm test
proc: fix proc-empty-vm test with vsyscall
fs/proc/base.c: remove unneeded semicolon
do_io_accounting: use sig->stats_lock
do_io_accounting: use __for_each_thread()
ocfs2: replace BUG_ON() at ocfs2_num_free_extents() with ocfs2_error()
ocfs2: fix a typo in a comment
scripts/show_delta: add __main__ judgement before main code
treewide: mark stuff as __ro_after_init
fs: ocfs2: check status values
proc: test /proc/${pid}/statm
compiler.h: move __is_constexpr() to compiler.h
...
Separating the VGA console screen_info from the EFI one unfortunately
caused a build failure for footbridge that I had never caught
with randconfig builds:
arch/arm/kernel/setup.c:932:27: error: static declaration of 'vgacon_screen_info' follows non-static declaration
932 | static struct screen_info vgacon_screen_info = {
| ^~~~~~~~~~~~~~~~~~
In file included from arch/arm/kernel/setup.c:44:
arch/arm/include/asm/setup.h:40:27: note: previous declaration of 'vgacon_screen_info' with type 'struct screen_info'
40 | extern struct screen_info vgacon_screen_info;
| ^~~~~~~~~~~~~~~~~~
arm-linux-gnueabi-ld: drivers/video/console/dummycon.o: in function `dummycon_init':
dummycon.c:(.text+0xe4): undefined reference to `screen_info'
Make sure the variable is global to avoid the conflict with the extern
declaration, and make it work in dummycon.c
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20231017093947.3627976-2-arnd@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To prepare for completely separating the VGA console screen_info from
the one used in EFI/sysfb, rename the vgacon instances and make them
local as much as possible.
ia64 and arm both have confurations with vgacon and efi, but the contents
never overlaps because ia64 has no EFI framebuffer, and arm only has
vga console on legacy platforms without EFI. Renaming these is required
before the EFI screen_info can be moved into drivers/firmware.
The ia64 vga console is actually registered in two places from
setup_arch(), but one of them is wrong, so drop the one in pcdp.c and
fix the one in setup.c to use the correct conditional.
x86 has to keep them together, as the boot protocol is used to switch
between VGA text console and framebuffer through the screen_info data.
Acked-by: Javier Martinez Canillas <javierm@redhat.com>
Acked-by: Khalid Aziz <khalid@gonehiking.org>
Acked-by: Helge Deller <deller@gmx.de>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20231009211845.3136536-7-arnd@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The dummycon default console size used to be determined by architecture,
but now this is a Kconfig setting on everything except ARM. Tracing this
back in the historic git trees, this was used to match the size of VGA
console or VGA framebuffer on early machines, but nowadays that code is
no longer used, except probably on the old footbridge/netwinder since
that is the only one that supports vgacon.
On machines with a framebuffer, booting with DT so far results in always
using the hardcoded 80x30 size in dummycon, while on ATAGS the setting
can come from a bootloader specific override. Both seem to be worse
choices than the Kconfig setting, since the actual text size for fbcon
also depends on the selected font.
Make this work the same way as everywhere else and use the normal
Kconfig setting, except for the footbridge with vgacon, which keeps
using the traditional code. If vgacon is disabled, footbridge can
also ignore the setting. This means the screen_info only has to be
provided when either vgacon or EFI are enabled now.
To limit the amount of surprises on Arm, change the Kconfig default
to the previously used 80x30 setting instead of the usual 80x25.
Reviewed-by: Thomas Zimmermann <tzimmermann@suse.de>
Tested-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Acked-by: Helge Deller <deller@gmx.de>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20231009211845.3136536-4-arnd@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The DT of_device.h and of_platform.h date back to the separate
of_platform_bus_type before it as merged into the regular platform bus.
As part of that merge prepping Arm DT support 13 years ago, they
"temporarily" include each other. They also include platform_device.h
and of.h. As a result, there's a pretty much random mix of those include
files used throughout the tree. In order to detangle these headers and
replace the implicit includes with struct declarations, users need to
explicitly include the correct includes.
Signed-off-by: Rob Herring <robh@kernel.org>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
A couple of functions are declared in arch/arm/mm/mmu.c rather than in a header,
which causes W=1 build warnings:
arch/arm/mm/init.c:97:13: error: no previous prototype for 'setup_dma_zone' [-Werror=missing-prototypes]
arch/arm/mm/mmu.c:118:13: error: no previous prototype for 'init_default_cache_policy' [-Werror=missing-prototypes]
arch/arm/mm/mmu.c:1195:13: error: no previous prototype for 'adjust_lowmem_bounds' [-Werror=missing-prototypes]
arch/arm/mm/mmu.c:1761:13: error: no previous prototype for 'paging_init' [-Werror=missing-prototypes]
arch/arm/mm/mmu.c:1794:13: error: no previous prototype for 'early_mm_init' [-Werror=missing-prototypes]
Move the declaratsion to asm/setup.h so they can be seen by the compiler
while building the definition.
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
ARM exclusively uses GENERIC_IRQ_MULTI_HANDLER, so at some point
set_handle_irq() needs to be called to handle system-wide
interrupts.
For all DT-enabled boards, this call happens down in the
drivers/irqchip subsystem, after locating the target irqchip
driver from the device tree.
We still have a few instances of the boardfiles with machine
descriptors passing a machine-specific .handle_irq() to the
ARM kernel core.
Get rid of this by letting the few remaining machines consistently
call set_handle_irq() from the end of the .init_irq() callback
instead and diet down one member from the machine descriptor.
Cc: Marc Zyngier <maz@kernel.org>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Pull ARM updates from Russell King:
- update unwinder to cope with module PLTs
- enable UBSAN on ARM
- improve kernel fault message
- update UEFI runtime page tables dump
- avoid clang's __aeabi_uldivmod generated in NWFPE code
- disable FIQs on CPU shutdown paths
- update XOR register usage
- a number of build updates (using .arch, thread pointer, removal of
lazy evaluation in Makefile)
- conversion of stacktrace code to stackwalk
- findbit assembly updates
- hwcap feature updates for ARMv8 CPUs
- instruction dump updates for big-endian platforms
- support for function error injection
* tag 'for-linus' of git://git.armlinux.org.uk/~rmk/linux-arm: (31 commits)
ARM: 9279/1: support function error injection
ARM: 9277/1: Make the dumped instructions are consistent with the disassembled ones
ARM: 9276/1: Refactor dump_instr()
ARM: 9275/1: Drop '-mthumb' from AFLAGS_ISA
ARM: 9274/1: Add hwcap for Speculative Store Bypassing Safe
ARM: 9273/1: Add hwcap for Speculation Barrier(SB)
ARM: 9272/1: vfp: Add hwcap for FEAT_AA32I8MM
ARM: 9271/1: vfp: Add hwcap for FEAT_AA32BF16
ARM: 9270/1: vfp: Add hwcap for FEAT_FHM
ARM: 9269/1: vfp: Add hwcap for FEAT_DotProd
ARM: 9268/1: vfp: Add hwcap FPHP and ASIMDHP for FEAT_FP16
ARM: 9267/1: Define Armv8 registers in AArch32 state
ARM: findbit: add unwinder information
ARM: findbit: operate by words
ARM: findbit: convert to macros
ARM: findbit: provide more efficient ARMv7 implementation
ARM: findbit: document ARMv5 bit offset calculation
ARM: 9259/1: stacktrace: Convert stacktrace to generic ARCH_STACKWALK
ARM: 9258/1: stacktrace: Make stack walk callback consistent with generic code
ARM: 9265/1: pass -march= only to compiler
...
Speculative Store Bypassing Safe(FEAT_SSBS) is a feature present in
AArch32 state for Armv8 and is represented by ID_PFR2_EL1.SSBS
identification register.
This feature denotes the presence of PSTATE.ssbs bit and hence adding a
hwcap will enable the userspace to check it before trying to set/unset
this PSTATE.
This commit adds the ID feature bit detection, and uses elf_hwcap2
accordingly.
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Amit Daniel Kachhap <amit.kachhap@arm.com>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Speculation Barrier(FEAT_SB) is a feature present in AArch32 state for
Armv8 and is represented by ISAR6.SB identification register.
This feature denotes the presence of SB instruction and hence adding a
hwcap will enable the userspace to check it before trying to use this
instruction.
This commit adds the ID feature bit detection, and uses elf_hwcap2
accordingly.
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Amit Daniel Kachhap <amit.kachhap@arm.com>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Int8 matrix multiplication (FEAT_AA32I8MM) is a feature present in AArch32 state for Armv8 and is represented by ISAR6.I8MM identification register.
This feature denotes the presence of VSMMLA, VSUDOT, VUMMLA, VUSMMLA and
VUSDOT instructions and hence adding a hwcap will enable the userspace
to check it before trying to use those instructions.
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Amit Daniel Kachhap <amit.kachhap@arm.com>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Advanced SIMD BFloat16 (FEAT_AA32BF16) is a feature present in AArch32
state for Armv8 and is represented by ISAR6.BF16 identification
register.
This feature denotes the presence of VCVT, VCVTB, VCVTT, VDOT, VFMAB,
VFMAT and VMMLA instructions and hence adding a hwcap will enable the
userspace to check it before trying to use those instructions.
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Amit Daniel Kachhap <amit.kachhap@arm.com>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Floating-point half-precision multiplication (FHM) is a feature present
in AArch32 state for Armv8 and is represented by ISAR6.FHM identification register.
This feature denotes the presence of VFMAL and VMFSL instructions and
hence adding a hwcap will enable the userspace to check it before
trying to use those instructions.
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Amit Daniel Kachhap <amit.kachhap@arm.com>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Advanced Dot product is a feature present in AArch32 state for Armv8 and
is represented by ISAR6 identification register.
This feature denotes the presence of UDOT and SDOT instructions and hence adding a hwcap will enable the userspace to check it before trying to use those instructions.
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Amit Daniel Kachhap <amit.kachhap@arm.com>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Floating point half-precision (FPHP) and Advanced SIMD half-precision
(ASIMDHP) are VFP features (FEAT_FP16) represented by MVFR1 identification register. These capabilities can optionally exist with VFPv3 and mandatory with VFPv4. Both these new features exist for Armv8 architecture in AArch32 state.
These hwcaps may be useful for the userspace to add conditional check
before trying to use FEAT_FP16 feature specific instructions.
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Amit Daniel Kachhap <amit.kachhap@arm.com>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Move some code that is only reachable when IS_ENABLED(CONFIG_ARM) into
the ARM EFI arch code.
Cc: Russell King <linux@armlinux.org.uk>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
