On Tue, 25 Jan 2011, Yauheni Kaliuta wrote:
> Hi!
>
> file_storage.c has the following code in do_write():
>
> nwritten = vfs_write(curlun->filp,
> (char __user *) bh->buf,
> amount, &file_offset_tmp);
> VLDBG(curlun, "file write %u @ %llu -> %d\n", amount,
> (unsigned long long) file_offset,
> (int) nwritten);
> if (signal_pending(current))
> return -EINTR; // Interrupted!
>
> if (nwritten < 0) {
> LDBG(curlun, "error in file write: %d\n",
> (int) nwritten);
> nwritten = 0;
> } else if (nwritten < amount) {
> LDBG(curlun, "partial file write: %d/%u\n",
> (int) nwritten, amount);
> nwritten -= (nwritten & 511);
> // Round down to a block
> }
> file_offset += nwritten;
> amount_left_to_write -= nwritten;
> fsg->residue -= nwritten;
>
> /* If an error occurred, report it and its position */
> if (nwritten < amount) {
> curlun->sense_data = SS_WRITE_ERROR;
> curlun->sense_data_info = file_offset >> 9;
> curlun->info_valid = 1;
> break;
> }
>
> In case of some data written (nwritten > 0), shouldn't it try to write
> the rest before error reporting?
In theory, perhaps it should. I haven't run across any situations
where nwritten was > 0 and < amount. Likewise with reading.
Alan Stern
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
