diff options
Diffstat (limited to 'include/qemu')
| -rw-r--r-- | include/qemu/coroutine-tls.h | 165 | ||||
| -rw-r--r-- | include/qemu/cpuid.h | 20 | ||||
| -rw-r--r-- | include/qemu/hbitmap.h | 12 | ||||
| -rw-r--r-- | include/qemu/job.h | 22 | ||||
| -rw-r--r-- | include/qemu/main-loop.h | 42 | ||||
| -rw-r--r-- | include/qemu/memalign.h | 61 | ||||
| -rw-r--r-- | include/qemu/osdep.h | 31 | ||||
| -rw-r--r-- | include/qemu/rcu.h | 7 | ||||
| -rw-r--r-- | include/qemu/typedefs.h | 2 | ||||
| -rw-r--r-- | include/qemu/xattr.h | 4 |
10 files changed, 328 insertions, 38 deletions
diff --git a/include/qemu/coroutine-tls.h b/include/qemu/coroutine-tls.h new file mode 100644 index 0000000000..1558a826aa --- /dev/null +++ b/include/qemu/coroutine-tls.h @@ -0,0 +1,165 @@ +/* + * QEMU Thread Local Storage for coroutines + * + * Copyright Red Hat + * + * SPDX-License-Identifier: LGPL-2.1-or-later + * + * This work is licensed under the terms of the GNU LGPL, version 2.1 or later. + * See the COPYING.LIB file in the top-level directory. + * + * It is forbidden to access Thread Local Storage in coroutines because + * compiler optimizations may cause values to be cached across coroutine + * re-entry. Coroutines can run in more than one thread through the course of + * their life, leading bugs when stale TLS values from the wrong thread are + * used as a result of compiler optimization. + * + * An example is: + * + * ..code-block:: c + * :caption: A coroutine that may see the wrong TLS value + * + * static __thread AioContext *current_aio_context; + * ... + * static void coroutine_fn foo(void) + * { + * aio_notify(current_aio_context); + * qemu_coroutine_yield(); + * aio_notify(current_aio_context); // <-- may be stale after yielding! + * } + * + * This header provides macros for safely defining variables in Thread Local + * Storage: + * + * ..code-block:: c + * :caption: A coroutine that safely uses TLS + * + * QEMU_DEFINE_STATIC_CO_TLS(AioContext *, current_aio_context) + * ... + * static void coroutine_fn foo(void) + * { + * aio_notify(get_current_aio_context()); + * qemu_coroutine_yield(); + * aio_notify(get_current_aio_context()); // <-- safe + * } + */ + +#ifndef QEMU_COROUTINE_TLS_H +#define QEMU_COROUTINE_TLS_H + +/* + * To stop the compiler from caching TLS values we define accessor functions + * with __attribute__((noinline)) plus asm volatile("") to prevent + * optimizations that override noinline. + * + * The compiler can still analyze noinline code and make optimizations based on + * that knowledge, so an inline asm output operand is used to prevent + * optimizations that make assumptions about the address of the TLS variable. + * + * This is fragile and ultimately needs to be solved by a mechanism that is + * guaranteed to work by the compiler (e.g. stackless coroutines), but for now + * we use this approach to prevent issues. + */ + +/** + * QEMU_DECLARE_CO_TLS: + * @type: the variable's C type + * @var: the variable name + * + * Declare an extern variable in Thread Local Storage from a header file: + * + * .. code-block:: c + * :caption: Declaring an extern variable in Thread Local Storage + * + * QEMU_DECLARE_CO_TLS(int, my_count) + * ... + * int c = get_my_count(); + * set_my_count(c + 1); + * *get_ptr_my_count() = 0; + * + * This is a coroutine-safe replacement for the __thread keyword and is + * equivalent to the following code: + * + * .. code-block:: c + * :caption: Declaring a TLS variable using __thread + * + * extern __thread int my_count; + * ... + * int c = my_count; + * my_count = c + 1; + * *(&my_count) = 0; + */ +#define QEMU_DECLARE_CO_TLS(type, var) \ + __attribute__((noinline)) type get_##var(void); \ + __attribute__((noinline)) void set_##var(type v); \ + __attribute__((noinline)) type *get_ptr_##var(void); + +/** + * QEMU_DEFINE_CO_TLS: + * @type: the variable's C type + * @var: the variable name + * + * Define a variable in Thread Local Storage that was previously declared from + * a header file with QEMU_DECLARE_CO_TLS(): + * + * .. code-block:: c + * :caption: Defining a variable in Thread Local Storage + * + * QEMU_DEFINE_CO_TLS(int, my_count) + * + * This is a coroutine-safe replacement for the __thread keyword and is + * equivalent to the following code: + * + * .. code-block:: c + * :caption: Defining a TLS variable using __thread + * + * __thread int my_count; + */ +#define QEMU_DEFINE_CO_TLS(type, var) \ + static __thread type co_tls_##var; \ + type get_##var(void) { asm volatile(""); return co_tls_##var; } \ + void set_##var(type v) { asm volatile(""); co_tls_##var = v; } \ + type *get_ptr_##var(void) \ + { type *ptr = &co_tls_##var; asm volatile("" : "+rm" (ptr)); return ptr; } + +/** + * QEMU_DEFINE_STATIC_CO_TLS: + * @type: the variable's C type + * @var: the variable name + * + * Define a static variable in Thread Local Storage: + * + * .. code-block:: c + * :caption: Defining a static variable in Thread Local Storage + * + * QEMU_DEFINE_STATIC_CO_TLS(int, my_count) + * ... + * int c = get_my_count(); + * set_my_count(c + 1); + * *get_ptr_my_count() = 0; + * + * This is a coroutine-safe replacement for the __thread keyword and is + * equivalent to the following code: + * + * .. code-block:: c + * :caption: Defining a static TLS variable using __thread + * + * static __thread int my_count; + * ... + * int c = my_count; + * my_count = c + 1; + * *(&my_count) = 0; + */ +#define QEMU_DEFINE_STATIC_CO_TLS(type, var) \ + static __thread type co_tls_##var; \ + static __attribute__((noinline, unused)) \ + type get_##var(void) \ + { asm volatile(""); return co_tls_##var; } \ + static __attribute__((noinline, unused)) \ + void set_##var(type v) \ + { asm volatile(""); co_tls_##var = v; } \ + static __attribute__((noinline, unused)) \ + type *get_ptr_##var(void) \ + { type *ptr = &co_tls_##var; asm volatile("" : "+rm" (ptr)); return ptr; } + +#endif /* QEMU_COROUTINE_TLS_H */ diff --git a/include/qemu/cpuid.h b/include/qemu/cpuid.h index 09fc245b91..7adb12d320 100644 --- a/include/qemu/cpuid.h +++ b/include/qemu/cpuid.h @@ -45,12 +45,26 @@ #ifndef bit_AVX2 #define bit_AVX2 (1 << 5) #endif -#ifndef bit_AVX512F -#define bit_AVX512F (1 << 16) -#endif #ifndef bit_BMI2 #define bit_BMI2 (1 << 8) #endif +#ifndef bit_AVX512F +#define bit_AVX512F (1 << 16) +#endif +#ifndef bit_AVX512DQ +#define bit_AVX512DQ (1 << 17) +#endif +#ifndef bit_AVX512BW +#define bit_AVX512BW (1 << 30) +#endif +#ifndef bit_AVX512VL +#define bit_AVX512VL (1u << 31) +#endif + +/* Leaf 7, %ecx */ +#ifndef bit_AVX512VBMI2 +#define bit_AVX512VBMI2 (1 << 6) +#endif /* Leaf 0x80000001, %ecx */ #ifndef bit_LZCNT diff --git a/include/qemu/hbitmap.h b/include/qemu/hbitmap.h index 5e71b6d6f7..5bd986aa44 100644 --- a/include/qemu/hbitmap.h +++ b/include/qemu/hbitmap.h @@ -340,6 +340,18 @@ bool hbitmap_next_dirty_area(const HBitmap *hb, int64_t start, int64_t end, int64_t max_dirty_count, int64_t *dirty_start, int64_t *dirty_count); +/* + * bdrv_dirty_bitmap_status: + * @hb: The HBitmap to operate on + * @start: The bit to start from + * @count: Number of bits to proceed + * @pnum: Out-parameter. How many bits has same value starting from @start + * + * Returns true if bitmap is dirty at @start, false otherwise. + */ +bool hbitmap_status(const HBitmap *hb, int64_t start, int64_t count, + int64_t *pnum); + /** * hbitmap_iter_next: * @hbi: HBitmapIter to operate on. diff --git a/include/qemu/job.h b/include/qemu/job.h index 6e67b6977f..c105b31076 100644 --- a/include/qemu/job.h +++ b/include/qemu/job.h @@ -169,6 +169,12 @@ typedef struct Job { * Callbacks and other information about a Job driver. */ struct JobDriver { + + /* + * These fields are initialized when this object is created, + * and are never changed afterwards + */ + /** Derived Job struct size */ size_t instance_size; @@ -184,9 +190,18 @@ struct JobDriver { * aborted. If it returns zero, the job moves into the WAITING state. If it * is the last job to complete in its transaction, all jobs in the * transaction move from WAITING to PENDING. + * + * This callback must be run in the job's context. */ int coroutine_fn (*run)(Job *job, Error **errp); + /* + * Functions run without regard to the BQL that may run in any + * arbitrary thread. These functions do not need to be thread-safe + * because the caller ensures that they are invoked from one + * thread at time. + */ + /** * If the callback is not NULL, it will be invoked when the job transitions * into the paused state. Paused jobs must not perform any asynchronous @@ -201,6 +216,13 @@ struct JobDriver { */ void coroutine_fn (*resume)(Job *job); + /* + * Global state (GS) API. These functions run under the BQL. + * + * See include/block/block-global-state.h for more information about + * the GS API. + */ + /** * Called when the job is resumed by the user (i.e. user_paused becomes * false). .user_resume is called before .resume. diff --git a/include/qemu/main-loop.h b/include/qemu/main-loop.h index 8dbc6fcb89..7a4d6a0920 100644 --- a/include/qemu/main-loop.h +++ b/include/qemu/main-loop.h @@ -242,10 +242,52 @@ AioContext *iohandler_get_aio_context(void); * must always be taken outside other locks. This function helps * functions take different paths depending on whether the current * thread is running within the main loop mutex. + * + * This function should never be used in the block layer, because + * unit tests, block layer tools and qemu-storage-daemon do not + * have a BQL. + * Please instead refer to qemu_in_main_thread(). */ bool qemu_mutex_iothread_locked(void); /** + * qemu_in_main_thread: return whether it's possible to safely access + * the global state of the block layer. + * + * Global state of the block layer is not accessible from I/O threads + * or worker threads; only from threads that "own" the default + * AioContext that qemu_get_aio_context() returns. For tests, block + * layer tools and qemu-storage-daemon there is a designated thread that + * runs the event loop for qemu_get_aio_context(), and that is the + * main thread. + * + * For emulators, however, any thread that holds the BQL can act + * as the block layer main thread; this will be any of the actual + * main thread, the vCPU threads or the RCU thread. + * + * For clarity, do not use this function outside the block layer. + */ +bool qemu_in_main_thread(void); + +/* Mark and check that the function is part of the global state API. */ +#define GLOBAL_STATE_CODE() \ + do { \ + assert(qemu_in_main_thread()); \ + } while (0) + +/* Mark and check that the function is part of the I/O API. */ +#define IO_CODE() \ + do { \ + /* nop */ \ + } while (0) + +/* Mark and check that the function is part of the "I/O OR GS" API. */ +#define IO_OR_GS_CODE() \ + do { \ + /* nop */ \ + } while (0) + +/** * qemu_mutex_lock_iothread: Lock the main loop mutex. * * This function locks the main loop mutex. The mutex is taken by diff --git a/include/qemu/memalign.h b/include/qemu/memalign.h new file mode 100644 index 0000000000..fa299f3bf6 --- /dev/null +++ b/include/qemu/memalign.h @@ -0,0 +1,61 @@ +/* + * Allocation and free functions for aligned memory + * + * This work is licensed under the terms of the GNU GPL, version 2 or later. + * See the COPYING file in the top-level directory. + */ + +#ifndef QEMU_MEMALIGN_H +#define QEMU_MEMALIGN_H + +/** + * qemu_try_memalign: Allocate aligned memory + * @alignment: required alignment, in bytes + * @size: size of allocation, in bytes + * + * Allocate memory on an aligned boundary (i.e. the returned + * address will be an exact multiple of @alignment). + * @alignment must be a power of 2, or the function will assert(). + * On success, returns allocated memory; on failure, returns NULL. + * + * The memory allocated through this function must be freed via + * qemu_vfree() (and not via free()). + */ +void *qemu_try_memalign(size_t alignment, size_t size); +/** + * qemu_memalign: Allocate aligned memory, without failing + * @alignment: required alignment, in bytes + * @size: size of allocation, in bytes + * + * Allocate memory in the same way as qemu_try_memalign(), but + * abort() with an error message if the memory allocation fails. + * + * The memory allocated through this function must be freed via + * qemu_vfree() (and not via free()). + */ +void *qemu_memalign(size_t alignment, size_t size); +/** + * qemu_vfree: Free memory allocated through qemu_memalign + * @ptr: memory to free + * + * This function must be used to free memory allocated via qemu_memalign() + * or qemu_try_memalign(). (Using the wrong free function will cause + * subtle bugs on Windows hosts.) + */ +void qemu_vfree(void *ptr); +/* + * It's an analog of GLIB's g_autoptr_cleanup_generic_gfree(), used to define + * g_autofree macro. + */ +static inline void qemu_cleanup_generic_vfree(void *p) +{ + void **pp = (void **)p; + qemu_vfree(*pp); +} + +/* + * Analog of g_autofree, but qemu_vfree is called on cleanup instead of g_free. + */ +#define QEMU_AUTO_VFREE __attribute__((cleanup(qemu_cleanup_generic_vfree))) + +#endif diff --git a/include/qemu/osdep.h b/include/qemu/osdep.h index 7bcce3bceb..c9ec7830c9 100644 --- a/include/qemu/osdep.h +++ b/include/qemu/osdep.h @@ -379,28 +379,10 @@ extern "C" { #endif int qemu_daemon(int nochdir, int noclose); -void *qemu_try_memalign(size_t alignment, size_t size); -void *qemu_memalign(size_t alignment, size_t size); void *qemu_anon_ram_alloc(size_t size, uint64_t *align, bool shared, bool noreserve); -void qemu_vfree(void *ptr); void qemu_anon_ram_free(void *ptr, size_t size); -/* - * It's an analog of GLIB's g_autoptr_cleanup_generic_gfree(), used to define - * g_autofree macro. - */ -static inline void qemu_cleanup_generic_vfree(void *p) -{ - void **pp = (void **)p; - qemu_vfree(*pp); -} - -/* - * Analog of g_autofree, but qemu_vfree is called on cleanup instead of g_free. - */ -#define QEMU_AUTO_VFREE __attribute__((cleanup(qemu_cleanup_generic_vfree))) - #ifdef _WIN32 #define HAVE_CHARDEV_SERIAL 1 #elif defined(__linux__) || defined(__sun__) || defined(__FreeBSD__) \ @@ -673,19 +655,6 @@ static inline int platform_does_not_support_system(const char *command) } #endif /* !HAVE_SYSTEM_FUNCTION */ -/** - * Duplicate directory entry @dent. - * - * It is highly recommended to use this function instead of open coding - * duplication of @c dirent objects, because the actual @c struct @c dirent - * size may be bigger or shorter than @c sizeof(struct dirent) and correct - * handling is platform specific (see gitlab issue #841). - * - * @dent - original directory entry to be duplicated - * @returns duplicated directory entry which should be freed with g_free() - */ -struct dirent *qemu_dirent_dup(struct dirent *dent); - #ifdef __cplusplus } #endif diff --git a/include/qemu/rcu.h b/include/qemu/rcu.h index e69efbd47f..b063c6fde8 100644 --- a/include/qemu/rcu.h +++ b/include/qemu/rcu.h @@ -29,6 +29,7 @@ #include "qemu/atomic.h" #include "qemu/notify.h" #include "qemu/sys_membarrier.h" +#include "qemu/coroutine-tls.h" #ifdef __cplusplus extern "C" { @@ -76,11 +77,11 @@ struct rcu_reader_data { NotifierList force_rcu; }; -extern __thread struct rcu_reader_data rcu_reader; +QEMU_DECLARE_CO_TLS(struct rcu_reader_data, rcu_reader) static inline void rcu_read_lock(void) { - struct rcu_reader_data *p_rcu_reader = &rcu_reader; + struct rcu_reader_data *p_rcu_reader = get_ptr_rcu_reader(); unsigned ctr; if (p_rcu_reader->depth++ > 0) { @@ -96,7 +97,7 @@ static inline void rcu_read_lock(void) static inline void rcu_read_unlock(void) { - struct rcu_reader_data *p_rcu_reader = &rcu_reader; + struct rcu_reader_data *p_rcu_reader = get_ptr_rcu_reader(); assert(p_rcu_reader->depth != 0); if (--p_rcu_reader->depth > 0) { diff --git a/include/qemu/typedefs.h b/include/qemu/typedefs.h index 5b302cb214..42f4ceb701 100644 --- a/include/qemu/typedefs.h +++ b/include/qemu/typedefs.h @@ -26,6 +26,7 @@ typedef struct AddressSpace AddressSpace; typedef struct AioContext AioContext; typedef struct Aml Aml; typedef struct AnnounceTimer AnnounceTimer; +typedef struct ArchCPU ArchCPU; typedef struct BdrvDirtyBitmap BdrvDirtyBitmap; typedef struct BdrvDirtyBitmapIter BdrvDirtyBitmapIter; typedef struct BlockBackend BlockBackend; @@ -39,6 +40,7 @@ typedef struct CompatProperty CompatProperty; typedef struct CoMutex CoMutex; typedef struct ConfidentialGuestSupport ConfidentialGuestSupport; typedef struct CPUAddressSpace CPUAddressSpace; +typedef struct CPUArchState CPUArchState; typedef struct CPUState CPUState; typedef struct DeviceListener DeviceListener; typedef struct DeviceState DeviceState; diff --git a/include/qemu/xattr.h b/include/qemu/xattr.h index a83fe8e749..f1d0f7be74 100644 --- a/include/qemu/xattr.h +++ b/include/qemu/xattr.h @@ -22,7 +22,9 @@ #ifdef CONFIG_LIBATTR # include <attr/xattr.h> #else -# define ENOATTR ENODATA +# if !defined(ENOATTR) +# define ENOATTR ENODATA +# endif # include <sys/xattr.h> #endif |