1 ## Note: this is my personal todo and ideas list (alexh@)
2 * dm pdev bounds checking
6 foocrypt: 1 dependencies : (0, 0)
7 foocrypt2: 1 dependencies : (0, 0)
8 foocrypt3: 1 dependencies : (0, 0)
11 foocrypt: 0 1021944 crypt aes-cbc-essiv:sha256 0000000000000000000000000000000000000000000000000000000000000000 0 /dev/ad0s0a 2056
12 foocrypt2: 0 1021944 crypt aes-cbc-essiv:sha256 0000000000000000000000000000000000000000000000000000000000000000 0 /dev/ad0s0a 2056
13 foocrypt3: 0 1021944 crypt aes-cbc-essiv:sha256 0000000000000000000000000000000000000000000000000000000000000000 0 /dev/ad0s0a 2056
16 * add a non-persistent unionfs VFS
17 - while this does not fully replace unionfs, it would deal with a few of the situations where unionfs is useful
18 - additionally the complexity is much lower, since everything can be kept in memory
20 * bring the samba3 hammer shadow copy foo to maturity
22 * add a communication channel mechanism to dm
23 - essentially a way to send messages and receive responses to dm target instances
26 - Look at plain64 iv support, what it implies vs plain.
27 - Think about benbi iv support
29 * Add TrueCrypt support
30 - Simply add a userland tool that sets dm_target_crypt up with the right parameters, extracted from the TrueCrypt header
31 - https://github.com/bwalex/tc-play <- done!
32 - needs OpenSSL with XTS support; probably next major release
37 - http://www.openbsd.org/cgi-bin/cvsweb/src/sys/crypto/
38 - think about support for other ciphers, such as Serpent and Twofish
40 * Port hwpmc & dig into (boot-up) performance
44 - separate out common arch parts (linprocfs, for example)
46 * Fix the crash analysis script (or rather the programs it calls [some segfault])
48 * Take a look at updating lvm/dm/libdevicemapper
50 * Take a look at importing libdm from NetBSD
51 - would allow tc-play and similarly newly developed foo to avoid GPL license issues
53 * rip out the disk partitioning from the disk subsystem and implement it in a more general fashion
54 - crazy idea: as dm targets with an auto-configuration option!
55 - would require to be able to create dm targets with an arbitrary name and not in /dev/mapper
57 * ATA (automatic) spindown (see FreeBSD current)
60 http://svn.freebsd.org/viewvc/base?view=revision&revision=127969
64 - some incorrect accounting going on, don't remember details :)
67 - make it work without whiteout
74 - Added VT6105M specific register definitions. VT6105M has the following hardware capabilities.
75 - Tx/Rx IP/TCP/UDP checksum offload.
76 - VLAN hardware tag insertion/extraction. Due to lack of information
77 for getting extracted VLAN tag in Rx path, VLAN hardware support
78 was not implemented yet.
79 - CAM(Content Addressable Memory) based 32 entry perfect multicast/
82 o Implemented CAM based 32 entry perfect multicast filtering for
83 VT6105M. If number of multicast entry is greater than 32, vr(4)
84 uses traditional hash based filtering.
86 * RedZone, a buffer corruption protection for the kernel malloc(9) facility has been implemented.
87 - This detects both buffer underflows and overflows at runtime on free(9) and realloc(9),
88 and prints backtraces from where memory was allocated and from where it was freed.
91 * port uart driver (?)
93 * port wscons (?) or update syscons
94 - probably way too much effort (wscons)
98 - wrapper is included for userland; should be easy to port
99 - http://svn.freebsd.org/viewvc/base?view=revision&revision=184610
100 - http://turbocat.net/~hselasky/usb4bsd/
101 - http://gitweb.dragonflybsd.org/~polachok/dragonfly.git/shortlog/refs/heads/usb2
103 * suspend/resume for SMP x86
104 - http://lists.freebsd.org/pipermail/freebsd-acpi/2008-May/004879.html
106 * AMD64 suspend/resume
107 - http://svn.freebsd.org/viewvc/base?view=revision&revision=189903
114 [alexh@leaf:~/home] $ roundup-server -p 8080 bt=bugtracker
117 -05:48- : dillon@: no, double frees to the object cache are nasty. It can't detect them. the object
118 winds up in the magazine array twice
119 -05:48- : dillon@: (and possibly different magazines, too)
120 -05:49- : alexh@: can't I just write some magic to a free object on the first objcache_put and check
121 if it's there on objcache_put?
122 -05:49- : alexh@: and clear it on objcache_get, anyways
123 -05:50- : dillon@: no, because the object is still may have live-initialized fields
124 -05:50- : dillon@: because it hasn't been dtor'ed yet (one of the features of the objcache, to avoid
125 having to reinitialize objects every time)
126 -05:50- : dillon@: the mbuf code uses that feature I think, probably other bits too
127 -05:51- : dillon@: theoretically we could allocate slightly larger objects and store a magic number at
128 offset [-1] or something like that, but it gets a little iffy doing that
129 -05:52- : dillon@: the objcache with the objcache malloc default could probably do something like that
131 -05:52- : dillon@: I don't consider memory tracking to be a huge issue w/ dragonfly, though I like the
132 idea of being able to do it. It is a much bigger problem in FreeBSD due to the
133 large number of committers
136 -05:55- : dillon@: For the slab allocator you may be able to do something using the Zone header.
137 -05:55- : dillon@: the slab allocator in fact I think already has optional code to allocate a tracking
138 bitmap to detect double-frees
139 -05:56- : dillon@: sorry, I just remembered the bit about the power-of-2 allocations
140 -05:56- : dillon@: for example, power-of-2-sized allocations are guaranteed not only to be aligned on
141 that particular size boundary, but also to not cross a PAGE_BOUNDARY (unless the
143 -05:57- : dillon@: various subsystems such as AHCI depend on that behavior to allocate system
144 structures for which the chipsets only allow one DMA descriptor.
145 -05:59- : alexh@: http://svn.freebsd.org/viewvc/base/head/sys/vm/redzone.c?view=markup&pathrev=155086
146 < this is redzone. it basically calls redzone_addr_ntor() to increase the size in
147 malloc(), and then redzone_setup() just before returning the chunk
148 -06:02- : dillon@: jeeze. that looks horrible.
149 -06:03- : alexh@: I don't quite get that nsize + redzone_roundup(nsize)
150 -06:03- : dillon@: I don't get it either. It would completely break power-of-2-sized alignments in the
152 -06:04- : dillon@: hmmm. well, no it won't break them, but the results are oging to be weird
153 -06:04- : dillon@: ick.
155 -06:15- : dillon@: if the original request is a power of 2 the redzone adjusted request must be a power
157 -06:15- : dillon@: basically
158 -06:16- : dillon@: so original request 64, redzone request must be 128, 256, 512, 1024, etc.
159 -06:16- : alexh@: yah, k
160 -06:16- : dillon@: original request 32, current redzone code would be 32+128 which is WRONG.
161 -06:16- : alexh@: how big is PAGE_SIZE ?
162 -06:16- : dillon@: 4096 on i386 and amd64
163 -06:17- : alexh@: and one single malloc can't be bigger than that?
164 -06:17- : dillon@: I'm fairly sure our kmalloc does not guarantee alignment past PAGE_SIZE (that is,
165 the alignment will be only PAGE_SIZE eve if you allocate PAGE_SIZE*2)
166 -06:17- : dillon@: a single kmalloc can be larger then PAGE_SIZe
167 -06:18- : dillon@: it will use the zone up to around 1/2 the zone size (~64KB I think), after which it
168 allocates pages directly with the kernel kvm allocator
169 -06:18- : dillon@: if you look at the kmalloc code you will see the check for oversized allocations
170 -06:18- : alexh@: yah, saw that
171 -06:18- : alexh@: "handle large allocations directly"
172 -06:19- : alexh@: not sure how to do this, really, as the size is obviously also changed in
174 -06:20- : alexh@: but kmem_slab_alloc isn't called always, is it?
175 -06:20- : alexh@: only if the req doesn't fit into an existant zone
176 -06:20- : dillon@: right
177 -06:20- : dillon@: you don't want to redzone the zone allocation itself