ddi_dma_req(9S) (man pages section 9: DDI and DKI Properties and Data Structures)

man pages section 9: DDI and DKI Properties and Data Structures

ddi_dma_req(9S)

NAME

ddi_dma_req– DMA Request structure

SYNOPSIS

#include <sys/ddidmareq.h>

INTERFACE LEVEL

Solaris DDI specific (Solaris DDI).

DESCRIPTION

A ddi_dma_req structure describes a request for DMA resources. A driver can use it to describe forms of allocations and ways to allocate DMA resources for a DMA request.

STRUCTURE MEMBERS

ddi_dma_lim_t   *dmar_limits;         /* Caller's dma engine's */
                                      /* constraints */
uint_t          dmar_flags;           /* Contains information for */
                                      /* mapping routines */
int             (*dmar_fp)(caddr_t);  /* Callback function */
caddr_t         dmar_arg;             /* Callback function's argument */
ddi_dma_obj_t   dmar_object;          /* Description of the object */
                                      /* to be mapped */

For the definition of the DMA limits structure, which dmar_limits points to, see ddi_dma_lim_sparc(9S) or ddi_dma_lim_x86(9S).

Valid values for dmar_flags are:

DDI_DMA_WRITE         /* Direction memory --> IO */
DDI_DMA_READ          /* Direction IO --> memory */
DDI_DMA_RDWR          /* Both read and write */
DDI_DMA_REDZONE       /* Establish an MMU redzone at end of mapping */
DDI_DMA_PARTIAL       /* Partial mapping is allowed */
DDI_DMA_CONSISTENT    /* Byte consistent access wanted */
DDI_DMA_SBUS_64BIT    /* Use 64 bit capability on SBus */

DDI_DMA_WRITE, DDI_DMA_READ, and DDI_DMA_RDWR describe the intended direction of the DMA transfer. Some implementations might explicitly disallow DDI_DMA_RDWR.

DDI_DMA_REDZONE asks the system to establish a protected red zone after the object. The DMA resource allocation functions do not guarantee the success of this request, as some implementations might not have the hardware ability to support it.

DDI_DMA_PARTIAL lets the system know that the caller can accept partial mapping. That is, if the size of the object exceeds the resources available, the system allocates only a portion of the object and returns status indicating this partial allocation. At a later point, the caller can use ddi_dma_curwin(9F) and ddi_dma_movwin(9F) to change the valid portion of the object that has resources allocated.

DDI_DMA_CONSISTENT gives a hint to the system that the object should be mapped for byte consistent access. Normal data transfers usually use a streaming mode of operation. They start at a specific point, transfer a fairly large amount of data sequentially, and then stop, usually on an aligned boundary. Control mode data transfers for memory-resident device control blocks (for example, Ethernet message descriptors) do not access memory in such a sequential fashion. Instead, they tend to modify a few words or bytes, move around and maybe modify a few more.

Many machine implementations make this non-sequential memory access difficult to control in a generic and seamless fashion. Therefore, explicit synchronization steps using ddi_dma_sync(9F) or ddi_dma_free(9F) are required to make the view of a memory object shared between a CPU and a DMA device consistent. However, proper use of the DDI_DMA_CONSISTENT flag can create a condition in which a system will pick resources in a way that makes these synchronization steps are as efficient as possible.

DDI_DMA_SBUS_64BIT tells the system that the device can perform 64–bit transfers on a 64–bit SBus. If the SBus does not support 64–bit data transfers, data will be transferred in 32–bit mode.

The callback function specified by the member dmar_fp indicates how a caller to one of the DMA resource allocation functions wants to deal with the possibility of resources not being available. (See ddi_dma_setup(9F).) If dmar_fp is set to DDI_DMA_DONTWAIT, then the caller does not care if the allocation fails, and can deal with an allocation failure appropriately. Setting dmar_fp to DDI_DMA_SLEEP indicates the caller wants to have the allocation routines wait for resources to become available. If any other value is set, and a DMA resource allocation fails, this value is assumed to be a function to call later, when resources become available. When the specified function is called, it is passed the value set in the structure member dmar_arg. The specified callback function must return either:

0: Indicating that it attempted to allocate a DMA resource but failed to do so, again, in which case the callback function will be put back on a list to be called again later.
1: Indicating either success at allocating DMA resources or that it no longer wants to retry.

The callback function is called in interrupt context. Therefore, only system functions and contexts that are accessible from interrupt context are available. The callback function must take whatever steps necessary to protect its critical resources, data structures, and queues.

It is possible that a call to ddi_dma_free(9F), which frees DMA resources, might cause a callback function to be called and, unless some care is taken, an undesired recursion can occur. This can cause an undesired recursive mutex_enter(9F), which makes the system panic.

dmar_object Structure

The dmar_object member of the ddi_dma_req structure is itself a complex and extensible structure:

uint_t            dmao_size;     /* size, in bytes, of the object */
ddi_dma_atyp_t    dmao_type;     /* type of object */
ddi_dma_aobj_t    dmao_obj;      /* the object described */

The dmao_size element is the size, in bytes, of the object resources allocated for DMA.

The dmao_type element selects the kind of object described by dmao_obj. It can be set to DMA_OTYP_VADDR, indicating virtual addresses.

The last element, dmao_obj, consists of the virtual address type:

struct v_address virt_obj;

It is specified as:

struct v_address {
      caddr_t      v_addr;   /* base virtual address */
      struct as    *v_as;    /* pointer to address space */
      void         *v_priv;  /* priv data for shadow I/O */
};