When the system memory is exhausted, linux will trigger kswapd to
shrink memory page cache. We found the csky's .text file mapping
pages would be reclaimed earlier than arm's elf. Because csky
doesn't give _PAGE_ACCESSED for default pgprot and in zap_pte_range

if (pte_young(ptent) &&
    likely(!(vma->vm_flags & VM_SEQ_READ)))
        mark_page_accessed(page);

mark_page_accessed will put the pages into active lru list.

csky will throw the pages at first and keep them in active lru
list later after real accessed, but arm would keep them in active
lru list at the beginning.

The following are statistics of different architecture styles:

Default _PAGE_ACCESSED: alpha, arm, arm64, ia64, m68k, microblaze,
			openrisc, powerpc, riscv, sh, um, x86,
			xtensa
Not def _PAGE_ACCESSED: arc, c6x, h8300, hexgon, mips, s390, nds32,
			nios2, parisc, sparc

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>
Co-developed-by: Xu Kai <xukai@nationalchip.com>
Signed-off-by: Xu Kai <xukai@nationalchip.com>

When the timer counts to the upper limit, an overflow interrupt is
generated, and the count is reset with the value in the TIME_INI
register. But the software expects to start counting from 0 when
the count overflows, so it forces TIME_INI to 0 to solve the
potential interrupt storm problem.

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>
Tested-by: Xu Kai <xukai@nationalchip.com>
Cc: Daniel Lezcano <daniel.lezcano@linaro.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/1597735877-71115-1-git-send-email-guoren@kernel.org

We can't directly use CPUHP_AP_PERF_ONLINE as cpuhp_setup_state,
it will cause return -EBUSY. Define CPUHP_AP_PERF_CSKY_ONLINE just
as X86's will solve the problem.

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>
Tested-by: Chen Pei <hanyu.cp@alibaba-inc.com>
Reported-by: Xiang Yuanchao <xiangzhi.xyc@alibaba-inc.com>

We need sync.is to prevent potential ptw before tlb operation. And
let tlb_tfatal handler retry the epc.

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

There is no fixup or feature in the patch, we only cleanup with:

 - Remove unnecessary reg used (r11, r12), just use r9 & r10 &
   syscallid regs as temp useage.
 - Add _TIF_SYSCALL_WORK and _TIF_WORK_MASK to gather macros.

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

Current implementation could destory a4 & a5 when strace, so we need to get them
from pt_regs by SAVE_ALL.

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

log:
[    0.13373200] Calibrating delay loop...
[    0.14077600] ------------[ cut here ]------------
[    0.14116700] WARNING: CPU: 0 PID: 0 at kernel/sched/core.c:3790 preempt_count_add+0xc8/0x11c
[    0.14348000] DEBUG_LOCKS_WARN_ON((preempt_count() < 0))Modules linked in:
[    0.14395100] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 5.6.0 #7
[    0.14410800]
[    0.14427400] Call Trace:
[    0.14450700] [<807cd226>] dump_stack+0x8a/0xe4
[    0.14473500] [<80072792>] __warn+0x10e/0x15c
[    0.14495900] [<80072852>] warn_slowpath_fmt+0x72/0xc0
[    0.14518600] [<800a5240>] preempt_count_add+0xc8/0x11c
[    0.14544900] [<807ef918>] _raw_spin_lock+0x28/0x68
[    0.14572600] [<800e0eb8>] vprintk_emit+0x84/0x2d8
[    0.14599000] [<800e113a>] vprintk_default+0x2e/0x44
[    0.14625100] [<800e2042>] vprintk_func+0x12a/0x1d0
[    0.14651300] [<800e1804>] printk+0x30/0x48
[    0.14677600] [<80008052>] lockdep_init+0x12/0xb0
[    0.14703800] [<80002080>] start_kernel+0x558/0x7f8
[    0.14730000] [<800052bc>] csky_start+0x58/0x94
[    0.14756600] irq event stamp: 34
[    0.14775100] hardirqs last  enabled at (33): [<80067370>] ret_from_exception+0x2c/0x72
[    0.14793700] hardirqs last disabled at (34): [<800e0eae>] vprintk_emit+0x7a/0x2d8
[    0.14812300] softirqs last  enabled at (32): [<800655b0>] __do_softirq+0x578/0x6d8
[    0.14830800] softirqs last disabled at (25): [<8007b3b8>] irq_exit+0xec/0x128

The preempt_count of reg could be destroyed after csky_do_IRQ without reload
from memory.

After reference to other architectures (arm64, riscv), we move preempt entry
into ret_from_exception and disable irq at the beginning of
ret_from_exception instead of RESTORE_ALL.

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>
Reported-by: Lu Baoquan <lu.baoquan@intellif.com>

If raw_copy_from_user(to, from, N) returns K, callers expect
the first N - K bytes starting at to to have been replaced with
the contents of corresponding area starting at from and the last
K bytes of destination *left* *unmodified*.

What arch/sky/lib/usercopy.c is doing is broken - it can lead to e.g.
data corruption on write(2).

raw_copy_to_user() is inaccurate about return value, which is a bug,
but consequences are less drastic than for raw_copy_from_user().
And just what are those access_ok() doing in there?  I mean, look into
linux/uaccess.h; that's where we do that check (as well as zero tail
on failure in the callers that need zeroing).

AFAICS, all of that shouldn't be hard to fix; something like a patch
below might make a useful starting point.

I would suggest moving these macros into usercopy.c (they are never
used anywhere else) and possibly expanding them there; if you leave
them alive, please at least rename __copy_user_zeroing(). Again,
it must not zero anything on failed read.

Said that, I'm not sure we won't be better off simply turning
usercopy.c into usercopy.S - all that is left there is a couple of
functions, each consisting only of inline asm.

Guo Ren reply:

Yes, raw_copy_from_user is wrong, it's no need zeroing code.

unsigned long _copy_from_user(void *to, const void __user *from,
unsigned long n)
{
        unsigned long res = n;
        might_fault();
        if (likely(access_ok(from, n))) {
                kasan_check_write(to, n);
                res = raw_copy_from_user(to, from, n);
        }
        if (unlikely(res))
                memset(to + (n - res), 0, res);
        return res;
}
EXPORT_SYMBOL(_copy_from_user);

You are right and access_ok() should be removed.

but, how about:
do {
...
        "2:     stw     %3, (%1, 0)     \n"             \
+       "       subi    %0, 4          \n"               \
        "9:     stw     %4, (%1, 4)     \n"             \
+       "       subi    %0, 4          \n"               \
        "10:    stw     %5, (%1, 8)     \n"             \
+       "       subi    %0, 4          \n"               \
        "11:    stw     %6, (%1, 12)    \n"             \
+       "       subi    %0, 4          \n"               \
        "       addi    %2, 16          \n"             \
        "       addi    %1, 16          \n"             \

Don't expand __ex_table

AI Viro reply:

Hey, I've no idea about the instruction scheduling on csky -
if that doesn't slow the things down, all the better.  It's just
that copy_to_user() and friends are on fairly hot codepaths,
and in quite a few situations they will dominate the speed of
e.g. read(2).  So I tried to keep the fast path unchanged.
Up to the architecture maintainers, obviously.  Which would be
you...

As for the fixups size increase (__ex_table size is unchanged)...
You have each of those macros expanded exactly once.
So the size is not a serious argument, IMO - useless complexity
would be, if it is, in fact, useless; the size... not really,
especially since those extra subi will at least offset it.

Again, up to you - asm optimizations of (essentially)
memcpy()-style loops are tricky and can depend upon the
fairly subtle details of architecture.  So even on something
I know reasonably well I would resort to direct experiments
if I can't pass the buck to architecture maintainers.

It *is* worth optimizing - this is where read() from a file
that is already in page cache spends most of the time, etc.

Guo Ren reply:

Thx, after fixup some typo “sub %0, 4”, apply the patch.

TODO:
 - user copy/from codes are still need optimizing.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

The gdbmacros.txt use sp in thread_struct, but csky use ksp. This
cause bttnobp fail to excute.

TODO:
 - Still couldn't display the contents of stack.

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

All processes' PSR could success from SETUP_MMU, so need set it
in INIT_THREAD again.

And use a3 instead of r7 in __switch_to for code convention.

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

Interrupt has been disabled in __schedule() with local_irq_disable()
and enabled in finish_task_switch->finish_lock_switch() with
local_irq_enabled(), So needn't to disable irq here.

Signed-off-by: Liu Yibin <jiulong@linux.alibaba.com>
Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

The implementation of show_stack will panic with wrong fp:

addr    = *fp++;

because the fp isn't checked properly.

The current implementations of show_stack, wchan and stack_trace
haven't been designed properly, so just deprecate them.

This patch is a reference to riscv's way, all codes are modified from
arm's. The patch is passed with:

 - cat /proc/<pid>/stack
 - cat /proc/<pid>/wchan
 - echo c > /proc/sysrq-trigger

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

 [ 5221.974084] Unable to handle kernel paging request at virtual address 0xfffff000, pc: 0x8002c18e
 [ 5221.985929] Oops: 00000000
 [ 5221.989488]
 [ 5221.989488] CURRENT PROCESS:
 [ 5221.989488]
 [ 5221.992877] COMM=callchain_test PID=11962
 [ 5221.995213] TEXT=00008000-000087e0 DATA=00009f1c-0000a018 BSS=0000a018-0000b000
 [ 5221.999037] USER-STACK=7fc18e20  KERNEL-STACK=be204680
 [ 5221.999037]
 [ 5222.003292] PC: 0x8002c18e (perf_callchain_kernel+0x3e/0xd4)
 [ 5222.007957] LR: 0x8002c198 (perf_callchain_kernel+0x48/0xd4)
 [ 5222.074873] Call Trace:
 [ 5222.074873] [<800a248e>] get_perf_callchain+0x20a/0x29c
 [ 5222.074873] [<8009d964>] perf_callchain+0x64/0x80
 [ 5222.074873] [<8009dc1c>] perf_prepare_sample+0x29c/0x4b8
 [ 5222.074873] [<8009de6e>] perf_event_output_forward+0x36/0x98
 [ 5222.074873] [<800497e0>] search_exception_tables+0x20/0x44
 [ 5222.074873] [<8002cbb6>] do_page_fault+0x92/0x378
 [ 5222.074873] [<80098608>] __perf_event_overflow+0x54/0xdc
 [ 5222.074873] [<80098778>] perf_swevent_hrtimer+0xe8/0x164
 [ 5222.074873] [<8002ddd0>] update_mmu_cache+0x0/0xd8
 [ 5222.074873] [<8002c014>] user_backtrace+0x58/0xc4
 [ 5222.074873] [<8002c0b4>] perf_callchain_user+0x34/0xd0
 [ 5222.074873] [<800a2442>] get_perf_callchain+0x1be/0x29c
 [ 5222.074873] [<8009d964>] perf_callchain+0x64/0x80
 [ 5222.074873] [<8009d834>] perf_output_sample+0x78c/0x858
 [ 5222.074873] [<8009dc1c>] perf_prepare_sample+0x29c/0x4b8
 [ 5222.074873] [<8009de94>] perf_event_output_forward+0x5c/0x98
 [ 5222.097846]
 [ 5222.097846] [<800a0300>] perf_event_exit_task+0x58/0x43c
 [ 5222.097846] [<8006c874>] hrtimer_interrupt+0x104/0x2ec
 [ 5222.097846] [<800a0300>] perf_event_exit_task+0x58/0x43c
 [ 5222.097846] [<80437bb6>] dw_apb_clockevent_irq+0x2a/0x4c
 [ 5222.097846] [<8006c770>] hrtimer_interrupt+0x0/0x2ec
 [ 5222.097846] [<8005f2e4>] __handle_irq_event_percpu+0xac/0x19c
 [ 5222.097846] [<80437bb6>] dw_apb_clockevent_irq+0x2a/0x4c
 [ 5222.097846] [<8005f408>] handle_irq_event_percpu+0x34/0x88
 [ 5222.097846] [<8005f480>] handle_irq_event+0x24/0x64
 [ 5222.097846] [<8006218c>] handle_level_irq+0x68/0xdc
 [ 5222.097846] [<8005ec76>] __handle_domain_irq+0x56/0xa8
 [ 5222.097846] [<80450e90>] ck_irq_handler+0xac/0xe4
 [ 5222.097846] [<80029012>] csky_do_IRQ+0x12/0x24
 [ 5222.097846] [<8002a3a0>] csky_irq+0x70/0x80
 [ 5222.097846] [<800ca612>] alloc_set_pte+0xd2/0x238
 [ 5222.097846] [<8002ddd0>] update_mmu_cache+0x0/0xd8
 [ 5222.097846] [<800a0340>] perf_event_exit_task+0x98/0x43c

The original fp check doesn't base on the real kernal stack region.
Invalid fp address may cause kernel panic.

Signed-off-by: Mao Han <han_mao@linux.alibaba.com>
Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

Just as comment mentioned, the msa format:

 cr<30/31, 15> MSA register format:
 31 - 29 | 28 - 9 | 8 | 7 | 6 | 5 | 4 | 3 | 2 | 1 | 0
   BA     Reserved  SH  WA  B   SO SEC  C   D   V

So we should shift 29 bits not 28 bits for mask

Signed-off-by: Liu Yibin <jiulong@linux.alibaba.com>
Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

case:
 # perf probe -x /lib/libc-2.28.9000.so memcpy
 # perf record -e probe_libc:memcpy -aR sleep 1

System hangup and cpu get in trap_c loop, because our hardware
singlestep state could still get interrupt signal. When we get in
uprobe_xol singlestep slot, we should disable irq in pt_regs->psr.

And is_swbp_insn() need a csky arch implementation with a low 16bit
mask.

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

610 don't support KPROBES

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

It's different from linux-next

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

When CONFIG_DYNAMIC_FTRACE is enabled, static ftrace will fail to
boot up and compile. It's a carelessness when developing "dynamic
ftrace" and "ftrace with regs".

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

This patch adds support for uprobes on csky architecture.

Just like kprobe, it support single-step and simulate instructions.

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Steven Rostedt (VMware) <rostedt@goodmis.org>

This patch enable kprobes, kretprobes, ftrace interface. It utilized
software breakpoint and single step debug exceptions, instructions
simulation on csky.

We use USR_BKPT replace origin instruction, and the kprobe handler
prepares an excutable memory slot for out-of-line execution with a
copy of the original instruction being probed. Most of instructions
could be executed by single-step, but some instructions need origin
pc value to execute and we need software simulate these instructions.

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Steven Rostedt (VMware) <rostedt@goodmis.org>

Lockdep is needed by proving the spinlocks and rwlocks. Currently,
we only put trace_hardirqs_on/off with csky_irq and
ret_from_exception.

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

We should get psr value from regs->psr in stack, not directly get
it from phyiscal register then save the vector number in
tsk->trap_no.

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

If ICACHE_INS is not supported, we use IPI to sync icache on each
core. But ftrace_modify_code is called from stop_machine from default
implementation of arch_ftrace_update_code and stop_machine callback
is irq_disabled. When you call ipi with irq_disabled, a deadlock will
happen.

We couldn't use icache_flush with irq_disabled, but startup make_nop
is specific case and it needn't ipi other cores.

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

This patch implements FTRACE_WITH_REGS for csky, which allows a traced
function's arguments (and some other registers) to be captured into a
struct pt_regs, allowing these to be inspected and/or modified.

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

Needed for kprobes support. Copied and adapted from Patrick's patch,
but it has been modified for csky's pt_regs.

ref:
https://lore.kernel.org/linux-riscv/1572919114-3886-2-git-send-email-vincent.chen@sifive.com/raw



Signed-off-by: Guo Ren <ren_guo@c-sky.com>
Cc: Patrick Staehlin <me@packi.ch>

WARNING: vmlinux.o(.text+0x2366): Section mismatch in reference from the
function csky_start_secondary() to the function .init.text:init_fpu()

The function csky_start_secondary() references
the function __init init_fpu().
This is often because csky_start_secondary lacks a __init
annotation or the annotation of init_fpu is wrong.

Reported-by: Lu Chongzhi <chongzhi.lcz@alibaba-inc.com>
Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

For the memory size ( > 512MB, < 1GB), the MSA setting is:

 - SSEG0: PHY_START        , PHY_START + 512MB
 - SSEG1: PHY_START + 512MB, PHY_START + 1GB

But the real memory is no more than 1GB, there is a gap between the
end size of memory and border of 1GB. CPU could speculatively
execute to that gap and if the gap of the bus couldn't respond to
the CPU request, then the crash will happen.

Now make the setting with:

 - SSEG0: PHY_START        , PHY_START + 512MB (no change)
 - SSEG1: Disabled (We use highmem to use the memory of 512MB~1GB)

We also deprecated zhole_szie[] settings, it's only used by arm
style CPUs. All memory gap should use Reserved setting of dts in
csky system.

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

Support the gcov function in csky architecture

Signed-off-by: Ma Jun <majun258@linux.alibaba.com>

TO use the defconfig of csky arch,  the KBUILD_DEFCONFIG definition
is added. orelse there is compile error.

Signed-off-by: Ma Jun <majun258@linux.alibaba.com>

Add the pci related code for csky arch to support basic pci virtual
function, such as qemu virt-pci-9pfs.

Signed-off-by: Ma Jun <majun258@linux.alibaba.com>
Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

Some bsp (eg: buildroot) has defconfig.fragment design to add more
configs into the defconfig in linux source code tree. For example,
we could put different cpu configs into different defconfig.fragments,
but they all use the same defconfig in Linux.

Signed-off-by: Ma Jun <majun258@linux.alibaba.com>
Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

We should give some necessary check for initrd just like other
architectures and it seems that setup_initrd() could be a common
code for all architectures.

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

During ftrace init, linux will replace all function prologues
(call_mcout) with nops, but it need flush_dcache and
invalidate_icache to make it work. So flush_cache functions
couldn't be nested called by ftrace framework.

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

Only for coding convention, these codes are unnecessary for abiv2.

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

Some CPUs don't support icache.va instruction to maintain the whole
smp cores' icache. Using icache.all + IPI casue a lot on performace
and using defer mechanism could reduce the number of calling icache
_flush_all functions.

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

Only when vma is for VM_EXEC, we need sync dcache & icache. eg:
 - gdb ptrace modify user space instruction code area.

Add VM_EXEC condition to reduce unnecessary cache flush.

The abiv1 cpus' cache are all VIPT, so we still need to deal with
dcache aliasing problem. But there is optimized way to use cache
color, just like what's done in arch/csky/abiv1/inc/abi/page.h.

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

Instead of flushing cache per update_mmu_cache() called, we use
flush_dcache_page to reduce the frequency of flashing the cache.

As abiv2 cpus are all PIPT for icache & dcache, we needn't handle
dcache aliasing problem. But their icache can't snoop dcache, so
we still need sync_icache_dcache in update_mmu_cache().

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>

The abiv2 CPUs are all PIPT cache, so there is no need to implement
flush_icache_page function.

The function flush_icache_user_range hasn't been used, so just
remove it.

The function flush_cache_range is not necessary for PIPT cache when
tlb mapping changed.

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>