[00:25:55] --- Simon Wilkinson has left [01:08:57] --- lars.malinowsky has become available [01:36:18] --- lars.malinowsky has left [01:36:27] --- lars.malinowsky has become available [02:29:50] --- pod has become available [06:10:06] --- jakllsch has become available [06:21:52] --- reuteras has left [06:32:57] --- ksumner has become available [06:53:12] --- mfelliott has become available [07:32:14] --- deason has become available [08:10:23] --- lars.malinowsky has left [08:14:50] --- mfelliott has left [09:19:52] --- asedeno has become available [09:22:47] --- steven.jenkins has left [09:23:48] --- steven.jenkins has become available [10:08:35] --- asedeno has left [10:16:06] --- steven.jenkins has left [10:16:57] --- steven.jenkins has become available [10:20:25] --- rra has become available [10:39:42] --- jakllsch has left [10:45:38] --- jakllsch has become available [11:22:57] --- phalenor has left [11:23:02] --- phalenor has become available [11:23:13] --- mfelliott has become available [11:38:33] --- ksumner has left: Lost connection [11:40:13] --- lars.malinowsky has become available [11:41:20] --- ksumner has become available [11:54:50] --- kaduk@mit.edu/barnowl has become available [11:54:50] --- andersk has become available [12:04:37] [test 1] [12:04:59] it passed or failed depending how harshly i am judging [12:05:21] jmuc presence still looks wonky, but it is maybe getting better. [12:15:01] --- asedeno has become available [12:16:54] hm. does the openbsd cache manager port have "faulting-in" support? [14:17:31] --- ksumner has left [14:23:29] --- shadow@gmail.com/barnowlA9683AD7 has left [14:23:39] --- shadow@gmail.com/barnowlA9683AD7 has become available [14:37:01] --- shadow@gmail.com/barnowlA9683AD7 has left [14:37:18] --- shadow@gmail.com/barnowlA9683AD7 has become available [15:15:05] --- lars.malinowsky has left [15:15:10] --- lars.malinowsky has become available [16:01:18] --- deason has left [16:06:13] --- shadow@gmail.com/barnowlA9683AD7 has left [16:06:28] --- shadow@gmail.com/barnowlA9683AD7 has become available [17:15:11] --- deason has become available [17:47:37] so, it turns out that the fun part is actually memory mapping the vnode [17:48:37] Yup. [17:48:54] The NFS implementation is probably your friend. [17:50:20] --- rra has left: Disconnected [17:51:46] how important is VOP_BMAP? [17:51:59] * jakllsch greps [17:52:46] grep is perhaps not the best way to check. I would look at what the default implementation is, and whether it is mandatory for a filesystem to implement. [17:53:18] --- steven.jenkins has left [17:53:53] --- steven.jenkins has become available [17:57:14] i'm only seeing pseudo filesystems like kernfs not implementing something [17:59:03] arla nnpfs does a 1:1 translation, coda does nothing with the block number [17:59:41] and puffs does a 1:1 with the addition of a_runp getting filled in [17:59:56] kern/vfs_default.c has a default implementation, pretty simple. I see XXX comments "needs good comment and manpage" [18:00:20] I would probably recommend doing nothing for now and seeing what breaks. [18:00:52] at the moment, i can't read (via mmap) beyond 1 page [18:01:14] or, at least that's what happens with this Makefile, a statically linked binary managed to execute [18:02:21] Shouldn't this be the realm of vop_{get,put}pages? [18:03:19] i'm using the generic genfs-supplied versions of those [18:03:37] (which are used by many other file systems, often directly) [18:07:31] the key to getting as far as i have on this issue was uvm_vnp_setsize() [18:08:15] --- Russ has become available [18:09:24] Have you written a simple test to see if mmap() succeeds, if you can successfully read from a mmaped file, write from one, etc? [18:09:41] no, i should :-) [18:10:18] It sounds a bit like you are grasping at straws, and would benefit from some analysis of what is actually failing. [18:29:35] --- Russ has left: Disconnected [18:30:35] --- Russ has become available [19:12:07] Does 'make dest' install a server, too, or just a client? [19:16:25] installs everything [19:17:19] Bah. Does it canonically install an rc script for bosserver? [19:18:20] canonically? no. [19:20:15] Yippee. [20:34:23] --- mfelliott has left [20:52:18] --- jakllsch has left [21:52:02] --- deason has left [21:52:02] --- phalenor has left [21:53:02] --- phalenor has become available [21:58:34] --- Russ has left: Disconnected [22:13:11] --- reuteras has become available [23:20:22] --- lars.malinowsky has left