maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/lldb/source/Target/Memory.cpp

462 lines
16 KiB
C++
Raw Normal View History

Added an allocated memory cache to avoid having to allocate memory over and over when running JITed expressions. The allocated memory cache will cache allocate memory a page at a time for each permission combination and divvy up the memory and hand it out in 16 byte increments. llvm-svn: 131453
2011-05-17 11:37:42 +08:00
//===-- Memory.cpp ----------------------------------------------*- C++ -*-===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
#include "lldb/Target/Memory.h"
// C Includes
// C++ Includes
// Other libraries and framework includes
// Project includes
#include "lldb/Core/DataBufferHeap.h"
#include "lldb/Core/State.h"
#include "lldb/Core/Log.h"
#include "lldb/Target/Process.h"
using namespace lldb;
using namespace lldb_private;
//----------------------------------------------------------------------
// MemoryCache constructor
//----------------------------------------------------------------------
MemoryCache::MemoryCache(Process &process) :
m_process (process),
m_cache_line_byte_size (512),
For stepping performance I added the ability to outlaw all memory accesseses to the __PAGEZERO segment on darwin. The dynamic loader now correctly doesn't slide __PAGEZERO and it also registers it as an invalid region of memory. This allows us to not make any memory requests from the local or remote debug session for any addresses in this region. Stepping performance can improve when uninitialized local variables that point to locations in __PAGEZERO are attempted to be read from memory as we won't even make the memory read or write request. llvm-svn: 151128
2012-02-22 12:37:26 +08:00
m_mutex (Mutex::eMutexTypeRecursive),
m_cache (),
m_invalid_ranges ()
Added an allocated memory cache to avoid having to allocate memory over and over when running JITed expressions. The allocated memory cache will cache allocate memory a page at a time for each permission combination and divvy up the memory and hand it out in 16 byte increments. llvm-svn: 131453
2011-05-17 11:37:42 +08:00
{
}
//----------------------------------------------------------------------
// Destructor
//----------------------------------------------------------------------
MemoryCache::~MemoryCache()
{
}
void
<rdar://problem/13892516> LLDB can now debug across calls to exec when the architecture changes from say i386 to x86_64 (on darwin). llvm-svn: 182345
2013-05-21 09:00:52 +08:00
MemoryCache::Clear(bool clear_invalid_ranges)
Added an allocated memory cache to avoid having to allocate memory over and over when running JITed expressions. The allocated memory cache will cache allocate memory a page at a time for each permission combination and divvy up the memory and hand it out in 16 byte increments. llvm-svn: 131453
2011-05-17 11:37:42 +08:00
{
For stepping performance I added the ability to outlaw all memory accesseses to the __PAGEZERO segment on darwin. The dynamic loader now correctly doesn't slide __PAGEZERO and it also registers it as an invalid region of memory. This allows us to not make any memory requests from the local or remote debug session for any addresses in this region. Stepping performance can improve when uninitialized local variables that point to locations in __PAGEZERO are attempted to be read from memory as we won't even make the memory read or write request. llvm-svn: 151128
2012-02-22 12:37:26 +08:00
Mutex::Locker locker (m_mutex);
Added an allocated memory cache to avoid having to allocate memory over and over when running JITed expressions. The allocated memory cache will cache allocate memory a page at a time for each permission combination and divvy up the memory and hand it out in 16 byte increments. llvm-svn: 131453
2011-05-17 11:37:42 +08:00
m_cache.clear();
<rdar://problem/13892516> LLDB can now debug across calls to exec when the architecture changes from say i386 to x86_64 (on darwin). llvm-svn: 182345
2013-05-21 09:00:52 +08:00
if (clear_invalid_ranges)
m_invalid_ranges.Clear();
Added an allocated memory cache to avoid having to allocate memory over and over when running JITed expressions. The allocated memory cache will cache allocate memory a page at a time for each permission combination and divvy up the memory and hand it out in 16 byte increments. llvm-svn: 131453
2011-05-17 11:37:42 +08:00
}
void
MemoryCache::Flush (addr_t addr, size_t size)
{
if (size == 0)
return;
<rdar://problem/11193466> Fixed an error where lldb would hang when writing memory near the end of the addres space due to an unsigned overflow. llvm-svn: 154697
2012-04-14 04:37:20 +08:00
For stepping performance I added the ability to outlaw all memory accesseses to the __PAGEZERO segment on darwin. The dynamic loader now correctly doesn't slide __PAGEZERO and it also registers it as an invalid region of memory. This allows us to not make any memory requests from the local or remote debug session for any addresses in this region. Stepping performance can improve when uninitialized local variables that point to locations in __PAGEZERO are attempted to be read from memory as we won't even make the memory read or write request. llvm-svn: 151128
2012-02-22 12:37:26 +08:00
Mutex::Locker locker (m_mutex);
Added an allocated memory cache to avoid having to allocate memory over and over when running JITed expressions. The allocated memory cache will cache allocate memory a page at a time for each permission combination and divvy up the memory and hand it out in 16 byte increments. llvm-svn: 131453
2011-05-17 11:37:42 +08:00
if (m_cache.empty())
return;
<rdar://problem/11193466> Fixed an error where lldb would hang when writing memory near the end of the addres space due to an unsigned overflow. llvm-svn: 154697
2012-04-14 04:37:20 +08:00
const uint32_t cache_line_byte_size = m_cache_line_byte_size;
const addr_t end_addr = (addr + size - 1);
const addr_t first_cache_line_addr = addr - (addr % cache_line_byte_size);
const addr_t last_cache_line_addr = end_addr - (end_addr % cache_line_byte_size);
// Watch for overflow where size will cause us to go off the end of the
// 64 bit address space
uint32_t num_cache_lines;
if (last_cache_line_addr >= first_cache_line_addr)
num_cache_lines = ((last_cache_line_addr - first_cache_line_addr)/cache_line_byte_size) + 1;
else
num_cache_lines = (UINT64_MAX - first_cache_line_addr + 1)/cache_line_byte_size;
uint32_t cache_idx = 0;
for (addr_t curr_addr = first_cache_line_addr;
cache_idx < num_cache_lines;
curr_addr += cache_line_byte_size, ++cache_idx)
Added an allocated memory cache to avoid having to allocate memory over and over when running JITed expressions. The allocated memory cache will cache allocate memory a page at a time for each permission combination and divvy up the memory and hand it out in 16 byte increments. llvm-svn: 131453
2011-05-17 11:37:42 +08:00
{
For stepping performance I added the ability to outlaw all memory accesseses to the __PAGEZERO segment on darwin. The dynamic loader now correctly doesn't slide __PAGEZERO and it also registers it as an invalid region of memory. This allows us to not make any memory requests from the local or remote debug session for any addresses in this region. Stepping performance can improve when uninitialized local variables that point to locations in __PAGEZERO are attempted to be read from memory as we won't even make the memory read or write request. llvm-svn: 151128
2012-02-22 12:37:26 +08:00
BlockMap::iterator pos = m_cache.find (curr_addr);
Added an allocated memory cache to avoid having to allocate memory over and over when running JITed expressions. The allocated memory cache will cache allocate memory a page at a time for each permission combination and divvy up the memory and hand it out in 16 byte increments. llvm-svn: 131453
2011-05-17 11:37:42 +08:00
if (pos != m_cache.end())
m_cache.erase(pos);
}
}
For stepping performance I added the ability to outlaw all memory accesseses to the __PAGEZERO segment on darwin. The dynamic loader now correctly doesn't slide __PAGEZERO and it also registers it as an invalid region of memory. This allows us to not make any memory requests from the local or remote debug session for any addresses in this region. Stepping performance can improve when uninitialized local variables that point to locations in __PAGEZERO are attempted to be read from memory as we won't even make the memory read or write request. llvm-svn: 151128
2012-02-22 12:37:26 +08:00
void
MemoryCache::AddInvalidRange (lldb::addr_t base_addr, lldb::addr_t byte_size)
{
if (byte_size > 0)
{
Mutex::Locker locker (m_mutex);
InvalidRanges::Entry range (base_addr, byte_size);
m_invalid_ranges.Append(range);
m_invalid_ranges.Sort();
}
}
bool
MemoryCache::RemoveInvalidRange (lldb::addr_t base_addr, lldb::addr_t byte_size)
{
if (byte_size > 0)
{
Mutex::Locker locker (m_mutex);
const uint32_t idx = m_invalid_ranges.FindEntryIndexThatContains(base_addr);
if (idx != UINT32_MAX)
{
const InvalidRanges::Entry *entry = m_invalid_ranges.GetEntryAtIndex (idx);
if (entry->GetRangeBase() == base_addr && entry->GetByteSize() == byte_size)
return m_invalid_ranges.RemoveEntrtAtIndex (idx);
}
}
return false;
}
Added an allocated memory cache to avoid having to allocate memory over and over when running JITed expressions. The allocated memory cache will cache allocate memory a page at a time for each permission combination and divvy up the memory and hand it out in 16 byte increments. llvm-svn: 131453
2011-05-17 11:37:42 +08:00
size_t
MemoryCache::Read (addr_t addr,
void *dst,
size_t dst_len,
Error &error)
{
size_t bytes_left = dst_len;
if (dst && bytes_left > 0)
{
const uint32_t cache_line_byte_size = m_cache_line_byte_size;
uint8_t *dst_buf = (uint8_t *)dst;
addr_t curr_addr = addr - (addr % cache_line_byte_size);
addr_t cache_offset = addr - curr_addr;
For stepping performance I added the ability to outlaw all memory accesseses to the __PAGEZERO segment on darwin. The dynamic loader now correctly doesn't slide __PAGEZERO and it also registers it as an invalid region of memory. This allows us to not make any memory requests from the local or remote debug session for any addresses in this region. Stepping performance can improve when uninitialized local variables that point to locations in __PAGEZERO are attempted to be read from memory as we won't even make the memory read or write request. llvm-svn: 151128
2012-02-22 12:37:26 +08:00
Mutex::Locker locker (m_mutex);
Added an allocated memory cache to avoid having to allocate memory over and over when running JITed expressions. The allocated memory cache will cache allocate memory a page at a time for each permission combination and divvy up the memory and hand it out in 16 byte increments. llvm-svn: 131453
2011-05-17 11:37:42 +08:00
while (bytes_left > 0)
{
For stepping performance I added the ability to outlaw all memory accesseses to the __PAGEZERO segment on darwin. The dynamic loader now correctly doesn't slide __PAGEZERO and it also registers it as an invalid region of memory. This allows us to not make any memory requests from the local or remote debug session for any addresses in this region. Stepping performance can improve when uninitialized local variables that point to locations in __PAGEZERO are attempted to be read from memory as we won't even make the memory read or write request. llvm-svn: 151128
2012-02-22 12:37:26 +08:00
if (m_invalid_ranges.FindEntryThatContains(curr_addr))
If the MemoryCache is asked to look up an address that is in an invalid range, it should set the error in the Error passed into the function. llvm-svn: 181386
2013-05-08 09:20:53 +08:00
{
error.SetErrorStringWithFormat("memory read failed for 0x%" PRIx64, curr_addr);
For stepping performance I added the ability to outlaw all memory accesseses to the __PAGEZERO segment on darwin. The dynamic loader now correctly doesn't slide __PAGEZERO and it also registers it as an invalid region of memory. This allows us to not make any memory requests from the local or remote debug session for any addresses in this region. Stepping performance can improve when uninitialized local variables that point to locations in __PAGEZERO are attempted to be read from memory as we won't even make the memory read or write request. llvm-svn: 151128
2012-02-22 12:37:26 +08:00
return dst_len - bytes_left;
If the MemoryCache is asked to look up an address that is in an invalid range, it should set the error in the Error passed into the function. llvm-svn: 181386
2013-05-08 09:20:53 +08:00
}
For stepping performance I added the ability to outlaw all memory accesseses to the __PAGEZERO segment on darwin. The dynamic loader now correctly doesn't slide __PAGEZERO and it also registers it as an invalid region of memory. This allows us to not make any memory requests from the local or remote debug session for any addresses in this region. Stepping performance can improve when uninitialized local variables that point to locations in __PAGEZERO are attempted to be read from memory as we won't even make the memory read or write request. llvm-svn: 151128
2012-02-22 12:37:26 +08:00
BlockMap::const_iterator pos = m_cache.find (curr_addr);
BlockMap::const_iterator end = m_cache.end ();
Added an allocated memory cache to avoid having to allocate memory over and over when running JITed expressions. The allocated memory cache will cache allocate memory a page at a time for each permission combination and divvy up the memory and hand it out in 16 byte increments. llvm-svn: 131453
2011-05-17 11:37:42 +08:00
if (pos != end)
{
size_t curr_read_size = cache_line_byte_size - cache_offset;
if (curr_read_size > bytes_left)
curr_read_size = bytes_left;
memcpy (dst_buf + dst_len - bytes_left, pos->second->GetBytes() + cache_offset, curr_read_size);
bytes_left -= curr_read_size;
curr_addr += curr_read_size + cache_offset;
cache_offset = 0;
if (bytes_left > 0)
{
// Get sequential cache page hits
for (++pos; (pos != end) && (bytes_left > 0); ++pos)
{
assert ((curr_addr % cache_line_byte_size) == 0);
if (pos->first != curr_addr)
break;
curr_read_size = pos->second->GetByteSize();
if (curr_read_size > bytes_left)
curr_read_size = bytes_left;
memcpy (dst_buf + dst_len - bytes_left, pos->second->GetBytes(), curr_read_size);
bytes_left -= curr_read_size;
curr_addr += curr_read_size;
// We have a cache page that succeeded to read some bytes
// but not an entire page. If this happens, we must cap
// off how much data we are able to read...
if (pos->second->GetByteSize() != cache_line_byte_size)
return dst_len - bytes_left;
}
}
}
// We need to read from the process
if (bytes_left > 0)
{
assert ((curr_addr % cache_line_byte_size) == 0);
After discussing with Chris Lattner, we require C++11, so lets get rid of the macros and just use C++11. llvm-svn: 179805
2013-04-19 06:45:39 +08:00
std::unique_ptr<DataBufferHeap> data_buffer_heap_ap(new DataBufferHeap (cache_line_byte_size, 0));
Added an allocated memory cache to avoid having to allocate memory over and over when running JITed expressions. The allocated memory cache will cache allocate memory a page at a time for each permission combination and divvy up the memory and hand it out in 16 byte increments. llvm-svn: 131453
2011-05-17 11:37:42 +08:00
size_t process_bytes_read = m_process.ReadMemoryFromInferior (curr_addr,
data_buffer_heap_ap->GetBytes(),
data_buffer_heap_ap->GetByteSize(),
error);
if (process_bytes_read == 0)
return dst_len - bytes_left;
if (process_bytes_read != cache_line_byte_size)
data_buffer_heap_ap->SetByteSize (process_bytes_read);
m_cache[curr_addr] = DataBufferSP (data_buffer_heap_ap.release());
// We have read data and put it into the cache, continue through the
// loop again to get the data out of the cache...
}
}
}
return dst_len - bytes_left;
}
AllocatedBlock::AllocatedBlock (lldb::addr_t addr,
uint32_t byte_size,
uint32_t permissions,
uint32_t chunk_size) :
m_addr (addr),
m_byte_size (byte_size),
m_permissions (permissions),
m_chunk_size (chunk_size),
m_offset_to_chunk_size ()
// m_allocated (byte_size / chunk_size)
{
assert (byte_size > chunk_size);
}
AllocatedBlock::~AllocatedBlock ()
{
}
lldb::addr_t
AllocatedBlock::ReserveBlock (uint32_t size)
{
addr_t addr = LLDB_INVALID_ADDRESS;
if (size <= m_byte_size)
{
const uint32_t needed_chunks = CalculateChunksNeededForSize (size);
<rdar://problem/13521159> LLDB is crashing when logging is enabled from lldb-perf-clang. This has to do with the global destructor chain as the process and its threads are being torn down. All logging channels now make one and only one instance that is kept in a global pointer which is never freed. This guarantees that logging can correctly continue as the process tears itself down. llvm-svn: 178191
2013-03-28 07:08:40 +08:00
Log *log (GetLogIfAllCategoriesSet (LIBLLDB_LOG_PROCESS | LIBLLDB_LOG_VERBOSE));
Added an allocated memory cache to avoid having to allocate memory over and over when running JITed expressions. The allocated memory cache will cache allocate memory a page at a time for each permission combination and divvy up the memory and hand it out in 16 byte increments. llvm-svn: 131453
2011-05-17 11:37:42 +08:00
if (m_offset_to_chunk_size.empty())
{
m_offset_to_chunk_size[0] = needed_chunks;
if (log)
log->Printf ("[1] AllocatedBlock::ReserveBlock (size = %u (0x%x)) => offset = 0x%x, %u %u bit chunks", size, size, 0, needed_chunks, m_chunk_size);
addr = m_addr;
}
else
{
uint32_t last_offset = 0;
OffsetToChunkSize::const_iterator pos = m_offset_to_chunk_size.begin();
OffsetToChunkSize::const_iterator end = m_offset_to_chunk_size.end();
while (pos != end)
{
if (pos->first > last_offset)
{
const uint32_t bytes_available = pos->first - last_offset;
const uint32_t num_chunks = CalculateChunksNeededForSize (bytes_available);
if (num_chunks >= needed_chunks)
{
m_offset_to_chunk_size[last_offset] = needed_chunks;
if (log)
log->Printf ("[2] AllocatedBlock::ReserveBlock (size = %u (0x%x)) => offset = 0x%x, %u %u bit chunks", size, size, last_offset, needed_chunks, m_chunk_size);
addr = m_addr + last_offset;
break;
}
}
last_offset = pos->first + pos->second * m_chunk_size;
if (++pos == end)
{
// Last entry...
const uint32_t chunks_left = CalculateChunksNeededForSize (m_byte_size - last_offset);
if (chunks_left >= needed_chunks)
{
m_offset_to_chunk_size[last_offset] = needed_chunks;
if (log)
log->Printf ("[3] AllocatedBlock::ReserveBlock (size = %u (0x%x)) => offset = 0x%x, %u %u bit chunks", size, size, last_offset, needed_chunks, m_chunk_size);
addr = m_addr + last_offset;
break;
}
}
}
}
// const uint32_t total_chunks = m_allocated.size ();
// uint32_t unallocated_idx = 0;
// uint32_t allocated_idx = m_allocated.find_first();
// uint32_t first_chunk_idx = UINT32_MAX;
// uint32_t num_chunks;
// while (1)
// {
// if (allocated_idx == UINT32_MAX)
// {
// // No more bits are set starting from unallocated_idx, so we
// // either have enough chunks for the request, or we don't.
// // Eiter way we break out of the while loop...
// num_chunks = total_chunks - unallocated_idx;
// if (needed_chunks <= num_chunks)
// first_chunk_idx = unallocated_idx;
// break;
// }
// else if (allocated_idx > unallocated_idx)
// {
// // We have some allocated chunks, check if there are enough
// // free chunks to satisfy the request?
// num_chunks = allocated_idx - unallocated_idx;
// if (needed_chunks <= num_chunks)
// {
// // Yep, we have enough!
// first_chunk_idx = unallocated_idx;
// break;
// }
// }
//
// while (unallocated_idx < total_chunks)
// {
// if (m_allocated[unallocated_idx])
// ++unallocated_idx;
// else
// break;
// }
//
// if (unallocated_idx >= total_chunks)
// break;
//
// allocated_idx = m_allocated.find_next(unallocated_idx);
// }
//
// if (first_chunk_idx != UINT32_MAX)
// {
// const uint32_t end_bit_idx = unallocated_idx + needed_chunks;
// for (uint32_t idx = first_chunk_idx; idx < end_bit_idx; ++idx)
// m_allocated.set(idx);
// return m_addr + m_chunk_size * first_chunk_idx;
// }
}
<rdar://problem/13521159> LLDB is crashing when logging is enabled from lldb-perf-clang. This has to do with the global destructor chain as the process and its threads are being torn down. All logging channels now make one and only one instance that is kept in a global pointer which is never freed. This guarantees that logging can correctly continue as the process tears itself down. llvm-svn: 178191
2013-03-28 07:08:40 +08:00
Log *log (GetLogIfAllCategoriesSet (LIBLLDB_LOG_PROCESS | LIBLLDB_LOG_VERBOSE));
Added an allocated memory cache to avoid having to allocate memory over and over when running JITed expressions. The allocated memory cache will cache allocate memory a page at a time for each permission combination and divvy up the memory and hand it out in 16 byte increments. llvm-svn: 131453
2011-05-17 11:37:42 +08:00
if (log)
Resolve printf formatting warnings on Linux: - use macros from inttypes.h for format strings instead of OS-specific types Patch from Matt Kopec! llvm-svn: 168945
2012-11-30 05:49:15 +08:00
log->Printf ("AllocatedBlock::ReserveBlock (size = %u (0x%x)) => 0x%16.16" PRIx64, size, size, (uint64_t)addr);
Added an allocated memory cache to avoid having to allocate memory over and over when running JITed expressions. The allocated memory cache will cache allocate memory a page at a time for each permission combination and divvy up the memory and hand it out in 16 byte increments. llvm-svn: 131453
2011-05-17 11:37:42 +08:00
return addr;
}
bool
AllocatedBlock::FreeBlock (addr_t addr)
{
uint32_t offset = addr - m_addr;
OffsetToChunkSize::iterator pos = m_offset_to_chunk_size.find (offset);
bool success = false;
if (pos != m_offset_to_chunk_size.end())
{
m_offset_to_chunk_size.erase (pos);
success = true;
}
<rdar://problem/13521159> LLDB is crashing when logging is enabled from lldb-perf-clang. This has to do with the global destructor chain as the process and its threads are being torn down. All logging channels now make one and only one instance that is kept in a global pointer which is never freed. This guarantees that logging can correctly continue as the process tears itself down. llvm-svn: 178191
2013-03-28 07:08:40 +08:00
Log *log (GetLogIfAllCategoriesSet (LIBLLDB_LOG_PROCESS | LIBLLDB_LOG_VERBOSE));
Added an allocated memory cache to avoid having to allocate memory over and over when running JITed expressions. The allocated memory cache will cache allocate memory a page at a time for each permission combination and divvy up the memory and hand it out in 16 byte increments. llvm-svn: 131453
2011-05-17 11:37:42 +08:00
if (log)
Resolve printf formatting warnings on Linux: - use macros from inttypes.h for format strings instead of OS-specific types Patch from Matt Kopec! llvm-svn: 168945
2012-11-30 05:49:15 +08:00
log->Printf ("AllocatedBlock::FreeBlock (addr = 0x%16.16" PRIx64 ") => %i", (uint64_t)addr, success);
Added an allocated memory cache to avoid having to allocate memory over and over when running JITed expressions. The allocated memory cache will cache allocate memory a page at a time for each permission combination and divvy up the memory and hand it out in 16 byte increments. llvm-svn: 131453
2011-05-17 11:37:42 +08:00
return success;
}
AllocatedMemoryCache::AllocatedMemoryCache (Process &process) :
m_process (process),
m_mutex (Mutex::eMutexTypeRecursive),
m_memory_map()
{
}
AllocatedMemoryCache::~AllocatedMemoryCache ()
{
}
void
AllocatedMemoryCache::Clear()
{
Mutex::Locker locker (m_mutex);
if (m_process.IsAlive())
{
PermissionsToBlockMap::iterator pos, end = m_memory_map.end();
for (pos = m_memory_map.begin(); pos != end; ++pos)
m_process.DoDeallocateMemory(pos->second->GetBaseAddress());
}
m_memory_map.clear();
}
AllocatedMemoryCache::AllocatedBlockSP
AllocatedMemoryCache::AllocatePage (uint32_t byte_size,
uint32_t permissions,
uint32_t chunk_size,
Error &error)
{
AllocatedBlockSP block_sp;
const size_t page_size = 4096;
const size_t num_pages = (byte_size + page_size - 1) / page_size;
const size_t page_byte_size = num_pages * page_size;
addr_t addr = m_process.DoAllocateMemory(page_byte_size, permissions, error);
<rdar://problem/13521159> LLDB is crashing when logging is enabled from lldb-perf-clang. This has to do with the global destructor chain as the process and its threads are being torn down. All logging channels now make one and only one instance that is kept in a global pointer which is never freed. This guarantees that logging can correctly continue as the process tears itself down. llvm-svn: 178191
2013-03-28 07:08:40 +08:00
Log *log (GetLogIfAllCategoriesSet (LIBLLDB_LOG_PROCESS));
Added an allocated memory cache to avoid having to allocate memory over and over when running JITed expressions. The allocated memory cache will cache allocate memory a page at a time for each permission combination and divvy up the memory and hand it out in 16 byte increments. llvm-svn: 131453
2011-05-17 11:37:42 +08:00
if (log)
{
Resolve printf formatting warnings on Linux: - use macros from inttypes.h for format strings instead of OS-specific types Patch from Matt Kopec! llvm-svn: 168945
2012-11-30 05:49:15 +08:00
log->Printf ("Process::DoAllocateMemory (byte_size = 0x%8.8zx, permissions = %s) => 0x%16.16" PRIx64,
Added an allocated memory cache to avoid having to allocate memory over and over when running JITed expressions. The allocated memory cache will cache allocate memory a page at a time for each permission combination and divvy up the memory and hand it out in 16 byte increments. llvm-svn: 131453
2011-05-17 11:37:42 +08:00
page_byte_size,
GetPermissionsAsCString(permissions),
(uint64_t)addr);
}
if (addr != LLDB_INVALID_ADDRESS)
{
block_sp.reset (new AllocatedBlock (addr, page_byte_size, permissions, chunk_size));
m_memory_map.insert (std::make_pair (permissions, block_sp));
}
return block_sp;
}
lldb::addr_t
AllocatedMemoryCache::AllocateMemory (size_t byte_size,
uint32_t permissions,
Error &error)
{
Mutex::Locker locker (m_mutex);
addr_t addr = LLDB_INVALID_ADDRESS;
std::pair<PermissionsToBlockMap::iterator, PermissionsToBlockMap::iterator> range = m_memory_map.equal_range (permissions);
for (PermissionsToBlockMap::iterator pos = range.first; pos != range.second; ++pos)
{
addr = (*pos).second->ReserveBlock (byte_size);
}
if (addr == LLDB_INVALID_ADDRESS)
{
AllocatedBlockSP block_sp (AllocatePage (byte_size, permissions, 16, error));
if (block_sp)
addr = block_sp->ReserveBlock (byte_size);
}
<rdar://problem/13521159> LLDB is crashing when logging is enabled from lldb-perf-clang. This has to do with the global destructor chain as the process and its threads are being torn down. All logging channels now make one and only one instance that is kept in a global pointer which is never freed. This guarantees that logging can correctly continue as the process tears itself down. llvm-svn: 178191
2013-03-28 07:08:40 +08:00
Log *log (GetLogIfAllCategoriesSet (LIBLLDB_LOG_PROCESS));
Added an allocated memory cache to avoid having to allocate memory over and over when running JITed expressions. The allocated memory cache will cache allocate memory a page at a time for each permission combination and divvy up the memory and hand it out in 16 byte increments. llvm-svn: 131453
2011-05-17 11:37:42 +08:00
if (log)
Resolve printf formatting warnings on Linux: - use macros from inttypes.h for format strings instead of OS-specific types Patch from Matt Kopec! llvm-svn: 168945
2012-11-30 05:49:15 +08:00
log->Printf ("AllocatedMemoryCache::AllocateMemory (byte_size = 0x%8.8zx, permissions = %s) => 0x%16.16" PRIx64, byte_size, GetPermissionsAsCString(permissions), (uint64_t)addr);
Added an allocated memory cache to avoid having to allocate memory over and over when running JITed expressions. The allocated memory cache will cache allocate memory a page at a time for each permission combination and divvy up the memory and hand it out in 16 byte increments. llvm-svn: 131453
2011-05-17 11:37:42 +08:00
return addr;
}
bool
AllocatedMemoryCache::DeallocateMemory (lldb::addr_t addr)
{
Mutex::Locker locker (m_mutex);
PermissionsToBlockMap::iterator pos, end = m_memory_map.end();
bool success = false;
for (pos = m_memory_map.begin(); pos != end; ++pos)
{
if (pos->second->Contains (addr))
{
success = pos->second->FreeBlock (addr);
break;
}
}
<rdar://problem/13521159> LLDB is crashing when logging is enabled from lldb-perf-clang. This has to do with the global destructor chain as the process and its threads are being torn down. All logging channels now make one and only one instance that is kept in a global pointer which is never freed. This guarantees that logging can correctly continue as the process tears itself down. llvm-svn: 178191
2013-03-28 07:08:40 +08:00
Log *log (GetLogIfAllCategoriesSet (LIBLLDB_LOG_PROCESS));
Added an allocated memory cache to avoid having to allocate memory over and over when running JITed expressions. The allocated memory cache will cache allocate memory a page at a time for each permission combination and divvy up the memory and hand it out in 16 byte increments. llvm-svn: 131453
2011-05-17 11:37:42 +08:00
if (log)
Resolve printf formatting warnings on Linux: - use macros from inttypes.h for format strings instead of OS-specific types Patch from Matt Kopec! llvm-svn: 168945
2012-11-30 05:49:15 +08:00
log->Printf("AllocatedMemoryCache::DeallocateMemory (addr = 0x%16.16" PRIx64 ") => %i", (uint64_t)addr, success);
Added an allocated memory cache to avoid having to allocate memory over and over when running JITed expressions. The allocated memory cache will cache allocate memory a page at a time for each permission combination and divvy up the memory and hand it out in 16 byte increments. llvm-svn: 131453
2011-05-17 11:37:42 +08:00
return success;
}