Fix compat32 for ntp_adjtime(2).
ClosedPublic
Actions

Authored by kib on Dec 3 2020, 10:00 PM.

Details

Reviewers

• ian
phk
cy
brooks
emaste

Commits

rS368343: Fix compat32 for ntp_adjtime(2).
rS368342: Add kern_ntp_adjtime(9).

Summary

Add kern_ntp_adjtime(9).
struct timex is not 32-bit safe, it uses longs for members. Provide translation.

Diff Detail

Repository

rS FreeBSD src repository - subversion

Lint

Lint Skipped

Unit

Tests Skipped

Build Status

Buildable 35182

Event Timeline

kib created this revision.Dec 3 2020, 10:00 PM

Herald added a reviewer: brooks. · View Herald TranscriptDec 3 2020, 10:00 PM

Herald added a subscriber: imp. · View Herald Transcript

kib requested review of this revision.Dec 3 2020, 10:00 PM

Harbormaster completed remote builds in B35180: Diff 80286.Dec 3 2020, 10:00 PM

kib added a reviewer: emaste.Dec 3 2020, 10:01 PM

I don't see any issues worth noting.

This revision is now accepted and ready to land.Dec 3 2020, 11:12 PM

I'd be tempted to pass timex_copyin and timex_copyout function pointers to kern_ntp_adjtime rather than having duplicate code in sys_ and freebsd32_. That would also simplify a future compat syscall implementation.

sys/kern/kern_ntptime.c
1089–1109	This looks like it's part of a future syscall revision, not the compat change?

In D27471#613779, @brooks wrote:

I'd be tempted to pass timex_copyin and timex_copyout function pointers to kern_ntp_adjtime rather than having duplicate code in sys_ and freebsd32_. That would also simplify a future compat syscall implementation.

Right now it seems there will be no compat syscall, ie struct timex is not going to change.

Do you still want copyin callbacks there ? It is more overhead in code for rather trivial wrappers IMO.

Remove stray chunk.

This revision now requires review to proceed.Dec 3 2020, 11:42 PM

Harbormaster completed remote builds in B35182: Diff 80292.Dec 3 2020, 11:42 PM

Without a compat syscall it seems fine to leave this as is.

brooks accepted this revision.Dec 4 2020, 4:39 PM

This revision is now accepted and ready to land.Dec 4 2020, 4:39 PM

Closed by commit rS368342: Add kern_ntp_adjtime(9). (authored by kib). · Explain WhyDec 4 2020, 6:57 PM

This revision was automatically updated to reflect the committed changes.

kib added a commit: rS368342: Add kern_ntp_adjtime(9)..

kib added a commit: rS368343: Fix compat32 for ntp_adjtime(2)..

sak_codered.rmta.org added a subscriber: sak_codered.rmta.org.Dec 9 2020, 12:24 AM

Revision Contents
Changeset List

Path

Size

sys/

compat/

freebsd32/

freebsd32.h

20 lines

freebsd32_misc.c

66 lines

syscalls.master

3 lines

kern/

kern_ntptime.c

112 lines

sys/

syscallsubr.h

2 lines

Diff 80292

View Options

sys/compat/freebsd32/freebsd32.h

/*-		/*-
* Copyright (c) 2016 Ruslan Bukin <br@bsdpad.com>		* Copyright (c) 2016-2017 Ruslan Bukin <br@bsdpad.com>
* All rights reserved.		* All rights reserved.
*		*
* This software was developed by SRI International and the University of		* This software was developed by SRI International and the University of
* Cambridge Computer Laboratory under DARPA/AFRL contract FA8750-10-C-0237		* Cambridge Computer Laboratory under DARPA/AFRL contract FA8750-10-C-0237
* ("CTSRD"), as part of the DARPA CRASH research programme.		* ("CTSRD"), as part of the DARPA CRASH research programme.
*		*
* Redistribution and use in source and binary forms, with or without		* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions		* modification, are permitted provided that the following conditions
Show All 20 Lines
#include <sys/cdefs.h>		#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");		__FBSDID("$FreeBSD$");

#include "opt_platform.h"		#include "opt_platform.h"
#include <sys/param.h>		#include <sys/param.h>
#include <sys/conf.h>		#include <sys/conf.h>
#include <sys/bus.h>		#include <sys/bus.h>
#include <sys/kernel.h>		#include <sys/kernel.h>
		#include <sys/kthread.h>
#include <sys/queue.h>		#include <sys/queue.h>
#include <sys/kobj.h>		#include <sys/kobj.h>
#include <sys/malloc.h>		#include <sys/malloc.h>
#include <sys/mutex.h>		#include <sys/mutex.h>
		#include <sys/mbuf.h>
#include <sys/limits.h>		#include <sys/limits.h>
#include <sys/lock.h>		#include <sys/lock.h>
#include <sys/sysctl.h>		#include <sys/sysctl.h>
#include <sys/systm.h>		#include <sys/systm.h>
#include <sys/sx.h>		#include <sys/sx.h>
#include <sys/bus_dma.h>		#include <sys/bus_dma.h>

#include <machine/bus.h>		#include <machine/bus.h>

#ifdef FDT		#ifdef FDT
#include <dev/fdt/fdt_common.h>		#include <dev/fdt/fdt_common.h>
#include <dev/ofw/ofw_bus.h>		#include <dev/ofw/ofw_bus.h>
#include <dev/ofw/ofw_bus_subr.h>		#include <dev/ofw/ofw_bus_subr.h>
#endif		#endif

#include <dev/xdma/xdma.h>		#include <dev/xdma/xdma.h>

#include <xdma_if.h>		#include <xdma_if.h>

MALLOC_DEFINE(M_XDMA, "xdma", "xDMA framework");

/*		/*
* Multiple xDMA controllers may work with single DMA device,		* Multiple xDMA controllers may work with single DMA device,
* so we have global lock for physical channel management.		* so we have global lock for physical channel management.
*/		*/
static struct mtx xdma_mtx;		static struct mtx xdma_mtx;

#define XDMA_LOCK() mtx_lock(&xdma_mtx)		#define XDMA_LOCK() mtx_lock(&xdma_mtx)
#define XDMA_UNLOCK() mtx_unlock(&xdma_mtx)		#define XDMA_UNLOCK() mtx_unlock(&xdma_mtx)
#define XDMA_ASSERT_LOCKED() mtx_assert(&xdma_mtx, MA_OWNED)		#define XDMA_ASSERT_LOCKED() mtx_assert(&xdma_mtx, MA_OWNED)

/*		/*
* Per channel locks.		* Per channel locks.
*/		*/
#define XCHAN_LOCK(xchan) mtx_lock(&(xchan)->mtx_lock)		#define XCHAN_LOCK(xchan) mtx_lock(&(xchan)->mtx_lock)
#define XCHAN_UNLOCK(xchan) mtx_unlock(&(xchan)->mtx_lock)		#define XCHAN_UNLOCK(xchan) mtx_unlock(&(xchan)->mtx_lock)
#define XCHAN_ASSERT_LOCKED(xchan) mtx_assert(&(xchan)->mtx_lock, MA_OWNED)		#define XCHAN_ASSERT_LOCKED(xchan) mtx_assert(&(xchan)->mtx_lock, MA_OWNED)

		static int xchan_bufs_alloc(xdma_channel_t *xchan);
		static int xchan_bufs_free(xdma_channel_t *xchan);

		static void
		xdma_task(void *arg)
		{
		xdma_controller_t *xdma;
		xdma_channel_t *xchan_tmp;
		xdma_channel_t *xchan;

		xdma = arg;

		for (;;) {
		mtx_lock(&xdma->proc_mtx);
		msleep(xdma, &xdma->proc_mtx, PRIBIO, "jobqueue", hz);
		mtx_unlock(&xdma->proc_mtx);

		if (TAILQ_EMPTY(&xdma->channels)) {
		continue;
		}

		TAILQ_FOREACH_SAFE(xchan, &xdma->channels, xchan_next, xchan_tmp) {
		if (xchan->flags & XCHAN_TYPE_SG) {
		xdma_queue_submit(xchan);
		}
		}
		}
		}

/*		/*
* Allocate virtual xDMA channel.		* Allocate virtual xDMA channel.
*/		*/
xdma_channel_t *		xdma_channel_t *
xdma_channel_alloc(xdma_controller_t *xdma)		xdma_channel_alloc(xdma_controller_t *xdma, uint32_t caps)
{		{
xdma_channel_t *xchan;		xdma_channel_t *xchan;
int ret;		int ret;

xchan = malloc(sizeof(xdma_channel_t), M_XDMA, M_WAITOK \| M_ZERO);		xchan = malloc(sizeof(xdma_channel_t), M_XDMA, M_WAITOK \| M_ZERO);
if (xchan == NULL) {		if (xchan == NULL) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't allocate memory for channel.\n", __func__);		"%s: Can't allocate memory for channel.\n", __func__);
return (NULL);		return (NULL);
}		}
xchan->xdma = xdma;		xchan->xdma = xdma;
		xchan->caps = caps;

XDMA_LOCK();		XDMA_LOCK();

/* Request a real channel from hardware driver. */		/* Request a real channel from hardware driver. */
ret = XDMA_CHANNEL_ALLOC(xdma->dma_dev, xchan);		ret = XDMA_CHANNEL_ALLOC(xdma->dma_dev, xchan);
if (ret != 0) {		if (ret != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't request hardware channel.\n", __func__);		"%s: Can't request hardware channel.\n", __func__);
Show All 29 Lines	if (err != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't free real hw channel.\n", __func__);		"%s: Can't free real hw channel.\n", __func__);
XDMA_UNLOCK();		XDMA_UNLOCK();
return (-1);		return (-1);
}		}

xdma_teardown_all_intr(xchan);		xdma_teardown_all_intr(xchan);

/* Deallocate descriptors, if any. */		/* Deallocate bufs, if any. */
xdma_desc_free(xchan);		xchan_bufs_free(xchan);
		xchan_sglist_free(xchan);

		if (xchan->flags & XCHAN_TYPE_SG) {
		free(xchan->xr, M_XDMA);
		}

mtx_destroy(&xchan->mtx_lock);		mtx_destroy(&xchan->mtx_lock);

TAILQ_REMOVE(&xdma->channels, xchan, xchan_next);		TAILQ_REMOVE(&xdma->channels, xchan, xchan_next);

free(xchan, M_XDMA);		free(xchan, M_XDMA);

XDMA_UNLOCK();		XDMA_UNLOCK();

return (0);		return (0);
}		}

int		int
xdma_setup_intr(xdma_channel_t xchan, int (cb)(void ), void arg,		xdma_setup_intr(xdma_channel_t *xchan,
void **ihandler)		int (cb)(void , xdma_transfer_status_t *),
		void arg, void *ihandler)
{		{
struct xdma_intr_handler *ih;		struct xdma_intr_handler *ih;
xdma_controller_t *xdma;		xdma_controller_t *xdma;

xdma = xchan->xdma;		xdma = xchan->xdma;
KASSERT(xdma != NULL, ("xdma is NULL"));		KASSERT(xdma != NULL, ("xdma is NULL"));

/* Sanity check. */		/* Sanity check. */
▲ Show 20 Lines • Show All 61 Lines • ▼ Show 20 Lines	xdma_teardown_all_intr(xdma_channel_t *xchan)
TAILQ_FOREACH_SAFE(ih, &xchan->ie_handlers, ih_next, ih_tmp) {		TAILQ_FOREACH_SAFE(ih, &xchan->ie_handlers, ih_next, ih_tmp) {
TAILQ_REMOVE(&xchan->ie_handlers, ih, ih_next);		TAILQ_REMOVE(&xchan->ie_handlers, ih, ih_next);
free(ih, M_XDMA);		free(ih, M_XDMA);
}		}

return (0);		return (0);
}		}

static void		static int
xdma_dmamap_cb(void arg, bus_dma_segment_t segs, int nseg, int err)		xdma_bufs_alloc_no_busdma(xdma_channel_t *xchan)
		kanUnsubmitted Done Inline Actions This is a xdma_desc_dmamap_cb in , the name is too short and nondescript. I do not think descriptor allocation belongs in xdma at all, but this is wrong nonetheless. kan: This is a xdma_desc_dmamap_cb in , the name is too short and nondescript. I do not think…
		brAuthorUnsubmitted Not Done Inline Actions I moved descriptors allocation to DMA engine drivers. Thanks br: I moved descriptors allocation to DMA engine drivers. Thanks
{		{
xdma_channel_t *xchan;		xdma_controller_t *xdma;
		int nsegments;
int i;		int i;

xchan = (xdma_channel_t *)arg;		xdma = xchan->xdma;
KASSERT(xchan != NULL, ("xchan is NULL"));

if (err) {		nsegments = xchan->bufs_num;
xchan->map_err = 1;
return;		xchan->bufs = malloc(nsegments * sizeof(struct xchan_buf),
		M_XDMA, (M_WAITOK \| M_ZERO));
		if (xchan->bufs == NULL) {
		device_printf(xdma->dev,
		"%s: Can't allocate memory.\n", __func__);
		return (-1);
}		}

for (i = 0; i < nseg; i++) {		for (i = 0; i < nsegments; i++) {
xchan->descs_phys[i].ds_addr = segs[i].ds_addr;		xchan->bufs[i].cbuf = contigmalloc(xchan->maxsegsize,
xchan->descs_phys[i].ds_len = segs[i].ds_len;		M_XDMA, 0, 0, ~0, PAGE_SIZE, 0);
}		}

		return (0);
}		}

static int		static int
xdma_desc_alloc_bus_dma(xdma_channel_t *xchan, uint32_t desc_size,		xdma_bufs_alloc_busdma(xdma_channel_t *xchan)
uint32_t align)
{		{
xdma_controller_t *xdma;		xdma_controller_t *xdma;
bus_size_t all_desc_sz;
xdma_config_t *conf;
int nsegments;		int nsegments;
int err;		int err;
		int i;

xdma = xchan->xdma;		xdma = xchan->xdma;
conf = &xchan->conf;

nsegments = conf->block_num;		nsegments = xchan->bufs_num;
all_desc_sz = (nsegments * desc_size);

		xchan->bufs = malloc(nsegments * sizeof(struct xchan_buf),
		M_XDMA, (M_WAITOK \| M_ZERO));
		if (xchan->bufs == NULL) {
		device_printf(xdma->dev,
		"%s: Can't allocate memory.\n", __func__);
		return (-1);
		}

err = bus_dma_tag_create(		err = bus_dma_tag_create(
bus_get_dma_tag(xdma->dev),		bus_get_dma_tag(xdma->dev), /* Parent tag. */
align, desc_size, /* alignment, boundary */		xchan->alignment, 0, /* alignment, boundary */
BUS_SPACE_MAXADDR_32BIT, /* lowaddr */		BUS_SPACE_MAXADDR_32BIT, /* lowaddr */
		kanUnsubmitted Done Inline Actions Why? How xchan knows boundaries handed by the underlying DMA device? How does it even know if bus_get_dma_tag(xdma->dev) is even a right tag to be used for descriptors? Descriptors often have much stricter requirements on alignment and locations of descriptors and just 'uint32_t desc_size, uint32_t align' cannot possibly describe them. The operation does not belong in generic xchan layer at all and is better left to an actual pdma implementation. kan: Why? How xchan knows boundaries handed by the underlying DMA device? How does it even know if…
		brAuthorUnsubmitted Not Done Inline Actions Sounds right, thanks. I moved all the descriptors allocation/handling to DMA engine driver. br: Sounds right, thanks. I moved all the descriptors allocation/handling to DMA engine driver.
BUS_SPACE_MAXADDR, /* highaddr */		BUS_SPACE_MAXADDR, /* highaddr */
NULL, NULL, /* filter, filterarg */		NULL, NULL, /* filter, filterarg */
all_desc_sz, nsegments, /* maxsize, nsegments*/		xchan->maxsegsize * xchan->maxnsegs, /* maxsize */
desc_size, 0, /* maxsegsize, flags */		xchan->maxnsegs, /* nsegments */
		xchan->maxsegsize, /* maxsegsize */
		0, /* flags */
NULL, NULL, /* lockfunc, lockarg */		NULL, NULL, /* lockfunc, lockarg */
&xchan->dma_tag);		&xchan->dma_tag_bufs);
if (err) {		if (err != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't create bus_dma tag.\n", __func__);		"%s: Can't create bus_dma tag.\n", __func__);
return (-1);		return (-1);
}		}

err = bus_dmamem_alloc(xchan->dma_tag, (void **)&xchan->descs,		for (i = 0; i < nsegments; i++) {
BUS_DMA_WAITOK \| BUS_DMA_COHERENT, &xchan->dma_map);		err = bus_dmamap_create(xchan->dma_tag_bufs, 0,
if (err) {		&xchan->bufs[i].map);
		if (err != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't allocate memory for descriptors.\n", __func__);		"%s: Can't create buf DMA map.\n", __func__);
return (-1);		return (-1);
}		}
		}

xchan->descs_phys = malloc(nsegments * sizeof(xdma_descriptor_t), M_XDMA,		return (0);
(M_WAITOK \| M_ZERO));		}

xchan->map_err = 0;		static int
err = bus_dmamap_load(xchan->dma_tag, xchan->dma_map, xchan->descs,		xchan_bufs_alloc(xdma_channel_t *xchan)
all_desc_sz, xdma_dmamap_cb, xchan, BUS_DMA_WAITOK);		{
if (err) {		xdma_controller_t *xdma;
		int ret;

		xdma = xchan->xdma;
		if (xdma == NULL) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't load DMA map.\n", __func__);		"%s: Channel was not allocated properly.\n", __func__);
return (-1);		return (-1);
}		}

if (xchan->map_err != 0) {		if (xchan->caps & XCHAN_CAP_BUSDMA) {
		ret = xdma_bufs_alloc_busdma(xchan);
		} else {
		ret = xdma_bufs_alloc_no_busdma(xchan);
		}
		kanUnsubmitted Not Done Inline Actions Error handling, in particular, freeing already allocated memory and descriptors and unmapping mapped descriptors is considered futile? kan: Error handling, in particular, freeing already allocated memory and descriptors and unmapping…
		if (ret != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't load DMA map.\n", __func__);		"%s: Can't setup busdma.\n",
		__func__);
return (-1);		return (-1);
}		}

		xchan->flags \|= XCHAN_BUFS_ALLOCATED;

return (0);		return (0);
}		}

/*		static int
* This function called by DMA controller driver.		xchan_bufs_free(xdma_channel_t *xchan)
*/		{
		xchan_buf_t *b;
		kanUnsubmitted Done Inline Actions The name of the function and later use of MCLBYTES do correspond to each other rather weakly. It if an mbuf-specific functionality, name better should reflect that. kan: The name of the function and later use of MCLBYTES do correspond to each other rather weakly.
		int i;
		kanUnsubmitted Done Inline Actions Same thing. This seems to be very mbuf specific and ridden with arbitrary limitations. kan: Same thing. This seems to be very mbuf specific and ridden with arbitrary limitations.
		brAuthorUnsubmitted Not Done Inline Actions tag parameters now specified by transfer preparation functions, e.g. xdma_prep_sg br: tag parameters now specified by transfer preparation functions, e.g. xdma_prep_sg

		if ((xchan->flags & XCHAN_BUFS_ALLOCATED) == 0) {
		/* No bufs allocated. */
		return (-1);
		}

		if (xchan->caps & XCHAN_CAP_BUSDMA) {
		for (i = 0; i < xchan->bufs_num; i++) {
		b = &xchan->bufs[i];
		bus_dmamap_destroy(xchan->dma_tag_bufs, b->map);
		}
		bus_dma_tag_destroy(xchan->dma_tag_bufs);
		} else {
		for (i = 0; i < xchan->bufs_num; i++) {
		contigfree(xchan->bufs[i].cbuf, xchan->maxsegsize, M_XDMA);
		}
		}

		free(xchan->bufs, M_XDMA);
		xchan->flags &= ~XCHAN_BUFS_ALLOCATED;

		return (0);
		}
		kanUnsubmitted Done Inline Actions Again, WHY??? kan: Again, WHY???

int		int
xdma_desc_alloc(xdma_channel_t *xchan, uint32_t desc_size, uint32_t align)		xdma_prep_memcpy(xdma_channel_t *xchan, uintptr_t src_addr,
		uintptr_t dst_addr, size_t len)
{		{
xdma_controller_t *xdma;		xdma_controller_t *xdma;
xdma_config_t *conf;		xdma_config_t *conf;
int ret;		int ret;

XCHAN_ASSERT_LOCKED(xchan);		xdma = xchan->xdma;
		KASSERT(xdma != NULL, ("xdma is NULL"));

		conf = &xchan->conf;
		conf->direction = XDMA_MEM_TO_MEM;
		conf->src_addr = src_addr;
		conf->dst_addr = dst_addr;
		conf->block_len = len;
		conf->block_num = 1;

		xchan->flags \|= (XCHAN_CONFIGURED \| XCHAN_TYPE_MEMCPY);

		XCHAN_LOCK(xchan);

		ret = XDMA_CHANNEL_PREP_MEMCPY(xdma->dma_dev, xchan);
		if (ret != 0) {
		device_printf(xdma->dev,
		"%s: Can't prepare memcpy transfer.\n", __func__);
		XCHAN_UNLOCK(xchan);

		return (-1);
		}

		XCHAN_UNLOCK(xchan);

		kanUnsubmitted Not Done Inline Actions BUS_DMA_COHERENT??? kan: BUS_DMA_COHERENT???
		return (0);
		}

		/*
		* xr_num - xdma requests queue size,
		* maxsegsize - maximum allowed scatter-gather list element size in bytes
		*/
		int
		xdma_prep_sg(xdma_channel_t *xchan, uint32_t xr_num,
		uint32_t maxsegsize, uint32_t alignment)
		{
		xdma_controller_t *xdma;
		int ret;

xdma = xchan->xdma;		xdma = xchan->xdma;
if (xdma == NULL) {
		KASSERT(xdma != NULL, ("xdma is NULL"));

		if (xchan->flags & XCHAN_CONFIGURED) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Channel was not allocated properly.\n", __func__);		"%s: Channel is already configured.\n", __func__);
return (-1);		return (-1);
}		}

if (xchan->flags & XCHAN_DESC_ALLOCATED) {		xchan->maxsegsize = maxsegsize;
		xchan->alignment = alignment;
		xchan->maxnsegs = 8;
		xchan->bufs_num = xr_num;
		xchan->xr_num = xr_num;

		/* Allocate sglist. */
		ret = xchan_sglist_init(xchan);
		if (ret != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Descriptors already allocated.\n", __func__);		"%s: Can't allocate sglist.\n", __func__);
return (-1);		return (-1);
}		}

if ((xchan->flags & XCHAN_CONFIGURED) == 0) {		/* Allocate request queue. */
		xchan->xr = malloc(sizeof(struct xdma_request) * xr_num,
		M_XDMA, M_WAITOK \| M_ZERO);
		if (xchan->xr == NULL) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Channel has no configuration.\n", __func__);		"%s: Can't allocate request queue.\n", __func__);
return (-1);		return (-1);
}		}

conf = &xchan->conf;		xchan->flags \|= (XCHAN_CONFIGURED \| XCHAN_TYPE_SG);

XCHAN_UNLOCK(xchan);
ret = xdma_desc_alloc_bus_dma(xchan, desc_size, align);
XCHAN_LOCK(xchan);		XCHAN_LOCK(xchan);

		/* Deallocate bufs, if any. */
		xchan_bufs_free(xchan);

		ret = XDMA_CHANNEL_PREP_SG(xdma->dma_dev, xchan);
if (ret != 0) {		if (ret != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't allocate memory for descriptors.\n",		"%s: Can't prepare SG transfer.\n", __func__);
__func__);		XCHAN_UNLOCK(xchan);

return (-1);		return (-1);
}		}

xchan->flags \|= XCHAN_DESC_ALLOCATED;		/* Allocate bufs */
		ret = xchan_bufs_alloc(xchan);
		if (ret != 0) {
		device_printf(xdma->dev,
		"%s: Can't allocate bufs.\n", __func__);
		return (-1);
		}

/* We are going to write to descriptors. */		XCHAN_UNLOCK(xchan);
bus_dmamap_sync(xchan->dma_tag, xchan->dma_map, BUS_DMASYNC_PREWRITE);

return (0);		return (0);
}		}

int		int
xdma_desc_free(xdma_channel_t *xchan)		xdma_dequeue(xdma_channel_t xchan, void *user,
		xdma_transfer_status_t *status)
{		{
		struct xdma_request *xr;

if ((xchan->flags & XCHAN_DESC_ALLOCATED) == 0) {		if (xchan->xr_tail == xchan->xr_processed) {
/* No descriptors allocated. */
return (-1);		return (-1);
}		}

bus_dmamap_unload(xchan->dma_tag, xchan->dma_map);		xr = &xchan->xr[xchan->xr_tail];
bus_dmamem_free(xchan->dma_tag, xchan->descs, xchan->dma_map);		if (xr->done == 0) {
bus_dma_tag_destroy(xchan->dma_tag);		return (-1);
free(xchan->descs_phys, M_XDMA);		}

xchan->flags &= ~(XCHAN_DESC_ALLOCATED);		*user = xr->user;
		status->error = xr->status.error;
		status->transferred = xr->status.transferred;
		xchan->xr_tail = xchan_next_req(xchan, xchan->xr_tail);
		atomic_subtract_int(&xchan->xr_count, 1);
		kanUnsubmitted Done Inline Actions You did not protect XCHAN_DESC_ALLOCATED with ()'s above, why now? kan: You did not protect XCHAN_DESC_ALLOCATED with ()'s above, why now?

return (0);		return (0);
}		}

int		int
xdma_prep_memcpy(xdma_channel_t *xchan, uintptr_t src_addr,		xdma_enqueue(xdma_channel_t *xchan, uintptr_t src, uintptr_t dst,
uintptr_t dst_addr, size_t len)		uint8_t src_width, uint8_t dst_width, bus_size_t len,
		enum xdma_direction dir, void *user)
{		{
		struct xdma_request *xr;
xdma_controller_t *xdma;		xdma_controller_t *xdma;
xdma_config_t *conf;
		xdma = xchan->xdma;

		if (xchan->xr_count >= (xchan->xr_num - 1)) {
		/* No space is available yet. */
		kanUnsubmitted Done Inline Actions This is unconditional? You did not call map_create for !XCHAN_CAP_BUSDMA case. kan: This is unconditional? You did not call map_create for !XCHAN_CAP_BUSDMA case.
		brAuthorUnsubmitted Not Done Inline Actions Fixed. thanks! br: Fixed. thanks!
		return (-1);
		}

		xr = &xchan->xr[xchan->xr_head];
		xr->user = user;
		kanUnsubmitted Done Inline Actions You did not create tag either. kan: You did not create tag either.
		xr->direction = dir;
		xr->m = NULL;
		xr->len = len;
		kanUnsubmitted Done Inline Actions Same here? Why ()'? kan: Same here? Why ()'?
		xr->type = 0;
		xr->src_addr = src;
		xr->dst_addr = dst;
		xr->src_width = src_width;
		xr->dst_width = dst_width;
		xr->done = 0;

		xchan->xr_head = xchan_next_req(xchan, xchan->xr_head);
		atomic_add_int(&xchan->xr_count, 1);

		return (0);
		}

		struct seg_load_request {
		struct bus_dma_segment *seg;
		uint32_t nsegs;
		uint32_t error;
		};

		static void
		xdma_dmamap_cb(void arg, bus_dma_segment_t segs, int nsegs, int error)
		{
		struct seg_load_request *slr;
		struct bus_dma_segment *seg;
		int i;

		slr = arg;
		seg = slr->seg;

		if (error != 0) {
		slr->error = error;
		return;
		}

		slr->nsegs = nsegs;

		for (i = 0; i < nsegs; i++) {
		seg[i].ds_addr = segs[i].ds_addr;
		seg[i].ds_len = segs[i].ds_len;
		}
		}

		static int
		xdma_load_busdma(xdma_channel_t xchan, struct xdma_request xr,
		struct bus_dma_segment *seg, uint32_t i)
		{
		xdma_controller_t *xdma;
		struct seg_load_request slr;
		uint32_t nsegs;
		void *addr;
		int error;

		xdma = xchan->xdma;

		error = 0;
		nsegs = 0;

		switch (xr->type) {
		case XR_TYPE_MBUF:
		error = bus_dmamap_load_mbuf_sg(xchan->dma_tag_bufs,
		xchan->bufs[i].map, xr->m, seg, &nsegs, BUS_DMA_NOWAIT);
		break;
		case XR_TYPE_BIO:
		slr.nsegs = 0;
		slr.error = 0;
		slr.seg = seg;
		error = bus_dmamap_load_bio(xchan->dma_tag_bufs,
		xchan->bufs[i].map, xr->bp, xdma_dmamap_cb, &slr, BUS_DMA_NOWAIT);
		if (slr.error != 0) {
		device_printf(xdma->dma_dev,
		"%s: bus_dmamap_load failed, err %d\n",
		__func__, slr.error);
		return (0);
		}
		nsegs = slr.nsegs;
		break;
		case XR_TYPE_ADDR:
		switch (xr->direction) {
		case XDMA_MEM_TO_DEV:
		addr = (void *)xr->src_addr;
		break;
		case XDMA_DEV_TO_MEM:
		addr = (void *)xr->dst_addr;
		break;
		default:
		device_printf(xdma->dma_dev,
		"%s: Direction is not supported\n", __func__);
		return (0);
		}
		slr.nsegs = 0;
		slr.error = 0;
		slr.seg = seg;
		error = bus_dmamap_load(xchan->dma_tag_bufs, xchan->bufs[i].map,
		addr, xr->len, xdma_dmamap_cb, &slr, BUS_DMA_NOWAIT);
		if (slr.error != 0) {
		device_printf(xdma->dma_dev,
		"%s: bus_dmamap_load failed, err %d\n",
		__func__, slr.error);
		return (0);
		}
		nsegs = slr.nsegs;
		break;
		default:
		break;
		}

		if (error != 0) {
		if (error == ENOMEM) {
		/*
		* Out of memory. Try again later.
		* TODO: count errors.
		*/
		} else {
		device_printf(xdma->dma_dev,
		"%s: bus_dmamap_load failed with err %d\n",
		__func__, error);
		}
		return (0);
		}

		if (xr->direction == XDMA_MEM_TO_DEV) {
		bus_dmamap_sync(xchan->dma_tag_bufs, xchan->bufs[i].map,
		BUS_DMASYNC_PREWRITE);
		} else {
		bus_dmamap_sync(xchan->dma_tag_bufs, xchan->bufs[i].map,
		BUS_DMASYNC_PREREAD);
		}

		return (nsegs);
		}

		static int
		xdma_load_no_busdma(xdma_channel_t xchan, struct xdma_request xr,
		struct bus_dma_segment *seg, uint32_t i)
		{
		xdma_controller_t *xdma;
		struct mbuf *m;
		uint32_t nsegs;

		xdma = xchan->xdma;

		m = xr->m;

		nsegs = 1;

		switch (xr->type) {
		case XR_TYPE_MBUF:
		if (xr->direction == XDMA_MEM_TO_DEV) {
		m_copydata(m, 0, m->m_pkthdr.len, xchan->bufs[i].cbuf);
		seg[0].ds_addr = (bus_addr_t)xchan->bufs[i].cbuf;
		seg[0].ds_len = m->m_pkthdr.len;
		} else {
		seg[0].ds_addr = mtod(m, bus_addr_t);
		seg[0].ds_len = m->m_pkthdr.len;
		}
		break;
		case XR_TYPE_BIO:
		case XR_TYPE_ADDR:
		default:
		panic("implement me\n");
		}

		return (nsegs);
		}

		static int
		xdma_sglist_prepare_one(xdma_channel_t *xchan,
		struct xdma_request xr, struct bus_dma_segment seg)
		{
		xdma_controller_t *xdma;
		int error;
		int nsegs;
		int i;

		xdma = xchan->xdma;

		error = 0;
		nsegs = 0;

		i = xchan->buf_head;

		if (xchan->caps & XCHAN_CAP_BUSDMA) {
		nsegs = xdma_load_busdma(xchan, xr, seg, i);
		} else {
		nsegs = xdma_load_no_busdma(xchan, xr, seg, i);
		}
		if (nsegs == 0) {
		/* Try again later. */
		return (0);
		}

		xchan->bufs[i].xr = xr;
		xchan->bufs[i].nsegs = nsegs;
		xchan->bufs[i].nsegs_left = nsegs;

		xchan->buf_head = xchan_next_buf(xchan, xchan->buf_head);

		return (nsegs);
		}

		static int
		xdma_sglist_prepare(xdma_channel_t *xchan,
		struct xdma_sglist *sg)
		{
		struct bus_dma_segment seg[128];
		struct xdma_request *xr;
		xdma_controller_t *xdma;
		uint32_t capacity;
		uint32_t n;
		uint32_t c;
		int nsegs;
int ret;		int ret;

xdma = xchan->xdma;		xdma = xchan->xdma;

		n = 0;

		ret = XDMA_CHANNEL_CAPACITY(xdma->dma_dev, xchan, &capacity);
		if (ret != 0) {
		device_printf(xdma->dev,
		"%s: Can't get DMA controller capacity.\n", __func__);
		return (-1);
		}

		for (;;) {
		if (xchan->xr_processed == xchan->xr_head) {
		/* All the requests processed. */
		break;
		}
		xr = &xchan->xr[xchan->xr_processed];

		switch (xr->type) {
		case XR_TYPE_MBUF:
		c = xdma_mbuf_defrag(xchan, xr);
		break;
		case XR_TYPE_BIO:
		case XR_TYPE_ADDR:
		default:
		c = 1;
		}

		if (capacity <= (c + n)) {
		/*
		* No space yet available for the entire
		* request in the DMA engine.
		*/
		break;
		}

		nsegs = xdma_sglist_prepare_one(xchan, xr, seg);
		if (nsegs == 0) {
		break;
		}

		xdma_sglist_add(&sg[n], seg, nsegs, xr);
		n += nsegs;

		xchan->xr_processed = xchan_next_req(xchan, xchan->xr_processed);
		}

		return (n);
		}

		int
		xdma_queue_submit(xdma_channel_t *xchan)
		{
		struct xdma_sglist *sg;
		xdma_controller_t *xdma;
		uint32_t sg_n;
		int ret;

		xdma = xchan->xdma;
KASSERT(xdma != NULL, ("xdma is NULL"));		KASSERT(xdma != NULL, ("xdma is NULL"));

conf = &xchan->conf;		sg = xchan->sg;
conf->direction = XDMA_MEM_TO_MEM;
conf->src_addr = src_addr;
conf->dst_addr = dst_addr;
conf->block_len = len;
conf->block_num = 1;

xchan->flags \|= (XCHAN_CONFIGURED \| XCHAN_TYPE_MEMCPY);		if ((xchan->flags & XCHAN_BUFS_ALLOCATED) == 0) {
		device_printf(xdma->dev,
		"%s: Can't submit SG transfer: no bufs\n",
		__func__);
		return (-1);
		}

XCHAN_LOCK(xchan);		XCHAN_LOCK(xchan);

/* Deallocate old descriptors, if any. */		sg_n = xdma_sglist_prepare(xchan, sg);
xdma_desc_free(xchan);		if (sg_n == 0) {
		/* Nothing to submit */
		XCHAN_UNLOCK(xchan);
		return (0);
		}

ret = XDMA_CHANNEL_PREP_MEMCPY(xdma->dma_dev, xchan);		/* Now submit xdma_sglist to DMA engine driver. */

		ret = XDMA_CHANNEL_SUBMIT_SG(xdma->dma_dev, xchan, sg, sg_n);
if (ret != 0) {		if (ret != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't prepare memcpy transfer.\n", __func__);		"%s: Can't submit SG transfer.\n", __func__);

XCHAN_UNLOCK(xchan);		XCHAN_UNLOCK(xchan);

return (-1);		return (-1);
}		}

if (xchan->flags & XCHAN_DESC_ALLOCATED) {
/* Driver created xDMA descriptors. */
bus_dmamap_sync(xchan->dma_tag, xchan->dma_map,
BUS_DMASYNC_POSTWRITE);
}

XCHAN_UNLOCK(xchan);		XCHAN_UNLOCK(xchan);

return (0);		return (0);
}		}

int		int
xdma_prep_cyclic(xdma_channel_t *xchan, enum xdma_direction dir,		xdma_prep_cyclic(xdma_channel_t *xchan, enum xdma_direction dir,
uintptr_t src_addr, uintptr_t dst_addr, int block_len,		uintptr_t src_addr, uintptr_t dst_addr, int block_len,
Show All 14 Lines	xdma_prep_cyclic(xdma_channel_t *xchan, enum xdma_direction dir,
conf->block_num = block_num;		conf->block_num = block_num;
conf->src_width = src_width;		conf->src_width = src_width;
conf->dst_width = dst_width;		conf->dst_width = dst_width;

xchan->flags \|= (XCHAN_CONFIGURED \| XCHAN_TYPE_CYCLIC);		xchan->flags \|= (XCHAN_CONFIGURED \| XCHAN_TYPE_CYCLIC);

XCHAN_LOCK(xchan);		XCHAN_LOCK(xchan);

/* Deallocate old descriptors, if any. */		/* Deallocate bufs, if any. */
xdma_desc_free(xchan);		xchan_bufs_free(xchan);

ret = XDMA_CHANNEL_PREP_CYCLIC(xdma->dma_dev, xchan);		ret = XDMA_CHANNEL_PREP_CYCLIC(xdma->dma_dev, xchan);
if (ret != 0) {		if (ret != 0) {
		kanUnsubmitted Done Inline Actions By using circular buffer here you encode the requirement that dma requests should be completed by the controller in the order they were submitted, but that is not necessarily the case. I guess same applies by sglists. kan: By using circular buffer here you encode the requirement that dma requests should be completed…
		brAuthorUnsubmitted Not Done Inline Actions I switched back to linked lists. Works a bit slower due to locking, but easier to improve /add features in future. thanks! br: I switched back to linked lists. Works a bit slower due to locking, but easier to improve /add…
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't prepare cyclic transfer.\n", __func__);		"%s: Can't prepare cyclic transfer.\n", __func__);
XCHAN_UNLOCK(xchan);		XCHAN_UNLOCK(xchan);

return (-1);		return (-1);
}		}

if (xchan->flags & XCHAN_DESC_ALLOCATED) {
/* Driver has created xDMA descriptors. */
bus_dmamap_sync(xchan->dma_tag, xchan->dma_map,
BUS_DMASYNC_POSTWRITE);
}

XCHAN_UNLOCK(xchan);		XCHAN_UNLOCK(xchan);

return (0);		return (0);
}		}

int		int
		kanUnsubmitted Done Inline Actions ()'s again kan: ()'s again
xdma_begin(xdma_channel_t *xchan)		xdma_begin(xdma_channel_t *xchan)
{		{
xdma_controller_t *xdma;		xdma_controller_t *xdma;
int ret;		int ret;

xdma = xchan->xdma;		xdma = xchan->xdma;

		if (xchan->flags & XCHAN_TYPE_SG) {
		/* Not valid. */
		return (0);
		};

ret = XDMA_CHANNEL_CONTROL(xdma->dma_dev, xchan, XDMA_CMD_BEGIN);		ret = XDMA_CHANNEL_CONTROL(xdma->dma_dev, xchan, XDMA_CMD_BEGIN);
if (ret != 0) {		if (ret != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't begin the channel operation.\n", __func__);		"%s: Can't begin the channel operation.\n", __func__);
return (-1);		return (-1);
}		}

return (0);		return (0);
Show All 30 Lines	if (ret != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't pause the channel operation.\n", __func__);		"%s: Can't pause the channel operation.\n", __func__);
return (-1);		return (-1);
}		}

return (ret);		return (ret);
}		}

int		int
		kanUnsubmitted Not Done Inline Actions Is there a direct relationship of number of mbuf fragments and number of physical addresses it can result in? kan: Is there a direct relationship of number of mbuf fragments and number of physical addresses it…
xdma_callback(xdma_channel_t *xchan)		xchan_seg_done(xdma_channel_t *xchan,
		struct xdma_transfer_status *st)
{		{
struct xdma_intr_handler *ih_tmp;		struct xdma_request *xr;
struct xdma_intr_handler *ih;		xdma_controller_t *xdma;
		xchan_buf_t *b;

TAILQ_FOREACH_SAFE(ih, &xchan->ie_handlers, ih_next, ih_tmp) {		xdma = xchan->xdma;
if (ih->cb != NULL) {
ih->cb(ih->cb_user);		b = &xchan->bufs[xchan->buf_tail];
		xr = b->xr;

		atomic_subtract_int(&b->nsegs_left, 1);

		if (b->nsegs_left == 0) {
		kanUnsubmitted Not Done Inline Actions What happens if segment in the middle completes with the error? Also, what is st_trasferred and in what units it is specified? You invoke callback in per-desc (WHY? DMA driver is not allowed to do desc chaining?), but then only report status on the last desc up the consumer chain. kan: What happens if segment in the middle completes with the error? Also, what is st_trasferred and…
		if (xchan->caps & XCHAN_CAP_BUSDMA) {
		if (xr->direction == XDMA_MEM_TO_DEV) {
		bus_dmamap_sync(xchan->dma_tag_bufs, b->map,
		BUS_DMASYNC_POSTWRITE);
		} else {
		bus_dmamap_sync(xchan->dma_tag_bufs, b->map,
		BUS_DMASYNC_POSTREAD);
}		}

		bus_dmamap_unload(xchan->dma_tag_bufs, b->map);
}		}
		xr->status.error = st->error;
		xr->status.transferred = st->transferred;
		xr->done = 1;

		xchan->buf_tail = xchan_next_buf(xchan, xchan->buf_tail);
		}

return (0);		return (0);
}		}

void		int
xdma_assert_locked(void)		xdma_callback(xdma_channel_t xchan, xdma_transfer_status_t status)
{		{
		struct xdma_intr_handler *ih_tmp;
		struct xdma_intr_handler *ih;
		xdma_controller_t *xdma;

XDMA_ASSERT_LOCKED();		xdma = xchan->xdma;

		TAILQ_FOREACH_SAFE(ih, &xchan->ie_handlers, ih_next, ih_tmp) {
		if (ih->cb != NULL) {
		ih->cb(ih->cb_user, status);
}		}
		}

		wakeup(xdma);

		return (0);
		}

#ifdef FDT		#ifdef FDT
/*		/*
* Notify the DMA driver we have machine-dependent data in FDT.		* Notify the DMA driver we have machine-dependent data in FDT.
*/		*/
static int		static int
xdma_ofw_md_data(xdma_controller_t xdma, pcell_t cells, int ncells)		xdma_ofw_md_data(xdma_controller_t xdma, pcell_t cells, int ncells)
{		{
uint32_t ret;		uint32_t ret;

ret = XDMA_OFW_MD_DATA(xdma->dma_dev, cells, ncells, (void **)&xdma->data);		ret = XDMA_OFW_MD_DATA(xdma->dma_dev,
		cells, ncells, (void **)&xdma->data);

return (ret);		return (ret);
}		}

/*		/*
* Allocate xdma controller.		* Allocate xdma controller.
*/		*/
xdma_controller_t *		xdma_controller_t *
▲ Show 20 Lines • Show All 59 Lines • ▼ Show 20 Lines	xdma_ofw_get(device_t dev, const char *prop)
xdma->dev = dev;		xdma->dev = dev;
xdma->dma_dev = dma_dev;		xdma->dma_dev = dma_dev;

TAILQ_INIT(&xdma->channels);		TAILQ_INIT(&xdma->channels);

xdma_ofw_md_data(xdma, cells, ncells);		xdma_ofw_md_data(xdma, cells, ncells);
free(cells, M_OFWPROP);		free(cells, M_OFWPROP);

		mtx_init(&xdma->proc_mtx, "xDMA ofw controller", NULL, MTX_DEF);
		kproc_create(&xdma_task, xdma, &xdma->xdma_proc, 0, 0, "xdma drainer");

return (xdma);		return (xdma);
}		}
#endif		#endif

/*		/*
* Free xDMA controller object.		* Free xDMA controller object.
*/		*/
int		int
xdma_put(xdma_controller_t *xdma)		xdma_put(xdma_controller_t *xdma)
{		{

XDMA_LOCK();		XDMA_LOCK();

/* Ensure no channels allocated. */		/* Ensure no channels allocated. */
if (!TAILQ_EMPTY(&xdma->channels)) {		if (!TAILQ_EMPTY(&xdma->channels)) {
device_printf(xdma->dev, "%s: Can't free xDMA\n", __func__);		device_printf(xdma->dev, "%s: Can't free xDMA\n", __func__);
return (-1);		return (-1);
}		}

		kproc_shutdown(&xdma->xdma_proc, 0);
		mtx_destroy(&xdma->proc_mtx);

free(xdma->data, M_DEVBUF);		free(xdma->data, M_DEVBUF);
free(xdma, M_XDMA);		free(xdma, M_XDMA);

XDMA_UNLOCK();		XDMA_UNLOCK();

return (0);		return (0);
}		}
Show All 9 Lines

View Options

sys/compat/freebsd32/freebsd32_misc.c

/*-		/*-
* Copyright (c) 2016 Ruslan Bukin <br@bsdpad.com>		* Copyright (c) 2016-2017 Ruslan Bukin <br@bsdpad.com>
* All rights reserved.		* All rights reserved.
*		*
* This software was developed by SRI International and the University of		* This software was developed by SRI International and the University of
* Cambridge Computer Laboratory under DARPA/AFRL contract FA8750-10-C-0237		* Cambridge Computer Laboratory under DARPA/AFRL contract FA8750-10-C-0237
* ("CTSRD"), as part of the DARPA CRASH research programme.		* ("CTSRD"), as part of the DARPA CRASH research programme.
*		*
* Redistribution and use in source and binary forms, with or without		* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions		* modification, are permitted provided that the following conditions
Show All 20 Lines
#include <sys/cdefs.h>		#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");		__FBSDID("$FreeBSD$");

#include "opt_platform.h"		#include "opt_platform.h"
#include <sys/param.h>		#include <sys/param.h>
#include <sys/conf.h>		#include <sys/conf.h>
#include <sys/bus.h>		#include <sys/bus.h>
#include <sys/kernel.h>		#include <sys/kernel.h>
		#include <sys/kthread.h>
#include <sys/queue.h>		#include <sys/queue.h>
#include <sys/kobj.h>		#include <sys/kobj.h>
#include <sys/malloc.h>		#include <sys/malloc.h>
#include <sys/mutex.h>		#include <sys/mutex.h>
		#include <sys/mbuf.h>
#include <sys/limits.h>		#include <sys/limits.h>
#include <sys/lock.h>		#include <sys/lock.h>
#include <sys/sysctl.h>		#include <sys/sysctl.h>
#include <sys/systm.h>		#include <sys/systm.h>
#include <sys/sx.h>		#include <sys/sx.h>
#include <sys/bus_dma.h>		#include <sys/bus_dma.h>

#include <machine/bus.h>		#include <machine/bus.h>

#ifdef FDT		#ifdef FDT
#include <dev/fdt/fdt_common.h>		#include <dev/fdt/fdt_common.h>
#include <dev/ofw/ofw_bus.h>		#include <dev/ofw/ofw_bus.h>
#include <dev/ofw/ofw_bus_subr.h>		#include <dev/ofw/ofw_bus_subr.h>
#endif		#endif

#include <dev/xdma/xdma.h>		#include <dev/xdma/xdma.h>

#include <xdma_if.h>		#include <xdma_if.h>

MALLOC_DEFINE(M_XDMA, "xdma", "xDMA framework");

/*		/*
* Multiple xDMA controllers may work with single DMA device,		* Multiple xDMA controllers may work with single DMA device,
* so we have global lock for physical channel management.		* so we have global lock for physical channel management.
*/		*/
static struct mtx xdma_mtx;		static struct mtx xdma_mtx;

#define XDMA_LOCK() mtx_lock(&xdma_mtx)		#define XDMA_LOCK() mtx_lock(&xdma_mtx)
#define XDMA_UNLOCK() mtx_unlock(&xdma_mtx)		#define XDMA_UNLOCK() mtx_unlock(&xdma_mtx)
#define XDMA_ASSERT_LOCKED() mtx_assert(&xdma_mtx, MA_OWNED)		#define XDMA_ASSERT_LOCKED() mtx_assert(&xdma_mtx, MA_OWNED)

/*		/*
* Per channel locks.		* Per channel locks.
*/		*/
#define XCHAN_LOCK(xchan) mtx_lock(&(xchan)->mtx_lock)		#define XCHAN_LOCK(xchan) mtx_lock(&(xchan)->mtx_lock)
#define XCHAN_UNLOCK(xchan) mtx_unlock(&(xchan)->mtx_lock)		#define XCHAN_UNLOCK(xchan) mtx_unlock(&(xchan)->mtx_lock)
#define XCHAN_ASSERT_LOCKED(xchan) mtx_assert(&(xchan)->mtx_lock, MA_OWNED)		#define XCHAN_ASSERT_LOCKED(xchan) mtx_assert(&(xchan)->mtx_lock, MA_OWNED)

		static int xchan_bufs_alloc(xdma_channel_t *xchan);
		static int xchan_bufs_free(xdma_channel_t *xchan);

		static void
		xdma_task(void *arg)
		{
		xdma_controller_t *xdma;
		xdma_channel_t *xchan_tmp;
		xdma_channel_t *xchan;

		xdma = arg;

		for (;;) {
		mtx_lock(&xdma->proc_mtx);
		msleep(xdma, &xdma->proc_mtx, PRIBIO, "jobqueue", hz);
		mtx_unlock(&xdma->proc_mtx);

		if (TAILQ_EMPTY(&xdma->channels)) {
		continue;
		}

		TAILQ_FOREACH_SAFE(xchan, &xdma->channels, xchan_next, xchan_tmp) {
		if (xchan->flags & XCHAN_TYPE_SG) {
		xdma_queue_submit(xchan);
		}
		}
		}
		}

/*		/*
* Allocate virtual xDMA channel.		* Allocate virtual xDMA channel.
*/		*/
xdma_channel_t *		xdma_channel_t *
xdma_channel_alloc(xdma_controller_t *xdma)		xdma_channel_alloc(xdma_controller_t *xdma, uint32_t caps)
{		{
xdma_channel_t *xchan;		xdma_channel_t *xchan;
int ret;		int ret;

xchan = malloc(sizeof(xdma_channel_t), M_XDMA, M_WAITOK \| M_ZERO);		xchan = malloc(sizeof(xdma_channel_t), M_XDMA, M_WAITOK \| M_ZERO);
if (xchan == NULL) {		if (xchan == NULL) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't allocate memory for channel.\n", __func__);		"%s: Can't allocate memory for channel.\n", __func__);
return (NULL);		return (NULL);
}		}
xchan->xdma = xdma;		xchan->xdma = xdma;
		xchan->caps = caps;

XDMA_LOCK();		XDMA_LOCK();

/* Request a real channel from hardware driver. */		/* Request a real channel from hardware driver. */
ret = XDMA_CHANNEL_ALLOC(xdma->dma_dev, xchan);		ret = XDMA_CHANNEL_ALLOC(xdma->dma_dev, xchan);
if (ret != 0) {		if (ret != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't request hardware channel.\n", __func__);		"%s: Can't request hardware channel.\n", __func__);
Show All 29 Lines	if (err != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't free real hw channel.\n", __func__);		"%s: Can't free real hw channel.\n", __func__);
XDMA_UNLOCK();		XDMA_UNLOCK();
return (-1);		return (-1);
}		}

xdma_teardown_all_intr(xchan);		xdma_teardown_all_intr(xchan);

/* Deallocate descriptors, if any. */		/* Deallocate bufs, if any. */
xdma_desc_free(xchan);		xchan_bufs_free(xchan);
		xchan_sglist_free(xchan);

		if (xchan->flags & XCHAN_TYPE_SG) {
		free(xchan->xr, M_XDMA);
		}

mtx_destroy(&xchan->mtx_lock);		mtx_destroy(&xchan->mtx_lock);

TAILQ_REMOVE(&xdma->channels, xchan, xchan_next);		TAILQ_REMOVE(&xdma->channels, xchan, xchan_next);

free(xchan, M_XDMA);		free(xchan, M_XDMA);

XDMA_UNLOCK();		XDMA_UNLOCK();

return (0);		return (0);
}		}

int		int
xdma_setup_intr(xdma_channel_t xchan, int (cb)(void ), void arg,		xdma_setup_intr(xdma_channel_t *xchan,
void **ihandler)		int (cb)(void , xdma_transfer_status_t *),
		void arg, void *ihandler)
{		{
struct xdma_intr_handler *ih;		struct xdma_intr_handler *ih;
xdma_controller_t *xdma;		xdma_controller_t *xdma;

xdma = xchan->xdma;		xdma = xchan->xdma;
KASSERT(xdma != NULL, ("xdma is NULL"));		KASSERT(xdma != NULL, ("xdma is NULL"));

/* Sanity check. */		/* Sanity check. */
▲ Show 20 Lines • Show All 61 Lines • ▼ Show 20 Lines	xdma_teardown_all_intr(xdma_channel_t *xchan)
TAILQ_FOREACH_SAFE(ih, &xchan->ie_handlers, ih_next, ih_tmp) {		TAILQ_FOREACH_SAFE(ih, &xchan->ie_handlers, ih_next, ih_tmp) {
TAILQ_REMOVE(&xchan->ie_handlers, ih, ih_next);		TAILQ_REMOVE(&xchan->ie_handlers, ih, ih_next);
free(ih, M_XDMA);		free(ih, M_XDMA);
}		}

return (0);		return (0);
}		}

static void		static int
xdma_dmamap_cb(void arg, bus_dma_segment_t segs, int nseg, int err)		xdma_bufs_alloc_no_busdma(xdma_channel_t *xchan)
		kanUnsubmitted Done Inline Actions This is a xdma_desc_dmamap_cb in , the name is too short and nondescript. I do not think descriptor allocation belongs in xdma at all, but this is wrong nonetheless. kan: This is a xdma_desc_dmamap_cb in , the name is too short and nondescript. I do not think…
		brAuthorUnsubmitted Not Done Inline Actions I moved descriptors allocation to DMA engine drivers. Thanks br: I moved descriptors allocation to DMA engine drivers. Thanks
{		{
xdma_channel_t *xchan;		xdma_controller_t *xdma;
		int nsegments;
int i;		int i;

xchan = (xdma_channel_t *)arg;		xdma = xchan->xdma;
KASSERT(xchan != NULL, ("xchan is NULL"));

if (err) {		nsegments = xchan->bufs_num;
xchan->map_err = 1;
return;		xchan->bufs = malloc(nsegments * sizeof(struct xchan_buf),
		M_XDMA, (M_WAITOK \| M_ZERO));
		if (xchan->bufs == NULL) {
		device_printf(xdma->dev,
		"%s: Can't allocate memory.\n", __func__);
		return (-1);
}		}

for (i = 0; i < nseg; i++) {		for (i = 0; i < nsegments; i++) {
xchan->descs_phys[i].ds_addr = segs[i].ds_addr;		xchan->bufs[i].cbuf = contigmalloc(xchan->maxsegsize,
xchan->descs_phys[i].ds_len = segs[i].ds_len;		M_XDMA, 0, 0, ~0, PAGE_SIZE, 0);
}		}

		return (0);
}		}

static int		static int
xdma_desc_alloc_bus_dma(xdma_channel_t *xchan, uint32_t desc_size,		xdma_bufs_alloc_busdma(xdma_channel_t *xchan)
uint32_t align)
{		{
xdma_controller_t *xdma;		xdma_controller_t *xdma;
bus_size_t all_desc_sz;
xdma_config_t *conf;
int nsegments;		int nsegments;
int err;		int err;
		int i;

xdma = xchan->xdma;		xdma = xchan->xdma;
conf = &xchan->conf;

nsegments = conf->block_num;		nsegments = xchan->bufs_num;
all_desc_sz = (nsegments * desc_size);

		xchan->bufs = malloc(nsegments * sizeof(struct xchan_buf),
		M_XDMA, (M_WAITOK \| M_ZERO));
		if (xchan->bufs == NULL) {
		device_printf(xdma->dev,
		"%s: Can't allocate memory.\n", __func__);
		return (-1);
		}

err = bus_dma_tag_create(		err = bus_dma_tag_create(
bus_get_dma_tag(xdma->dev),		bus_get_dma_tag(xdma->dev), /* Parent tag. */
align, desc_size, /* alignment, boundary */		xchan->alignment, 0, /* alignment, boundary */
BUS_SPACE_MAXADDR_32BIT, /* lowaddr */		BUS_SPACE_MAXADDR_32BIT, /* lowaddr */
		kanUnsubmitted Done Inline Actions Why? How xchan knows boundaries handed by the underlying DMA device? How does it even know if bus_get_dma_tag(xdma->dev) is even a right tag to be used for descriptors? Descriptors often have much stricter requirements on alignment and locations of descriptors and just 'uint32_t desc_size, uint32_t align' cannot possibly describe them. The operation does not belong in generic xchan layer at all and is better left to an actual pdma implementation. kan: Why? How xchan knows boundaries handed by the underlying DMA device? How does it even know if…
		brAuthorUnsubmitted Not Done Inline Actions Sounds right, thanks. I moved all the descriptors allocation/handling to DMA engine driver. br: Sounds right, thanks. I moved all the descriptors allocation/handling to DMA engine driver.
BUS_SPACE_MAXADDR, /* highaddr */		BUS_SPACE_MAXADDR, /* highaddr */
NULL, NULL, /* filter, filterarg */		NULL, NULL, /* filter, filterarg */
all_desc_sz, nsegments, /* maxsize, nsegments*/		xchan->maxsegsize * xchan->maxnsegs, /* maxsize */
desc_size, 0, /* maxsegsize, flags */		xchan->maxnsegs, /* nsegments */
		xchan->maxsegsize, /* maxsegsize */
		0, /* flags */
NULL, NULL, /* lockfunc, lockarg */		NULL, NULL, /* lockfunc, lockarg */
&xchan->dma_tag);		&xchan->dma_tag_bufs);
if (err) {		if (err != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't create bus_dma tag.\n", __func__);		"%s: Can't create bus_dma tag.\n", __func__);
return (-1);		return (-1);
}		}

err = bus_dmamem_alloc(xchan->dma_tag, (void **)&xchan->descs,		for (i = 0; i < nsegments; i++) {
BUS_DMA_WAITOK \| BUS_DMA_COHERENT, &xchan->dma_map);		err = bus_dmamap_create(xchan->dma_tag_bufs, 0,
if (err) {		&xchan->bufs[i].map);
		if (err != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't allocate memory for descriptors.\n", __func__);		"%s: Can't create buf DMA map.\n", __func__);
return (-1);		return (-1);
}		}
		}

xchan->descs_phys = malloc(nsegments * sizeof(xdma_descriptor_t), M_XDMA,		return (0);
(M_WAITOK \| M_ZERO));		}

xchan->map_err = 0;		static int
err = bus_dmamap_load(xchan->dma_tag, xchan->dma_map, xchan->descs,		xchan_bufs_alloc(xdma_channel_t *xchan)
all_desc_sz, xdma_dmamap_cb, xchan, BUS_DMA_WAITOK);		{
if (err) {		xdma_controller_t *xdma;
		int ret;

		xdma = xchan->xdma;
		if (xdma == NULL) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't load DMA map.\n", __func__);		"%s: Channel was not allocated properly.\n", __func__);
return (-1);		return (-1);
}		}

if (xchan->map_err != 0) {		if (xchan->caps & XCHAN_CAP_BUSDMA) {
		ret = xdma_bufs_alloc_busdma(xchan);
		} else {
		ret = xdma_bufs_alloc_no_busdma(xchan);
		}
		kanUnsubmitted Not Done Inline Actions Error handling, in particular, freeing already allocated memory and descriptors and unmapping mapped descriptors is considered futile? kan: Error handling, in particular, freeing already allocated memory and descriptors and unmapping…
		if (ret != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't load DMA map.\n", __func__);		"%s: Can't setup busdma.\n",
		__func__);
return (-1);		return (-1);
}		}

		xchan->flags \|= XCHAN_BUFS_ALLOCATED;

return (0);		return (0);
}		}

/*		static int
* This function called by DMA controller driver.		xchan_bufs_free(xdma_channel_t *xchan)
*/		{
		xchan_buf_t *b;
		kanUnsubmitted Done Inline Actions The name of the function and later use of MCLBYTES do correspond to each other rather weakly. It if an mbuf-specific functionality, name better should reflect that. kan: The name of the function and later use of MCLBYTES do correspond to each other rather weakly.
		int i;
		kanUnsubmitted Done Inline Actions Same thing. This seems to be very mbuf specific and ridden with arbitrary limitations. kan: Same thing. This seems to be very mbuf specific and ridden with arbitrary limitations.
		brAuthorUnsubmitted Not Done Inline Actions tag parameters now specified by transfer preparation functions, e.g. xdma_prep_sg br: tag parameters now specified by transfer preparation functions, e.g. xdma_prep_sg

		if ((xchan->flags & XCHAN_BUFS_ALLOCATED) == 0) {
		/* No bufs allocated. */
		return (-1);
		}

		if (xchan->caps & XCHAN_CAP_BUSDMA) {
		for (i = 0; i < xchan->bufs_num; i++) {
		b = &xchan->bufs[i];
		bus_dmamap_destroy(xchan->dma_tag_bufs, b->map);
		}
		bus_dma_tag_destroy(xchan->dma_tag_bufs);
		} else {
		for (i = 0; i < xchan->bufs_num; i++) {
		contigfree(xchan->bufs[i].cbuf, xchan->maxsegsize, M_XDMA);
		}
		}

		free(xchan->bufs, M_XDMA);
		xchan->flags &= ~XCHAN_BUFS_ALLOCATED;

		return (0);
		}
		kanUnsubmitted Done Inline Actions Again, WHY??? kan: Again, WHY???

int		int
xdma_desc_alloc(xdma_channel_t *xchan, uint32_t desc_size, uint32_t align)		xdma_prep_memcpy(xdma_channel_t *xchan, uintptr_t src_addr,
		uintptr_t dst_addr, size_t len)
{		{
xdma_controller_t *xdma;		xdma_controller_t *xdma;
xdma_config_t *conf;		xdma_config_t *conf;
int ret;		int ret;

XCHAN_ASSERT_LOCKED(xchan);		xdma = xchan->xdma;
		KASSERT(xdma != NULL, ("xdma is NULL"));

		conf = &xchan->conf;
		conf->direction = XDMA_MEM_TO_MEM;
		conf->src_addr = src_addr;
		conf->dst_addr = dst_addr;
		conf->block_len = len;
		conf->block_num = 1;

		xchan->flags \|= (XCHAN_CONFIGURED \| XCHAN_TYPE_MEMCPY);

		XCHAN_LOCK(xchan);

		ret = XDMA_CHANNEL_PREP_MEMCPY(xdma->dma_dev, xchan);
		if (ret != 0) {
		device_printf(xdma->dev,
		"%s: Can't prepare memcpy transfer.\n", __func__);
		XCHAN_UNLOCK(xchan);

		return (-1);
		}

		XCHAN_UNLOCK(xchan);

		kanUnsubmitted Not Done Inline Actions BUS_DMA_COHERENT??? kan: BUS_DMA_COHERENT???
		return (0);
		}

		/*
		* xr_num - xdma requests queue size,
		* maxsegsize - maximum allowed scatter-gather list element size in bytes
		*/
		int
		xdma_prep_sg(xdma_channel_t *xchan, uint32_t xr_num,
		uint32_t maxsegsize, uint32_t alignment)
		{
		xdma_controller_t *xdma;
		int ret;

xdma = xchan->xdma;		xdma = xchan->xdma;
if (xdma == NULL) {
		KASSERT(xdma != NULL, ("xdma is NULL"));

		if (xchan->flags & XCHAN_CONFIGURED) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Channel was not allocated properly.\n", __func__);		"%s: Channel is already configured.\n", __func__);
return (-1);		return (-1);
}		}

if (xchan->flags & XCHAN_DESC_ALLOCATED) {		xchan->maxsegsize = maxsegsize;
		xchan->alignment = alignment;
		xchan->maxnsegs = 8;
		xchan->bufs_num = xr_num;
		xchan->xr_num = xr_num;

		/* Allocate sglist. */
		ret = xchan_sglist_init(xchan);
		if (ret != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Descriptors already allocated.\n", __func__);		"%s: Can't allocate sglist.\n", __func__);
return (-1);		return (-1);
}		}

if ((xchan->flags & XCHAN_CONFIGURED) == 0) {		/* Allocate request queue. */
		xchan->xr = malloc(sizeof(struct xdma_request) * xr_num,
		M_XDMA, M_WAITOK \| M_ZERO);
		if (xchan->xr == NULL) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Channel has no configuration.\n", __func__);		"%s: Can't allocate request queue.\n", __func__);
return (-1);		return (-1);
}		}

conf = &xchan->conf;		xchan->flags \|= (XCHAN_CONFIGURED \| XCHAN_TYPE_SG);

XCHAN_UNLOCK(xchan);
ret = xdma_desc_alloc_bus_dma(xchan, desc_size, align);
XCHAN_LOCK(xchan);		XCHAN_LOCK(xchan);

		/* Deallocate bufs, if any. */
		xchan_bufs_free(xchan);

		ret = XDMA_CHANNEL_PREP_SG(xdma->dma_dev, xchan);
if (ret != 0) {		if (ret != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't allocate memory for descriptors.\n",		"%s: Can't prepare SG transfer.\n", __func__);
__func__);		XCHAN_UNLOCK(xchan);

return (-1);		return (-1);
}		}

xchan->flags \|= XCHAN_DESC_ALLOCATED;		/* Allocate bufs */
		ret = xchan_bufs_alloc(xchan);
		if (ret != 0) {
		device_printf(xdma->dev,
		"%s: Can't allocate bufs.\n", __func__);
		return (-1);
		}

/* We are going to write to descriptors. */		XCHAN_UNLOCK(xchan);
bus_dmamap_sync(xchan->dma_tag, xchan->dma_map, BUS_DMASYNC_PREWRITE);

return (0);		return (0);
}		}

int		int
xdma_desc_free(xdma_channel_t *xchan)		xdma_dequeue(xdma_channel_t xchan, void *user,
		xdma_transfer_status_t *status)
{		{
		struct xdma_request *xr;

if ((xchan->flags & XCHAN_DESC_ALLOCATED) == 0) {		if (xchan->xr_tail == xchan->xr_processed) {
/* No descriptors allocated. */
return (-1);		return (-1);
}		}

bus_dmamap_unload(xchan->dma_tag, xchan->dma_map);		xr = &xchan->xr[xchan->xr_tail];
bus_dmamem_free(xchan->dma_tag, xchan->descs, xchan->dma_map);		if (xr->done == 0) {
bus_dma_tag_destroy(xchan->dma_tag);		return (-1);
free(xchan->descs_phys, M_XDMA);		}

xchan->flags &= ~(XCHAN_DESC_ALLOCATED);		*user = xr->user;
		status->error = xr->status.error;
		status->transferred = xr->status.transferred;
		xchan->xr_tail = xchan_next_req(xchan, xchan->xr_tail);
		atomic_subtract_int(&xchan->xr_count, 1);
		kanUnsubmitted Done Inline Actions You did not protect XCHAN_DESC_ALLOCATED with ()'s above, why now? kan: You did not protect XCHAN_DESC_ALLOCATED with ()'s above, why now?

return (0);		return (0);
}		}

int		int
xdma_prep_memcpy(xdma_channel_t *xchan, uintptr_t src_addr,		xdma_enqueue(xdma_channel_t *xchan, uintptr_t src, uintptr_t dst,
uintptr_t dst_addr, size_t len)		uint8_t src_width, uint8_t dst_width, bus_size_t len,
		enum xdma_direction dir, void *user)
{		{
		struct xdma_request *xr;
xdma_controller_t *xdma;		xdma_controller_t *xdma;
xdma_config_t *conf;
		xdma = xchan->xdma;

		if (xchan->xr_count >= (xchan->xr_num - 1)) {
		/* No space is available yet. */
		kanUnsubmitted Done Inline Actions This is unconditional? You did not call map_create for !XCHAN_CAP_BUSDMA case. kan: This is unconditional? You did not call map_create for !XCHAN_CAP_BUSDMA case.
		brAuthorUnsubmitted Not Done Inline Actions Fixed. thanks! br: Fixed. thanks!
		return (-1);
		}

		xr = &xchan->xr[xchan->xr_head];
		xr->user = user;
		kanUnsubmitted Done Inline Actions You did not create tag either. kan: You did not create tag either.
		xr->direction = dir;
		xr->m = NULL;
		xr->len = len;
		kanUnsubmitted Done Inline Actions Same here? Why ()'? kan: Same here? Why ()'?
		xr->type = 0;
		xr->src_addr = src;
		xr->dst_addr = dst;
		xr->src_width = src_width;
		xr->dst_width = dst_width;
		xr->done = 0;

		xchan->xr_head = xchan_next_req(xchan, xchan->xr_head);
		atomic_add_int(&xchan->xr_count, 1);

		return (0);
		}

		struct seg_load_request {
		struct bus_dma_segment *seg;
		uint32_t nsegs;
		uint32_t error;
		};

		static void
		xdma_dmamap_cb(void arg, bus_dma_segment_t segs, int nsegs, int error)
		{
		struct seg_load_request *slr;
		struct bus_dma_segment *seg;
		int i;

		slr = arg;
		seg = slr->seg;

		if (error != 0) {
		slr->error = error;
		return;
		}

		slr->nsegs = nsegs;

		for (i = 0; i < nsegs; i++) {
		seg[i].ds_addr = segs[i].ds_addr;
		seg[i].ds_len = segs[i].ds_len;
		}
		}

		static int
		xdma_load_busdma(xdma_channel_t xchan, struct xdma_request xr,
		struct bus_dma_segment *seg, uint32_t i)
		{
		xdma_controller_t *xdma;
		struct seg_load_request slr;
		uint32_t nsegs;
		void *addr;
		int error;

		xdma = xchan->xdma;

		error = 0;
		nsegs = 0;

		switch (xr->type) {
		case XR_TYPE_MBUF:
		error = bus_dmamap_load_mbuf_sg(xchan->dma_tag_bufs,
		xchan->bufs[i].map, xr->m, seg, &nsegs, BUS_DMA_NOWAIT);
		break;
		case XR_TYPE_BIO:
		slr.nsegs = 0;
		slr.error = 0;
		slr.seg = seg;
		error = bus_dmamap_load_bio(xchan->dma_tag_bufs,
		xchan->bufs[i].map, xr->bp, xdma_dmamap_cb, &slr, BUS_DMA_NOWAIT);
		if (slr.error != 0) {
		device_printf(xdma->dma_dev,
		"%s: bus_dmamap_load failed, err %d\n",
		__func__, slr.error);
		return (0);
		}
		nsegs = slr.nsegs;
		break;
		case XR_TYPE_ADDR:
		switch (xr->direction) {
		case XDMA_MEM_TO_DEV:
		addr = (void *)xr->src_addr;
		break;
		case XDMA_DEV_TO_MEM:
		addr = (void *)xr->dst_addr;
		break;
		default:
		device_printf(xdma->dma_dev,
		"%s: Direction is not supported\n", __func__);
		return (0);
		}
		slr.nsegs = 0;
		slr.error = 0;
		slr.seg = seg;
		error = bus_dmamap_load(xchan->dma_tag_bufs, xchan->bufs[i].map,
		addr, xr->len, xdma_dmamap_cb, &slr, BUS_DMA_NOWAIT);
		if (slr.error != 0) {
		device_printf(xdma->dma_dev,
		"%s: bus_dmamap_load failed, err %d\n",
		__func__, slr.error);
		return (0);
		}
		nsegs = slr.nsegs;
		break;
		default:
		break;
		}

		if (error != 0) {
		if (error == ENOMEM) {
		/*
		* Out of memory. Try again later.
		* TODO: count errors.
		*/
		} else {
		device_printf(xdma->dma_dev,
		"%s: bus_dmamap_load failed with err %d\n",
		__func__, error);
		}
		return (0);
		}

		if (xr->direction == XDMA_MEM_TO_DEV) {
		bus_dmamap_sync(xchan->dma_tag_bufs, xchan->bufs[i].map,
		BUS_DMASYNC_PREWRITE);
		} else {
		bus_dmamap_sync(xchan->dma_tag_bufs, xchan->bufs[i].map,
		BUS_DMASYNC_PREREAD);
		}

		return (nsegs);
		}

		static int
		xdma_load_no_busdma(xdma_channel_t xchan, struct xdma_request xr,
		struct bus_dma_segment *seg, uint32_t i)
		{
		xdma_controller_t *xdma;
		struct mbuf *m;
		uint32_t nsegs;

		xdma = xchan->xdma;

		m = xr->m;

		nsegs = 1;

		switch (xr->type) {
		case XR_TYPE_MBUF:
		if (xr->direction == XDMA_MEM_TO_DEV) {
		m_copydata(m, 0, m->m_pkthdr.len, xchan->bufs[i].cbuf);
		seg[0].ds_addr = (bus_addr_t)xchan->bufs[i].cbuf;
		seg[0].ds_len = m->m_pkthdr.len;
		} else {
		seg[0].ds_addr = mtod(m, bus_addr_t);
		seg[0].ds_len = m->m_pkthdr.len;
		}
		break;
		case XR_TYPE_BIO:
		case XR_TYPE_ADDR:
		default:
		panic("implement me\n");
		}

		return (nsegs);
		}

		static int
		xdma_sglist_prepare_one(xdma_channel_t *xchan,
		struct xdma_request xr, struct bus_dma_segment seg)
		{
		xdma_controller_t *xdma;
		int error;
		int nsegs;
		int i;

		xdma = xchan->xdma;

		error = 0;
		nsegs = 0;

		i = xchan->buf_head;

		if (xchan->caps & XCHAN_CAP_BUSDMA) {
		nsegs = xdma_load_busdma(xchan, xr, seg, i);
		} else {
		nsegs = xdma_load_no_busdma(xchan, xr, seg, i);
		}
		if (nsegs == 0) {
		/* Try again later. */
		return (0);
		}

		xchan->bufs[i].xr = xr;
		xchan->bufs[i].nsegs = nsegs;
		xchan->bufs[i].nsegs_left = nsegs;

		xchan->buf_head = xchan_next_buf(xchan, xchan->buf_head);

		return (nsegs);
		}

		static int
		xdma_sglist_prepare(xdma_channel_t *xchan,
		struct xdma_sglist *sg)
		{
		struct bus_dma_segment seg[128];
		struct xdma_request *xr;
		xdma_controller_t *xdma;
		uint32_t capacity;
		uint32_t n;
		uint32_t c;
		int nsegs;
int ret;		int ret;

xdma = xchan->xdma;		xdma = xchan->xdma;

		n = 0;

		ret = XDMA_CHANNEL_CAPACITY(xdma->dma_dev, xchan, &capacity);
		if (ret != 0) {
		device_printf(xdma->dev,
		"%s: Can't get DMA controller capacity.\n", __func__);
		return (-1);
		}

		for (;;) {
		if (xchan->xr_processed == xchan->xr_head) {
		/* All the requests processed. */
		break;
		}
		xr = &xchan->xr[xchan->xr_processed];

		switch (xr->type) {
		case XR_TYPE_MBUF:
		c = xdma_mbuf_defrag(xchan, xr);
		break;
		case XR_TYPE_BIO:
		case XR_TYPE_ADDR:
		default:
		c = 1;
		}

		if (capacity <= (c + n)) {
		/*
		* No space yet available for the entire
		* request in the DMA engine.
		*/
		break;
		}

		nsegs = xdma_sglist_prepare_one(xchan, xr, seg);
		if (nsegs == 0) {
		break;
		}

		xdma_sglist_add(&sg[n], seg, nsegs, xr);
		n += nsegs;

		xchan->xr_processed = xchan_next_req(xchan, xchan->xr_processed);
		}

		return (n);
		}

		int
		xdma_queue_submit(xdma_channel_t *xchan)
		{
		struct xdma_sglist *sg;
		xdma_controller_t *xdma;
		uint32_t sg_n;
		int ret;

		xdma = xchan->xdma;
KASSERT(xdma != NULL, ("xdma is NULL"));		KASSERT(xdma != NULL, ("xdma is NULL"));

conf = &xchan->conf;		sg = xchan->sg;
conf->direction = XDMA_MEM_TO_MEM;
conf->src_addr = src_addr;
conf->dst_addr = dst_addr;
conf->block_len = len;
conf->block_num = 1;

xchan->flags \|= (XCHAN_CONFIGURED \| XCHAN_TYPE_MEMCPY);		if ((xchan->flags & XCHAN_BUFS_ALLOCATED) == 0) {
		device_printf(xdma->dev,
		"%s: Can't submit SG transfer: no bufs\n",
		__func__);
		return (-1);
		}

XCHAN_LOCK(xchan);		XCHAN_LOCK(xchan);

/* Deallocate old descriptors, if any. */		sg_n = xdma_sglist_prepare(xchan, sg);
xdma_desc_free(xchan);		if (sg_n == 0) {
		/* Nothing to submit */
		XCHAN_UNLOCK(xchan);
		return (0);
		}

ret = XDMA_CHANNEL_PREP_MEMCPY(xdma->dma_dev, xchan);		/* Now submit xdma_sglist to DMA engine driver. */

		ret = XDMA_CHANNEL_SUBMIT_SG(xdma->dma_dev, xchan, sg, sg_n);
if (ret != 0) {		if (ret != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't prepare memcpy transfer.\n", __func__);		"%s: Can't submit SG transfer.\n", __func__);

XCHAN_UNLOCK(xchan);		XCHAN_UNLOCK(xchan);

return (-1);		return (-1);
}		}

if (xchan->flags & XCHAN_DESC_ALLOCATED) {
/* Driver created xDMA descriptors. */
bus_dmamap_sync(xchan->dma_tag, xchan->dma_map,
BUS_DMASYNC_POSTWRITE);
}

XCHAN_UNLOCK(xchan);		XCHAN_UNLOCK(xchan);

return (0);		return (0);
}		}

int		int
xdma_prep_cyclic(xdma_channel_t *xchan, enum xdma_direction dir,		xdma_prep_cyclic(xdma_channel_t *xchan, enum xdma_direction dir,
uintptr_t src_addr, uintptr_t dst_addr, int block_len,		uintptr_t src_addr, uintptr_t dst_addr, int block_len,
Show All 14 Lines	xdma_prep_cyclic(xdma_channel_t *xchan, enum xdma_direction dir,
conf->block_num = block_num;		conf->block_num = block_num;
conf->src_width = src_width;		conf->src_width = src_width;
conf->dst_width = dst_width;		conf->dst_width = dst_width;

xchan->flags \|= (XCHAN_CONFIGURED \| XCHAN_TYPE_CYCLIC);		xchan->flags \|= (XCHAN_CONFIGURED \| XCHAN_TYPE_CYCLIC);

XCHAN_LOCK(xchan);		XCHAN_LOCK(xchan);

/* Deallocate old descriptors, if any. */		/* Deallocate bufs, if any. */
xdma_desc_free(xchan);		xchan_bufs_free(xchan);

ret = XDMA_CHANNEL_PREP_CYCLIC(xdma->dma_dev, xchan);		ret = XDMA_CHANNEL_PREP_CYCLIC(xdma->dma_dev, xchan);
if (ret != 0) {		if (ret != 0) {
		kanUnsubmitted Done Inline Actions By using circular buffer here you encode the requirement that dma requests should be completed by the controller in the order they were submitted, but that is not necessarily the case. I guess same applies by sglists. kan: By using circular buffer here you encode the requirement that dma requests should be completed…
		brAuthorUnsubmitted Not Done Inline Actions I switched back to linked lists. Works a bit slower due to locking, but easier to improve /add features in future. thanks! br: I switched back to linked lists. Works a bit slower due to locking, but easier to improve /add…
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't prepare cyclic transfer.\n", __func__);		"%s: Can't prepare cyclic transfer.\n", __func__);
XCHAN_UNLOCK(xchan);		XCHAN_UNLOCK(xchan);

return (-1);		return (-1);
}		}

if (xchan->flags & XCHAN_DESC_ALLOCATED) {
/* Driver has created xDMA descriptors. */
bus_dmamap_sync(xchan->dma_tag, xchan->dma_map,
BUS_DMASYNC_POSTWRITE);
}

XCHAN_UNLOCK(xchan);		XCHAN_UNLOCK(xchan);

return (0);		return (0);
}		}

int		int
		kanUnsubmitted Done Inline Actions ()'s again kan: ()'s again
xdma_begin(xdma_channel_t *xchan)		xdma_begin(xdma_channel_t *xchan)
{		{
xdma_controller_t *xdma;		xdma_controller_t *xdma;
int ret;		int ret;

xdma = xchan->xdma;		xdma = xchan->xdma;

		if (xchan->flags & XCHAN_TYPE_SG) {
		/* Not valid. */
		return (0);
		};

ret = XDMA_CHANNEL_CONTROL(xdma->dma_dev, xchan, XDMA_CMD_BEGIN);		ret = XDMA_CHANNEL_CONTROL(xdma->dma_dev, xchan, XDMA_CMD_BEGIN);
if (ret != 0) {		if (ret != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't begin the channel operation.\n", __func__);		"%s: Can't begin the channel operation.\n", __func__);
return (-1);		return (-1);
}		}

return (0);		return (0);
Show All 30 Lines	if (ret != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't pause the channel operation.\n", __func__);		"%s: Can't pause the channel operation.\n", __func__);
return (-1);		return (-1);
}		}

return (ret);		return (ret);
}		}

int		int
		kanUnsubmitted Not Done Inline Actions Is there a direct relationship of number of mbuf fragments and number of physical addresses it can result in? kan: Is there a direct relationship of number of mbuf fragments and number of physical addresses it…
xdma_callback(xdma_channel_t *xchan)		xchan_seg_done(xdma_channel_t *xchan,
		struct xdma_transfer_status *st)
{		{
struct xdma_intr_handler *ih_tmp;		struct xdma_request *xr;
struct xdma_intr_handler *ih;		xdma_controller_t *xdma;
		xchan_buf_t *b;

TAILQ_FOREACH_SAFE(ih, &xchan->ie_handlers, ih_next, ih_tmp) {		xdma = xchan->xdma;
if (ih->cb != NULL) {
ih->cb(ih->cb_user);		b = &xchan->bufs[xchan->buf_tail];
		xr = b->xr;

		atomic_subtract_int(&b->nsegs_left, 1);

		if (b->nsegs_left == 0) {
		kanUnsubmitted Not Done Inline Actions What happens if segment in the middle completes with the error? Also, what is st_trasferred and in what units it is specified? You invoke callback in per-desc (WHY? DMA driver is not allowed to do desc chaining?), but then only report status on the last desc up the consumer chain. kan: What happens if segment in the middle completes with the error? Also, what is st_trasferred and…
		if (xchan->caps & XCHAN_CAP_BUSDMA) {
		if (xr->direction == XDMA_MEM_TO_DEV) {
		bus_dmamap_sync(xchan->dma_tag_bufs, b->map,
		BUS_DMASYNC_POSTWRITE);
		} else {
		bus_dmamap_sync(xchan->dma_tag_bufs, b->map,
		BUS_DMASYNC_POSTREAD);
}		}

		bus_dmamap_unload(xchan->dma_tag_bufs, b->map);
}		}
		xr->status.error = st->error;
		xr->status.transferred = st->transferred;
		xr->done = 1;

		xchan->buf_tail = xchan_next_buf(xchan, xchan->buf_tail);
		}

return (0);		return (0);
}		}

void		int
xdma_assert_locked(void)		xdma_callback(xdma_channel_t xchan, xdma_transfer_status_t status)
{		{
		struct xdma_intr_handler *ih_tmp;
		struct xdma_intr_handler *ih;
		xdma_controller_t *xdma;

XDMA_ASSERT_LOCKED();		xdma = xchan->xdma;

		TAILQ_FOREACH_SAFE(ih, &xchan->ie_handlers, ih_next, ih_tmp) {
		if (ih->cb != NULL) {
		ih->cb(ih->cb_user, status);
}		}
		}

		wakeup(xdma);

		return (0);
		}

#ifdef FDT		#ifdef FDT
/*		/*
* Notify the DMA driver we have machine-dependent data in FDT.		* Notify the DMA driver we have machine-dependent data in FDT.
*/		*/
static int		static int
xdma_ofw_md_data(xdma_controller_t xdma, pcell_t cells, int ncells)		xdma_ofw_md_data(xdma_controller_t xdma, pcell_t cells, int ncells)
{		{
uint32_t ret;		uint32_t ret;

ret = XDMA_OFW_MD_DATA(xdma->dma_dev, cells, ncells, (void **)&xdma->data);		ret = XDMA_OFW_MD_DATA(xdma->dma_dev,
		cells, ncells, (void **)&xdma->data);

return (ret);		return (ret);
}		}

/*		/*
* Allocate xdma controller.		* Allocate xdma controller.
*/		*/
xdma_controller_t *		xdma_controller_t *
▲ Show 20 Lines • Show All 59 Lines • ▼ Show 20 Lines	xdma_ofw_get(device_t dev, const char *prop)
xdma->dev = dev;		xdma->dev = dev;
xdma->dma_dev = dma_dev;		xdma->dma_dev = dma_dev;

TAILQ_INIT(&xdma->channels);		TAILQ_INIT(&xdma->channels);

xdma_ofw_md_data(xdma, cells, ncells);		xdma_ofw_md_data(xdma, cells, ncells);
free(cells, M_OFWPROP);		free(cells, M_OFWPROP);

		mtx_init(&xdma->proc_mtx, "xDMA ofw controller", NULL, MTX_DEF);
		kproc_create(&xdma_task, xdma, &xdma->xdma_proc, 0, 0, "xdma drainer");

return (xdma);		return (xdma);
}		}
#endif		#endif

/*		/*
* Free xDMA controller object.		* Free xDMA controller object.
*/		*/
int		int
xdma_put(xdma_controller_t *xdma)		xdma_put(xdma_controller_t *xdma)
{		{

XDMA_LOCK();		XDMA_LOCK();

/* Ensure no channels allocated. */		/* Ensure no channels allocated. */
if (!TAILQ_EMPTY(&xdma->channels)) {		if (!TAILQ_EMPTY(&xdma->channels)) {
device_printf(xdma->dev, "%s: Can't free xDMA\n", __func__);		device_printf(xdma->dev, "%s: Can't free xDMA\n", __func__);
return (-1);		return (-1);
}		}

		kproc_shutdown(&xdma->xdma_proc, 0);
		mtx_destroy(&xdma->proc_mtx);

free(xdma->data, M_DEVBUF);		free(xdma->data, M_DEVBUF);
free(xdma, M_XDMA);		free(xdma, M_XDMA);

XDMA_UNLOCK();		XDMA_UNLOCK();

return (0);		return (0);
}		}
Show All 9 Lines

View Options

sys/compat/freebsd32/syscalls.master

/*-		/*-
* Copyright (c) 2016 Ruslan Bukin <br@bsdpad.com>		* Copyright (c) 2016-2017 Ruslan Bukin <br@bsdpad.com>
* All rights reserved.		* All rights reserved.
*		*
* This software was developed by SRI International and the University of		* This software was developed by SRI International and the University of
* Cambridge Computer Laboratory under DARPA/AFRL contract FA8750-10-C-0237		* Cambridge Computer Laboratory under DARPA/AFRL contract FA8750-10-C-0237
* ("CTSRD"), as part of the DARPA CRASH research programme.		* ("CTSRD"), as part of the DARPA CRASH research programme.
*		*
* Redistribution and use in source and binary forms, with or without		* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions		* modification, are permitted provided that the following conditions
Show All 20 Lines
#include <sys/cdefs.h>		#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");		__FBSDID("$FreeBSD$");

#include "opt_platform.h"		#include "opt_platform.h"
#include <sys/param.h>		#include <sys/param.h>
#include <sys/conf.h>		#include <sys/conf.h>
#include <sys/bus.h>		#include <sys/bus.h>
#include <sys/kernel.h>		#include <sys/kernel.h>
		#include <sys/kthread.h>
#include <sys/queue.h>		#include <sys/queue.h>
#include <sys/kobj.h>		#include <sys/kobj.h>
#include <sys/malloc.h>		#include <sys/malloc.h>
#include <sys/mutex.h>		#include <sys/mutex.h>
		#include <sys/mbuf.h>
#include <sys/limits.h>		#include <sys/limits.h>
#include <sys/lock.h>		#include <sys/lock.h>
#include <sys/sysctl.h>		#include <sys/sysctl.h>
#include <sys/systm.h>		#include <sys/systm.h>
#include <sys/sx.h>		#include <sys/sx.h>
#include <sys/bus_dma.h>		#include <sys/bus_dma.h>

#include <machine/bus.h>		#include <machine/bus.h>

#ifdef FDT		#ifdef FDT
#include <dev/fdt/fdt_common.h>		#include <dev/fdt/fdt_common.h>
#include <dev/ofw/ofw_bus.h>		#include <dev/ofw/ofw_bus.h>
#include <dev/ofw/ofw_bus_subr.h>		#include <dev/ofw/ofw_bus_subr.h>
#endif		#endif

#include <dev/xdma/xdma.h>		#include <dev/xdma/xdma.h>

#include <xdma_if.h>		#include <xdma_if.h>

MALLOC_DEFINE(M_XDMA, "xdma", "xDMA framework");

/*		/*
* Multiple xDMA controllers may work with single DMA device,		* Multiple xDMA controllers may work with single DMA device,
* so we have global lock for physical channel management.		* so we have global lock for physical channel management.
*/		*/
static struct mtx xdma_mtx;		static struct mtx xdma_mtx;

#define XDMA_LOCK() mtx_lock(&xdma_mtx)		#define XDMA_LOCK() mtx_lock(&xdma_mtx)
#define XDMA_UNLOCK() mtx_unlock(&xdma_mtx)		#define XDMA_UNLOCK() mtx_unlock(&xdma_mtx)
#define XDMA_ASSERT_LOCKED() mtx_assert(&xdma_mtx, MA_OWNED)		#define XDMA_ASSERT_LOCKED() mtx_assert(&xdma_mtx, MA_OWNED)

/*		/*
* Per channel locks.		* Per channel locks.
*/		*/
#define XCHAN_LOCK(xchan) mtx_lock(&(xchan)->mtx_lock)		#define XCHAN_LOCK(xchan) mtx_lock(&(xchan)->mtx_lock)
#define XCHAN_UNLOCK(xchan) mtx_unlock(&(xchan)->mtx_lock)		#define XCHAN_UNLOCK(xchan) mtx_unlock(&(xchan)->mtx_lock)
#define XCHAN_ASSERT_LOCKED(xchan) mtx_assert(&(xchan)->mtx_lock, MA_OWNED)		#define XCHAN_ASSERT_LOCKED(xchan) mtx_assert(&(xchan)->mtx_lock, MA_OWNED)

		static int xchan_bufs_alloc(xdma_channel_t *xchan);
		static int xchan_bufs_free(xdma_channel_t *xchan);

		static void
		xdma_task(void *arg)
		{
		xdma_controller_t *xdma;
		xdma_channel_t *xchan_tmp;
		xdma_channel_t *xchan;

		xdma = arg;

		for (;;) {
		mtx_lock(&xdma->proc_mtx);
		msleep(xdma, &xdma->proc_mtx, PRIBIO, "jobqueue", hz);
		mtx_unlock(&xdma->proc_mtx);

		if (TAILQ_EMPTY(&xdma->channels)) {
		continue;
		}

		TAILQ_FOREACH_SAFE(xchan, &xdma->channels, xchan_next, xchan_tmp) {
		if (xchan->flags & XCHAN_TYPE_SG) {
		xdma_queue_submit(xchan);
		}
		}
		}
		}

/*		/*
* Allocate virtual xDMA channel.		* Allocate virtual xDMA channel.
*/		*/
xdma_channel_t *		xdma_channel_t *
xdma_channel_alloc(xdma_controller_t *xdma)		xdma_channel_alloc(xdma_controller_t *xdma, uint32_t caps)
{		{
xdma_channel_t *xchan;		xdma_channel_t *xchan;
int ret;		int ret;

xchan = malloc(sizeof(xdma_channel_t), M_XDMA, M_WAITOK \| M_ZERO);		xchan = malloc(sizeof(xdma_channel_t), M_XDMA, M_WAITOK \| M_ZERO);
if (xchan == NULL) {		if (xchan == NULL) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't allocate memory for channel.\n", __func__);		"%s: Can't allocate memory for channel.\n", __func__);
return (NULL);		return (NULL);
}		}
xchan->xdma = xdma;		xchan->xdma = xdma;
		xchan->caps = caps;

XDMA_LOCK();		XDMA_LOCK();

/* Request a real channel from hardware driver. */		/* Request a real channel from hardware driver. */
ret = XDMA_CHANNEL_ALLOC(xdma->dma_dev, xchan);		ret = XDMA_CHANNEL_ALLOC(xdma->dma_dev, xchan);
if (ret != 0) {		if (ret != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't request hardware channel.\n", __func__);		"%s: Can't request hardware channel.\n", __func__);
Show All 29 Lines	if (err != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't free real hw channel.\n", __func__);		"%s: Can't free real hw channel.\n", __func__);
XDMA_UNLOCK();		XDMA_UNLOCK();
return (-1);		return (-1);
}		}

xdma_teardown_all_intr(xchan);		xdma_teardown_all_intr(xchan);

/* Deallocate descriptors, if any. */		/* Deallocate bufs, if any. */
xdma_desc_free(xchan);		xchan_bufs_free(xchan);
		xchan_sglist_free(xchan);

		if (xchan->flags & XCHAN_TYPE_SG) {
		free(xchan->xr, M_XDMA);
		}

mtx_destroy(&xchan->mtx_lock);		mtx_destroy(&xchan->mtx_lock);

TAILQ_REMOVE(&xdma->channels, xchan, xchan_next);		TAILQ_REMOVE(&xdma->channels, xchan, xchan_next);

free(xchan, M_XDMA);		free(xchan, M_XDMA);

XDMA_UNLOCK();		XDMA_UNLOCK();

return (0);		return (0);
}		}

int		int
xdma_setup_intr(xdma_channel_t xchan, int (cb)(void ), void arg,		xdma_setup_intr(xdma_channel_t *xchan,
void **ihandler)		int (cb)(void , xdma_transfer_status_t *),
		void arg, void *ihandler)
{		{
struct xdma_intr_handler *ih;		struct xdma_intr_handler *ih;
xdma_controller_t *xdma;		xdma_controller_t *xdma;

xdma = xchan->xdma;		xdma = xchan->xdma;
KASSERT(xdma != NULL, ("xdma is NULL"));		KASSERT(xdma != NULL, ("xdma is NULL"));

/* Sanity check. */		/* Sanity check. */
▲ Show 20 Lines • Show All 61 Lines • ▼ Show 20 Lines	xdma_teardown_all_intr(xdma_channel_t *xchan)
TAILQ_FOREACH_SAFE(ih, &xchan->ie_handlers, ih_next, ih_tmp) {		TAILQ_FOREACH_SAFE(ih, &xchan->ie_handlers, ih_next, ih_tmp) {
TAILQ_REMOVE(&xchan->ie_handlers, ih, ih_next);		TAILQ_REMOVE(&xchan->ie_handlers, ih, ih_next);
free(ih, M_XDMA);		free(ih, M_XDMA);
}		}

return (0);		return (0);
}		}

static void		static int
xdma_dmamap_cb(void arg, bus_dma_segment_t segs, int nseg, int err)		xdma_bufs_alloc_no_busdma(xdma_channel_t *xchan)
		kanUnsubmitted Done Inline Actions This is a xdma_desc_dmamap_cb in , the name is too short and nondescript. I do not think descriptor allocation belongs in xdma at all, but this is wrong nonetheless. kan: This is a xdma_desc_dmamap_cb in , the name is too short and nondescript. I do not think…
		brAuthorUnsubmitted Not Done Inline Actions I moved descriptors allocation to DMA engine drivers. Thanks br: I moved descriptors allocation to DMA engine drivers. Thanks
{		{
xdma_channel_t *xchan;		xdma_controller_t *xdma;
		int nsegments;
int i;		int i;

xchan = (xdma_channel_t *)arg;		xdma = xchan->xdma;
KASSERT(xchan != NULL, ("xchan is NULL"));

if (err) {		nsegments = xchan->bufs_num;
xchan->map_err = 1;
return;		xchan->bufs = malloc(nsegments * sizeof(struct xchan_buf),
		M_XDMA, (M_WAITOK \| M_ZERO));
		if (xchan->bufs == NULL) {
		device_printf(xdma->dev,
		"%s: Can't allocate memory.\n", __func__);
		return (-1);
}		}

for (i = 0; i < nseg; i++) {		for (i = 0; i < nsegments; i++) {
xchan->descs_phys[i].ds_addr = segs[i].ds_addr;		xchan->bufs[i].cbuf = contigmalloc(xchan->maxsegsize,
xchan->descs_phys[i].ds_len = segs[i].ds_len;		M_XDMA, 0, 0, ~0, PAGE_SIZE, 0);
}		}

		return (0);
}		}

static int		static int
xdma_desc_alloc_bus_dma(xdma_channel_t *xchan, uint32_t desc_size,		xdma_bufs_alloc_busdma(xdma_channel_t *xchan)
uint32_t align)
{		{
xdma_controller_t *xdma;		xdma_controller_t *xdma;
bus_size_t all_desc_sz;
xdma_config_t *conf;
int nsegments;		int nsegments;
int err;		int err;
		int i;

xdma = xchan->xdma;		xdma = xchan->xdma;
conf = &xchan->conf;

nsegments = conf->block_num;		nsegments = xchan->bufs_num;
all_desc_sz = (nsegments * desc_size);

		xchan->bufs = malloc(nsegments * sizeof(struct xchan_buf),
		M_XDMA, (M_WAITOK \| M_ZERO));
		if (xchan->bufs == NULL) {
		device_printf(xdma->dev,
		"%s: Can't allocate memory.\n", __func__);
		return (-1);
		}

err = bus_dma_tag_create(		err = bus_dma_tag_create(
bus_get_dma_tag(xdma->dev),		bus_get_dma_tag(xdma->dev), /* Parent tag. */
align, desc_size, /* alignment, boundary */		xchan->alignment, 0, /* alignment, boundary */
BUS_SPACE_MAXADDR_32BIT, /* lowaddr */		BUS_SPACE_MAXADDR_32BIT, /* lowaddr */
		kanUnsubmitted Done Inline Actions Why? How xchan knows boundaries handed by the underlying DMA device? How does it even know if bus_get_dma_tag(xdma->dev) is even a right tag to be used for descriptors? Descriptors often have much stricter requirements on alignment and locations of descriptors and just 'uint32_t desc_size, uint32_t align' cannot possibly describe them. The operation does not belong in generic xchan layer at all and is better left to an actual pdma implementation. kan: Why? How xchan knows boundaries handed by the underlying DMA device? How does it even know if…
		brAuthorUnsubmitted Not Done Inline Actions Sounds right, thanks. I moved all the descriptors allocation/handling to DMA engine driver. br: Sounds right, thanks. I moved all the descriptors allocation/handling to DMA engine driver.
BUS_SPACE_MAXADDR, /* highaddr */		BUS_SPACE_MAXADDR, /* highaddr */
NULL, NULL, /* filter, filterarg */		NULL, NULL, /* filter, filterarg */
all_desc_sz, nsegments, /* maxsize, nsegments*/		xchan->maxsegsize * xchan->maxnsegs, /* maxsize */
desc_size, 0, /* maxsegsize, flags */		xchan->maxnsegs, /* nsegments */
		xchan->maxsegsize, /* maxsegsize */
		0, /* flags */
NULL, NULL, /* lockfunc, lockarg */		NULL, NULL, /* lockfunc, lockarg */
&xchan->dma_tag);		&xchan->dma_tag_bufs);
if (err) {		if (err != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't create bus_dma tag.\n", __func__);		"%s: Can't create bus_dma tag.\n", __func__);
return (-1);		return (-1);
}		}

err = bus_dmamem_alloc(xchan->dma_tag, (void **)&xchan->descs,		for (i = 0; i < nsegments; i++) {
BUS_DMA_WAITOK \| BUS_DMA_COHERENT, &xchan->dma_map);		err = bus_dmamap_create(xchan->dma_tag_bufs, 0,
if (err) {		&xchan->bufs[i].map);
		if (err != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't allocate memory for descriptors.\n", __func__);		"%s: Can't create buf DMA map.\n", __func__);
return (-1);		return (-1);
}		}
		}

xchan->descs_phys = malloc(nsegments * sizeof(xdma_descriptor_t), M_XDMA,		return (0);
(M_WAITOK \| M_ZERO));		}

xchan->map_err = 0;		static int
err = bus_dmamap_load(xchan->dma_tag, xchan->dma_map, xchan->descs,		xchan_bufs_alloc(xdma_channel_t *xchan)
all_desc_sz, xdma_dmamap_cb, xchan, BUS_DMA_WAITOK);		{
if (err) {		xdma_controller_t *xdma;
		int ret;

		xdma = xchan->xdma;
		if (xdma == NULL) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't load DMA map.\n", __func__);		"%s: Channel was not allocated properly.\n", __func__);
return (-1);		return (-1);
}		}

if (xchan->map_err != 0) {		if (xchan->caps & XCHAN_CAP_BUSDMA) {
		ret = xdma_bufs_alloc_busdma(xchan);
		} else {
		ret = xdma_bufs_alloc_no_busdma(xchan);
		}
		kanUnsubmitted Not Done Inline Actions Error handling, in particular, freeing already allocated memory and descriptors and unmapping mapped descriptors is considered futile? kan: Error handling, in particular, freeing already allocated memory and descriptors and unmapping…
		if (ret != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't load DMA map.\n", __func__);		"%s: Can't setup busdma.\n",
		__func__);
return (-1);		return (-1);
}		}

		xchan->flags \|= XCHAN_BUFS_ALLOCATED;

return (0);		return (0);
}		}

/*		static int
* This function called by DMA controller driver.		xchan_bufs_free(xdma_channel_t *xchan)
*/		{
		xchan_buf_t *b;
		kanUnsubmitted Done Inline Actions The name of the function and later use of MCLBYTES do correspond to each other rather weakly. It if an mbuf-specific functionality, name better should reflect that. kan: The name of the function and later use of MCLBYTES do correspond to each other rather weakly.
		int i;
		kanUnsubmitted Done Inline Actions Same thing. This seems to be very mbuf specific and ridden with arbitrary limitations. kan: Same thing. This seems to be very mbuf specific and ridden with arbitrary limitations.
		brAuthorUnsubmitted Not Done Inline Actions tag parameters now specified by transfer preparation functions, e.g. xdma_prep_sg br: tag parameters now specified by transfer preparation functions, e.g. xdma_prep_sg

		if ((xchan->flags & XCHAN_BUFS_ALLOCATED) == 0) {
		/* No bufs allocated. */
		return (-1);
		}

		if (xchan->caps & XCHAN_CAP_BUSDMA) {
		for (i = 0; i < xchan->bufs_num; i++) {
		b = &xchan->bufs[i];
		bus_dmamap_destroy(xchan->dma_tag_bufs, b->map);
		}
		bus_dma_tag_destroy(xchan->dma_tag_bufs);
		} else {
		for (i = 0; i < xchan->bufs_num; i++) {
		contigfree(xchan->bufs[i].cbuf, xchan->maxsegsize, M_XDMA);
		}
		}

		free(xchan->bufs, M_XDMA);
		xchan->flags &= ~XCHAN_BUFS_ALLOCATED;

		return (0);
		}
		kanUnsubmitted Done Inline Actions Again, WHY??? kan: Again, WHY???

int		int
xdma_desc_alloc(xdma_channel_t *xchan, uint32_t desc_size, uint32_t align)		xdma_prep_memcpy(xdma_channel_t *xchan, uintptr_t src_addr,
		uintptr_t dst_addr, size_t len)
{		{
xdma_controller_t *xdma;		xdma_controller_t *xdma;
xdma_config_t *conf;		xdma_config_t *conf;
int ret;		int ret;

XCHAN_ASSERT_LOCKED(xchan);		xdma = xchan->xdma;
		KASSERT(xdma != NULL, ("xdma is NULL"));

		conf = &xchan->conf;
		conf->direction = XDMA_MEM_TO_MEM;
		conf->src_addr = src_addr;
		conf->dst_addr = dst_addr;
		conf->block_len = len;
		conf->block_num = 1;

		xchan->flags \|= (XCHAN_CONFIGURED \| XCHAN_TYPE_MEMCPY);

		XCHAN_LOCK(xchan);

		ret = XDMA_CHANNEL_PREP_MEMCPY(xdma->dma_dev, xchan);
		if (ret != 0) {
		device_printf(xdma->dev,
		"%s: Can't prepare memcpy transfer.\n", __func__);
		XCHAN_UNLOCK(xchan);

		return (-1);
		}

		XCHAN_UNLOCK(xchan);

		kanUnsubmitted Not Done Inline Actions BUS_DMA_COHERENT??? kan: BUS_DMA_COHERENT???
		return (0);
		}

		/*
		* xr_num - xdma requests queue size,
		* maxsegsize - maximum allowed scatter-gather list element size in bytes
		*/
		int
		xdma_prep_sg(xdma_channel_t *xchan, uint32_t xr_num,
		uint32_t maxsegsize, uint32_t alignment)
		{
		xdma_controller_t *xdma;
		int ret;

xdma = xchan->xdma;		xdma = xchan->xdma;
if (xdma == NULL) {
		KASSERT(xdma != NULL, ("xdma is NULL"));

		if (xchan->flags & XCHAN_CONFIGURED) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Channel was not allocated properly.\n", __func__);		"%s: Channel is already configured.\n", __func__);
return (-1);		return (-1);
}		}

if (xchan->flags & XCHAN_DESC_ALLOCATED) {		xchan->maxsegsize = maxsegsize;
		xchan->alignment = alignment;
		xchan->maxnsegs = 8;
		xchan->bufs_num = xr_num;
		xchan->xr_num = xr_num;

		/* Allocate sglist. */
		ret = xchan_sglist_init(xchan);
		if (ret != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Descriptors already allocated.\n", __func__);		"%s: Can't allocate sglist.\n", __func__);
return (-1);		return (-1);
}		}

if ((xchan->flags & XCHAN_CONFIGURED) == 0) {		/* Allocate request queue. */
		xchan->xr = malloc(sizeof(struct xdma_request) * xr_num,
		M_XDMA, M_WAITOK \| M_ZERO);
		if (xchan->xr == NULL) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Channel has no configuration.\n", __func__);		"%s: Can't allocate request queue.\n", __func__);
return (-1);		return (-1);
}		}

conf = &xchan->conf;		xchan->flags \|= (XCHAN_CONFIGURED \| XCHAN_TYPE_SG);

XCHAN_UNLOCK(xchan);
ret = xdma_desc_alloc_bus_dma(xchan, desc_size, align);
XCHAN_LOCK(xchan);		XCHAN_LOCK(xchan);

		/* Deallocate bufs, if any. */
		xchan_bufs_free(xchan);

		ret = XDMA_CHANNEL_PREP_SG(xdma->dma_dev, xchan);
if (ret != 0) {		if (ret != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't allocate memory for descriptors.\n",		"%s: Can't prepare SG transfer.\n", __func__);
__func__);		XCHAN_UNLOCK(xchan);

return (-1);		return (-1);
}		}

xchan->flags \|= XCHAN_DESC_ALLOCATED;		/* Allocate bufs */
		ret = xchan_bufs_alloc(xchan);
		if (ret != 0) {
		device_printf(xdma->dev,
		"%s: Can't allocate bufs.\n", __func__);
		return (-1);
		}

/* We are going to write to descriptors. */		XCHAN_UNLOCK(xchan);
bus_dmamap_sync(xchan->dma_tag, xchan->dma_map, BUS_DMASYNC_PREWRITE);

return (0);		return (0);
}		}

int		int
xdma_desc_free(xdma_channel_t *xchan)		xdma_dequeue(xdma_channel_t xchan, void *user,
		xdma_transfer_status_t *status)
{		{
		struct xdma_request *xr;

if ((xchan->flags & XCHAN_DESC_ALLOCATED) == 0) {		if (xchan->xr_tail == xchan->xr_processed) {
/* No descriptors allocated. */
return (-1);		return (-1);
}		}

bus_dmamap_unload(xchan->dma_tag, xchan->dma_map);		xr = &xchan->xr[xchan->xr_tail];
bus_dmamem_free(xchan->dma_tag, xchan->descs, xchan->dma_map);		if (xr->done == 0) {
bus_dma_tag_destroy(xchan->dma_tag);		return (-1);
free(xchan->descs_phys, M_XDMA);		}

xchan->flags &= ~(XCHAN_DESC_ALLOCATED);		*user = xr->user;
		status->error = xr->status.error;
		status->transferred = xr->status.transferred;
		xchan->xr_tail = xchan_next_req(xchan, xchan->xr_tail);
		atomic_subtract_int(&xchan->xr_count, 1);
		kanUnsubmitted Done Inline Actions You did not protect XCHAN_DESC_ALLOCATED with ()'s above, why now? kan: You did not protect XCHAN_DESC_ALLOCATED with ()'s above, why now?

return (0);		return (0);
}		}

int		int
xdma_prep_memcpy(xdma_channel_t *xchan, uintptr_t src_addr,		xdma_enqueue(xdma_channel_t *xchan, uintptr_t src, uintptr_t dst,
uintptr_t dst_addr, size_t len)		uint8_t src_width, uint8_t dst_width, bus_size_t len,
		enum xdma_direction dir, void *user)
{		{
		struct xdma_request *xr;
xdma_controller_t *xdma;		xdma_controller_t *xdma;
xdma_config_t *conf;
		xdma = xchan->xdma;

		if (xchan->xr_count >= (xchan->xr_num - 1)) {
		/* No space is available yet. */
		kanUnsubmitted Done Inline Actions This is unconditional? You did not call map_create for !XCHAN_CAP_BUSDMA case. kan: This is unconditional? You did not call map_create for !XCHAN_CAP_BUSDMA case.
		brAuthorUnsubmitted Not Done Inline Actions Fixed. thanks! br: Fixed. thanks!
		return (-1);
		}

		xr = &xchan->xr[xchan->xr_head];
		xr->user = user;
		kanUnsubmitted Done Inline Actions You did not create tag either. kan: You did not create tag either.
		xr->direction = dir;
		xr->m = NULL;
		xr->len = len;
		kanUnsubmitted Done Inline Actions Same here? Why ()'? kan: Same here? Why ()'?
		xr->type = 0;
		xr->src_addr = src;
		xr->dst_addr = dst;
		xr->src_width = src_width;
		xr->dst_width = dst_width;
		xr->done = 0;

		xchan->xr_head = xchan_next_req(xchan, xchan->xr_head);
		atomic_add_int(&xchan->xr_count, 1);

		return (0);
		}

		struct seg_load_request {
		struct bus_dma_segment *seg;
		uint32_t nsegs;
		uint32_t error;
		};

		static void
		xdma_dmamap_cb(void arg, bus_dma_segment_t segs, int nsegs, int error)
		{
		struct seg_load_request *slr;
		struct bus_dma_segment *seg;
		int i;

		slr = arg;
		seg = slr->seg;

		if (error != 0) {
		slr->error = error;
		return;
		}

		slr->nsegs = nsegs;

		for (i = 0; i < nsegs; i++) {
		seg[i].ds_addr = segs[i].ds_addr;
		seg[i].ds_len = segs[i].ds_len;
		}
		}

		static int
		xdma_load_busdma(xdma_channel_t xchan, struct xdma_request xr,
		struct bus_dma_segment *seg, uint32_t i)
		{
		xdma_controller_t *xdma;
		struct seg_load_request slr;
		uint32_t nsegs;
		void *addr;
		int error;

		xdma = xchan->xdma;

		error = 0;
		nsegs = 0;

		switch (xr->type) {
		case XR_TYPE_MBUF:
		error = bus_dmamap_load_mbuf_sg(xchan->dma_tag_bufs,
		xchan->bufs[i].map, xr->m, seg, &nsegs, BUS_DMA_NOWAIT);
		break;
		case XR_TYPE_BIO:
		slr.nsegs = 0;
		slr.error = 0;
		slr.seg = seg;
		error = bus_dmamap_load_bio(xchan->dma_tag_bufs,
		xchan->bufs[i].map, xr->bp, xdma_dmamap_cb, &slr, BUS_DMA_NOWAIT);
		if (slr.error != 0) {
		device_printf(xdma->dma_dev,
		"%s: bus_dmamap_load failed, err %d\n",
		__func__, slr.error);
		return (0);
		}
		nsegs = slr.nsegs;
		break;
		case XR_TYPE_ADDR:
		switch (xr->direction) {
		case XDMA_MEM_TO_DEV:
		addr = (void *)xr->src_addr;
		break;
		case XDMA_DEV_TO_MEM:
		addr = (void *)xr->dst_addr;
		break;
		default:
		device_printf(xdma->dma_dev,
		"%s: Direction is not supported\n", __func__);
		return (0);
		}
		slr.nsegs = 0;
		slr.error = 0;
		slr.seg = seg;
		error = bus_dmamap_load(xchan->dma_tag_bufs, xchan->bufs[i].map,
		addr, xr->len, xdma_dmamap_cb, &slr, BUS_DMA_NOWAIT);
		if (slr.error != 0) {
		device_printf(xdma->dma_dev,
		"%s: bus_dmamap_load failed, err %d\n",
		__func__, slr.error);
		return (0);
		}
		nsegs = slr.nsegs;
		break;
		default:
		break;
		}

		if (error != 0) {
		if (error == ENOMEM) {
		/*
		* Out of memory. Try again later.
		* TODO: count errors.
		*/
		} else {
		device_printf(xdma->dma_dev,
		"%s: bus_dmamap_load failed with err %d\n",
		__func__, error);
		}
		return (0);
		}

		if (xr->direction == XDMA_MEM_TO_DEV) {
		bus_dmamap_sync(xchan->dma_tag_bufs, xchan->bufs[i].map,
		BUS_DMASYNC_PREWRITE);
		} else {
		bus_dmamap_sync(xchan->dma_tag_bufs, xchan->bufs[i].map,
		BUS_DMASYNC_PREREAD);
		}

		return (nsegs);
		}

		static int
		xdma_load_no_busdma(xdma_channel_t xchan, struct xdma_request xr,
		struct bus_dma_segment *seg, uint32_t i)
		{
		xdma_controller_t *xdma;
		struct mbuf *m;
		uint32_t nsegs;

		xdma = xchan->xdma;

		m = xr->m;

		nsegs = 1;

		switch (xr->type) {
		case XR_TYPE_MBUF:
		if (xr->direction == XDMA_MEM_TO_DEV) {
		m_copydata(m, 0, m->m_pkthdr.len, xchan->bufs[i].cbuf);
		seg[0].ds_addr = (bus_addr_t)xchan->bufs[i].cbuf;
		seg[0].ds_len = m->m_pkthdr.len;
		} else {
		seg[0].ds_addr = mtod(m, bus_addr_t);
		seg[0].ds_len = m->m_pkthdr.len;
		}
		break;
		case XR_TYPE_BIO:
		case XR_TYPE_ADDR:
		default:
		panic("implement me\n");
		}

		return (nsegs);
		}

		static int
		xdma_sglist_prepare_one(xdma_channel_t *xchan,
		struct xdma_request xr, struct bus_dma_segment seg)
		{
		xdma_controller_t *xdma;
		int error;
		int nsegs;
		int i;

		xdma = xchan->xdma;

		error = 0;
		nsegs = 0;

		i = xchan->buf_head;

		if (xchan->caps & XCHAN_CAP_BUSDMA) {
		nsegs = xdma_load_busdma(xchan, xr, seg, i);
		} else {
		nsegs = xdma_load_no_busdma(xchan, xr, seg, i);
		}
		if (nsegs == 0) {
		/* Try again later. */
		return (0);
		}

		xchan->bufs[i].xr = xr;
		xchan->bufs[i].nsegs = nsegs;
		xchan->bufs[i].nsegs_left = nsegs;

		xchan->buf_head = xchan_next_buf(xchan, xchan->buf_head);

		return (nsegs);
		}

		static int
		xdma_sglist_prepare(xdma_channel_t *xchan,
		struct xdma_sglist *sg)
		{
		struct bus_dma_segment seg[128];
		struct xdma_request *xr;
		xdma_controller_t *xdma;
		uint32_t capacity;
		uint32_t n;
		uint32_t c;
		int nsegs;
int ret;		int ret;

xdma = xchan->xdma;		xdma = xchan->xdma;

		n = 0;

		ret = XDMA_CHANNEL_CAPACITY(xdma->dma_dev, xchan, &capacity);
		if (ret != 0) {
		device_printf(xdma->dev,
		"%s: Can't get DMA controller capacity.\n", __func__);
		return (-1);
		}

		for (;;) {
		if (xchan->xr_processed == xchan->xr_head) {
		/* All the requests processed. */
		break;
		}
		xr = &xchan->xr[xchan->xr_processed];

		switch (xr->type) {
		case XR_TYPE_MBUF:
		c = xdma_mbuf_defrag(xchan, xr);
		break;
		case XR_TYPE_BIO:
		case XR_TYPE_ADDR:
		default:
		c = 1;
		}

		if (capacity <= (c + n)) {
		/*
		* No space yet available for the entire
		* request in the DMA engine.
		*/
		break;
		}

		nsegs = xdma_sglist_prepare_one(xchan, xr, seg);
		if (nsegs == 0) {
		break;
		}

		xdma_sglist_add(&sg[n], seg, nsegs, xr);
		n += nsegs;

		xchan->xr_processed = xchan_next_req(xchan, xchan->xr_processed);
		}

		return (n);
		}

		int
		xdma_queue_submit(xdma_channel_t *xchan)
		{
		struct xdma_sglist *sg;
		xdma_controller_t *xdma;
		uint32_t sg_n;
		int ret;

		xdma = xchan->xdma;
KASSERT(xdma != NULL, ("xdma is NULL"));		KASSERT(xdma != NULL, ("xdma is NULL"));

conf = &xchan->conf;		sg = xchan->sg;
conf->direction = XDMA_MEM_TO_MEM;
conf->src_addr = src_addr;
conf->dst_addr = dst_addr;
conf->block_len = len;
conf->block_num = 1;

xchan->flags \|= (XCHAN_CONFIGURED \| XCHAN_TYPE_MEMCPY);		if ((xchan->flags & XCHAN_BUFS_ALLOCATED) == 0) {
		device_printf(xdma->dev,
		"%s: Can't submit SG transfer: no bufs\n",
		__func__);
		return (-1);
		}

XCHAN_LOCK(xchan);		XCHAN_LOCK(xchan);

/* Deallocate old descriptors, if any. */		sg_n = xdma_sglist_prepare(xchan, sg);
xdma_desc_free(xchan);		if (sg_n == 0) {
		/* Nothing to submit */
		XCHAN_UNLOCK(xchan);
		return (0);
		}

ret = XDMA_CHANNEL_PREP_MEMCPY(xdma->dma_dev, xchan);		/* Now submit xdma_sglist to DMA engine driver. */

		ret = XDMA_CHANNEL_SUBMIT_SG(xdma->dma_dev, xchan, sg, sg_n);
if (ret != 0) {		if (ret != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't prepare memcpy transfer.\n", __func__);		"%s: Can't submit SG transfer.\n", __func__);

XCHAN_UNLOCK(xchan);		XCHAN_UNLOCK(xchan);

return (-1);		return (-1);
}		}

if (xchan->flags & XCHAN_DESC_ALLOCATED) {
/* Driver created xDMA descriptors. */
bus_dmamap_sync(xchan->dma_tag, xchan->dma_map,
BUS_DMASYNC_POSTWRITE);
}

XCHAN_UNLOCK(xchan);		XCHAN_UNLOCK(xchan);

return (0);		return (0);
}		}

int		int
xdma_prep_cyclic(xdma_channel_t *xchan, enum xdma_direction dir,		xdma_prep_cyclic(xdma_channel_t *xchan, enum xdma_direction dir,
uintptr_t src_addr, uintptr_t dst_addr, int block_len,		uintptr_t src_addr, uintptr_t dst_addr, int block_len,
Show All 14 Lines	xdma_prep_cyclic(xdma_channel_t *xchan, enum xdma_direction dir,
conf->block_num = block_num;		conf->block_num = block_num;
conf->src_width = src_width;		conf->src_width = src_width;
conf->dst_width = dst_width;		conf->dst_width = dst_width;

xchan->flags \|= (XCHAN_CONFIGURED \| XCHAN_TYPE_CYCLIC);		xchan->flags \|= (XCHAN_CONFIGURED \| XCHAN_TYPE_CYCLIC);

XCHAN_LOCK(xchan);		XCHAN_LOCK(xchan);

/* Deallocate old descriptors, if any. */		/* Deallocate bufs, if any. */
xdma_desc_free(xchan);		xchan_bufs_free(xchan);

ret = XDMA_CHANNEL_PREP_CYCLIC(xdma->dma_dev, xchan);		ret = XDMA_CHANNEL_PREP_CYCLIC(xdma->dma_dev, xchan);
if (ret != 0) {		if (ret != 0) {
		kanUnsubmitted Done Inline Actions By using circular buffer here you encode the requirement that dma requests should be completed by the controller in the order they were submitted, but that is not necessarily the case. I guess same applies by sglists. kan: By using circular buffer here you encode the requirement that dma requests should be completed…
		brAuthorUnsubmitted Not Done Inline Actions I switched back to linked lists. Works a bit slower due to locking, but easier to improve /add features in future. thanks! br: I switched back to linked lists. Works a bit slower due to locking, but easier to improve /add…
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't prepare cyclic transfer.\n", __func__);		"%s: Can't prepare cyclic transfer.\n", __func__);
XCHAN_UNLOCK(xchan);		XCHAN_UNLOCK(xchan);

return (-1);		return (-1);
}		}

if (xchan->flags & XCHAN_DESC_ALLOCATED) {
/* Driver has created xDMA descriptors. */
bus_dmamap_sync(xchan->dma_tag, xchan->dma_map,
BUS_DMASYNC_POSTWRITE);
}

XCHAN_UNLOCK(xchan);		XCHAN_UNLOCK(xchan);

return (0);		return (0);
}		}

int		int
		kanUnsubmitted Done Inline Actions ()'s again kan: ()'s again
xdma_begin(xdma_channel_t *xchan)		xdma_begin(xdma_channel_t *xchan)
{		{
xdma_controller_t *xdma;		xdma_controller_t *xdma;
int ret;		int ret;

xdma = xchan->xdma;		xdma = xchan->xdma;

		if (xchan->flags & XCHAN_TYPE_SG) {
		/* Not valid. */
		return (0);
		};

ret = XDMA_CHANNEL_CONTROL(xdma->dma_dev, xchan, XDMA_CMD_BEGIN);		ret = XDMA_CHANNEL_CONTROL(xdma->dma_dev, xchan, XDMA_CMD_BEGIN);
if (ret != 0) {		if (ret != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't begin the channel operation.\n", __func__);		"%s: Can't begin the channel operation.\n", __func__);
return (-1);		return (-1);
}		}

return (0);		return (0);
Show All 30 Lines	if (ret != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't pause the channel operation.\n", __func__);		"%s: Can't pause the channel operation.\n", __func__);
return (-1);		return (-1);
}		}

return (ret);		return (ret);
}		}

int		int
		kanUnsubmitted Not Done Inline Actions Is there a direct relationship of number of mbuf fragments and number of physical addresses it can result in? kan: Is there a direct relationship of number of mbuf fragments and number of physical addresses it…
xdma_callback(xdma_channel_t *xchan)		xchan_seg_done(xdma_channel_t *xchan,
		struct xdma_transfer_status *st)
{		{
struct xdma_intr_handler *ih_tmp;		struct xdma_request *xr;
struct xdma_intr_handler *ih;		xdma_controller_t *xdma;
		xchan_buf_t *b;

TAILQ_FOREACH_SAFE(ih, &xchan->ie_handlers, ih_next, ih_tmp) {		xdma = xchan->xdma;
if (ih->cb != NULL) {
ih->cb(ih->cb_user);		b = &xchan->bufs[xchan->buf_tail];
		xr = b->xr;

		atomic_subtract_int(&b->nsegs_left, 1);

		if (b->nsegs_left == 0) {
		kanUnsubmitted Not Done Inline Actions What happens if segment in the middle completes with the error? Also, what is st_trasferred and in what units it is specified? You invoke callback in per-desc (WHY? DMA driver is not allowed to do desc chaining?), but then only report status on the last desc up the consumer chain. kan: What happens if segment in the middle completes with the error? Also, what is st_trasferred and…
		if (xchan->caps & XCHAN_CAP_BUSDMA) {
		if (xr->direction == XDMA_MEM_TO_DEV) {
		bus_dmamap_sync(xchan->dma_tag_bufs, b->map,
		BUS_DMASYNC_POSTWRITE);
		} else {
		bus_dmamap_sync(xchan->dma_tag_bufs, b->map,
		BUS_DMASYNC_POSTREAD);
}		}

		bus_dmamap_unload(xchan->dma_tag_bufs, b->map);
}		}
		xr->status.error = st->error;
		xr->status.transferred = st->transferred;
		xr->done = 1;

		xchan->buf_tail = xchan_next_buf(xchan, xchan->buf_tail);
		}

return (0);		return (0);
}		}

void		int
xdma_assert_locked(void)		xdma_callback(xdma_channel_t xchan, xdma_transfer_status_t status)
{		{
		struct xdma_intr_handler *ih_tmp;
		struct xdma_intr_handler *ih;
		xdma_controller_t *xdma;

XDMA_ASSERT_LOCKED();		xdma = xchan->xdma;

		TAILQ_FOREACH_SAFE(ih, &xchan->ie_handlers, ih_next, ih_tmp) {
		if (ih->cb != NULL) {
		ih->cb(ih->cb_user, status);
}		}
		}

		wakeup(xdma);

		return (0);
		}

#ifdef FDT		#ifdef FDT
/*		/*
* Notify the DMA driver we have machine-dependent data in FDT.		* Notify the DMA driver we have machine-dependent data in FDT.
*/		*/
static int		static int
xdma_ofw_md_data(xdma_controller_t xdma, pcell_t cells, int ncells)		xdma_ofw_md_data(xdma_controller_t xdma, pcell_t cells, int ncells)
{		{
uint32_t ret;		uint32_t ret;

ret = XDMA_OFW_MD_DATA(xdma->dma_dev, cells, ncells, (void **)&xdma->data);		ret = XDMA_OFW_MD_DATA(xdma->dma_dev,
		cells, ncells, (void **)&xdma->data);

return (ret);		return (ret);
}		}

/*		/*
* Allocate xdma controller.		* Allocate xdma controller.
*/		*/
xdma_controller_t *		xdma_controller_t *
▲ Show 20 Lines • Show All 59 Lines • ▼ Show 20 Lines	xdma_ofw_get(device_t dev, const char *prop)
xdma->dev = dev;		xdma->dev = dev;
xdma->dma_dev = dma_dev;		xdma->dma_dev = dma_dev;

TAILQ_INIT(&xdma->channels);		TAILQ_INIT(&xdma->channels);

xdma_ofw_md_data(xdma, cells, ncells);		xdma_ofw_md_data(xdma, cells, ncells);
free(cells, M_OFWPROP);		free(cells, M_OFWPROP);

		mtx_init(&xdma->proc_mtx, "xDMA ofw controller", NULL, MTX_DEF);
		kproc_create(&xdma_task, xdma, &xdma->xdma_proc, 0, 0, "xdma drainer");

return (xdma);		return (xdma);
}		}
#endif		#endif

/*		/*
* Free xDMA controller object.		* Free xDMA controller object.
*/		*/
int		int
xdma_put(xdma_controller_t *xdma)		xdma_put(xdma_controller_t *xdma)
{		{

XDMA_LOCK();		XDMA_LOCK();

/* Ensure no channels allocated. */		/* Ensure no channels allocated. */
if (!TAILQ_EMPTY(&xdma->channels)) {		if (!TAILQ_EMPTY(&xdma->channels)) {
device_printf(xdma->dev, "%s: Can't free xDMA\n", __func__);		device_printf(xdma->dev, "%s: Can't free xDMA\n", __func__);
return (-1);		return (-1);
}		}

		kproc_shutdown(&xdma->xdma_proc, 0);
		mtx_destroy(&xdma->proc_mtx);

free(xdma->data, M_DEVBUF);		free(xdma->data, M_DEVBUF);
free(xdma, M_XDMA);		free(xdma, M_XDMA);

XDMA_UNLOCK();		XDMA_UNLOCK();

return (0);		return (0);
}		}
Show All 9 Lines

View Options

sys/kern/kern_ntptime.c

/*-		/*-
* Copyright (c) 2016 Ruslan Bukin <br@bsdpad.com>		* Copyright (c) 2016-2017 Ruslan Bukin <br@bsdpad.com>
* All rights reserved.		* All rights reserved.
*		*
* This software was developed by SRI International and the University of		* This software was developed by SRI International and the University of
* Cambridge Computer Laboratory under DARPA/AFRL contract FA8750-10-C-0237		* Cambridge Computer Laboratory under DARPA/AFRL contract FA8750-10-C-0237
* ("CTSRD"), as part of the DARPA CRASH research programme.		* ("CTSRD"), as part of the DARPA CRASH research programme.
*		*
* Redistribution and use in source and binary forms, with or without		* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions		* modification, are permitted provided that the following conditions
Show All 20 Lines
#include <sys/cdefs.h>		#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");		__FBSDID("$FreeBSD$");

#include "opt_platform.h"		#include "opt_platform.h"
#include <sys/param.h>		#include <sys/param.h>
#include <sys/conf.h>		#include <sys/conf.h>
#include <sys/bus.h>		#include <sys/bus.h>
#include <sys/kernel.h>		#include <sys/kernel.h>
		#include <sys/kthread.h>
#include <sys/queue.h>		#include <sys/queue.h>
#include <sys/kobj.h>		#include <sys/kobj.h>
#include <sys/malloc.h>		#include <sys/malloc.h>
#include <sys/mutex.h>		#include <sys/mutex.h>
		#include <sys/mbuf.h>
#include <sys/limits.h>		#include <sys/limits.h>
#include <sys/lock.h>		#include <sys/lock.h>
#include <sys/sysctl.h>		#include <sys/sysctl.h>
#include <sys/systm.h>		#include <sys/systm.h>
#include <sys/sx.h>		#include <sys/sx.h>
#include <sys/bus_dma.h>		#include <sys/bus_dma.h>

#include <machine/bus.h>		#include <machine/bus.h>

#ifdef FDT		#ifdef FDT
#include <dev/fdt/fdt_common.h>		#include <dev/fdt/fdt_common.h>
#include <dev/ofw/ofw_bus.h>		#include <dev/ofw/ofw_bus.h>
#include <dev/ofw/ofw_bus_subr.h>		#include <dev/ofw/ofw_bus_subr.h>
#endif		#endif

#include <dev/xdma/xdma.h>		#include <dev/xdma/xdma.h>

#include <xdma_if.h>		#include <xdma_if.h>

MALLOC_DEFINE(M_XDMA, "xdma", "xDMA framework");

/*		/*
* Multiple xDMA controllers may work with single DMA device,		* Multiple xDMA controllers may work with single DMA device,
* so we have global lock for physical channel management.		* so we have global lock for physical channel management.
*/		*/
static struct mtx xdma_mtx;		static struct mtx xdma_mtx;

#define XDMA_LOCK() mtx_lock(&xdma_mtx)		#define XDMA_LOCK() mtx_lock(&xdma_mtx)
#define XDMA_UNLOCK() mtx_unlock(&xdma_mtx)		#define XDMA_UNLOCK() mtx_unlock(&xdma_mtx)
#define XDMA_ASSERT_LOCKED() mtx_assert(&xdma_mtx, MA_OWNED)		#define XDMA_ASSERT_LOCKED() mtx_assert(&xdma_mtx, MA_OWNED)

/*		/*
* Per channel locks.		* Per channel locks.
*/		*/
#define XCHAN_LOCK(xchan) mtx_lock(&(xchan)->mtx_lock)		#define XCHAN_LOCK(xchan) mtx_lock(&(xchan)->mtx_lock)
#define XCHAN_UNLOCK(xchan) mtx_unlock(&(xchan)->mtx_lock)		#define XCHAN_UNLOCK(xchan) mtx_unlock(&(xchan)->mtx_lock)
#define XCHAN_ASSERT_LOCKED(xchan) mtx_assert(&(xchan)->mtx_lock, MA_OWNED)		#define XCHAN_ASSERT_LOCKED(xchan) mtx_assert(&(xchan)->mtx_lock, MA_OWNED)

		static int xchan_bufs_alloc(xdma_channel_t *xchan);
		static int xchan_bufs_free(xdma_channel_t *xchan);

		static void
		xdma_task(void *arg)
		{
		xdma_controller_t *xdma;
		xdma_channel_t *xchan_tmp;
		xdma_channel_t *xchan;

		xdma = arg;

		for (;;) {
		mtx_lock(&xdma->proc_mtx);
		msleep(xdma, &xdma->proc_mtx, PRIBIO, "jobqueue", hz);
		mtx_unlock(&xdma->proc_mtx);

		if (TAILQ_EMPTY(&xdma->channels)) {
		continue;
		}

		TAILQ_FOREACH_SAFE(xchan, &xdma->channels, xchan_next, xchan_tmp) {
		if (xchan->flags & XCHAN_TYPE_SG) {
		xdma_queue_submit(xchan);
		}
		}
		}
		}

/*		/*
* Allocate virtual xDMA channel.		* Allocate virtual xDMA channel.
*/		*/
xdma_channel_t *		xdma_channel_t *
xdma_channel_alloc(xdma_controller_t *xdma)		xdma_channel_alloc(xdma_controller_t *xdma, uint32_t caps)
{		{
xdma_channel_t *xchan;		xdma_channel_t *xchan;
int ret;		int ret;

xchan = malloc(sizeof(xdma_channel_t), M_XDMA, M_WAITOK \| M_ZERO);		xchan = malloc(sizeof(xdma_channel_t), M_XDMA, M_WAITOK \| M_ZERO);
if (xchan == NULL) {		if (xchan == NULL) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't allocate memory for channel.\n", __func__);		"%s: Can't allocate memory for channel.\n", __func__);
return (NULL);		return (NULL);
}		}
xchan->xdma = xdma;		xchan->xdma = xdma;
		xchan->caps = caps;

XDMA_LOCK();		XDMA_LOCK();

/* Request a real channel from hardware driver. */		/* Request a real channel from hardware driver. */
ret = XDMA_CHANNEL_ALLOC(xdma->dma_dev, xchan);		ret = XDMA_CHANNEL_ALLOC(xdma->dma_dev, xchan);
if (ret != 0) {		if (ret != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't request hardware channel.\n", __func__);		"%s: Can't request hardware channel.\n", __func__);
Show All 29 Lines	if (err != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't free real hw channel.\n", __func__);		"%s: Can't free real hw channel.\n", __func__);
XDMA_UNLOCK();		XDMA_UNLOCK();
return (-1);		return (-1);
}		}

xdma_teardown_all_intr(xchan);		xdma_teardown_all_intr(xchan);

/* Deallocate descriptors, if any. */		/* Deallocate bufs, if any. */
xdma_desc_free(xchan);		xchan_bufs_free(xchan);
		xchan_sglist_free(xchan);

		if (xchan->flags & XCHAN_TYPE_SG) {
		free(xchan->xr, M_XDMA);
		}

mtx_destroy(&xchan->mtx_lock);		mtx_destroy(&xchan->mtx_lock);

TAILQ_REMOVE(&xdma->channels, xchan, xchan_next);		TAILQ_REMOVE(&xdma->channels, xchan, xchan_next);

free(xchan, M_XDMA);		free(xchan, M_XDMA);

XDMA_UNLOCK();		XDMA_UNLOCK();

return (0);		return (0);
}		}

int		int
xdma_setup_intr(xdma_channel_t xchan, int (cb)(void ), void arg,		xdma_setup_intr(xdma_channel_t *xchan,
void **ihandler)		int (cb)(void , xdma_transfer_status_t *),
		void arg, void *ihandler)
{		{
struct xdma_intr_handler *ih;		struct xdma_intr_handler *ih;
xdma_controller_t *xdma;		xdma_controller_t *xdma;

xdma = xchan->xdma;		xdma = xchan->xdma;
KASSERT(xdma != NULL, ("xdma is NULL"));		KASSERT(xdma != NULL, ("xdma is NULL"));

/* Sanity check. */		/* Sanity check. */
▲ Show 20 Lines • Show All 61 Lines • ▼ Show 20 Lines	xdma_teardown_all_intr(xdma_channel_t *xchan)
TAILQ_FOREACH_SAFE(ih, &xchan->ie_handlers, ih_next, ih_tmp) {		TAILQ_FOREACH_SAFE(ih, &xchan->ie_handlers, ih_next, ih_tmp) {
TAILQ_REMOVE(&xchan->ie_handlers, ih, ih_next);		TAILQ_REMOVE(&xchan->ie_handlers, ih, ih_next);
free(ih, M_XDMA);		free(ih, M_XDMA);
}		}

return (0);		return (0);
}		}

static void		static int
xdma_dmamap_cb(void arg, bus_dma_segment_t segs, int nseg, int err)		xdma_bufs_alloc_no_busdma(xdma_channel_t *xchan)
		kanUnsubmitted Done Inline Actions This is a xdma_desc_dmamap_cb in , the name is too short and nondescript. I do not think descriptor allocation belongs in xdma at all, but this is wrong nonetheless. kan: This is a xdma_desc_dmamap_cb in , the name is too short and nondescript. I do not think…
		brAuthorUnsubmitted Not Done Inline Actions I moved descriptors allocation to DMA engine drivers. Thanks br: I moved descriptors allocation to DMA engine drivers. Thanks
{		{
xdma_channel_t *xchan;		xdma_controller_t *xdma;
		int nsegments;
int i;		int i;

xchan = (xdma_channel_t *)arg;		xdma = xchan->xdma;
KASSERT(xchan != NULL, ("xchan is NULL"));

if (err) {		nsegments = xchan->bufs_num;
xchan->map_err = 1;
return;		xchan->bufs = malloc(nsegments * sizeof(struct xchan_buf),
		M_XDMA, (M_WAITOK \| M_ZERO));
		if (xchan->bufs == NULL) {
		device_printf(xdma->dev,
		"%s: Can't allocate memory.\n", __func__);
		return (-1);
}		}

for (i = 0; i < nseg; i++) {		for (i = 0; i < nsegments; i++) {
xchan->descs_phys[i].ds_addr = segs[i].ds_addr;		xchan->bufs[i].cbuf = contigmalloc(xchan->maxsegsize,
xchan->descs_phys[i].ds_len = segs[i].ds_len;		M_XDMA, 0, 0, ~0, PAGE_SIZE, 0);
}		}

		return (0);
}		}

static int		static int
xdma_desc_alloc_bus_dma(xdma_channel_t *xchan, uint32_t desc_size,		xdma_bufs_alloc_busdma(xdma_channel_t *xchan)
uint32_t align)
{		{
xdma_controller_t *xdma;		xdma_controller_t *xdma;
bus_size_t all_desc_sz;
xdma_config_t *conf;
int nsegments;		int nsegments;
int err;		int err;
		int i;

xdma = xchan->xdma;		xdma = xchan->xdma;
conf = &xchan->conf;

nsegments = conf->block_num;		nsegments = xchan->bufs_num;
all_desc_sz = (nsegments * desc_size);

		xchan->bufs = malloc(nsegments * sizeof(struct xchan_buf),
		M_XDMA, (M_WAITOK \| M_ZERO));
		if (xchan->bufs == NULL) {
		device_printf(xdma->dev,
		"%s: Can't allocate memory.\n", __func__);
		return (-1);
		}

err = bus_dma_tag_create(		err = bus_dma_tag_create(
bus_get_dma_tag(xdma->dev),		bus_get_dma_tag(xdma->dev), /* Parent tag. */
align, desc_size, /* alignment, boundary */		xchan->alignment, 0, /* alignment, boundary */
BUS_SPACE_MAXADDR_32BIT, /* lowaddr */		BUS_SPACE_MAXADDR_32BIT, /* lowaddr */
		kanUnsubmitted Done Inline Actions Why? How xchan knows boundaries handed by the underlying DMA device? How does it even know if bus_get_dma_tag(xdma->dev) is even a right tag to be used for descriptors? Descriptors often have much stricter requirements on alignment and locations of descriptors and just 'uint32_t desc_size, uint32_t align' cannot possibly describe them. The operation does not belong in generic xchan layer at all and is better left to an actual pdma implementation. kan: Why? How xchan knows boundaries handed by the underlying DMA device? How does it even know if…
		brAuthorUnsubmitted Not Done Inline Actions Sounds right, thanks. I moved all the descriptors allocation/handling to DMA engine driver. br: Sounds right, thanks. I moved all the descriptors allocation/handling to DMA engine driver.
BUS_SPACE_MAXADDR, /* highaddr */		BUS_SPACE_MAXADDR, /* highaddr */
NULL, NULL, /* filter, filterarg */		NULL, NULL, /* filter, filterarg */
all_desc_sz, nsegments, /* maxsize, nsegments*/		xchan->maxsegsize * xchan->maxnsegs, /* maxsize */
desc_size, 0, /* maxsegsize, flags */		xchan->maxnsegs, /* nsegments */
		xchan->maxsegsize, /* maxsegsize */
		0, /* flags */
NULL, NULL, /* lockfunc, lockarg */		NULL, NULL, /* lockfunc, lockarg */
&xchan->dma_tag);		&xchan->dma_tag_bufs);
if (err) {		if (err != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't create bus_dma tag.\n", __func__);		"%s: Can't create bus_dma tag.\n", __func__);
return (-1);		return (-1);
}		}

err = bus_dmamem_alloc(xchan->dma_tag, (void **)&xchan->descs,		for (i = 0; i < nsegments; i++) {
BUS_DMA_WAITOK \| BUS_DMA_COHERENT, &xchan->dma_map);		err = bus_dmamap_create(xchan->dma_tag_bufs, 0,
if (err) {		&xchan->bufs[i].map);
		if (err != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't allocate memory for descriptors.\n", __func__);		"%s: Can't create buf DMA map.\n", __func__);
return (-1);		return (-1);
}		}
		}

xchan->descs_phys = malloc(nsegments * sizeof(xdma_descriptor_t), M_XDMA,		return (0);
(M_WAITOK \| M_ZERO));		}

xchan->map_err = 0;		static int
err = bus_dmamap_load(xchan->dma_tag, xchan->dma_map, xchan->descs,		xchan_bufs_alloc(xdma_channel_t *xchan)
all_desc_sz, xdma_dmamap_cb, xchan, BUS_DMA_WAITOK);		{
if (err) {		xdma_controller_t *xdma;
		int ret;

		xdma = xchan->xdma;
		if (xdma == NULL) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't load DMA map.\n", __func__);		"%s: Channel was not allocated properly.\n", __func__);
return (-1);		return (-1);
}		}

if (xchan->map_err != 0) {		if (xchan->caps & XCHAN_CAP_BUSDMA) {
		ret = xdma_bufs_alloc_busdma(xchan);
		} else {
		ret = xdma_bufs_alloc_no_busdma(xchan);
		}
		kanUnsubmitted Not Done Inline Actions Error handling, in particular, freeing already allocated memory and descriptors and unmapping mapped descriptors is considered futile? kan: Error handling, in particular, freeing already allocated memory and descriptors and unmapping…
		if (ret != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't load DMA map.\n", __func__);		"%s: Can't setup busdma.\n",
		__func__);
return (-1);		return (-1);
}		}

		xchan->flags \|= XCHAN_BUFS_ALLOCATED;

return (0);		return (0);
}		}

/*		static int
* This function called by DMA controller driver.		xchan_bufs_free(xdma_channel_t *xchan)
*/		{
		xchan_buf_t *b;
		kanUnsubmitted Done Inline Actions The name of the function and later use of MCLBYTES do correspond to each other rather weakly. It if an mbuf-specific functionality, name better should reflect that. kan: The name of the function and later use of MCLBYTES do correspond to each other rather weakly.
		int i;
		kanUnsubmitted Done Inline Actions Same thing. This seems to be very mbuf specific and ridden with arbitrary limitations. kan: Same thing. This seems to be very mbuf specific and ridden with arbitrary limitations.
		brAuthorUnsubmitted Not Done Inline Actions tag parameters now specified by transfer preparation functions, e.g. xdma_prep_sg br: tag parameters now specified by transfer preparation functions, e.g. xdma_prep_sg

		if ((xchan->flags & XCHAN_BUFS_ALLOCATED) == 0) {
		/* No bufs allocated. */
		return (-1);
		}

		if (xchan->caps & XCHAN_CAP_BUSDMA) {
		for (i = 0; i < xchan->bufs_num; i++) {
		b = &xchan->bufs[i];
		bus_dmamap_destroy(xchan->dma_tag_bufs, b->map);
		}
		bus_dma_tag_destroy(xchan->dma_tag_bufs);
		} else {
		for (i = 0; i < xchan->bufs_num; i++) {
		contigfree(xchan->bufs[i].cbuf, xchan->maxsegsize, M_XDMA);
		}
		}

		free(xchan->bufs, M_XDMA);
		xchan->flags &= ~XCHAN_BUFS_ALLOCATED;

		return (0);
		}
		kanUnsubmitted Done Inline Actions Again, WHY??? kan: Again, WHY???

int		int
xdma_desc_alloc(xdma_channel_t *xchan, uint32_t desc_size, uint32_t align)		xdma_prep_memcpy(xdma_channel_t *xchan, uintptr_t src_addr,
		uintptr_t dst_addr, size_t len)
{		{
xdma_controller_t *xdma;		xdma_controller_t *xdma;
xdma_config_t *conf;		xdma_config_t *conf;
int ret;		int ret;

XCHAN_ASSERT_LOCKED(xchan);		xdma = xchan->xdma;
		KASSERT(xdma != NULL, ("xdma is NULL"));

		conf = &xchan->conf;
		conf->direction = XDMA_MEM_TO_MEM;
		conf->src_addr = src_addr;
		conf->dst_addr = dst_addr;
		conf->block_len = len;
		conf->block_num = 1;

		xchan->flags \|= (XCHAN_CONFIGURED \| XCHAN_TYPE_MEMCPY);

		XCHAN_LOCK(xchan);

		ret = XDMA_CHANNEL_PREP_MEMCPY(xdma->dma_dev, xchan);
		if (ret != 0) {
		device_printf(xdma->dev,
		"%s: Can't prepare memcpy transfer.\n", __func__);
		XCHAN_UNLOCK(xchan);

		return (-1);
		}

		XCHAN_UNLOCK(xchan);

		kanUnsubmitted Not Done Inline Actions BUS_DMA_COHERENT??? kan: BUS_DMA_COHERENT???
		return (0);
		}

		/*
		* xr_num - xdma requests queue size,
		* maxsegsize - maximum allowed scatter-gather list element size in bytes
		*/
		int
		xdma_prep_sg(xdma_channel_t *xchan, uint32_t xr_num,
		uint32_t maxsegsize, uint32_t alignment)
		{
		xdma_controller_t *xdma;
		int ret;

xdma = xchan->xdma;		xdma = xchan->xdma;
if (xdma == NULL) {
		KASSERT(xdma != NULL, ("xdma is NULL"));

		if (xchan->flags & XCHAN_CONFIGURED) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Channel was not allocated properly.\n", __func__);		"%s: Channel is already configured.\n", __func__);
return (-1);		return (-1);
}		}

if (xchan->flags & XCHAN_DESC_ALLOCATED) {		xchan->maxsegsize = maxsegsize;
		xchan->alignment = alignment;
		xchan->maxnsegs = 8;
		xchan->bufs_num = xr_num;
		xchan->xr_num = xr_num;

		/* Allocate sglist. */
		ret = xchan_sglist_init(xchan);
		if (ret != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Descriptors already allocated.\n", __func__);		"%s: Can't allocate sglist.\n", __func__);
return (-1);		return (-1);
}		}

if ((xchan->flags & XCHAN_CONFIGURED) == 0) {		/* Allocate request queue. */
		xchan->xr = malloc(sizeof(struct xdma_request) * xr_num,
		M_XDMA, M_WAITOK \| M_ZERO);
		if (xchan->xr == NULL) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Channel has no configuration.\n", __func__);		"%s: Can't allocate request queue.\n", __func__);
return (-1);		return (-1);
}		}

conf = &xchan->conf;		xchan->flags \|= (XCHAN_CONFIGURED \| XCHAN_TYPE_SG);

XCHAN_UNLOCK(xchan);
ret = xdma_desc_alloc_bus_dma(xchan, desc_size, align);
XCHAN_LOCK(xchan);		XCHAN_LOCK(xchan);

		/* Deallocate bufs, if any. */
		xchan_bufs_free(xchan);

		ret = XDMA_CHANNEL_PREP_SG(xdma->dma_dev, xchan);
if (ret != 0) {		if (ret != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't allocate memory for descriptors.\n",		"%s: Can't prepare SG transfer.\n", __func__);
__func__);		XCHAN_UNLOCK(xchan);

return (-1);		return (-1);
}		}

xchan->flags \|= XCHAN_DESC_ALLOCATED;		/* Allocate bufs */
		ret = xchan_bufs_alloc(xchan);
		if (ret != 0) {
		device_printf(xdma->dev,
		"%s: Can't allocate bufs.\n", __func__);
		return (-1);
		}

/* We are going to write to descriptors. */		XCHAN_UNLOCK(xchan);
bus_dmamap_sync(xchan->dma_tag, xchan->dma_map, BUS_DMASYNC_PREWRITE);

return (0);		return (0);
}		}

int		int
xdma_desc_free(xdma_channel_t *xchan)		xdma_dequeue(xdma_channel_t xchan, void *user,
		xdma_transfer_status_t *status)
{		{
		struct xdma_request *xr;

if ((xchan->flags & XCHAN_DESC_ALLOCATED) == 0) {		if (xchan->xr_tail == xchan->xr_processed) {
/* No descriptors allocated. */
return (-1);		return (-1);
}		}

bus_dmamap_unload(xchan->dma_tag, xchan->dma_map);		xr = &xchan->xr[xchan->xr_tail];
bus_dmamem_free(xchan->dma_tag, xchan->descs, xchan->dma_map);		if (xr->done == 0) {
bus_dma_tag_destroy(xchan->dma_tag);		return (-1);
free(xchan->descs_phys, M_XDMA);		}

xchan->flags &= ~(XCHAN_DESC_ALLOCATED);		*user = xr->user;
		status->error = xr->status.error;
		status->transferred = xr->status.transferred;
		xchan->xr_tail = xchan_next_req(xchan, xchan->xr_tail);
		atomic_subtract_int(&xchan->xr_count, 1);
		kanUnsubmitted Done Inline Actions You did not protect XCHAN_DESC_ALLOCATED with ()'s above, why now? kan: You did not protect XCHAN_DESC_ALLOCATED with ()'s above, why now?

return (0);		return (0);
}		}

int		int
xdma_prep_memcpy(xdma_channel_t *xchan, uintptr_t src_addr,		xdma_enqueue(xdma_channel_t *xchan, uintptr_t src, uintptr_t dst,
uintptr_t dst_addr, size_t len)		uint8_t src_width, uint8_t dst_width, bus_size_t len,
		enum xdma_direction dir, void *user)
{		{
		struct xdma_request *xr;
xdma_controller_t *xdma;		xdma_controller_t *xdma;
xdma_config_t *conf;
		xdma = xchan->xdma;

		if (xchan->xr_count >= (xchan->xr_num - 1)) {
		/* No space is available yet. */
		kanUnsubmitted Done Inline Actions This is unconditional? You did not call map_create for !XCHAN_CAP_BUSDMA case. kan: This is unconditional? You did not call map_create for !XCHAN_CAP_BUSDMA case.
		brAuthorUnsubmitted Not Done Inline Actions Fixed. thanks! br: Fixed. thanks!
		return (-1);
		}

		xr = &xchan->xr[xchan->xr_head];
		xr->user = user;
		kanUnsubmitted Done Inline Actions You did not create tag either. kan: You did not create tag either.
		xr->direction = dir;
		xr->m = NULL;
		xr->len = len;
		kanUnsubmitted Done Inline Actions Same here? Why ()'? kan: Same here? Why ()'?
		xr->type = 0;
		xr->src_addr = src;
		xr->dst_addr = dst;
		xr->src_width = src_width;
		xr->dst_width = dst_width;
		xr->done = 0;

		xchan->xr_head = xchan_next_req(xchan, xchan->xr_head);
		atomic_add_int(&xchan->xr_count, 1);

		return (0);
		}

		struct seg_load_request {
		struct bus_dma_segment *seg;
		uint32_t nsegs;
		uint32_t error;
		};

		static void
		xdma_dmamap_cb(void arg, bus_dma_segment_t segs, int nsegs, int error)
		{
		struct seg_load_request *slr;
		struct bus_dma_segment *seg;
		int i;

		slr = arg;
		seg = slr->seg;

		if (error != 0) {
		slr->error = error;
		return;
		}

		slr->nsegs = nsegs;

		for (i = 0; i < nsegs; i++) {
		seg[i].ds_addr = segs[i].ds_addr;
		seg[i].ds_len = segs[i].ds_len;
		}
		}

		static int
		xdma_load_busdma(xdma_channel_t xchan, struct xdma_request xr,
		struct bus_dma_segment *seg, uint32_t i)
		{
		xdma_controller_t *xdma;
		struct seg_load_request slr;
		uint32_t nsegs;
		void *addr;
		int error;

		xdma = xchan->xdma;

		error = 0;
		nsegs = 0;

		switch (xr->type) {
		case XR_TYPE_MBUF:
		error = bus_dmamap_load_mbuf_sg(xchan->dma_tag_bufs,
		xchan->bufs[i].map, xr->m, seg, &nsegs, BUS_DMA_NOWAIT);
		break;
		case XR_TYPE_BIO:
		slr.nsegs = 0;
		slr.error = 0;
		slr.seg = seg;
		error = bus_dmamap_load_bio(xchan->dma_tag_bufs,
		xchan->bufs[i].map, xr->bp, xdma_dmamap_cb, &slr, BUS_DMA_NOWAIT);
		if (slr.error != 0) {
		device_printf(xdma->dma_dev,
		"%s: bus_dmamap_load failed, err %d\n",
		__func__, slr.error);
		return (0);
		}
		nsegs = slr.nsegs;
		break;
		case XR_TYPE_ADDR:
		switch (xr->direction) {
		case XDMA_MEM_TO_DEV:
		addr = (void *)xr->src_addr;
		break;
		case XDMA_DEV_TO_MEM:
		addr = (void *)xr->dst_addr;
		break;
		default:
		device_printf(xdma->dma_dev,
		"%s: Direction is not supported\n", __func__);
		return (0);
		}
		slr.nsegs = 0;
		slr.error = 0;
		slr.seg = seg;
		error = bus_dmamap_load(xchan->dma_tag_bufs, xchan->bufs[i].map,
		addr, xr->len, xdma_dmamap_cb, &slr, BUS_DMA_NOWAIT);
		if (slr.error != 0) {
		device_printf(xdma->dma_dev,
		"%s: bus_dmamap_load failed, err %d\n",
		__func__, slr.error);
		return (0);
		}
		nsegs = slr.nsegs;
		break;
		default:
		break;
		}

		if (error != 0) {
		if (error == ENOMEM) {
		/*
		* Out of memory. Try again later.
		* TODO: count errors.
		*/
		} else {
		device_printf(xdma->dma_dev,
		"%s: bus_dmamap_load failed with err %d\n",
		__func__, error);
		}
		return (0);
		}

		if (xr->direction == XDMA_MEM_TO_DEV) {
		bus_dmamap_sync(xchan->dma_tag_bufs, xchan->bufs[i].map,
		BUS_DMASYNC_PREWRITE);
		} else {
		bus_dmamap_sync(xchan->dma_tag_bufs, xchan->bufs[i].map,
		BUS_DMASYNC_PREREAD);
		}

		return (nsegs);
		}

		static int
		xdma_load_no_busdma(xdma_channel_t xchan, struct xdma_request xr,
		struct bus_dma_segment *seg, uint32_t i)
		{
		xdma_controller_t *xdma;
		struct mbuf *m;
		uint32_t nsegs;

		xdma = xchan->xdma;

		m = xr->m;

		nsegs = 1;

		switch (xr->type) {
		case XR_TYPE_MBUF:
		if (xr->direction == XDMA_MEM_TO_DEV) {
		m_copydata(m, 0, m->m_pkthdr.len, xchan->bufs[i].cbuf);
		seg[0].ds_addr = (bus_addr_t)xchan->bufs[i].cbuf;
		seg[0].ds_len = m->m_pkthdr.len;
		} else {
		seg[0].ds_addr = mtod(m, bus_addr_t);
		seg[0].ds_len = m->m_pkthdr.len;
		}
		break;
		case XR_TYPE_BIO:
		case XR_TYPE_ADDR:
		default:
		panic("implement me\n");
		}

		return (nsegs);
		}

		static int
		xdma_sglist_prepare_one(xdma_channel_t *xchan,
		struct xdma_request xr, struct bus_dma_segment seg)
		{
		xdma_controller_t *xdma;
		int error;
		int nsegs;
		int i;

		xdma = xchan->xdma;

		error = 0;
		nsegs = 0;

		i = xchan->buf_head;

		if (xchan->caps & XCHAN_CAP_BUSDMA) {
		nsegs = xdma_load_busdma(xchan, xr, seg, i);
		} else {
		nsegs = xdma_load_no_busdma(xchan, xr, seg, i);
		}
		if (nsegs == 0) {
		/* Try again later. */
		return (0);
		}

		xchan->bufs[i].xr = xr;
		xchan->bufs[i].nsegs = nsegs;
		xchan->bufs[i].nsegs_left = nsegs;

		xchan->buf_head = xchan_next_buf(xchan, xchan->buf_head);

		return (nsegs);
		}

		static int
		xdma_sglist_prepare(xdma_channel_t *xchan,
		struct xdma_sglist *sg)
		{
		struct bus_dma_segment seg[128];
		struct xdma_request *xr;
		xdma_controller_t *xdma;
		uint32_t capacity;
		uint32_t n;
		uint32_t c;
		int nsegs;
int ret;		int ret;

xdma = xchan->xdma;		xdma = xchan->xdma;

		n = 0;

		ret = XDMA_CHANNEL_CAPACITY(xdma->dma_dev, xchan, &capacity);
		if (ret != 0) {
		device_printf(xdma->dev,
		"%s: Can't get DMA controller capacity.\n", __func__);
		return (-1);
		}

		for (;;) {
		if (xchan->xr_processed == xchan->xr_head) {
		/* All the requests processed. */
		break;
		}
		xr = &xchan->xr[xchan->xr_processed];

		switch (xr->type) {
		case XR_TYPE_MBUF:
		c = xdma_mbuf_defrag(xchan, xr);
		break;
		case XR_TYPE_BIO:
		case XR_TYPE_ADDR:
		default:
		c = 1;
		}

		if (capacity <= (c + n)) {
		/*
		* No space yet available for the entire
		* request in the DMA engine.
		*/
		break;
		}

		nsegs = xdma_sglist_prepare_one(xchan, xr, seg);
		if (nsegs == 0) {
		break;
		}

		xdma_sglist_add(&sg[n], seg, nsegs, xr);
		n += nsegs;

		xchan->xr_processed = xchan_next_req(xchan, xchan->xr_processed);
		}

		return (n);
		}

		int
		xdma_queue_submit(xdma_channel_t *xchan)
		{
		struct xdma_sglist *sg;
		xdma_controller_t *xdma;
		uint32_t sg_n;
		int ret;

		xdma = xchan->xdma;
KASSERT(xdma != NULL, ("xdma is NULL"));		KASSERT(xdma != NULL, ("xdma is NULL"));

conf = &xchan->conf;		sg = xchan->sg;
conf->direction = XDMA_MEM_TO_MEM;
conf->src_addr = src_addr;
conf->dst_addr = dst_addr;
conf->block_len = len;
conf->block_num = 1;

xchan->flags \|= (XCHAN_CONFIGURED \| XCHAN_TYPE_MEMCPY);		if ((xchan->flags & XCHAN_BUFS_ALLOCATED) == 0) {
		device_printf(xdma->dev,
		"%s: Can't submit SG transfer: no bufs\n",
		__func__);
		return (-1);
		}

XCHAN_LOCK(xchan);		XCHAN_LOCK(xchan);

/* Deallocate old descriptors, if any. */		sg_n = xdma_sglist_prepare(xchan, sg);
xdma_desc_free(xchan);		if (sg_n == 0) {
		/* Nothing to submit */
		XCHAN_UNLOCK(xchan);
		return (0);
		}

ret = XDMA_CHANNEL_PREP_MEMCPY(xdma->dma_dev, xchan);		/* Now submit xdma_sglist to DMA engine driver. */

		ret = XDMA_CHANNEL_SUBMIT_SG(xdma->dma_dev, xchan, sg, sg_n);
if (ret != 0) {		if (ret != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't prepare memcpy transfer.\n", __func__);		"%s: Can't submit SG transfer.\n", __func__);

XCHAN_UNLOCK(xchan);		XCHAN_UNLOCK(xchan);

return (-1);		return (-1);
}		}

if (xchan->flags & XCHAN_DESC_ALLOCATED) {
/* Driver created xDMA descriptors. */
bus_dmamap_sync(xchan->dma_tag, xchan->dma_map,
BUS_DMASYNC_POSTWRITE);
}

XCHAN_UNLOCK(xchan);		XCHAN_UNLOCK(xchan);

return (0);		return (0);
}		}

int		int
xdma_prep_cyclic(xdma_channel_t *xchan, enum xdma_direction dir,		xdma_prep_cyclic(xdma_channel_t *xchan, enum xdma_direction dir,
uintptr_t src_addr, uintptr_t dst_addr, int block_len,		uintptr_t src_addr, uintptr_t dst_addr, int block_len,
Show All 14 Lines	xdma_prep_cyclic(xdma_channel_t *xchan, enum xdma_direction dir,
conf->block_num = block_num;		conf->block_num = block_num;
conf->src_width = src_width;		conf->src_width = src_width;
conf->dst_width = dst_width;		conf->dst_width = dst_width;

xchan->flags \|= (XCHAN_CONFIGURED \| XCHAN_TYPE_CYCLIC);		xchan->flags \|= (XCHAN_CONFIGURED \| XCHAN_TYPE_CYCLIC);

XCHAN_LOCK(xchan);		XCHAN_LOCK(xchan);

/* Deallocate old descriptors, if any. */		/* Deallocate bufs, if any. */
xdma_desc_free(xchan);		xchan_bufs_free(xchan);

ret = XDMA_CHANNEL_PREP_CYCLIC(xdma->dma_dev, xchan);		ret = XDMA_CHANNEL_PREP_CYCLIC(xdma->dma_dev, xchan);
if (ret != 0) {		if (ret != 0) {
		kanUnsubmitted Done Inline Actions By using circular buffer here you encode the requirement that dma requests should be completed by the controller in the order they were submitted, but that is not necessarily the case. I guess same applies by sglists. kan: By using circular buffer here you encode the requirement that dma requests should be completed…
		brAuthorUnsubmitted Not Done Inline Actions I switched back to linked lists. Works a bit slower due to locking, but easier to improve /add features in future. thanks! br: I switched back to linked lists. Works a bit slower due to locking, but easier to improve /add…
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't prepare cyclic transfer.\n", __func__);		"%s: Can't prepare cyclic transfer.\n", __func__);
XCHAN_UNLOCK(xchan);		XCHAN_UNLOCK(xchan);

return (-1);		return (-1);
}		}

if (xchan->flags & XCHAN_DESC_ALLOCATED) {
/* Driver has created xDMA descriptors. */
bus_dmamap_sync(xchan->dma_tag, xchan->dma_map,
BUS_DMASYNC_POSTWRITE);
}

XCHAN_UNLOCK(xchan);		XCHAN_UNLOCK(xchan);

return (0);		return (0);
}		}

int		int
		kanUnsubmitted Done Inline Actions ()'s again kan: ()'s again
xdma_begin(xdma_channel_t *xchan)		xdma_begin(xdma_channel_t *xchan)
{		{
xdma_controller_t *xdma;		xdma_controller_t *xdma;
int ret;		int ret;

xdma = xchan->xdma;		xdma = xchan->xdma;

		if (xchan->flags & XCHAN_TYPE_SG) {
		/* Not valid. */
		return (0);
		};

ret = XDMA_CHANNEL_CONTROL(xdma->dma_dev, xchan, XDMA_CMD_BEGIN);		ret = XDMA_CHANNEL_CONTROL(xdma->dma_dev, xchan, XDMA_CMD_BEGIN);
if (ret != 0) {		if (ret != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't begin the channel operation.\n", __func__);		"%s: Can't begin the channel operation.\n", __func__);
return (-1);		return (-1);
}		}

return (0);		return (0);
Show All 30 Lines	if (ret != 0) {
device_printf(xdma->dev,		device_printf(xdma->dev,
"%s: Can't pause the channel operation.\n", __func__);		"%s: Can't pause the channel operation.\n", __func__);
return (-1);		return (-1);
}		}

return (ret);		return (ret);
}		}

int		int
		kanUnsubmitted Not Done Inline Actions Is there a direct relationship of number of mbuf fragments and number of physical addresses it can result in? kan: Is there a direct relationship of number of mbuf fragments and number of physical addresses it…
xdma_callback(xdma_channel_t *xchan)		xchan_seg_done(xdma_channel_t *xchan,
		struct xdma_transfer_status *st)
{		{
struct xdma_intr_handler *ih_tmp;		struct xdma_request *xr;
struct xdma_intr_handler *ih;		xdma_controller_t *xdma;
		xchan_buf_t *b;

TAILQ_FOREACH_SAFE(ih, &xchan->ie_handlers, ih_next, ih_tmp) {		xdma = xchan->xdma;
if (ih->cb != NULL) {
ih->cb(ih->cb_user);		b = &xchan->bufs[xchan->buf_tail];
		xr = b->xr;

		atomic_subtract_int(&b->nsegs_left, 1);

		if (b->nsegs_left == 0) {
		kanUnsubmitted Not Done Inline Actions What happens if segment in the middle completes with the error? Also, what is st_trasferred and in what units it is specified? You invoke callback in per-desc (WHY? DMA driver is not allowed to do desc chaining?), but then only report status on the last desc up the consumer chain. kan: What happens if segment in the middle completes with the error? Also, what is st_trasferred and…
		if (xchan->caps & XCHAN_CAP_BUSDMA) {
		if (xr->direction == XDMA_MEM_TO_DEV) {
		bus_dmamap_sync(xchan->dma_tag_bufs, b->map,
		BUS_DMASYNC_POSTWRITE);
		} else {
		bus_dmamap_sync(xchan->dma_tag_bufs, b->map,
		BUS_DMASYNC_POSTREAD);
}		}

		bus_dmamap_unload(xchan->dma_tag_bufs, b->map);
}		}
		xr->status.error = st->error;
		xr->status.transferred = st->transferred;
		xr->done = 1;

		xchan->buf_tail = xchan_next_buf(xchan, xchan->buf_tail);
		}

return (0);		return (0);
}		}

void		int
xdma_assert_locked(void)		xdma_callback(xdma_channel_t xchan, xdma_transfer_status_t status)
{		{
		struct xdma_intr_handler *ih_tmp;
		struct xdma_intr_handler *ih;
		xdma_controller_t *xdma;

XDMA_ASSERT_LOCKED();		xdma = xchan->xdma;

		TAILQ_FOREACH_SAFE(ih, &xchan->ie_handlers, ih_next, ih_tmp) {
		if (ih->cb != NULL) {
		ih->cb(ih->cb_user, status);
}		}
		}

		wakeup(xdma);

		return (0);
		}

#ifdef FDT		#ifdef FDT
/*		/*
* Notify the DMA driver we have machine-dependent data in FDT.		* Notify the DMA driver we have machine-dependent data in FDT.
*/		*/
static int		static int
xdma_ofw_md_data(xdma_controller_t xdma, pcell_t cells, int ncells)		xdma_ofw_md_data(xdma_controller_t xdma, pcell_t cells, int ncells)
{		{
uint32_t ret;		uint32_t ret;

ret = XDMA_OFW_MD_DATA(xdma->dma_dev, cells, ncells, (void **)&xdma->data);		ret = XDMA_OFW_MD_DATA(xdma->dma_dev,
		cells, ncells, (void **)&xdma->data);

return (ret);		return (ret);
}		}

/*		/*
* Allocate xdma controller.		* Allocate xdma controller.
*/		*/
xdma_controller_t *		xdma_controller_t *
▲ Show 20 Lines • Show All 59 Lines • ▼ Show 20 Lines	xdma_ofw_get(device_t dev, const char *prop)
xdma->dev = dev;		xdma->dev = dev;
xdma->dma_dev = dma_dev;		xdma->dma_dev = dma_dev;

TAILQ_INIT(&xdma->channels);		TAILQ_INIT(&xdma->channels);

xdma_ofw_md_data(xdma, cells, ncells);		xdma_ofw_md_data(xdma, cells, ncells);
free(cells, M_OFWPROP);		free(cells, M_OFWPROP);

		mtx_init(&xdma->proc_mtx, "xDMA ofw controller", NULL, MTX_DEF);
		kproc_create(&xdma_task, xdma, &xdma->xdma_proc, 0, 0, "xdma drainer");

return (xdma);		return (xdma);
}		}
#endif		#endif

/*		/*
* Free xDMA controller object.		* Free xDMA controller object.
*/		*/
int		int
xdma_put(xdma_controller_t *xdma)		xdma_put(xdma_controller_t *xdma)
{		{

XDMA_LOCK();		XDMA_LOCK();

/* Ensure no channels allocated. */		/* Ensure no channels allocated. */
if (!TAILQ_EMPTY(&xdma->channels)) {		if (!TAILQ_EMPTY(&xdma->channels)) {
device_printf(xdma->dev, "%s: Can't free xDMA\n", __func__);		device_printf(xdma->dev, "%s: Can't free xDMA\n", __func__);
return (-1);		return (-1);
}		}

		kproc_shutdown(&xdma->xdma_proc, 0);
		mtx_destroy(&xdma->proc_mtx);

free(xdma->data, M_DEVBUF);		free(xdma->data, M_DEVBUF);
free(xdma, M_XDMA);		free(xdma, M_XDMA);

XDMA_UNLOCK();		XDMA_UNLOCK();

return (0);		return (0);
}		}
Show All 9 Lines

View Options

Fix compat32 for ntp_adjtime(2).ClosedPublicActions

Details

Diff Detail

Event Timeline

Revision ContentsChangeset List

Diff 80292

sys/compat/freebsd32/freebsd32.h

sys/compat/freebsd32/freebsd32_misc.c

sys/compat/freebsd32/syscalls.master

sys/kern/kern_ntptime.c

sys/sys/syscallsubr.h

Fix compat32 for ntp_adjtime(2).
ClosedPublic
Actions

Revision Contents
Changeset List