linux/drivers/s390/cio/chsc_sch.c
Al Viro cb787f4ac0 [tree-wide] finally take no_llseek out
no_llseek had been defined to NULL two years ago, in commit 868941b144
("fs: remove no_llseek")

To quote that commit,

  At -rc1 we'll need do a mechanical removal of no_llseek -

  git grep -l -w no_llseek | grep -v porting.rst | while read i; do
	sed -i '/\<no_llseek\>/d' $i
  done

  would do it.

Unfortunately, that hadn't been done.  Linus, could you do that now, so
that we could finally put that thing to rest? All instances are of the
form
	.llseek = no_llseek,
so it's obviously safe.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2024-09-27 08:18:43 -07:00

978 lines
22 KiB
C

// SPDX-License-Identifier: GPL-2.0
/*
* Driver for s390 chsc subchannels
*
* Copyright IBM Corp. 2008, 2011
*
* Author(s): Cornelia Huck <cornelia.huck@de.ibm.com>
*
*/
#include <linux/slab.h>
#include <linux/compat.h>
#include <linux/device.h>
#include <linux/io.h>
#include <linux/module.h>
#include <linux/uaccess.h>
#include <linux/miscdevice.h>
#include <linux/kernel_stat.h>
#include <asm/cio.h>
#include <asm/chsc.h>
#include <asm/isc.h>
#include "cio.h"
#include "cio_debug.h"
#include "css.h"
#include "chsc_sch.h"
#include "ioasm.h"
static debug_info_t *chsc_debug_msg_id;
static debug_info_t *chsc_debug_log_id;
static struct chsc_request *on_close_request;
static struct chsc_async_area *on_close_chsc_area;
static DEFINE_MUTEX(on_close_mutex);
#define CHSC_MSG(imp, args...) do { \
debug_sprintf_event(chsc_debug_msg_id, imp , ##args); \
} while (0)
#define CHSC_LOG(imp, txt) do { \
debug_text_event(chsc_debug_log_id, imp , txt); \
} while (0)
static void CHSC_LOG_HEX(int level, void *data, int length)
{
debug_event(chsc_debug_log_id, level, data, length);
}
MODULE_AUTHOR("IBM Corporation");
MODULE_DESCRIPTION("driver for s390 chsc subchannels");
MODULE_LICENSE("GPL");
static void chsc_subchannel_irq(struct subchannel *sch)
{
struct chsc_private *private = dev_get_drvdata(&sch->dev);
struct chsc_request *request = private->request;
struct irb *irb = this_cpu_ptr(&cio_irb);
CHSC_LOG(4, "irb");
CHSC_LOG_HEX(4, irb, sizeof(*irb));
inc_irq_stat(IRQIO_CSC);
/* Copy irb to provided request and set done. */
if (!request) {
CHSC_MSG(0, "Interrupt on sch 0.%x.%04x with no request\n",
sch->schid.ssid, sch->schid.sch_no);
return;
}
private->request = NULL;
memcpy(&request->irb, irb, sizeof(*irb));
cio_update_schib(sch);
complete(&request->completion);
put_device(&sch->dev);
}
static int chsc_subchannel_probe(struct subchannel *sch)
{
struct chsc_private *private;
int ret;
CHSC_MSG(6, "Detected chsc subchannel 0.%x.%04x\n",
sch->schid.ssid, sch->schid.sch_no);
sch->isc = CHSC_SCH_ISC;
private = kzalloc(sizeof(*private), GFP_KERNEL);
if (!private)
return -ENOMEM;
dev_set_drvdata(&sch->dev, private);
ret = cio_enable_subchannel(sch, (u32)virt_to_phys(sch));
if (ret) {
CHSC_MSG(0, "Failed to enable 0.%x.%04x: %d\n",
sch->schid.ssid, sch->schid.sch_no, ret);
dev_set_drvdata(&sch->dev, NULL);
kfree(private);
}
return ret;
}
static void chsc_subchannel_remove(struct subchannel *sch)
{
struct chsc_private *private;
cio_disable_subchannel(sch);
private = dev_get_drvdata(&sch->dev);
dev_set_drvdata(&sch->dev, NULL);
if (private->request) {
complete(&private->request->completion);
put_device(&sch->dev);
}
kfree(private);
}
static void chsc_subchannel_shutdown(struct subchannel *sch)
{
cio_disable_subchannel(sch);
}
static struct css_device_id chsc_subchannel_ids[] = {
{ .match_flags = 0x1, .type =SUBCHANNEL_TYPE_CHSC, },
{ /* end of list */ },
};
MODULE_DEVICE_TABLE(css, chsc_subchannel_ids);
static struct css_driver chsc_subchannel_driver = {
.drv = {
.owner = THIS_MODULE,
.name = "chsc_subchannel",
},
.subchannel_type = chsc_subchannel_ids,
.irq = chsc_subchannel_irq,
.probe = chsc_subchannel_probe,
.remove = chsc_subchannel_remove,
.shutdown = chsc_subchannel_shutdown,
};
static int __init chsc_init_dbfs(void)
{
chsc_debug_msg_id = debug_register("chsc_msg", 8, 1, 4 * sizeof(long));
if (!chsc_debug_msg_id)
goto out;
debug_register_view(chsc_debug_msg_id, &debug_sprintf_view);
debug_set_level(chsc_debug_msg_id, 2);
chsc_debug_log_id = debug_register("chsc_log", 16, 1, 16);
if (!chsc_debug_log_id)
goto out;
debug_register_view(chsc_debug_log_id, &debug_hex_ascii_view);
debug_set_level(chsc_debug_log_id, 2);
return 0;
out:
debug_unregister(chsc_debug_msg_id);
return -ENOMEM;
}
static void chsc_remove_dbfs(void)
{
debug_unregister(chsc_debug_log_id);
debug_unregister(chsc_debug_msg_id);
}
static int __init chsc_init_sch_driver(void)
{
return css_driver_register(&chsc_subchannel_driver);
}
static void chsc_cleanup_sch_driver(void)
{
css_driver_unregister(&chsc_subchannel_driver);
}
static DEFINE_SPINLOCK(chsc_lock);
static int chsc_subchannel_match_next_free(struct device *dev, const void *data)
{
struct subchannel *sch = to_subchannel(dev);
return sch->schib.pmcw.ena && !scsw_fctl(&sch->schib.scsw);
}
static struct subchannel *chsc_get_next_subchannel(struct subchannel *sch)
{
struct device *dev;
dev = driver_find_device(&chsc_subchannel_driver.drv,
sch ? &sch->dev : NULL, NULL,
chsc_subchannel_match_next_free);
return dev ? to_subchannel(dev) : NULL;
}
/**
* chsc_async() - try to start a chsc request asynchronously
* @chsc_area: request to be started
* @request: request structure to associate
*
* Tries to start a chsc request on one of the existing chsc subchannels.
* Returns:
* %0 if the request was performed synchronously
* %-EINPROGRESS if the request was successfully started
* %-EBUSY if all chsc subchannels are busy
* %-ENODEV if no chsc subchannels are available
* Context:
* interrupts disabled, chsc_lock held
*/
static int chsc_async(struct chsc_async_area *chsc_area,
struct chsc_request *request)
{
int cc;
struct chsc_private *private;
struct subchannel *sch = NULL;
int ret = -ENODEV;
char dbf[10];
chsc_area->header.key = PAGE_DEFAULT_KEY >> 4;
while ((sch = chsc_get_next_subchannel(sch))) {
spin_lock(&sch->lock);
private = dev_get_drvdata(&sch->dev);
if (private->request) {
spin_unlock(&sch->lock);
ret = -EBUSY;
continue;
}
chsc_area->header.sid = sch->schid;
CHSC_LOG(2, "schid");
CHSC_LOG_HEX(2, &sch->schid, sizeof(sch->schid));
cc = chsc(chsc_area);
snprintf(dbf, sizeof(dbf), "cc:%d", cc);
CHSC_LOG(2, dbf);
switch (cc) {
case 0:
ret = 0;
break;
case 1:
sch->schib.scsw.cmd.fctl |= SCSW_FCTL_START_FUNC;
ret = -EINPROGRESS;
private->request = request;
break;
case 2:
ret = -EBUSY;
break;
default:
ret = -ENODEV;
}
spin_unlock(&sch->lock);
CHSC_MSG(2, "chsc on 0.%x.%04x returned cc=%d\n",
sch->schid.ssid, sch->schid.sch_no, cc);
if (ret == -EINPROGRESS)
return -EINPROGRESS;
put_device(&sch->dev);
if (ret == 0)
return 0;
}
return ret;
}
static void chsc_log_command(void *chsc_area)
{
char dbf[10];
snprintf(dbf, sizeof(dbf), "CHSC:%x", ((uint16_t *)chsc_area)[1]);
CHSC_LOG(0, dbf);
CHSC_LOG_HEX(0, chsc_area, 32);
}
static int chsc_examine_irb(struct chsc_request *request)
{
int backed_up;
if (!(scsw_stctl(&request->irb.scsw) & SCSW_STCTL_STATUS_PEND))
return -EIO;
backed_up = scsw_cstat(&request->irb.scsw) & SCHN_STAT_CHAIN_CHECK;
request->irb.scsw.cmd.cstat &= ~SCHN_STAT_CHAIN_CHECK;
if (scsw_cstat(&request->irb.scsw) == 0)
return 0;
if (!backed_up)
return 0;
if (scsw_cstat(&request->irb.scsw) & SCHN_STAT_PROG_CHECK)
return -EIO;
if (scsw_cstat(&request->irb.scsw) & SCHN_STAT_PROT_CHECK)
return -EPERM;
if (scsw_cstat(&request->irb.scsw) & SCHN_STAT_CHN_DATA_CHK)
return -EAGAIN;
if (scsw_cstat(&request->irb.scsw) & SCHN_STAT_CHN_CTRL_CHK)
return -EAGAIN;
return -EIO;
}
static int chsc_ioctl_start(void __user *user_area)
{
struct chsc_request *request;
struct chsc_async_area *chsc_area;
int ret;
char dbf[10];
if (!css_general_characteristics.dynio)
/* It makes no sense to try. */
return -EOPNOTSUPP;
chsc_area = (void *)get_zeroed_page(GFP_KERNEL);
if (!chsc_area)
return -ENOMEM;
request = kzalloc(sizeof(*request), GFP_KERNEL);
if (!request) {
ret = -ENOMEM;
goto out_free;
}
init_completion(&request->completion);
if (copy_from_user(chsc_area, user_area, PAGE_SIZE)) {
ret = -EFAULT;
goto out_free;
}
chsc_log_command(chsc_area);
spin_lock_irq(&chsc_lock);
ret = chsc_async(chsc_area, request);
spin_unlock_irq(&chsc_lock);
if (ret == -EINPROGRESS) {
wait_for_completion(&request->completion);
ret = chsc_examine_irb(request);
}
/* copy area back to user */
if (!ret)
if (copy_to_user(user_area, chsc_area, PAGE_SIZE))
ret = -EFAULT;
out_free:
snprintf(dbf, sizeof(dbf), "ret:%d", ret);
CHSC_LOG(0, dbf);
kfree(request);
free_page((unsigned long)chsc_area);
return ret;
}
static int chsc_ioctl_on_close_set(void __user *user_area)
{
char dbf[13];
int ret;
mutex_lock(&on_close_mutex);
if (on_close_chsc_area) {
ret = -EBUSY;
goto out_unlock;
}
on_close_request = kzalloc(sizeof(*on_close_request), GFP_KERNEL);
if (!on_close_request) {
ret = -ENOMEM;
goto out_unlock;
}
on_close_chsc_area = (void *)get_zeroed_page(GFP_KERNEL);
if (!on_close_chsc_area) {
ret = -ENOMEM;
goto out_free_request;
}
if (copy_from_user(on_close_chsc_area, user_area, PAGE_SIZE)) {
ret = -EFAULT;
goto out_free_chsc;
}
ret = 0;
goto out_unlock;
out_free_chsc:
free_page((unsigned long)on_close_chsc_area);
on_close_chsc_area = NULL;
out_free_request:
kfree(on_close_request);
on_close_request = NULL;
out_unlock:
mutex_unlock(&on_close_mutex);
snprintf(dbf, sizeof(dbf), "ocsret:%d", ret);
CHSC_LOG(0, dbf);
return ret;
}
static int chsc_ioctl_on_close_remove(void)
{
char dbf[13];
int ret;
mutex_lock(&on_close_mutex);
if (!on_close_chsc_area) {
ret = -ENOENT;
goto out_unlock;
}
free_page((unsigned long)on_close_chsc_area);
on_close_chsc_area = NULL;
kfree(on_close_request);
on_close_request = NULL;
ret = 0;
out_unlock:
mutex_unlock(&on_close_mutex);
snprintf(dbf, sizeof(dbf), "ocrret:%d", ret);
CHSC_LOG(0, dbf);
return ret;
}
static int chsc_ioctl_start_sync(void __user *user_area)
{
struct chsc_sync_area *chsc_area;
int ret, ccode;
chsc_area = (void *)get_zeroed_page(GFP_KERNEL);
if (!chsc_area)
return -ENOMEM;
if (copy_from_user(chsc_area, user_area, PAGE_SIZE)) {
ret = -EFAULT;
goto out_free;
}
if (chsc_area->header.code & 0x4000) {
ret = -EINVAL;
goto out_free;
}
chsc_log_command(chsc_area);
ccode = chsc(chsc_area);
if (ccode != 0) {
ret = -EIO;
goto out_free;
}
if (copy_to_user(user_area, chsc_area, PAGE_SIZE))
ret = -EFAULT;
else
ret = 0;
out_free:
free_page((unsigned long)chsc_area);
return ret;
}
static int chsc_ioctl_info_channel_path(void __user *user_cd)
{
struct chsc_chp_cd *cd;
int ret, ccode;
struct {
struct chsc_header request;
u32 : 2;
u32 m : 1;
u32 : 1;
u32 fmt1 : 4;
u32 cssid : 8;
u32 : 8;
u32 first_chpid : 8;
u32 : 24;
u32 last_chpid : 8;
u32 : 32;
struct chsc_header response;
u8 data[PAGE_SIZE - 20];
} __attribute__ ((packed)) *scpcd_area;
scpcd_area = (void *)get_zeroed_page(GFP_KERNEL);
if (!scpcd_area)
return -ENOMEM;
cd = kzalloc(sizeof(*cd), GFP_KERNEL);
if (!cd) {
ret = -ENOMEM;
goto out_free;
}
if (copy_from_user(cd, user_cd, sizeof(*cd))) {
ret = -EFAULT;
goto out_free;
}
scpcd_area->request.length = 0x0010;
scpcd_area->request.code = 0x0028;
scpcd_area->m = cd->m;
scpcd_area->fmt1 = cd->fmt;
scpcd_area->cssid = cd->chpid.cssid;
scpcd_area->first_chpid = cd->chpid.id;
scpcd_area->last_chpid = cd->chpid.id;
ccode = chsc(scpcd_area);
if (ccode != 0) {
ret = -EIO;
goto out_free;
}
if (scpcd_area->response.code != 0x0001) {
ret = -EIO;
CHSC_MSG(0, "scpcd: response code=%x\n",
scpcd_area->response.code);
goto out_free;
}
memcpy(&cd->cpcb, &scpcd_area->response, scpcd_area->response.length);
if (copy_to_user(user_cd, cd, sizeof(*cd)))
ret = -EFAULT;
else
ret = 0;
out_free:
kfree(cd);
free_page((unsigned long)scpcd_area);
return ret;
}
static int chsc_ioctl_info_cu(void __user *user_cd)
{
struct chsc_cu_cd *cd;
int ret, ccode;
struct {
struct chsc_header request;
u32 : 2;
u32 m : 1;
u32 : 1;
u32 fmt1 : 4;
u32 cssid : 8;
u32 : 8;
u32 first_cun : 8;
u32 : 24;
u32 last_cun : 8;
u32 : 32;
struct chsc_header response;
u8 data[PAGE_SIZE - 20];
} __attribute__ ((packed)) *scucd_area;
scucd_area = (void *)get_zeroed_page(GFP_KERNEL);
if (!scucd_area)
return -ENOMEM;
cd = kzalloc(sizeof(*cd), GFP_KERNEL);
if (!cd) {
ret = -ENOMEM;
goto out_free;
}
if (copy_from_user(cd, user_cd, sizeof(*cd))) {
ret = -EFAULT;
goto out_free;
}
scucd_area->request.length = 0x0010;
scucd_area->request.code = 0x0026;
scucd_area->m = cd->m;
scucd_area->fmt1 = cd->fmt;
scucd_area->cssid = cd->cssid;
scucd_area->first_cun = cd->cun;
scucd_area->last_cun = cd->cun;
ccode = chsc(scucd_area);
if (ccode != 0) {
ret = -EIO;
goto out_free;
}
if (scucd_area->response.code != 0x0001) {
ret = -EIO;
CHSC_MSG(0, "scucd: response code=%x\n",
scucd_area->response.code);
goto out_free;
}
memcpy(&cd->cucb, &scucd_area->response, scucd_area->response.length);
if (copy_to_user(user_cd, cd, sizeof(*cd)))
ret = -EFAULT;
else
ret = 0;
out_free:
kfree(cd);
free_page((unsigned long)scucd_area);
return ret;
}
static int chsc_ioctl_info_sch_cu(void __user *user_cud)
{
struct chsc_sch_cud *cud;
int ret, ccode;
struct {
struct chsc_header request;
u32 : 2;
u32 m : 1;
u32 : 5;
u32 fmt1 : 4;
u32 : 2;
u32 ssid : 2;
u32 first_sch : 16;
u32 : 8;
u32 cssid : 8;
u32 last_sch : 16;
u32 : 32;
struct chsc_header response;
u8 data[PAGE_SIZE - 20];
} __attribute__ ((packed)) *sscud_area;
sscud_area = (void *)get_zeroed_page(GFP_KERNEL);
if (!sscud_area)
return -ENOMEM;
cud = kzalloc(sizeof(*cud), GFP_KERNEL);
if (!cud) {
ret = -ENOMEM;
goto out_free;
}
if (copy_from_user(cud, user_cud, sizeof(*cud))) {
ret = -EFAULT;
goto out_free;
}
sscud_area->request.length = 0x0010;
sscud_area->request.code = 0x0006;
sscud_area->m = cud->schid.m;
sscud_area->fmt1 = cud->fmt;
sscud_area->ssid = cud->schid.ssid;
sscud_area->first_sch = cud->schid.sch_no;
sscud_area->cssid = cud->schid.cssid;
sscud_area->last_sch = cud->schid.sch_no;
ccode = chsc(sscud_area);
if (ccode != 0) {
ret = -EIO;
goto out_free;
}
if (sscud_area->response.code != 0x0001) {
ret = -EIO;
CHSC_MSG(0, "sscud: response code=%x\n",
sscud_area->response.code);
goto out_free;
}
memcpy(&cud->scub, &sscud_area->response, sscud_area->response.length);
if (copy_to_user(user_cud, cud, sizeof(*cud)))
ret = -EFAULT;
else
ret = 0;
out_free:
kfree(cud);
free_page((unsigned long)sscud_area);
return ret;
}
static int chsc_ioctl_conf_info(void __user *user_ci)
{
struct chsc_conf_info *ci;
int ret, ccode;
struct {
struct chsc_header request;
u32 : 2;
u32 m : 1;
u32 : 1;
u32 fmt1 : 4;
u32 cssid : 8;
u32 : 6;
u32 ssid : 2;
u32 : 8;
u64 : 64;
struct chsc_header response;
u8 data[PAGE_SIZE - 20];
} __attribute__ ((packed)) *sci_area;
sci_area = (void *)get_zeroed_page(GFP_KERNEL);
if (!sci_area)
return -ENOMEM;
ci = kzalloc(sizeof(*ci), GFP_KERNEL);
if (!ci) {
ret = -ENOMEM;
goto out_free;
}
if (copy_from_user(ci, user_ci, sizeof(*ci))) {
ret = -EFAULT;
goto out_free;
}
sci_area->request.length = 0x0010;
sci_area->request.code = 0x0012;
sci_area->m = ci->id.m;
sci_area->fmt1 = ci->fmt;
sci_area->cssid = ci->id.cssid;
sci_area->ssid = ci->id.ssid;
ccode = chsc(sci_area);
if (ccode != 0) {
ret = -EIO;
goto out_free;
}
if (sci_area->response.code != 0x0001) {
ret = -EIO;
CHSC_MSG(0, "sci: response code=%x\n",
sci_area->response.code);
goto out_free;
}
memcpy(&ci->scid, &sci_area->response, sci_area->response.length);
if (copy_to_user(user_ci, ci, sizeof(*ci)))
ret = -EFAULT;
else
ret = 0;
out_free:
kfree(ci);
free_page((unsigned long)sci_area);
return ret;
}
static int chsc_ioctl_conf_comp_list(void __user *user_ccl)
{
struct chsc_comp_list *ccl;
int ret, ccode;
struct {
struct chsc_header request;
u32 ctype : 8;
u32 : 4;
u32 fmt : 4;
u32 : 16;
u64 : 64;
u32 list_parm[2];
u64 : 64;
struct chsc_header response;
u8 data[PAGE_SIZE - 36];
} __attribute__ ((packed)) *sccl_area;
struct {
u32 m : 1;
u32 : 31;
u32 cssid : 8;
u32 : 16;
u32 chpid : 8;
} __attribute__ ((packed)) *chpid_parm;
struct {
u32 f_cssid : 8;
u32 l_cssid : 8;
u32 : 16;
u32 res;
} __attribute__ ((packed)) *cssids_parm;
sccl_area = (void *)get_zeroed_page(GFP_KERNEL);
if (!sccl_area)
return -ENOMEM;
ccl = kzalloc(sizeof(*ccl), GFP_KERNEL);
if (!ccl) {
ret = -ENOMEM;
goto out_free;
}
if (copy_from_user(ccl, user_ccl, sizeof(*ccl))) {
ret = -EFAULT;
goto out_free;
}
sccl_area->request.length = 0x0020;
sccl_area->request.code = 0x0030;
sccl_area->fmt = ccl->req.fmt;
sccl_area->ctype = ccl->req.ctype;
switch (sccl_area->ctype) {
case CCL_CU_ON_CHP:
case CCL_IOP_CHP:
chpid_parm = (void *)&sccl_area->list_parm;
chpid_parm->m = ccl->req.chpid.m;
chpid_parm->cssid = ccl->req.chpid.chp.cssid;
chpid_parm->chpid = ccl->req.chpid.chp.id;
break;
case CCL_CSS_IMG:
case CCL_CSS_IMG_CONF_CHAR:
cssids_parm = (void *)&sccl_area->list_parm;
cssids_parm->f_cssid = ccl->req.cssids.f_cssid;
cssids_parm->l_cssid = ccl->req.cssids.l_cssid;
break;
}
ccode = chsc(sccl_area);
if (ccode != 0) {
ret = -EIO;
goto out_free;
}
if (sccl_area->response.code != 0x0001) {
ret = -EIO;
CHSC_MSG(0, "sccl: response code=%x\n",
sccl_area->response.code);
goto out_free;
}
memcpy(&ccl->sccl, &sccl_area->response, sccl_area->response.length);
if (copy_to_user(user_ccl, ccl, sizeof(*ccl)))
ret = -EFAULT;
else
ret = 0;
out_free:
kfree(ccl);
free_page((unsigned long)sccl_area);
return ret;
}
static int chsc_ioctl_chpd(void __user *user_chpd)
{
struct chsc_scpd *scpd_area;
struct chsc_cpd_info *chpd;
int ret;
chpd = kzalloc(sizeof(*chpd), GFP_KERNEL);
scpd_area = (void *)get_zeroed_page(GFP_KERNEL);
if (!scpd_area || !chpd) {
ret = -ENOMEM;
goto out_free;
}
if (copy_from_user(chpd, user_chpd, sizeof(*chpd))) {
ret = -EFAULT;
goto out_free;
}
ret = chsc_determine_channel_path_desc(chpd->chpid, chpd->fmt,
chpd->rfmt, chpd->c, chpd->m,
scpd_area);
if (ret)
goto out_free;
memcpy(&chpd->chpdb, &scpd_area->response, scpd_area->response.length);
if (copy_to_user(user_chpd, chpd, sizeof(*chpd)))
ret = -EFAULT;
out_free:
kfree(chpd);
free_page((unsigned long)scpd_area);
return ret;
}
static int chsc_ioctl_dcal(void __user *user_dcal)
{
struct chsc_dcal *dcal;
int ret, ccode;
struct {
struct chsc_header request;
u32 atype : 8;
u32 : 4;
u32 fmt : 4;
u32 : 16;
u32 res0[2];
u32 list_parm[2];
u32 res1[2];
struct chsc_header response;
u8 data[PAGE_SIZE - 36];
} __attribute__ ((packed)) *sdcal_area;
sdcal_area = (void *)get_zeroed_page(GFP_KERNEL);
if (!sdcal_area)
return -ENOMEM;
dcal = kzalloc(sizeof(*dcal), GFP_KERNEL);
if (!dcal) {
ret = -ENOMEM;
goto out_free;
}
if (copy_from_user(dcal, user_dcal, sizeof(*dcal))) {
ret = -EFAULT;
goto out_free;
}
sdcal_area->request.length = 0x0020;
sdcal_area->request.code = 0x0034;
sdcal_area->atype = dcal->req.atype;
sdcal_area->fmt = dcal->req.fmt;
memcpy(&sdcal_area->list_parm, &dcal->req.list_parm,
sizeof(sdcal_area->list_parm));
ccode = chsc(sdcal_area);
if (ccode != 0) {
ret = -EIO;
goto out_free;
}
if (sdcal_area->response.code != 0x0001) {
ret = -EIO;
CHSC_MSG(0, "sdcal: response code=%x\n",
sdcal_area->response.code);
goto out_free;
}
memcpy(&dcal->sdcal, &sdcal_area->response,
sdcal_area->response.length);
if (copy_to_user(user_dcal, dcal, sizeof(*dcal)))
ret = -EFAULT;
else
ret = 0;
out_free:
kfree(dcal);
free_page((unsigned long)sdcal_area);
return ret;
}
static long chsc_ioctl(struct file *filp, unsigned int cmd,
unsigned long arg)
{
void __user *argp;
CHSC_MSG(2, "chsc_ioctl called, cmd=%x\n", cmd);
if (is_compat_task())
argp = compat_ptr(arg);
else
argp = (void __user *)arg;
switch (cmd) {
case CHSC_START:
return chsc_ioctl_start(argp);
case CHSC_START_SYNC:
return chsc_ioctl_start_sync(argp);
case CHSC_INFO_CHANNEL_PATH:
return chsc_ioctl_info_channel_path(argp);
case CHSC_INFO_CU:
return chsc_ioctl_info_cu(argp);
case CHSC_INFO_SCH_CU:
return chsc_ioctl_info_sch_cu(argp);
case CHSC_INFO_CI:
return chsc_ioctl_conf_info(argp);
case CHSC_INFO_CCL:
return chsc_ioctl_conf_comp_list(argp);
case CHSC_INFO_CPD:
return chsc_ioctl_chpd(argp);
case CHSC_INFO_DCAL:
return chsc_ioctl_dcal(argp);
case CHSC_ON_CLOSE_SET:
return chsc_ioctl_on_close_set(argp);
case CHSC_ON_CLOSE_REMOVE:
return chsc_ioctl_on_close_remove();
default: /* unknown ioctl number */
return -ENOIOCTLCMD;
}
}
static atomic_t chsc_ready_for_use = ATOMIC_INIT(1);
static int chsc_open(struct inode *inode, struct file *file)
{
if (!atomic_dec_and_test(&chsc_ready_for_use)) {
atomic_inc(&chsc_ready_for_use);
return -EBUSY;
}
return nonseekable_open(inode, file);
}
static int chsc_release(struct inode *inode, struct file *filp)
{
char dbf[13];
int ret;
mutex_lock(&on_close_mutex);
if (!on_close_chsc_area)
goto out_unlock;
init_completion(&on_close_request->completion);
CHSC_LOG(0, "on_close");
chsc_log_command(on_close_chsc_area);
spin_lock_irq(&chsc_lock);
ret = chsc_async(on_close_chsc_area, on_close_request);
spin_unlock_irq(&chsc_lock);
if (ret == -EINPROGRESS) {
wait_for_completion(&on_close_request->completion);
ret = chsc_examine_irb(on_close_request);
}
snprintf(dbf, sizeof(dbf), "relret:%d", ret);
CHSC_LOG(0, dbf);
free_page((unsigned long)on_close_chsc_area);
on_close_chsc_area = NULL;
kfree(on_close_request);
on_close_request = NULL;
out_unlock:
mutex_unlock(&on_close_mutex);
atomic_inc(&chsc_ready_for_use);
return 0;
}
static const struct file_operations chsc_fops = {
.owner = THIS_MODULE,
.open = chsc_open,
.release = chsc_release,
.unlocked_ioctl = chsc_ioctl,
.compat_ioctl = chsc_ioctl,
};
static struct miscdevice chsc_misc_device = {
.minor = MISC_DYNAMIC_MINOR,
.name = "chsc",
.fops = &chsc_fops,
};
static int __init chsc_misc_init(void)
{
return misc_register(&chsc_misc_device);
}
static void chsc_misc_cleanup(void)
{
misc_deregister(&chsc_misc_device);
}
static int __init chsc_sch_init(void)
{
int ret;
ret = chsc_init_dbfs();
if (ret)
return ret;
isc_register(CHSC_SCH_ISC);
ret = chsc_init_sch_driver();
if (ret)
goto out_dbf;
ret = chsc_misc_init();
if (ret)
goto out_driver;
return ret;
out_driver:
chsc_cleanup_sch_driver();
out_dbf:
isc_unregister(CHSC_SCH_ISC);
chsc_remove_dbfs();
return ret;
}
static void __exit chsc_sch_exit(void)
{
chsc_misc_cleanup();
chsc_cleanup_sch_driver();
isc_unregister(CHSC_SCH_ISC);
chsc_remove_dbfs();
}
module_init(chsc_sch_init);
module_exit(chsc_sch_exit);