/* * $Id: mtdchar.c,v 1.21.2.3 2001/01/09 00:18:31 dwmw2 Exp $ * * Character-device access to raw MTD devices. * */ #include #include #include #include #include #include #ifdef CONFIG_DEVFS_FS #include static void mtd_notify_add(struct mtd_info* mtd); static void mtd_notify_remove(struct mtd_info* mtd); static struct mtd_notifier notifier = { mtd_notify_add, mtd_notify_remove, NULL }; static devfs_handle_t devfs_dir_handle = NULL; static devfs_handle_t devfs_rw_handle[MAX_MTD_DEVICES]; static devfs_handle_t devfs_ro_handle[MAX_MTD_DEVICES]; #endif #if LINUX_VERSION_CODE >= KERNEL_VERSION(2,2,0) static loff_t mtd_lseek (struct file *file, loff_t offset, int orig) #else static int mtd_lseek (struct inode *inode, struct file *file, off_t offset, int orig) #endif { struct mtd_info *mtd=(struct mtd_info *)file->private_data; switch (orig) { case 0: /* SEEK_SET */ file->f_pos = offset; break; case 1: /* SEEK_CUR */ file->f_pos += offset; break; case 2: /* SEEK_END */ file->f_pos =mtd->size + offset; break; default: return -EINVAL; } if (file->f_pos < 0) file->f_pos = 0; else if (file->f_pos >= mtd->size) file->f_pos = mtd->size - 1; return file->f_pos; } static int mtd_open(struct inode *inode, struct file *file) { int minor = MINOR(inode->i_rdev); int devnum = minor >> 1; struct mtd_info *mtd; DEBUG(MTD_DEBUG_LEVEL0, "MTD_open\n"); if (devnum >= MAX_MTD_DEVICES) return -ENODEV; /* You can't open the RO devices RW */ if ((file->f_mode & 2) && (minor & 1)) return -EACCES; mtd = get_mtd_device(NULL, devnum); if (!mtd) return -ENODEV; file->private_data = mtd; /* You can't open it RW if it's not a writeable device */ if ((file->f_mode & 2) && !(mtd->flags & MTD_WRITEABLE)) { put_mtd_device(mtd); return -EACCES; } return 0; } /* mtd_open */ /*====================================================================*/ static release_t mtd_close(struct inode *inode, struct file *file) { struct mtd_info *mtd; DEBUG(MTD_DEBUG_LEVEL0, "MTD_close\n"); mtd = (struct mtd_info *)file->private_data; if (mtd->sync) mtd->sync(mtd); put_mtd_device(mtd); release_return(0); } /* mtd_close */ #if LINUX_VERSION_CODE >= KERNEL_VERSION(2,2,0) #define FILE_POS *ppos #else #define FILE_POS file->f_pos #endif #if LINUX_VERSION_CODE >= KERNEL_VERSION(2,2,0) static ssize_t mtd_read(struct file *file, char *buf, size_t count,loff_t *ppos) #else static int mtd_read(struct inode *inode,struct file *file, char *buf, int count) #endif { struct mtd_info *mtd = (struct mtd_info *)file->private_data; size_t retlen=0; int ret=0; char *kbuf; DEBUG(MTD_DEBUG_LEVEL0,"MTD_read\n"); if (FILE_POS + count > mtd->size) count = mtd->size - FILE_POS; if (!count) return 0; /* FIXME: Use kiovec in 2.3 or 2.2+rawio, or at * least split the IO into smaller chunks. */ kbuf = vmalloc(count); if (!kbuf) return -ENOMEM; ret = MTD_READ(mtd, FILE_POS, count, &retlen, kbuf); if (!ret) { FILE_POS += retlen; if (copy_to_user(buf, kbuf, retlen)) ret = -EFAULT; else ret = retlen; } vfree(kbuf); return ret; } /* mtd_read */ #if LINUX_VERSION_CODE >= KERNEL_VERSION(2,2,0) static ssize_t mtd_write(struct file *file, const char *buf, size_t count,loff_t *ppos) #else static read_write_t mtd_write(struct inode *inode,struct file *file, const char *buf, count_t count) #endif { struct mtd_info *mtd = (struct mtd_info *)file->private_data; char *kbuf; size_t retlen; int ret=0; DEBUG(MTD_DEBUG_LEVEL0,"MTD_write\n"); if (FILE_POS == mtd->size) return -ENOSPC; if (FILE_POS + count > mtd->size) count = mtd->size - FILE_POS; if (!count) return 0; kbuf=vmalloc(count); if (!kbuf) return -ENOMEM; if (copy_from_user(kbuf, buf, count)) { vfree(kbuf); return -EFAULT; } ret = (*(mtd->write))(mtd, FILE_POS, count, &retlen, buf); if (!ret) { FILE_POS += retlen; ret = retlen; } vfree(kbuf); return ret; } /* mtd_write */ /*====================================================================== IOCTL calls for getting device parameters. ======================================================================*/ static void mtd_erase_callback (struct erase_info *instr) { wake_up((wait_queue_head_t *)instr->priv); } static int mtd_ioctl(struct inode *inode, struct file *file, u_int cmd, u_long arg) { struct mtd_info *mtd = (struct mtd_info *)file->private_data; int ret = 0; u_long size; DEBUG(MTD_DEBUG_LEVEL0, "MTD_ioctl\n"); size = (cmd & IOCSIZE_MASK) >> IOCSIZE_SHIFT; if (cmd & IOC_IN) { ret = verify_area(VERIFY_READ, (char *)arg, size); if (ret) return ret; } if (cmd & IOC_OUT) { ret = verify_area(VERIFY_WRITE, (char *)arg, size); if (ret) return ret; } switch (cmd) { case MEMGETINFO: if (copy_to_user((struct mtd_info *)arg, mtd, sizeof(struct mtd_info_user))) return -EFAULT; break; case MEMERASE: { struct erase_info *erase=kmalloc(sizeof(struct erase_info),GFP_KERNEL); if (!erase) ret = -ENOMEM; else { wait_queue_head_t waitq; DECLARE_WAITQUEUE(wait, current); init_waitqueue_head(&waitq); memset (erase,0,sizeof(struct erase_info)); if (copy_from_user(&erase->addr, (u_long *)arg, 2 * sizeof(u_long))) { kfree(erase); return -EFAULT; } erase->mtd = mtd; erase->callback = mtd_erase_callback; erase->priv = (unsigned long)&waitq; /* FIXME: Allow INTERRUPTIBLE. Which means not having the wait_queue head on the stack. If the wq_head is on the stack, and we leave because we got interrupted, then the wq_head is no longer there when the callback routine tries to wake us up. */ current->state = TASK_UNINTERRUPTIBLE; add_wait_queue(&waitq, &wait); ret = mtd->erase(mtd, erase); if (!ret) schedule(); remove_wait_queue(&waitq, &wait); current->state = TASK_RUNNING; if (!ret) ret = (erase->state == MTD_ERASE_FAILED); kfree(erase); } break; } case MEMWRITEOOB: { struct mtd_oob_buf buf; void *databuf; ssize_t retlen; if (copy_from_user(&buf, (struct mtd_oob_buf *)arg, sizeof(struct mtd_oob_buf))) return -EFAULT; if (buf.length > 0x4096) return -EINVAL; if (!mtd->write_oob) ret = -EOPNOTSUPP; else ret = verify_area(VERIFY_READ, (char *)buf.ptr, buf.length); if (ret) return ret; databuf = kmalloc(buf.length, GFP_KERNEL); if (!databuf) return -ENOMEM; if (copy_from_user(databuf, buf.ptr, buf.length)) return -EFAULT; ret = (mtd->write_oob)(mtd, buf.start, buf.length, &retlen, databuf); if (copy_to_user((void *)arg + sizeof(loff_t), &retlen, sizeof(ssize_t))) ret = -EFAULT; kfree(databuf); break; } case MEMREADOOB: { struct mtd_oob_buf buf; void *databuf; ssize_t retlen; if (copy_from_user(&buf, (struct mtd_oob_buf *)arg, sizeof(struct mtd_oob_buf))) return -EFAULT; if (buf.length > 0x4096) return -EINVAL; if (!mtd->read_oob) ret = -EOPNOTSUPP; else ret = verify_area(VERIFY_WRITE, (char *)buf.ptr, buf.length); if (ret) return ret; databuf = kmalloc(buf.length, GFP_KERNEL); if (!databuf) return -ENOMEM; ret = (mtd->read_oob)(mtd, buf.start, buf.length, &retlen, databuf); if (copy_to_user((void *)arg + sizeof(loff_t), &retlen, sizeof(ssize_t))) ret = -EFAULT; else if (retlen && copy_to_user(buf.ptr, databuf, retlen)) ret = -EFAULT; kfree(databuf); break; } case MEMLOCK: { unsigned long adrs[2]; if (copy_from_user(adrs ,(void *)arg, 2* sizeof(unsigned long))) return -EFAULT; if (!mtd->lock) ret = -EOPNOTSUPP; else ret = mtd->lock(mtd, adrs[0], adrs[1]); } case MEMUNLOCK: { unsigned long adrs[2]; if (copy_from_user(adrs, (void *)arg, 2* sizeof(unsigned long))) return -EFAULT; if (!mtd->unlock) ret = -EOPNOTSUPP; else ret = mtd->unlock(mtd, adrs[0], adrs[1]); } default: printk("Invalid ioctl %x (MEMGETINFO = %x)\n",cmd, MEMGETINFO); ret = -EINVAL; } return ret; } /* memory_ioctl */ static struct file_operations mtd_fops = { owner: THIS_MODULE, llseek: mtd_lseek, /* lseek */ read: mtd_read, /* read */ write: mtd_write, /* write */ ioctl: mtd_ioctl, /* ioctl */ open: mtd_open, /* open */ release: mtd_close, /* release */ }; #ifdef CONFIG_DEVFS_FS /* Notification that a new device has been added. Create the devfs entry for * it. */ static void mtd_notify_add(struct mtd_info* mtd) { char name[8]; if (!mtd) return; sprintf(name, "%d", mtd->index); devfs_rw_handle[mtd->index] = devfs_register(devfs_dir_handle, name, DEVFS_FL_DEFAULT, MTD_CHAR_MAJOR, mtd->index*2, S_IFCHR | S_IRUGO | S_IWUGO, &mtd_fops, NULL); sprintf(name, "%dro", mtd->index); devfs_ro_handle[mtd->index] = devfs_register(devfs_dir_handle, name, DEVFS_FL_DEFAULT, MTD_CHAR_MAJOR, mtd->index*2+1, S_IFCHR | S_IRUGO | S_IWUGO, &mtd_fops, NULL); } static void mtd_notify_remove(struct mtd_info* mtd) { if (!mtd) return; devfs_unregister(devfs_rw_handle[mtd->index]); devfs_unregister(devfs_ro_handle[mtd->index]); } #endif #if LINUX_VERSION_CODE < 0x20212 && defined(MODULE) #define init_mtdchar init_module #define cleanup_mtdchar cleanup_module #endif mod_init_t init_mtdchar(void) { #ifdef CONFIG_DEVFS_FS if (devfs_register_chrdev(MTD_CHAR_MAJOR, "mtd", &mtd_fops)) { printk(KERN_NOTICE "Can't allocate major number %d for Memory Technology Devices.\n", MTD_CHAR_MAJOR); return -EAGAIN; } devfs_dir_handle = devfs_mk_dir(NULL, "mtd", NULL); register_mtd_user(¬ifier); #else if (register_chrdev(MTD_CHAR_MAJOR, "mtd", &mtd_fops)) { printk(KERN_NOTICE "Can't allocate major number %d for Memory Technology Devices.\n", MTD_CHAR_MAJOR); return -EAGAIN; } #endif return 0; } mod_exit_t cleanup_mtdchar(void) { #ifdef CONFIG_DEVFS_FS unregister_mtd_user(¬ifier); devfs_unregister(devfs_dir_handle); devfs_unregister_chrdev(MTD_CHAR_MAJOR, "mtd"); #else unregister_chrdev(MTD_CHAR_MAJOR, "mtd"); #endif } module_init(init_mtdchar); module_exit(cleanup_mtdchar);