patch-2.2.15 linux/drivers/i2o/i2o_config.c
Next file: linux/drivers/i2o/i2o_core.c
Previous file: linux/drivers/i2o/i2o_block.c
Back to the patch index
Back to the overall index
- Lines: 956
- Date:
Fri Apr 21 12:46:07 2000
- Orig file:
v2.2.14/drivers/i2o/i2o_config.c
- Orig date:
Thu Jan 1 01:00:00 1970
diff -u --new-file --recursive --exclude-from ../../exclude v2.2.14/drivers/i2o/i2o_config.c linux/drivers/i2o/i2o_config.c
@@ -0,0 +1,955 @@
+/*
+ * I2O Configuration Interface Driver
+ *
+ * (C) Copyright 1999 Red Hat Software
+ *
+ * Written by Alan Cox, Building Number Three Ltd
+ *
+ * Modified 04/20/1999 by Deepak Saxena
+ * - Added basic ioctl() support
+ * Modified 06/07/1999 by Deepak Saxena
+ * - Added software download ioctl (still testing)
+ * Modified 09/10/1999 by Auvo Häkkinen
+ * - Changes to i2o_cfg_reply(), ioctl_parms()
+ * - Added ioct_validate()
+ * Modified 09/30/1999 by Taneli Vähäkangas
+ * - Fixed ioctl_swdl()
+ * Modified 10/04/1999 by Taneli Vähäkangas
+ * - Changed ioctl_swdl(), implemented ioctl_swul() and ioctl_swdel()
+ * Modified 11/18/199 by Deepak Saxena
+ * - Added event managmenet support
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version
+ * 2 of the License, or (at your option) any later version.
+ */
+
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/pci.h>
+#include <linux/i2o.h>
+#include <linux/errno.h>
+#include <linux/init.h>
+#include <linux/malloc.h>
+#include <linux/miscdevice.h>
+#include <linux/mm.h>
+
+#include <asm/spinlock.h>
+#include <asm/uaccess.h>
+#include <asm/io.h>
+
+static int i2o_cfg_context = -1;
+static void *page_buf;
+static void *i2o_buffer;
+static spinlock_t i2o_config_lock = SPIN_LOCK_UNLOCKED;
+struct wait_queue *i2o_wait_queue;
+
+#define MODINC(x,y) (x = x++ % y)
+
+struct i2o_cfg_info
+{
+ struct file* fp;
+ struct fasync_struct *fasync;
+ struct i2o_evt_info event_q[I2O_EVT_Q_LEN];
+ u16 q_in; // Queue head index
+ u16 q_out; // Queue tail index
+ u16 q_len; // Queue length
+ u16 q_lost; // Number of lost events
+ u32 q_id; // Event queue ID...used as tx_context
+ struct i2o_cfg_info *next;
+};
+static struct i2o_cfg_info *open_files = NULL;
+static int i2o_cfg_info_id = 0;
+
+static int ioctl_getiops(unsigned long);
+static int ioctl_gethrt(unsigned long);
+static int ioctl_getlct(unsigned long);
+static int ioctl_parms(unsigned long, unsigned int);
+static int ioctl_html(unsigned long);
+static int ioctl_swdl(unsigned long);
+static int ioctl_swul(unsigned long);
+static int ioctl_swdel(unsigned long);
+static int ioctl_validate(unsigned long);
+static int ioctl_evt_reg(unsigned long, struct file *);
+static int ioctl_evt_get(unsigned long, struct file *);
+static int cfg_fasync(int, struct file*, int);
+
+/*
+ * This is the callback for any message we have posted. The message itself
+ * will be returned to the message pool when we return from the IRQ
+ *
+ * This runs in irq context so be short and sweet.
+ */
+static void i2o_cfg_reply(struct i2o_handler *h, struct i2o_controller *c, struct i2o_message *m)
+{
+ u32 *msg = (u32 *)m;
+
+ if (msg[0] & (1<<13))
+ printk(KERN_ERR "i2o_config: IOP failed to process the msg.\n");
+
+ if (msg[4] >> 24) // RegStatus != SUCCESS
+ i2o_report_status(KERN_INFO,"i2o_config",msg);
+
+ if(m->function == I2O_CMD_UTIL_EVT_REGISTER)
+ {
+ struct i2o_cfg_info *inf;
+
+ for(inf = open_files; inf; inf = inf->next)
+ if(inf->q_id == msg[3])
+ break;
+
+ //
+ // If this is the case, it means that we're getting
+ // events for a file descriptor that's been close()'d
+ // w/o the user unregistering for events first.
+ // The code currently assumes that the user will
+ // take care of unregistering for events before closing
+ // a file.
+ //
+ // TODO:
+ // Should we track event registartion and deregister
+ // for events when a file is close()'d so this doesn't
+ // happen? That would get rid of the search through
+ // the linked list since file->private_data could point
+ // directly to the i2o_config_info data structure...but
+ // it would mean having all sorts of tables to track
+ // what each file is registered for...I think the
+ // current method is simpler. - DS
+ //
+ if(!inf)
+ return;
+
+ inf->event_q[inf->q_in].id.iop = c->unit;
+ inf->event_q[inf->q_in].id.tid = m->target_tid;
+ inf->event_q[inf->q_in].id.evt_mask = msg[4];
+
+ //
+ // Data size = msg size - reply header
+ //
+ inf->event_q[inf->q_in].data_size = (m->size - 5) * 4;
+ if(inf->event_q[inf->q_in].data_size)
+ memcpy(inf->event_q[inf->q_in].evt_data,
+ (unsigned char *)(msg + 5),
+ inf->event_q[inf->q_in].data_size);
+
+ spin_lock(&i2o_config_lock);
+ MODINC(inf->q_in, I2O_EVT_Q_LEN);
+ if(inf->q_len == I2O_EVT_Q_LEN)
+ {
+ MODINC(inf->q_out, I2O_EVT_Q_LEN);
+ inf->q_lost++;
+ }
+ else
+ {
+ // Keep I2OEVTGET on another CPU from touching this
+ inf->q_len++;
+ }
+ spin_unlock(&i2o_config_lock);
+
+
+// printk(KERN_INFO "File %p w/id %d has %d events\n",
+// inf->fp, inf->q_id, inf->q_len);
+
+ if(inf->fasync)
+ kill_fasync(inf->fasync, SIGIO);
+ }
+
+ return;
+}
+
+/*
+ * Each of these describes an i2o message handler. They are
+ * multiplexed by the i2o_core code
+ */
+
+struct i2o_handler cfg_handler=
+{
+ i2o_cfg_reply,
+ "Configuration",
+ 0,
+ 0xffffffff // All classes
+};
+
+static long long cfg_llseek(struct file *file, long long offset, int origin)
+{
+ return -ESPIPE;
+}
+
+
+static ssize_t cfg_write(struct file *file, const char *buf, size_t count, loff_t *ppos)
+{
+ printk(KERN_INFO "i2o_config write not yet supported\n");
+
+ return 0;
+}
+
+
+static ssize_t cfg_read(struct file *file, char *buf, size_t count, loff_t *ptr)
+{
+ return 0;
+}
+
+/*
+ * IOCTL Handler
+ */
+static int cfg_ioctl(struct inode *inode, struct file *fp, unsigned int cmd,
+ unsigned long arg)
+{
+ int ret;
+
+ switch(cmd)
+ {
+ case I2OGETIOPS:
+ ret = ioctl_getiops(arg);
+ break;
+
+ case I2OHRTGET:
+ ret = ioctl_gethrt(arg);
+ break;
+
+ case I2OLCTGET:
+ ret = ioctl_getlct(arg);
+ break;
+
+ case I2OPARMSET:
+ ret = ioctl_parms(arg, I2OPARMSET);
+ break;
+
+ case I2OPARMGET:
+ ret = ioctl_parms(arg, I2OPARMGET);
+ break;
+
+ case I2OSWDL:
+ ret = ioctl_swdl(arg);
+ break;
+
+ case I2OSWUL:
+ ret = ioctl_swul(arg);
+ break;
+
+ case I2OSWDEL:
+ ret = ioctl_swdel(arg);
+ break;
+
+ case I2OVALIDATE:
+ ret = ioctl_validate(arg);
+ break;
+
+ case I2OHTML:
+ ret = ioctl_html(arg);
+ break;
+
+ case I2OEVTREG:
+ ret = ioctl_evt_reg(arg, fp);
+ break;
+
+ case I2OEVTGET:
+ ret = ioctl_evt_get(arg, fp);
+ break;
+
+ default:
+ ret = -EINVAL;
+ }
+
+ return ret;
+}
+
+int ioctl_getiops(unsigned long arg)
+{
+ u8 *user_iop_table = (u8*)arg;
+ struct i2o_controller *c = NULL;
+ int i;
+ u8 foo[MAX_I2O_CONTROLLERS];
+
+ if(!access_ok(VERIFY_WRITE, user_iop_table, MAX_I2O_CONTROLLERS))
+ return -EFAULT;
+
+ for(i = 0; i < MAX_I2O_CONTROLLERS; i++)
+ {
+ c = i2o_find_controller(i);
+ if(c)
+ {
+ foo[i] = 1;
+ i2o_unlock_controller(c);
+ }
+ else
+ {
+ foo[i] = 0;
+ }
+ }
+
+ __copy_to_user(user_iop_table, foo, MAX_I2O_CONTROLLERS);
+ return 0;
+}
+
+int ioctl_gethrt(unsigned long arg)
+{
+ struct i2o_controller *c;
+ struct i2o_cmd_hrtlct *cmd = (struct i2o_cmd_hrtlct*)arg;
+ struct i2o_cmd_hrtlct kcmd;
+ i2o_hrt *hrt;
+ int len;
+ u32 reslen;
+ int ret = 0;
+
+ if(copy_from_user(&kcmd, cmd, sizeof(struct i2o_cmd_hrtlct)))
+ return -EFAULT;
+
+ if(get_user(reslen, kcmd.reslen) < 0)
+ return -EFAULT;
+
+ if(kcmd.resbuf == NULL)
+ return -EFAULT;
+
+ c = i2o_find_controller(kcmd.iop);
+ if(!c)
+ return -ENXIO;
+
+ hrt = (i2o_hrt *)c->hrt;
+
+ i2o_unlock_controller(c);
+
+ len = 8 + ((hrt->entry_len * hrt->num_entries) << 2);
+
+ /* We did a get user...so assuming mem is ok...is this bad? */
+ put_user(len, kcmd.reslen);
+ if(len > reslen)
+ ret = -ENOBUFS;
+ if(copy_to_user(kcmd.resbuf, (void*)hrt, len))
+ ret = -EFAULT;
+
+ return ret;
+}
+
+int ioctl_getlct(unsigned long arg)
+{
+ struct i2o_controller *c;
+ struct i2o_cmd_hrtlct *cmd = (struct i2o_cmd_hrtlct*)arg;
+ struct i2o_cmd_hrtlct kcmd;
+ i2o_lct *lct;
+ int len;
+ int ret = 0;
+ u32 reslen;
+
+ if(copy_from_user(&kcmd, cmd, sizeof(struct i2o_cmd_hrtlct)))
+ return -EFAULT;
+
+ if(get_user(reslen, kcmd.reslen) < 0)
+ return -EFAULT;
+
+ if(kcmd.resbuf == NULL)
+ return -EFAULT;
+
+ c = i2o_find_controller(kcmd.iop);
+ if(!c)
+ return -ENXIO;
+
+ lct = (i2o_lct *)c->lct;
+ i2o_unlock_controller(c);
+
+ len = (unsigned int)lct->table_size << 2;
+ put_user(len, kcmd.reslen);
+ if(len > reslen)
+ ret = -ENOBUFS;
+ else if(copy_to_user(kcmd.resbuf, (void*)lct, len))
+ ret = -EFAULT;
+
+ return ret;
+}
+
+static int ioctl_parms(unsigned long arg, unsigned int type)
+{
+ int ret = 0;
+ struct i2o_controller *c;
+ struct i2o_cmd_psetget *cmd = (struct i2o_cmd_psetget*)arg;
+ struct i2o_cmd_psetget kcmd;
+ u32 reslen;
+ u8 *ops;
+ u8 *res;
+ int len;
+
+ u32 i2o_cmd = (type == I2OPARMGET ?
+ I2O_CMD_UTIL_PARAMS_GET :
+ I2O_CMD_UTIL_PARAMS_SET);
+
+ if(copy_from_user(&kcmd, cmd, sizeof(struct i2o_cmd_psetget)))
+ return -EFAULT;
+
+ if(get_user(reslen, kcmd.reslen))
+ return -EFAULT;
+
+ c = i2o_find_controller(kcmd.iop);
+ if(!c)
+ return -ENXIO;
+
+ ops = (u8*)kmalloc(kcmd.oplen, GFP_KERNEL);
+ if(!ops)
+ {
+ i2o_unlock_controller(c);
+ return -ENOMEM;
+ }
+
+ if(copy_from_user(ops, kcmd.opbuf, kcmd.oplen))
+ {
+ i2o_unlock_controller(c);
+ kfree(ops);
+ return -EFAULT;
+ }
+
+ /*
+ * It's possible to have a _very_ large table
+ * and that the user asks for all of it at once...
+ */
+ res = (u8*)kmalloc(65536, GFP_KERNEL);
+ if(!res)
+ {
+ i2o_unlock_controller(c);
+ kfree(ops);
+ return -ENOMEM;
+ }
+
+ len = i2o_issue_params(i2o_cmd, c, kcmd.tid,
+ ops, kcmd.oplen, res, 65536);
+ i2o_unlock_controller(c);
+ kfree(ops);
+
+ if (len < 0) {
+ kfree(res);
+ return len; /* -DetailedStatus */
+ }
+
+ put_user(len, kcmd.reslen);
+ if(len > reslen)
+ ret = -ENOBUFS;
+ else if(copy_to_user(cmd->resbuf, res, len))
+ ret = -EFAULT;
+
+ kfree(res);
+
+ return ret;
+}
+
+int ioctl_html(unsigned long arg)
+{
+ struct i2o_html *cmd = (struct i2o_html*)arg;
+ struct i2o_html kcmd;
+ struct i2o_controller *c;
+ u8 *res = NULL;
+ void *query = NULL;
+ int ret = 0;
+ int token;
+ u32 len;
+ u32 reslen;
+ u32 msg[MSG_FRAME_SIZE/4];
+
+ if(copy_from_user(&kcmd, cmd, sizeof(struct i2o_html)))
+ {
+ printk(KERN_INFO "i2o_config: can't copy html cmd\n");
+ return -EFAULT;
+ }
+
+ if(get_user(reslen, kcmd.reslen) < 0)
+ {
+ printk(KERN_INFO "i2o_config: can't copy html reslen\n");
+ return -EFAULT;
+ }
+
+ if(!kcmd.resbuf)
+ {
+ printk(KERN_INFO "i2o_config: NULL html buffer\n");
+ return -EFAULT;
+ }
+
+ c = i2o_find_controller(kcmd.iop);
+ if(!c)
+ return -ENXIO;
+
+ if(kcmd.qlen) /* Check for post data */
+ {
+ query = kmalloc(kcmd.qlen, GFP_KERNEL);
+ if(!query)
+ {
+ i2o_unlock_controller(c);
+ return -ENOMEM;
+ }
+ if(copy_from_user(query, kcmd.qbuf, kcmd.qlen))
+ {
+ i2o_unlock_controller(c);
+ printk(KERN_INFO "i2o_config: could not get query\n");
+ kfree(query);
+ return -EFAULT;
+ }
+ }
+
+ res = kmalloc(65536, GFP_KERNEL);
+ if(!res)
+ {
+ i2o_unlock_controller(c);
+ return -ENOMEM;
+ }
+
+ msg[1] = (I2O_CMD_UTIL_CONFIG_DIALOG << 24)|HOST_TID<<12|kcmd.tid;
+ msg[2] = i2o_cfg_context;
+ msg[3] = 0;
+ msg[4] = kcmd.page;
+ msg[5] = 0xD0000000|65536;
+ msg[6] = virt_to_bus(res);
+ if(!kcmd.qlen) /* Check for post data */
+ msg[0] = SEVEN_WORD_MSG_SIZE|SGL_OFFSET_5;
+ else
+ {
+ msg[0] = NINE_WORD_MSG_SIZE|SGL_OFFSET_5;
+ msg[5] = 0x50000000|65536;
+ msg[7] = 0xD4000000|(kcmd.qlen);
+ msg[8] = virt_to_phys(query);
+ }
+
+ token = i2o_post_wait(c, msg, 9*4, 10);
+ if(token)
+ {
+ printk(KERN_DEBUG "token = %#10x\n", token);
+ i2o_unlock_controller(c);
+ kfree(res);
+ if(kcmd.qlen) kfree(query);
+
+ return -ETIMEDOUT;
+ }
+ i2o_unlock_controller(c);
+
+ len = strnlen(res, 65536);
+ put_user(len, kcmd.reslen);
+ if(len > reslen)
+ ret = -ENOMEM;
+ if(copy_to_user(kcmd.resbuf, res, len))
+ ret = -EFAULT;
+
+ kfree(res);
+ if(kcmd.qlen)
+ kfree(query);
+
+ return ret;
+}
+
+int ioctl_swdl(unsigned long arg)
+{
+ struct i2o_sw_xfer kxfer;
+ struct i2o_sw_xfer *pxfer = (struct i2o_sw_xfer *)arg;
+ unsigned char maxfrag = 0, curfrag = 1;
+ unsigned char *buffer;
+ u32 msg[9];
+ unsigned int status = 0, swlen = 0, fragsize = 8192;
+ struct i2o_controller *c;
+
+ if(copy_from_user(&kxfer, pxfer, sizeof(struct i2o_sw_xfer)))
+ return -EFAULT;
+
+ if(get_user(swlen, kxfer.swlen) < 0)
+ return -EFAULT;
+
+ if(get_user(maxfrag, kxfer.maxfrag) < 0)
+ return -EFAULT;
+
+ if(get_user(curfrag, kxfer.curfrag) < 0)
+ return -EFAULT;
+
+ if(curfrag==maxfrag) fragsize = swlen-(maxfrag-1)*8192;
+
+ if(!kxfer.buf || !access_ok(VERIFY_READ, kxfer.buf, fragsize))
+ return -EFAULT;
+
+ c = i2o_find_controller(kxfer.iop);
+ if(!c)
+ return -ENXIO;
+
+ buffer=kmalloc(fragsize, GFP_KERNEL);
+ if (buffer==NULL)
+ {
+ i2o_unlock_controller(c);
+ return -ENOMEM;
+ }
+ __copy_from_user(buffer, kxfer.buf, fragsize);
+
+ msg[0]= NINE_WORD_MSG_SIZE | SGL_OFFSET_7;
+ msg[1]= I2O_CMD_SW_DOWNLOAD<<24 | HOST_TID<<12 | ADAPTER_TID;
+ msg[2]= (u32)cfg_handler.context;
+ msg[3]= 0;
+ msg[4]= (((u32)kxfer.flags)<<24) | (((u32)kxfer.sw_type)<<16) |
+ (((u32)maxfrag)<<8) | (((u32)curfrag));
+ msg[5]= swlen;
+ msg[6]= kxfer.sw_id;
+ msg[7]= (0xD0000000 | fragsize);
+ msg[8]= virt_to_phys(buffer);
+
+// printk("i2o_config: swdl frag %d/%d (size %d)\n", curfrag, maxfrag, fragsize);
+ status = i2o_post_wait(c, msg, sizeof(msg), 60);
+
+ i2o_unlock_controller(c);
+ kfree(buffer);
+
+ if (status != I2O_POST_WAIT_OK)
+ {
+ // it fails if you try and send frags out of order
+ // and for some yet unknown reasons too
+ printk(KERN_INFO "i2o_config: swdl failed, DetailedStatus = %d\n", status);
+ return -ETIMEDOUT;
+ }
+
+ return 0;
+}
+
+int ioctl_swul(unsigned long arg)
+{
+ struct i2o_sw_xfer kxfer;
+ struct i2o_sw_xfer *pxfer = (struct i2o_sw_xfer *)arg;
+ unsigned char maxfrag = 0, curfrag = 1;
+ unsigned char *buffer;
+ u32 msg[9];
+ unsigned int status = 0, swlen = 0, fragsize = 8192;
+ struct i2o_controller *c;
+
+ if(copy_from_user(&kxfer, pxfer, sizeof(struct i2o_sw_xfer)))
+ return -EFAULT;
+
+ if(get_user(swlen, kxfer.swlen) < 0)
+ return -EFAULT;
+
+ if(get_user(maxfrag, kxfer.maxfrag) < 0)
+ return -EFAULT;
+
+ if(get_user(curfrag, kxfer.curfrag) < 0)
+ return -EFAULT;
+
+ if(curfrag==maxfrag) fragsize = swlen-(maxfrag-1)*8192;
+
+ if(!kxfer.buf || !access_ok(VERIFY_WRITE, kxfer.buf, fragsize))
+ return -EFAULT;
+
+ c = i2o_find_controller(kxfer.iop);
+ if(!c)
+ return -ENXIO;
+
+ buffer=kmalloc(fragsize, GFP_KERNEL);
+ if (buffer==NULL)
+ {
+ i2o_unlock_controller(c);
+ return -ENOMEM;
+ }
+
+ msg[0]= NINE_WORD_MSG_SIZE | SGL_OFFSET_7;
+ msg[1]= I2O_CMD_SW_UPLOAD<<24 | HOST_TID<<12 | ADAPTER_TID;
+ msg[2]= (u32)cfg_handler.context;
+ msg[3]= 0;
+ msg[4]= (u32)kxfer.flags<<24|(u32)kxfer.sw_type<<16|(u32)maxfrag<<8|(u32)curfrag;
+ msg[5]= swlen;
+ msg[6]= kxfer.sw_id;
+ msg[7]= (0xD0000000 | fragsize);
+ msg[8]= virt_to_bus(buffer);
+
+// printk("i2o_config: swul frag %d/%d (size %d)\n", curfrag, maxfrag, fragsize);
+ status = i2o_post_wait(c, msg, sizeof(msg), 60);
+ i2o_unlock_controller(c);
+
+ if (status != I2O_POST_WAIT_OK)
+ {
+ kfree(buffer);
+ printk(KERN_INFO "i2o_config: swul failed, DetailedStatus = %d\n", status);
+ return -ETIMEDOUT;
+ }
+
+ __copy_to_user(kxfer.buf, buffer, fragsize);
+ kfree(buffer);
+
+ return 0;
+}
+
+int ioctl_swdel(unsigned long arg)
+{
+ struct i2o_controller *c;
+ struct i2o_sw_xfer kxfer, *pxfer = (struct i2o_sw_xfer *)arg;
+ u32 msg[7];
+ unsigned int swlen;
+ int token;
+
+ if (copy_from_user(&kxfer, pxfer, sizeof(struct i2o_sw_xfer)))
+ return -EFAULT;
+
+ if (get_user(swlen, kxfer.swlen) < 0)
+ return -EFAULT;
+
+ c = i2o_find_controller(kxfer.iop);
+ if (!c)
+ return -ENXIO;
+
+ msg[0] = SEVEN_WORD_MSG_SIZE | SGL_OFFSET_0;
+ msg[1] = I2O_CMD_SW_REMOVE<<24 | HOST_TID<<12 | ADAPTER_TID;
+ msg[2] = (u32)i2o_cfg_context;
+ msg[3] = 0;
+ msg[4] = (u32)kxfer.flags<<24 | (u32)kxfer.sw_type<<16;
+ msg[5] = swlen;
+ msg[6] = kxfer.sw_id;
+
+ token = i2o_post_wait(c, msg, sizeof(msg), 10);
+ i2o_unlock_controller(c);
+
+ if (token != I2O_POST_WAIT_OK)
+ {
+ printk(KERN_INFO "i2o_config: swdel failed, DetailedStatus = %d\n", token);
+ return -ETIMEDOUT;
+ }
+
+ return 0;
+}
+
+int ioctl_validate(unsigned long arg)
+{
+ int token;
+ int iop = (int)arg;
+ u32 msg[4];
+ struct i2o_controller *c;
+
+ c=i2o_find_controller(iop);
+ if (!c)
+ return -ENXIO;
+
+ msg[0] = FOUR_WORD_MSG_SIZE|SGL_OFFSET_0;
+ msg[1] = I2O_CMD_CONFIG_VALIDATE<<24 | HOST_TID<<12 | iop;
+ msg[2] = (u32)i2o_cfg_context;
+ msg[3] = 0;
+
+ token = i2o_post_wait(c, msg, sizeof(msg), 10);
+ i2o_unlock_controller(c);
+
+ if (token != I2O_POST_WAIT_OK)
+ {
+ printk(KERN_INFO "Can't validate configuration, ErrorStatus = %d\n",
+ token);
+ return -ETIMEDOUT;
+ }
+
+ return 0;
+}
+
+static int ioctl_evt_reg(unsigned long arg, struct file *fp)
+{
+ u32 msg[5];
+ struct i2o_evt_id *pdesc = (struct i2o_evt_id *)arg;
+ struct i2o_evt_id kdesc;
+ struct i2o_controller *iop;
+ struct i2o_device *d;
+
+ if (copy_from_user(&kdesc, pdesc, sizeof(struct i2o_evt_id)))
+ return -EFAULT;
+
+ /* IOP exists? */
+ iop = i2o_find_controller(kdesc.iop);
+ if(!iop)
+ return -ENXIO;
+ i2o_unlock_controller(iop);
+
+ /* Device exists? */
+ for(d = iop->devices; d; d = d->next)
+ if(d->lct_data->tid == kdesc.tid)
+ break;
+
+ if(!d)
+ return -ENODEV;
+
+ msg[0] = FOUR_WORD_MSG_SIZE|SGL_OFFSET_0;
+ msg[1] = I2O_CMD_UTIL_EVT_REGISTER<<24 | HOST_TID<<12 | kdesc.tid;
+ msg[2] = (u32)i2o_cfg_context;
+ msg[3] = (u32)fp->private_data;
+ msg[4] = kdesc.evt_mask;
+
+ i2o_post_this(iop, msg, 20);
+
+ return 0;
+}
+
+static int ioctl_evt_get(unsigned long arg, struct file *fp)
+{
+ u32 id = (u32)fp->private_data;
+ struct i2o_cfg_info *p = NULL;
+ struct i2o_evt_get *uget = (struct i2o_evt_get*)arg;
+ struct i2o_evt_get kget;
+ unsigned int flags;
+
+ // access_ok doesn't check for NULL?!?!
+ if(!arg)
+ return -EFAULT;
+
+ if(!access_ok(VERIFY_WRITE, uget, sizeof(struct i2o_evt_get)))
+ return -EFAULT;
+
+ for(p = open_files; p; p = p->next)
+ if(p->q_id == id)
+ break;
+
+ if(!p->q_len)
+ {
+ return -ENOENT;
+ return 0;
+ }
+
+ memcpy(&kget.info, &p->event_q[p->q_out], sizeof(struct i2o_evt_info));
+ MODINC(p->q_out, I2O_EVT_Q_LEN);
+ spin_lock_irqsave(&i2o_config_lock, flags);
+ p->q_len--;
+ kget.pending = p->q_len;
+ kget.lost = p->q_lost;
+ spin_unlock_irqrestore(&i2o_config_lock, flags);
+
+ __copy_to_user(uget, &kget, sizeof(struct i2o_evt_get));
+
+ return 0;
+}
+
+static int cfg_open(struct inode *inode, struct file *file)
+{
+ struct i2o_cfg_info *tmp =
+ (struct i2o_cfg_info *)kmalloc(sizeof(struct i2o_cfg_info), GFP_KERNEL);
+ unsigned int flags;
+
+ if(!tmp)
+ return -ENOMEM;
+
+ file->private_data = (void*)(i2o_cfg_info_id++);
+ tmp->fp = file;
+ tmp->fasync = NULL;
+ tmp->q_id = (u32)file->private_data;
+ tmp->q_len = 0;
+ tmp->q_in = 0;
+ tmp->q_out = 0;
+ tmp->q_lost = 0;
+ tmp->next = open_files;
+
+ spin_lock_irqsave(&i2o_config_lock, flags);
+ open_files = tmp;
+ spin_unlock_irqrestore(&i2o_config_lock, flags);
+
+ MOD_INC_USE_COUNT;
+ return 0;
+}
+
+static int cfg_release(struct inode *inode, struct file *file)
+{
+ u32 id = (u32)file->private_data;
+ struct i2o_cfg_info *p1, *p2;
+ unsigned int flags;
+
+ p1 = p2 = NULL;
+
+ spin_lock_irqsave(&i2o_config_lock, flags);
+ for(p1 = open_files; p1; )
+ {
+ if(p1->q_id == id)
+ {
+
+ if(p1->fasync)
+ cfg_fasync(-1, file, 0);
+ if(p2)
+ p2->next = p1->next;
+ else
+ open_files = p1->next;
+
+ kfree(p1);
+ break;
+ }
+ p2 = p1;
+ p1 = p1->next;
+ }
+ spin_unlock_irqrestore(&i2o_config_lock, flags);
+
+ MOD_DEC_USE_COUNT;
+ return 0;
+}
+
+static int cfg_fasync(int fd, struct file *fp, int on)
+{
+ u32 id = (u32)fp->private_data;
+ struct i2o_cfg_info *p;
+
+ for(p = open_files; p; p = p->next)
+ if(p->q_id == id)
+ break;
+
+ if(!p)
+ return -EBADF;
+
+ return fasync_helper(fd, fp, on, &p->fasync);
+}
+
+static struct file_operations config_fops =
+{
+ llseek: cfg_llseek,
+ read: cfg_read,
+ write: cfg_write,
+ ioctl: cfg_ioctl,
+ open: cfg_open,
+ release: cfg_release,
+ fasync: cfg_fasync,
+};
+
+static struct miscdevice i2o_miscdev = {
+ I2O_MINOR,
+ "i2octl",
+ &config_fops
+};
+
+#ifdef MODULE
+int init_module(void)
+#else
+int __init i2o_config_init(void)
+#endif
+{
+ printk(KERN_INFO "I2O configuration manager v 0.04\n");
+ printk(KERN_INFO "(C) Copyright 1999 Red Hat Software\n");
+
+ if((page_buf = kmalloc(4096, GFP_KERNEL))==NULL)
+ {
+ printk(KERN_ERR "i2o_config: no memory for page buffer.\n");
+ return -ENOBUFS;
+ }
+ if(misc_register(&i2o_miscdev)==-1)
+ {
+ printk(KERN_ERR "i2o_config: can't register device.\n");
+ kfree(page_buf);
+ return -EBUSY;
+ }
+ /*
+ * Install our handler
+ */
+ if(i2o_install_handler(&cfg_handler)<0)
+ {
+ kfree(page_buf);
+ printk(KERN_ERR "i2o_config: handler register failed.\n");
+ misc_deregister(&i2o_miscdev);
+ return -EBUSY;
+ }
+ /*
+ * The low 16bits of the transaction context must match this
+ * for everything we post. Otherwise someone else gets our mail
+ */
+ i2o_cfg_context = cfg_handler.context;
+ return 0;
+}
+
+#ifdef MODULE
+
+void cleanup_module(void)
+{
+ misc_deregister(&i2o_miscdev);
+
+ if(page_buf)
+ kfree(page_buf);
+ if(i2o_cfg_context != -1)
+ i2o_remove_handler(&cfg_handler);
+ if(i2o_buffer)
+ kfree(i2o_buffer);
+}
+
+EXPORT_NO_SYMBOLS;
+MODULE_AUTHOR("Red Hat Software");
+MODULE_DESCRIPTION("I2O Configuration");
+
+#endif
FUNET's LINUX-ADM group, linux-adm@nic.funet.fi
TCL-scripts by Sam Shen (who was at: slshen@lbl.gov)