On 21-Nov-01 Alfred Perlstein wrote: > Jonathan Mini <[EMAIL PROTECTED]> sent me a variation of this > patch, it implments write support for UFS partitions from loader. > > Please review/test: > http://people.freebsd.org/~alfred/stand.diff
If bcache.c:write_strategy(), I wouldn't update the bcache with the new data unless the call to dv_strategy() suceeded. That is: err = dd->dv_strategy(...); if (err == 0) for (i = 0; i < nblk; i++) bcache_insert(buf + (i * bcache_blksize), blk + i); This may be part of my 4th ignorance, but why did you axe the create_null_terminated_sring functoin in support.4th? In bd_write(), why is the updating of 'p', 'dblk', and 'resid' #if 0'd. This would seem to break writing to a file when it can't be done all at once. In the 5th hunk of ufs.c, you don't free fp->f_buf if the block_map() fails. Other than that it looks great. :) Please fix those and I'll try to test it out and get it committed. -- John Baldwin <[EMAIL PROTECTED]> <>< http://www.FreeBSD.org/~jhb/ "Power Users Use the Power to Serve!" - http://www.FreeBSD.org/ To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message