(no commit message)
[ikiwiki.git] / docs / developer / alexh_todo / index.mdwn
CommitLineData
b475db1d 1## Note: this is my personal todo and ideas list (alexh@)
49a92aaf 2
fcd780c0 3* add a non-persistent unionfs VFS
4 - while this does not fully replace unionfs, it would deal with a few of the situations where unionfs is useful
5 - additionally the complexity is much lower, since everything can be kept in memory
6
7* bring the samba3 hammer shadow copy foo to maturity
8
d2e33566 9* add a communication channel mechanism to dm
10 - essentially a way to send messages and receive responses to dm target instances
11
6e280d9e 12* Add TrueCrypt support
13 - Simply add a userland tool that sets dm_target_crypt up with the right parameters, extracted from the TrueCrypt header
5d279a75 14 - https://github.com/bwalex/tc-play <- done!
15 - needs OpenSSL with XTS support; probably next major release
6e280d9e 16
e09e6796 17* Update cryptsetup
18
85fee687 19* Keep opencrypto up-to-date
744c3358 20 - http://www.openbsd.org/cgi-bin/cvsweb/src/sys/crypto/
21
e09e6796 22* Port hwpmc & dig into (boot-up) performance
23
1bf667d1 24* linuxulator
6c260381 25 - port to x86_64
26 - separate out common arch parts (linprocfs, for example)
53af51d2 27
e09e6796 28* Fix the crash analysis script (or rather the programs it calls [some segfault])
0d62372b 29
f3169700 30* Take a look at updating lvm/dm/libdevicemapper
53af51d2 31
f3169700 32* Take a look at importing libdm from NetBSD
33 - would allow tc-play and similarly newly developed foo to avoid GPL license issues
1bf667d1 34
6c260381 35* rip out the disk partitioning from the disk subsystem and implement it in a more general fashion
36 - crazy idea: as dm targets with an auto-configuration option!
b7038718 37 - would require to be able to create dm targets with an arbitrary name and not in /dev/mapper
6c260381 38
1bf667d1 39* ATA (automatic) spindown (see FreeBSD current)
0d62372b 40
41* Update callout
42 http://svn.freebsd.org/viewvc/base?view=revision&revision=127969
43
44* inv ctxsw rusage
45 - see irc logs
46 - some incorrect accounting going on, don't remember details :)
47
48* unionfs update
49 - make it work without whiteout
50
51
52
b475db1d 53###Boring:
1bf667d1 54
772830b5 55* Fix ipsec, get rid of old ipsec
56
f4ce1567 57* sync up vr
58 - Added VT6105M specific register definitions. VT6105M has the following hardware capabilities.
59 - Tx/Rx IP/TCP/UDP checksum offload.
60 - VLAN hardware tag insertion/extraction. Due to lack of information
61 for getting extracted VLAN tag in Rx path, VLAN hardware support
62 was not implemented yet.
63 - CAM(Content Addressable Memory) based 32 entry perfect multicast/
64 VLAN filtering.
65 - 8 priority queues.
66 o Implemented CAM based 32 entry perfect multicast filtering for
67 VT6105M. If number of multicast entry is greater than 32, vr(4)
68 uses traditional hash based filtering.
1bf667d1 69
70* RedZone, a buffer corruption protection for the kernel malloc(9) facility has been implemented.
71 - This detects both buffer underflows and overflows at runtime on free(9) and realloc(9),
72 and prints backtraces from where memory was allocated and from where it was freed.
73 - see irc log below.
74
75* port uart driver (?)
76
0d62372b 77* port wscons (?) or update syscons
78 - probably way too much effort (wscons)
79
80
81* port usb4bsd
82 - wrapper is included for userland; should be easy to port
83 - http://svn.freebsd.org/viewvc/base?view=revision&revision=184610
84 - http://turbocat.net/~hselasky/usb4bsd/
85 - http://gitweb.dragonflybsd.org/~polachok/dragonfly.git/shortlog/refs/heads/usb2
86
1bf667d1 87* suspend/resume for SMP x86
88 - http://lists.freebsd.org/pipermail/freebsd-acpi/2008-May/004879.html
89
1bf667d1 90* AMD64 suspend/resume
91 - http://svn.freebsd.org/viewvc/base?view=revision&revision=189903
6c260381 92
93
6c260381 94* text dumps
95
1bf667d1 96
1bf667d1 97
98[alexh@leaf:~/home] $ roundup-server -p 8080 bt=bugtracker
99
53af51d2 100<pre>
1bf667d1 101-05:48- : dillon@: no, double frees to the object cache are nasty. It can't detect them. the object
102 winds up in the magazine array twice
103-05:48- : dillon@: (and possibly different magazines, too)
104-05:49- : alexh@: can't I just write some magic to a free object on the first objcache_put and check
105 if it's there on objcache_put?
106-05:49- : alexh@: and clear it on objcache_get, anyways
107-05:50- : dillon@: no, because the object is still may have live-initialized fields
108-05:50- : dillon@: because it hasn't been dtor'ed yet (one of the features of the objcache, to avoid
109 having to reinitialize objects every time)
110-05:50- : dillon@: the mbuf code uses that feature I think, probably other bits too
111-05:51- : dillon@: theoretically we could allocate slightly larger objects and store a magic number at
112 offset [-1] or something like that, but it gets a little iffy doing that
113-05:52- : dillon@: the objcache with the objcache malloc default could probably do something like that
114 I guess.
115-05:52- : dillon@: I don't consider memory tracking to be a huge issue w/ dragonfly, though I like the
116 idea of being able to do it. It is a much bigger problem in FreeBSD due to the
117 large number of committers
118
119
120-05:55- : dillon@: For the slab allocator you may be able to do something using the Zone header.
121-05:55- : dillon@: the slab allocator in fact I think already has optional code to allocate a tracking
122 bitmap to detect double-frees
123-05:56- : dillon@: sorry, I just remembered the bit about the power-of-2 allocations
124-05:56- : dillon@: for example, power-of-2-sized allocations are guaranteed not only to be aligned on
125 that particular size boundary, but also to not cross a PAGE_BOUNDARY (unless the
126 size is > PAGE_SIZE)
127-05:57- : dillon@: various subsystems such as AHCI depend on that behavior to allocate system
128 structures for which the chipsets only allow one DMA descriptor.
129-05:59- : alexh@: http://svn.freebsd.org/viewvc/base/head/sys/vm/redzone.c?view=markup&pathrev=155086
130 < this is redzone. it basically calls redzone_addr_ntor() to increase the size in
131 malloc(), and then redzone_setup() just before returning the chunk
132-06:02- : dillon@: jeeze. that looks horrible.
133-06:03- : alexh@: I don't quite get that nsize + redzone_roundup(nsize)
134-06:03- : dillon@: I don't get it either. It would completely break power-of-2-sized alignments in the
135 original request
136-06:04- : dillon@: hmmm. well, no it won't break them, but the results are oging to be weird
137-06:04- : dillon@: ick.
138
139-06:15- : dillon@: if the original request is a power of 2 the redzone adjusted request must be a power
140 of 2
141-06:15- : dillon@: basically
142-06:16- : dillon@: so original request 64, redzone request must be 128, 256, 512, 1024, etc.
143-06:16- : alexh@: yah, k
144-06:16- : dillon@: original request 32, current redzone code would be 32+128 which is WRONG.
145-06:16- : alexh@: how big is PAGE_SIZE ?
146-06:16- : dillon@: 4096 on i386 and amd64
147-06:17- : alexh@: and one single malloc can't be bigger than that?
148-06:17- : dillon@: I'm fairly sure our kmalloc does not guarantee alignment past PAGE_SIZE (that is,
149 the alignment will be only PAGE_SIZE eve if you allocate PAGE_SIZE*2)
150-06:17- : dillon@: a single kmalloc can be larger then PAGE_SIZe
151-06:18- : dillon@: it will use the zone up to around 1/2 the zone size (~64KB I think), after which it
152 allocates pages directly with the kernel kvm allocator
153-06:18- : dillon@: if you look at the kmalloc code you will see the check for oversized allocations
154-06:18- : alexh@: yah, saw that
155-06:18- : alexh@: "handle large allocations directly"
156-06:19- : alexh@: not sure how to do this, really, as the size is obviously also changed in
157 kmem_slab_alloc
158-06:20- : alexh@: but kmem_slab_alloc isn't called always, is it?
159-06:20- : alexh@: only if the req doesn't fit into an existant zone
160-06:20- : dillon@: right
161-06:20- : dillon@: you don't want to redzone the zone allocation itself
53af51d2 162
163</pre>