This function simply returns the number of VPEs present in the current
core, or 1 if the core does not implement the MT ASE. In SMP kernels
this will typically equal smp_num_siblings, however it will also be
usable in UP kernels and helps prepare for the possibility of a
heterogenous system where the VPE count is not the same across all
cores.
Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/6665/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
This patch ensures that the kernel sets a sane base address for the
PIIX4 PM I/O register region during boot. Without this the kernel may
not successfully claim the region as a resource if the bootloader didn't
configure the region. With this patch the kernel will always succeed
with:
pci 0000:00:0a.3: quirk: [io 0x1000-0x103f] claimed by PIIX4 ACPI
The lack of the resource claiming is easily reproducible without this
patch using current versions of QEMU.
Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Tested-by: James Hogan <james.hogan@imgtec.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/6641/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Tips of Loongson's CPU hotplug:
1, To fully shutdown a core in Loongson 3, the target core should go to
CKSEG1 and flush all L1 cache entries at first. Then, another core
(usually Core 0) can safely disable the clock of the target core. So
play_dead() call loongson3_play_dead() via CKSEG1 (both uncached and
unmmaped).
2, The default clocksource of Loongson is MIPS. Since clock source is a
global device, timekeeping need the CP0' Count registers of each core
be synchronous. Thus, when a core is up, we use a SMP_ASK_C0COUNT IPI
to ask Core-0's Count.
Signed-off-by: Huacai Chen <chenhc@lemote.com>
Signed-off-by: Hongliang Tao <taohl@lemote.com>
Signed-off-by: Hua Yan <yanh@lemote.com>
Tested-by: Alex Smith <alex.smith@imgtec.com>
Reviewed-by: Alex Smith <alex.smith@imgtec.com>
Cc: John Crispin <john@phrozen.org>
Cc: Steven J. Hill <Steven.Hill@imgtec.com>
Cc: Aurelien Jarno <aurelien@aurel32.net>
Cc: linux-mips@linux-mips.org
Cc: Fuxin Zhang <zhangfx@lemote.com>
Cc: Zhangjin Wu <wuzhangjin@gmail.com>
Patchwork: https://patchwork.linux-mips.org/patch/6639
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
IPI registers of Loongson-3 include IPI_SET, IPI_CLEAR, IPI_STATUS,
IPI_EN and IPI_MAILBOX_BUF. Each bit of IPI_STATUS indicate a type of
IPI and IPI_EN indicate whether the IPI is enabled. The sender write 1
to IPI_SET bits generate IPIs in IPI_STATUS, and receiver write 1 to
bits of IPI_CLEAR to clear IPIs. IPI_MAILBOX_BUF are used to deliver
more information about IPIs.
Why we change code in arch/mips/loongson/common/setup.c?
If without this change, when SMP configured, system cannot boot since
it hang at printk() in cgroup_init_early(). The root cause is:
console_trylock()
\-->down_trylock(&console_sem)
\-->raw_spin_unlock_irqrestore(&sem->lock, flags)
\-->_raw_spin_unlock_irqrestore()(SMP/UP have different versions)
\-->__raw_spin_unlock_irqrestore() (following is the SMP case)
\-->do_raw_spin_unlock()
\-->arch_spin_unlock()
\-->nudge_writes()
\-->mb()
\-->wbflush()
\-->__wbflush()
In previous code __wbflush() is initialized in plat_mem_setup(), but
cgroup_init_early() is called before plat_mem_setup(). Therefore, In
this patch we make changes to avoid boot failure.
Signed-off-by: Huacai Chen <chenhc@lemote.com>
Signed-off-by: Hongliang Tao <taohl@lemote.com>
Signed-off-by: Hua Yan <yanh@lemote.com>
Tested-by: Alex Smith <alex.smith@imgtec.com>
Reviewed-by: Alex Smith <alex.smith@imgtec.com>
Cc: John Crispin <john@phrozen.org>
Cc: Steven J. Hill <Steven.Hill@imgtec.com>
Cc: Aurelien Jarno <aurelien@aurel32.net>
Cc: linux-mips@linux-mips.org
Cc: Fuxin Zhang <zhangfx@lemote.com>
Cc: Zhangjin Wu <wuzhangjin@gmail.com>
Patchwork: https://patchwork.linux-mips.org/patch/6638
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Loongson doesn't support DMA address above 4GB traditionally. If memory
is more than 4GB, CONFIG_SWIOTLB and ZONE_DMA32 should be selected. In
this way, DMA pages are allocated below 4GB preferably. However, if low
memory is not enough, high pages are allocated and swiotlb is used for
bouncing.
Moreover, we provide a platform-specific dma_map_ops::set_dma_mask() to
set a device's dma_mask and coherent_dma_mask. We use these masks to
distinguishes an allocated page can be used for DMA directly, or need
swiotlb to bounce.
Recently, we found that 32-bit DMA isn't a hardware bug, but a hardware
configuration issue. So, latest firmware has enable the DMA support as
high as 40-bit. To support all-memory DMA for all devices (besides the
Loongson platform limit, there are still some devices have their own
DMA32 limit), and also to be compatible with old firmware, we keep use
swiotlb.
Signed-off-by: Huacai Chen <chenhc@lemote.com>
Signed-off-by: Hongliang Tao <taohl@lemote.com>
Signed-off-by: Hua Yan <yanh@lemote.com>
Tested-by: Alex Smith <alex.smith@imgtec.com>
Reviewed-by: Alex Smith <alex.smith@imgtec.com>
Cc: John Crispin <john@phrozen.org>
Cc: Steven J. Hill <Steven.Hill@imgtec.com>
Cc: Aurelien Jarno <aurelien@aurel32.net>
Cc: linux-mips@linux-mips.org
Cc: Fuxin Zhang <zhangfx@lemote.com>
Cc: Zhangjin Wu <wuzhangjin@gmail.com>
Patchwork: https://patchwork.linux-mips.org/patch/6636
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
The new UEFI-like firmware interface (LEFI, i.e. Loongson Unified
Firmware Interface) has 3 advantages:
1, Firmware export a physical memory map which is similar to X86's
E820 map, so prom_init_memory() will be more elegant that #ifdef
clauses can be removed.
2, Firmware export a pci irq routing table, we no longer need pci
irq routing fixup in kernel's code.
3, Firmware has a built-in vga bios, and its address is exported,
the linux kernel no longer need an embedded blob.
With the LEFI interface, Loongson-3A/2G and all their successors can use
a unified kernel. All Loongson-based machines support this new interface
except 2E/2F series.
Signed-off-by: Huacai Chen <chenhc@lemote.com>
Signed-off-by: Hongliang Tao <taohl@lemote.com>
Signed-off-by: Hua Yan <yanh@lemote.com>
Tested-by: Alex Smith <alex.smith@imgtec.com>
Reviewed-by: Alex Smith <alex.smith@imgtec.com>
Cc: John Crispin <john@phrozen.org>
Cc: Steven J. Hill <Steven.Hill@imgtec.com>
Cc: Aurelien Jarno <aurelien@aurel32.net>
Cc: linux-mips@linux-mips.org
Cc: Fuxin Zhang <zhangfx@lemote.com>
Cc: Zhangjin Wu <wuzhangjin@gmail.com>
Patchwork: https://patchwork.linux-mips.org/patch/6632
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Loongson-3 is a multi-core MIPS family CPU, it support MIPS64R2 fully.
Loongson-3 has the same IMP field (0x6300) as Loongson-2.
Loongson-3 has a hardware-maintained cache, system software doesn't
need to maintain coherency.
Loongson-3A is the first revision of Loongson-3, and it is the quad-
core version of Loongson-2G. Loongson-3A has a simplified version named
Loongson-2Gq, the main difference between Loongson-3A/2Gq is 3A has two
HyperTransport controller but 2Gq has only one. HT0 is used for cross-
chip interconnection and HT1 is used to link PCI bus. Therefore, 2Gq
cannot support NUMA but 3A can. For software, Loongson-2Gq is simply
identified as Loongson-3A.
Exsisting Loongson family CPUs:
Loongson-1: Loongson-1A, Loongson-1B, they are 32-bit MIPS CPUs.
Loongson-2: Loongson-2E, Loongson-2F, Loongson-2G, they are 64-bit
single-core MIPS CPUs.
Loongson-3: Loongson-3A(including so-called Loongson-2Gq), they are
64-bit multi-core MIPS CPUs.
Signed-off-by: Huacai Chen <chenhc@lemote.com>
Signed-off-by: Hongliang Tao <taohl@lemote.com>
Signed-off-by: Hua Yan <yanh@lemote.com>
Tested-by: Alex Smith <alex.smith@imgtec.com>
Reviewed-by: Alex Smith <alex.smith@imgtec.com>
Cc: John Crispin <john@phrozen.org>
Cc: Steven J. Hill <Steven.Hill@imgtec.com>
Cc: Aurelien Jarno <aurelien@aurel32.net>
Cc: linux-mips@linux-mips.org
Cc: Fuxin Zhang <zhangfx@lemote.com>
Cc: Zhangjin Wu <wuzhangjin@gmail.com>
Patchwork: https://patchwork.linux-mips.org/patch/6629/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
And there are more CPUs or configurations that want to provide special
per-CPU information in /proc/cpuinfo. So I think there needs to be a
hook mechanism, such as a notifier.
This is a first cut only; I need to think about what sort of looking
the notifier needs to have. But I'd appreciate testing on MT hardware!
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Cc: Markos Chandras <markos.chandras@imgtec.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/6066/
The UART register names are identical to the ones in uapi/linux/serial_reg.h,
which causes build failures in various drivers when they indirectly pull in
the au1000.h header, for example via gpio.h:
In file included from arch/mips/include/asm/mach-au1x00/gpio.h:13:0,
from arch/mips/include/asm/gpio.h:4,
from include/linux/gpio.h:48,
from include/linux/ssb/ssb.h:9,
from drivers/ssb/driver_mipscore.c:11:
arch/mips/include/asm/mach-au1x00/au1000.h:1171:0: note: this is the location of the previous definition
#define UART_LSR 0x1C /* Line Status Register */
Get rid of the altogether, nothing in the core Alchemy code depends
on them any more.
Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Cc: Linux-MIPS <linux-mips@linux-mips.org>
Patchwork: https://patchwork.linux-mips.org/patch/6664/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
The Linksys WRT54G/GS/GL family uses the same boardtype numbers, and
the same gpio configuration. The boardtype numbers are changing with
the hardware versions, but these hardware numbers are different or each
model.
Detect them all as one device, this also worked in OpenWrt.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Cc: linux-mips@linux-mips.org
Cc: zajec5@gmail.com
Patchwork: https://patchwork.linux-mips.org/patch/6591/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Allow secondary cores to program their segment control registers
during smp bootstrap code. This enables EVA on Malta SMP
configurations
Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
The 'ememsize' variable is used to denote the real RAM which is
present on the Malta board. This is different compared to 'memsize'
which is capped to 256MB. The 'ememsize' is used to get the actual
physical memory when setting up the Malta memory layout. This only
makes sense in case the core operates in the EVA mode, and it's
ignored otherwise.
Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
Add a spaces.h file for Malta to override certain memory macros
when operating in EVA mode.
Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
The Malta board aliases 0x80000000 - 0xffffffff to 0x00000000
- 0x7fffffff ignoring the 256 MB IO hole in 0x10000000.
The physical memory is shifted to 0x80000000 so up to 2GB
can be used. Kuseg is expanded to 3GB (due to board limitations
only 2GB can be accessed) and lowmem (kernel space) is expanded to 2GB.
The Segment Control registers are programmed as follows:
Virtual memory Physical memory Mapping
0x00000000 - 0x7fffffff 0x80000000 - 0xfffffffff MUSUK (kuseg)
0x80000000 - 0x9fffffff 0x00000000 - 0x1ffffffff MUSUK (kseg0)
0xa0000000 - 0xbf000000 0x00000000 - 0x1ffffffff MUSUK (kseg1)
0xc0000000 - 0xdfffffff - MK (kseg2)
0xe0000000 - 0xffffffff - MK (kseg3)
The location of exception vectors remain the same since 0xbfc00000
(traditional exception base) still maps to 0x1fc00000 physical.
Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
A core in EVA mode can have any possible segment mapping, so the
default free_initmem_default() function may not always work as expected.
Therefore, add a callback that platforms can use to free up the init section.
Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
The MIPS *Aptiv family uses bit 28 in Config5 CP0 register to
indicate whether the core supports EVA or not.
Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
This will allow platforms to use an alternative way to get
the physical address of a symbol.
Signed-off-by: Leonid Yegoshin <Leonid.Yegoshin@imgtec.com>
Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
Add EVA cache flushing functions similar to non-EVA configurations.
Because the cache may or may not contain user virtual addresses, we
need to use the 'cache' or 'cachee' instruction based on whether we
flush the cache on behalf of kernel or user respectively.
Signed-off-by: Leonid Yegoshin <Leonid.Yegoshin@imgtec.com>
Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
Build code to invalidate an address range in the instruction cache
using the Hit Invalidate cache operation.
Signed-off-by: Leonid Yegoshin <Leonid.Yegoshin@imgtec.com>
Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
A MIPS specific csum_and_copy_from_user function is necessary because
the generic one from include/net/checksum.h will not work for EVA.
This is because the generic one will link to symbols from lib/checksum.c
which are not EVA aware.
Signed-off-by: Leonid Yegoshin <Leonid.Yegoshin@imgtec.com>
Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
In EVA mode, different instructions need to be used to read/write
from kernel and userland. In non-EVA mode, there is no functional
difference. The current address limit is checked to decide the
type of operation that will be performed.
Signed-off-by: Leonid Yegoshin <Leonid.Yegoshin@imgtec.com>
Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
The 'copy_user' symbol can be used to copy from or to
userland so we will use two different symbols for these
operations. This makes no difference in the existing code,
but when the core is operating in EVA mode, different instructions
need to be used to read and write to userland address space.
The old function has also been renamed to 'copy_kernel' to denote
that it is suitable for copy data to and from kernel space.
Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
The str*_user functions are used to securely access NULL terminated
strings from userland. Therefore, it's necessary to use the appropriate
EVA function. However, if the string is in kernel space, then the normal
instructions are being used to access it. The __str*_kernel_asm and
__str*_user_asm symbols are the same for non-EVA mode so there is no
functional change for the non-EVA kernels.
Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
Use the EVA specific functions from memcpy.S to perform
userspace operations. When get_fs() == get_ds() the usual load/store
instructions are used because the destination address is located in
the kernel address space region. Otherwise, the EVA specifc load/store
instructions are used which will go through th TLB to perform the virtual
to physical translation for the userspace address.
Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
The {get,put}_user_asm functions can be used to load data from
kernel or the user address space so rename them to avoid
confusion.
Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
Use the EVA instruction wrappers from asm.h to perform
read/write operations from userland.
Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
ulb, ulh, ulw are macros which emulate unaligned access for MIPS.
However, no such macros exist for EVA mode, so the only way to do
EVA unaligned accesses is in the ADE exception handler. As a result
of which, disable these macros for EVA.
Signed-off-by: Leonid Yegoshin <Leonid.Yegoshin@imgtec.com>
Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
Similar to __get_user_* functions, move common code to
__put_user_*_common so it can be shared among similar users.
Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
In preparation for EVA support, an instruction argument is needed
for the __get_user_asm{,_ll32} functions to allow instruction overrides in
EVA mode. Even though EVA only works for MIPS 32-bit, both codepaths are
changed (32-bit and 64-bit) for consistency reasons.
Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
Use LLE/SCE instructions for performing an address translation for
userspace when EVA is enabled.
Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
EVA uses specific instructions for accessing user memory.
Instead of polluting the kernel with numerous #ifdef CONFIG_EVA
we add wrappers for all the instructions that need special
handling when EVA is enabled.
Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
EVA can use the PREFE instruction to perform the virtual address
translation using the user mapping of the address rather than the
kernel mapping.
Signed-off-by: Leonid Yegoshin <Leonid.Yegoshin@imgtec.com>
Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
This patch extends sigcontext in order to hold the most significant 64
bits of each vector register in addition to the MSA control & status
register. The least significant 64 bits are already saved as the scalar
FP context. This makes things a little awkward since the least & most
significant 64 bits of each vector register are not contiguous in
memory. Thus the copy_u & insert instructions are used to transfer the
values of the most significant 64 bits via GP registers.
Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/6533/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
