caculate blk address when mempool_multiple_free
have a bug. need a real blocksize to caulate the
memory address.
Signed-off-by: anjiahao <anjiahao@xiaomi.com>
D:\archer\code\nuttx\mm\mempool\mempool_multiple.c(180,72): warning C4098: "mempool_multiple_free_callback":"void" void function returning a value
Compiler Warning C4098:
A function declared with return type void has a return statement that returns a value. The compiler assumes the function returns a value of type int.
Reference:
https://learn.microsoft.com/en-us/cpp/error-messages/compiler-warnings/compiler-warning-level-1-c4098?view=msvc-170
Signed-off-by: chao an <anchao@xiaomi.com>
NuttX kernel should not use the syscall functions, especially after
enabling CONFIG_SCHED_INSTRUMENTATION_SYSCALL, all system functions
will be traced to backend, which will impact system performance.
Signed-off-by: chao an <anchao@xiaomi.com>
Original code assumes the previous node in physical must be
ALLOCED, but other thread may free the previous node between
mm_malloc() and mm_lock(), and the original code didn't condsider
this, which will cause two adjacent free nodes situation and this
should not be happened.
This commit will merge the previous node and node if the previous
node is free to avoid the situation discribed above.
Signed-off-by: wangbowen6 <wangbowen6@xiaomi.com>
To avoid lto alias with override __asan_* symbols
undefined reference to `__asan_load4_noabort'
undefined reference to `__asan_load1_noabort'
undefined reference to `__asan_store1_noabort'
undefined reference to `__asan_load1_noabort'
undefined reference to `__asan_store1_noabort'
undefined reference to `__asan_load4_noabort'
undefined reference to `__asan_store4_noabort'
Signed-off-by: chao.an <anchao@xiaomi.com>
When adding more heap memory, the total heap size was not updated. This
results in a crash in mm_mallinfo:
DEBUGASSERT((size_t)info->uordblks + info->fordblks == heap->mm_heapsize);
This commit fixes this issue
Cast substraction arguments to FAR char *, which gives the same result as the
gcc extension on the original void * arithmetic.
Signed-off-by: Jukka Laitinen <jukkax@ssrc.tii.ae>
There are many small memory block in NuttX system, eg: struct tcb_s,
struct inode, etc, and several disadvantages about them:
1.Their frequent allocate and free cause the system memory fragmentation.
2.Since each memory block has an overhead, the utilization of small memory
blocks is relatively low, which will cause memory waste.
So we can use mempool to alloc smallo block, to improve alloc speed
and utilization, to reduce fragmentation.
Signed-off-by: dongjiuzhu1 <dongjiuzhu1@xiaomi.com>
do simple copy to instead of memset and memcpy operation because
they have been instrumented, if you access the posion area,
the system will crash.
Signed-off-by: dongjiuzhu1 <dongjiuzhu1@xiaomi.com>
The memalign is special to multiple mempool because multiple mempool
doesn't support split and shrink chunk operate. So When you alloc a
memory block and find an aligned address in this block, you need to
occupy 8 bytes before the address to save the address of the padding
size and pool to ensure correct use in realloc and free operations.
So we will use bit1 in the previous address of the address to represent
that it is applied by memalign.
Signed-off-by: dongjiuzhu1 <dongjiuzhu1@xiaomi.com>
kasan_report (addr=0xf3c68618, size=1, is_write=false) at
kasan/kasan.c:106
0x56585fbf in __asan_loadN_noabort (addr=0xf3c68618, size=1) at
kasan/kasan.c:300
0x565860ac in __asan_load1_noabort (addr=0xf3c68618) at
kasan/kasan.c:354
0x565843af in memcpy (dest=0xf3de9d6c, src=0xf3c685cc, n=3) at
string/lib_memcpy.c:44
0x56587ae8 in mempool_multiple_realloc (mpool=0xf3c670fc,
oldblk=0xf3c685cc, size=416) at mempool/mempool_multiple.c:218
0x5658707a in mm_realloc (heap=0xf3c67000, oldmem=0xf3c685cc,
size=416) at mm_heap/mm_realloc.c:98
0x5658524e in realloc (oldmem=0xf3c685cc, size=416) at
umm_heap/umm_realloc.c:97
Signed-off-by: dongjiuzhu1 <dongjiuzhu1@xiaomi.com>
This delta describes the relationship between the block size of each
mempool in multiple mempool by user initialized. It is automatically
detected by the mempool_multiple_init function. If the delta is not
equal to 0, the block size of the pool in the multiple mempool is an
arithmetic progressions, otherwise it is an increasing progressions.
Signed-off-by: dongjiuzhu1 <dongjiuzhu1@xiaomi.com>
Let's specify size instead of number, so that we can unify the size of
expansion memory in the multiple mempool.
Signed-off-by: dongjiuzhu1 <dongjiuzhu1@xiaomi.com>
the crash backtrace:
kasan_report (addr=0xf3d02fd4, size=4, is_write=false) at
kasan/kasan.c:106
0x5658518d in __asan_loadN_noabort (addr=0xf3d02fd4, size=4) at
kasan/kasan.c:300
0x565851ee in __asan_load4_noabort (addr=0xf3d02fd4) at
kasan/kasan.c:334
0x56580b02 in sq_remfirst (queue=0xf3d02b08) at
queue/sq_remfirst.c:45
0x565e0e0b in mempool_alloc (pool=0xf3d02aec) at
mempool/mempool.c:161
0x566033d2 in mempool_multiple_alloc (mpool=0xf3d02a30, size=16) at
mempool/mempool_multiple.c:147
Signed-off-by: dongjiuzhu1 <dongjiuzhu1@xiaomi.com>
- Also remove the nuttx private shm.h file nuttx/mm/shm.h, which became redundant
- Also remove the gran allocator initialization/release in binfmt since common
vpage allocator is initialized in group_create/group_leave
Signed-off-by: Jukka Laitinen <jukkax@ssrc.tii.ae>
The dependency should be vice versa; the MM_SHM should depend on the
existence of the virtual memory range allocator.
Create a new CONFIG flag CONFIG_ARCH_VMA_MAPPING, which will define that
there is a virtual memory range allocator. Make MM_SHM select that flag
Signed-off-by: Jukka Laitinen <jukkax@ssrc.tii.ae>
Replace static gs_vaddr with a new dynamic mapping list. Collecting all
this kind of virtual memory mappings into a single structure makes
things more consistent.
This still leaves the task group specific granule alloocator, gs_handle,
in the task group
Signed-off-by: Jukka Laitinen <jukkax@ssrc.tii.ae>
The task_group specific list can be used to store information about
mmappings.
For a driver or filesystem performing mmap can also enable munmap by
adding an item to this list using mm_map_add(). The item is then
returned in the corresponding munmap call.
Signed-off-by: Jukka Laitinen <jukkax@ssrc.tii.ae>
ubsan/ubsan.c: In function ‘get_signed_val’:
ubsan/ubsan.c:162:22: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast]
162 | uint64_t ret = (uint64_t)val & mask;
|
Signed-off-by: chao an <anchao@xiaomi.com>
This will allow changing the underlying backend, i.e. granule allocator,
to something more efficient, if this is needed later.
Also, it gives easier access to the common SHM book keeping.
User can ask for specific granules to be allocated.
This is useful for one thing only: when mmap() is called for a specific
vaddr. The parameter itself is non-sensical, this is just to satisfy
the POSIX standard.
In the past, header file paths were generated by the incdir command
Now they are generated by concatenating environment variables
In this way, when executing makefile, no shell command will be executed,
it will improve the speed of executing makfile
Signed-off-by: yinshengkai <yinshengkai@xiaomi.com>
In the past, predefined macros were generated by define.sh scripts
Now they are generated by concatenating environment variables
In this way, when executing makefile, no shell command will be executed,
it will improve the speed of executing makfile
Signed-off-by: yinshengkai <yinshengkai@xiaomi.com>
