Copyright (c) 2002 Sun Microsystems, Inc. All Rights Reserved.
The contents of this file are subject to the terms of the Common Development and Distribution License (the "License"). You may not use this file except in compliance with the License.
You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE or http://www.opensolaris.org/os/licensing. See the License for the specific language governing permissions and limitations under the License.
When distributing Covered Code, include this CDDL HEADER in each file and include the License file at usr/src/OPENSOLARIS.LICENSE. If applicable, add the following below this CDDL HEADER, with the fields enclosed by brackets "[]" replaced with your own identifying information: Portions Copyright [yyyy] [name of copyright owner]
#include <sys/ddi.h> #include <sys/sunddi.h> int ddi_dma_buf_bind_handle(ddi_dma_handle_t handle, struct buf *bp, uint_t flags, int (*callback)(caddr_t), caddr_t arg, ddi_dma_cookie_t *cookiep, uint_t *ccountp);
Solaris DDI specific (Solaris DDI).
The DMA handle previously allocated by a call to ddi_dma_alloc_handle(9F).
A pointer to a system buffer structure (see buf(9S)).
Valid flags include: DDI_DMA_WRITE
Transfer direction is from memory to I/O
Transfer direction is from I/O to memory
Both read and write
Establish an MMU redzone at end of the object.
Partial resource allocation
Nonsequential, random, and small block transfers.
Sequential, unidirectional, block-sized, and block-aligned transfers.
The address of a function to call back later if resources are not available now. The following special function addresses may also be used. DDI_DMA_SLEEP
Wait until resources are available.
Do not wait until resources are available and do not schedule a callback.
Argument to be passed to the callback function, callback, if such a function is specified.
A pointer to the first ddi_dma_cookie(9S) structure. This should be left as NULL in new callers.
Upon a successful return, ccountp points to a value representing the number of cookies for this DMA object. This can be left as NULL in new callers. The cookie count can be obtained by calling ddi_dma_ncookies(9F).
ddi_dma_buf_bind_handle() allocates DMA resources for a system buffer such that a device can perform DMA to or from the buffer. DMA resources are allocated considering the device's DMA attributes as expressed by ddi_dma_attr(9S) (see ddi_dma_alloc_handle(9F)).
ddi_dma_buf_bind_handle() allocates and associates a number of DMA cookies with handle. To get the total number of cookies, callers should use the ddi_dma_ncookies(9F) function. To get all of the cookies, callers should use the ddi_dma_cookie_iter(9F) or ddi_dma_cookie_get(9F) functions. Callers should pass NULL for cookiep and ccountp. These values are required if using the deprecated ddi_dma_nextcookie(9F) interface, in which case cookiep is filled in with the first ddi_dma_cookie(9S) structure.
When a DMA transfer completes, the driver should free up system DMA resources by calling ddi_dma_unbind_handle(9F).
The flags argument contains information for mapping routines. DDI_DMA_WRITE, DDI_DMA_READ, DDI_DMA_RDWR
These flags describe the intended direction of the DMA transfer.
This flag should be set if the device is doing sequential, unidirectional, block-sized, and block-aligned transfers to or from memory. The alignment and padding constraints specified by the minxfer and burstsizes fields in the DMA attribute structure, ddi_dma_attr(9S) (see ddi_dma_alloc_handle(9F)) is used to allocate the most effective hardware support for large transfers.
This flag should be set if the device accesses memory randomly, or if synchronization steps using ddi_dma_sync(9F) need to be as efficient as possible. I/O parameter blocks used for communication between a device and a driver should be allocated using DDI_DMA_CONSISTENT.
If this flag is set, the system attempts 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 may not have the hardware ability to support a red zone.
Setting this flag indicates the caller can accept resources for part of the object. That is, if the size of the object exceeds the resources available, only resources for a portion of the object are allocated. The system indicates this condition returning status DDI_DMA_PARTIAL_MAP. At a later point, the caller can use ddi_dma_getwin(9F) to change the valid portion of the object for which resources are allocated. If resources were allocated for only part of the object, ddi_dma_addr_bind_handle() returns resources for the first DMA window. Even when DDI_DMA_PARTIAL is set, the system may decide to allocate resources for the entire object (less overhead) in which case DDI_DMA_MAPPED is returned.
The callback function, callback, indicates how a caller wants to handle the possibility of resources not being available. If callback is set to DDI_DMA_DONTWAIT, the caller does not care if the allocation fails, and can handle an allocation failure appropriately. If callback is set to DDI_DMA_SLEEP, the caller wishes 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 the address of a function to call at a later time when resources may become available. When the specified function is called, it is passed arg as an argument. The specified callback function must return either DDI_DMA_CALLBACK_RUNOUT or DDI_DMA_CALLBACK_DONE. DDI_DMA_CALLBACK_RUNOUT indicates that the callback function attempted to allocate DMA resources but failed to do so. In this case the callback function is put back on a list to be called again later. DDI_DMA_CALLBACK_DONE indicates either a successful allocation of DMA resources or that the driver no longer wishes to retry.
The callback function is called in interrupt context. Therefore, only system functions accessible from interrupt context are be available. The callback function must take whatever steps necessary to protect its critical resources, data structures, queues, etc.
ddi_dma_buf_bind_handle() returns: DDI_DMA_MAPPED
Successfully allocated resources for the entire object.
Successfully allocated resources for a part of the object. This is acceptable when partial transfers are permitted by setting the DDI_DMA_PARTIAL flag in flags.
Another I/O transaction is using the DMA handle.
No resources are available at the present time.
The object cannot be reached by the device requesting the resources.
The object is too big. A request of this size can never be satisfied on this particular system. The maximum size varies depending on machine and configuration.
ddi_dma_buf_bind_handle() can be called from user, kernel, or interrupt context, except when callback is set to DDI_DMA_SLEEP, in which case it can be called from user or kernel context only.
ddi_dma_addr_bind_handle(9F), ddi_dma_alloc_handle(9F), ddi_dma_cookie_get(9F), ddi_dma_cookie_iter(9F), ddi_dma_free_handle(9F), ddi_dma_getwin(9F), ddi_dma_ncookies(9F), ddi_dma_sync(9F), ddi_dma_unbind_handle(9F), buf(9S), ddi_dma_attr(9S), ddi_dma_cookie(9S)
Writing Device Drivers
If the driver permits partial mapping with the DDI_DMA_PARTIAL flag, the number of cookies in each window may exceed the size of the device's scatter/gather list as specified in the dma_attr_sgllen field in the ddi_dma_attr(9S) structure. In this case, each set of cookies comprising a DMA window will satisfy the DMA attributes as described in the ddi_dma_attr(9S) structure in all aspects. The driver should set up its DMA engine and perform one transfer for each set of cookies sufficient for its scatter/gather list, up to the number of cookies for this window, before advancing to the next window using ddi_dma_getwin(9F).