1 ## Note: this is my personal todo and ideas list (alexh@)
3 * add a communication channel mechanism to dm
4 - essentially a way to send messages and receive responses to dm target instances
7 - Improve to support external scripts/programs providing passphrases
10 - Look at plain64 iv support, what it implies vs plain.
11 - Think about benbi iv support
13 * Add TrueCrypt support
14 - Simply add a userland tool that sets dm_target_crypt up with the right parameters, extracted from the TrueCrypt header
15 - https://github.com/bwalex/tc-play <- done!
16 - needs OpenSSL with XTS support; probably next major release
21 - http://www.openbsd.org/cgi-bin/cvsweb/src/sys/crypto/
22 - think about support for other ciphers, such as Serpent and Twofish
24 * Port hwpmc & dig into (boot-up) performance
28 - separate out common arch parts (linprocfs, for example)
30 * Fix the crash analysis script (or rather the programs it calls [some segfault])
32 * Take a look at updating lvm/dm/libdevicemapper
34 * Take a look at importing libdm from NetBSD
35 - would allow tc-play and similarly newly developed foo to avoid GPL license issues
38 o Added VT6105M specific register definitions. VT6105M has the
39 following hardware capabilities.
40 - Tx/Rx IP/TCP/UDP checksum offload.
41 - VLAN hardware tag insertion/extraction. Due to lack of information
42 for getting extracted VLAN tag in Rx path, VLAN hardware support
43 was not implemented yet.
44 - CAM(Content Addressable Memory) based 32 entry perfect multicast/
47 o Implemented CAM based 32 entry perfect multicast filtering for
48 VT6105M. If number of multicast entry is greater than 32, vr(4)
49 uses traditional hash based filtering.
51 * rip out the disk partitioning from the disk subsystem and implement it in a more general fashion
52 - crazy idea: as dm targets with an auto-configuration option!
54 * sync some more opencrypto from OpenBSD
56 * ATA (automatic) spindown (see FreeBSD current)
59 http://svn.freebsd.org/viewvc/base?view=revision&revision=127969
63 - some incorrect accounting going on, don't remember details :)
66 - make it work without whiteout
73 * RedZone, a buffer corruption protection for the kernel malloc(9) facility has been implemented.
74 - This detects both buffer underflows and overflows at runtime on free(9) and realloc(9),
75 and prints backtraces from where memory was allocated and from where it was freed.
78 * port uart driver (?)
80 * port wscons (?) or update syscons
81 - probably way too much effort (wscons)
85 - wrapper is included for userland; should be easy to port
86 - http://svn.freebsd.org/viewvc/base?view=revision&revision=184610
87 - http://turbocat.net/~hselasky/usb4bsd/
88 - http://gitweb.dragonflybsd.org/~polachok/dragonfly.git/shortlog/refs/heads/usb2
90 * suspend/resume for SMP x86
91 - http://lists.freebsd.org/pipermail/freebsd-acpi/2008-May/004879.html
93 * AMD64 suspend/resume
94 - http://svn.freebsd.org/viewvc/base?view=revision&revision=189903
101 [alexh@leaf:~/home] $ roundup-server -p 8080 bt=bugtracker
104 -05:48- : dillon@: no, double frees to the object cache are nasty. It can't detect them. the object
105 winds up in the magazine array twice
106 -05:48- : dillon@: (and possibly different magazines, too)
107 -05:49- : alexh@: can't I just write some magic to a free object on the first objcache_put and check
108 if it's there on objcache_put?
109 -05:49- : alexh@: and clear it on objcache_get, anyways
110 -05:50- : dillon@: no, because the object is still may have live-initialized fields
111 -05:50- : dillon@: because it hasn't been dtor'ed yet (one of the features of the objcache, to avoid
112 having to reinitialize objects every time)
113 -05:50- : dillon@: the mbuf code uses that feature I think, probably other bits too
114 -05:51- : dillon@: theoretically we could allocate slightly larger objects and store a magic number at
115 offset [-1] or something like that, but it gets a little iffy doing that
116 -05:52- : dillon@: the objcache with the objcache malloc default could probably do something like that
118 -05:52- : dillon@: I don't consider memory tracking to be a huge issue w/ dragonfly, though I like the
119 idea of being able to do it. It is a much bigger problem in FreeBSD due to the
120 large number of committers
123 -05:55- : dillon@: For the slab allocator you may be able to do something using the Zone header.
124 -05:55- : dillon@: the slab allocator in fact I think already has optional code to allocate a tracking
125 bitmap to detect double-frees
126 -05:56- : dillon@: sorry, I just remembered the bit about the power-of-2 allocations
127 -05:56- : dillon@: for example, power-of-2-sized allocations are guaranteed not only to be aligned on
128 that particular size boundary, but also to not cross a PAGE_BOUNDARY (unless the
130 -05:57- : dillon@: various subsystems such as AHCI depend on that behavior to allocate system
131 structures for which the chipsets only allow one DMA descriptor.
132 -05:59- : alexh@: http://svn.freebsd.org/viewvc/base/head/sys/vm/redzone.c?view=markup&pathrev=155086
133 < this is redzone. it basically calls redzone_addr_ntor() to increase the size in
134 malloc(), and then redzone_setup() just before returning the chunk
135 -06:02- : dillon@: jeeze. that looks horrible.
136 -06:03- : alexh@: I don't quite get that nsize + redzone_roundup(nsize)
137 -06:03- : dillon@: I don't get it either. It would completely break power-of-2-sized alignments in the
139 -06:04- : dillon@: hmmm. well, no it won't break them, but the results are oging to be weird
140 -06:04- : dillon@: ick.
142 -06:15- : dillon@: if the original request is a power of 2 the redzone adjusted request must be a power
144 -06:15- : dillon@: basically
145 -06:16- : dillon@: so original request 64, redzone request must be 128, 256, 512, 1024, etc.
146 -06:16- : alexh@: yah, k
147 -06:16- : dillon@: original request 32, current redzone code would be 32+128 which is WRONG.
148 -06:16- : alexh@: how big is PAGE_SIZE ?
149 -06:16- : dillon@: 4096 on i386 and amd64
150 -06:17- : alexh@: and one single malloc can't be bigger than that?
151 -06:17- : dillon@: I'm fairly sure our kmalloc does not guarantee alignment past PAGE_SIZE (that is,
152 the alignment will be only PAGE_SIZE eve if you allocate PAGE_SIZE*2)
153 -06:17- : dillon@: a single kmalloc can be larger then PAGE_SIZe
154 -06:18- : dillon@: it will use the zone up to around 1/2 the zone size (~64KB I think), after which it
155 allocates pages directly with the kernel kvm allocator
156 -06:18- : dillon@: if you look at the kmalloc code you will see the check for oversized allocations
157 -06:18- : alexh@: yah, saw that
158 -06:18- : alexh@: "handle large allocations directly"
159 -06:19- : alexh@: not sure how to do this, really, as the size is obviously also changed in
161 -06:20- : alexh@: but kmem_slab_alloc isn't called always, is it?
162 -06:20- : alexh@: only if the req doesn't fit into an existant zone
163 -06:20- : dillon@: right
164 -06:20- : dillon@: you don't want to redzone the zone allocation itself