1 ### Note this is my personal todo and ideas list, don't jump on me if you don't like something or think that it isn't important enough.
5 - separate out common arch parts (linprocfs, for example)
11 * Take a look at updating lvm/dm/libdm
14 * port wscons (?) or update syscons
15 - probably way too much effort (wscons)
19 - wrapper is included for userland; should be easy to port
20 - http://svn.freebsd.org/viewvc/base?view=revision&revision=184610
21 - http://turbocat.net/~hselasky/usb4bsd/
22 - http://gitweb.dragonflybsd.org/~polachok/dragonfly.git/shortlog/refs/heads/usb2
26 o Added VT6105M specific register definitions. VT6105M has the
27 following hardware capabilities.
28 - Tx/Rx IP/TCP/UDP checksum offload.
29 - VLAN hardware tag insertion/extraction. Due to lack of information
30 for getting extracted VLAN tag in Rx path, VLAN hardware support
31 was not implemented yet.
32 - CAM(Content Addressable Memory) based 32 entry perfect multicast/
35 o Implemented CAM based 32 entry perfect multicast filtering for
36 VT6105M. If number of multicast entry is greater than 32, vr(4)
37 uses traditional hash based filtering.
40 * rip out the disk partitioning from the disk subsystem and implement it in a more general fashion
41 - crazy idea: as dm targets with an auto-configuration option!
44 * sync some more opencrypto from OpenBSD
46 * ATA (automatic) spindown (see FreeBSD current)
49 * RedZone, a buffer corruption protection for the kernel malloc(9) facility has been implemented.
50 - This detects both buffer underflows and overflows at runtime on free(9) and realloc(9),
51 and prints backtraces from where memory was allocated and from where it was freed.
54 * port uart driver (?)
56 * suspend/resume for SMP x86
57 - http://lists.freebsd.org/pipermail/freebsd-acpi/2008-May/004879.html
59 * AMD64 suspend/resume
60 - http://svn.freebsd.org/viewvc/base?view=revision&revision=189903
69 http://svn.freebsd.org/viewvc/base?view=revision&revision=127969
73 - use namecache tricks as nullfs does
74 - make it work without whiteout
85 devclass_find_internal(const char *classname, const char *parentname,
100 [alexh@leaf:~/home] $ roundup-server -p 8080 bt=bugtracker
103 -05:48- : dillon@: no, double frees to the object cache are nasty. It can't detect them. the object
104 winds up in the magazine array twice
105 -05:48- : dillon@: (and possibly different magazines, too)
106 -05:49- : alexh@: can't I just write some magic to a free object on the first objcache_put and check
107 if it's there on objcache_put?
108 -05:49- : alexh@: and clear it on objcache_get, anyways
109 -05:50- : dillon@: no, because the object is still may have live-initialized fields
110 -05:50- : dillon@: because it hasn't been dtor'ed yet (one of the features of the objcache, to avoid
111 having to reinitialize objects every time)
112 -05:50- : dillon@: the mbuf code uses that feature I think, probably other bits too
113 -05:51- : dillon@: theoretically we could allocate slightly larger objects and store a magic number at
114 offset [-1] or something like that, but it gets a little iffy doing that
115 -05:52- : dillon@: the objcache with the objcache malloc default could probably do something like that
117 -05:52- : dillon@: I don't consider memory tracking to be a huge issue w/ dragonfly, though I like the
118 idea of being able to do it. It is a much bigger problem in FreeBSD due to the
119 large number of committers
122 -05:55- : dillon@: For the slab allocator you may be able to do something using the Zone header.
123 -05:55- : dillon@: the slab allocator in fact I think already has optional code to allocate a tracking
124 bitmap to detect double-frees
125 -05:56- : dillon@: sorry, I just remembered the bit about the power-of-2 allocations
126 -05:56- : dillon@: for example, power-of-2-sized allocations are guaranteed not only to be aligned on
127 that particular size boundary, but also to not cross a PAGE_BOUNDARY (unless the
129 -05:57- : dillon@: various subsystems such as AHCI depend on that behavior to allocate system
130 structures for which the chipsets only allow one DMA descriptor.
131 -05:59- : alexh@: http://svn.freebsd.org/viewvc/base/head/sys/vm/redzone.c?view=markup&pathrev=155086
132 < this is redzone. it basically calls redzone_addr_ntor() to increase the size in
133 malloc(), and then redzone_setup() just before returning the chunk
134 -06:02- : dillon@: jeeze. that looks horrible.
135 -06:03- : alexh@: I don't quite get that nsize + redzone_roundup(nsize)
136 -06:03- : dillon@: I don't get it either. It would completely break power-of-2-sized alignments in the
138 -06:04- : dillon@: hmmm. well, no it won't break them, but the results are oging to be weird
139 -06:04- : dillon@: ick.
141 -06:15- : dillon@: if the original request is a power of 2 the redzone adjusted request must be a power
143 -06:15- : dillon@: basically
144 -06:16- : dillon@: so original request 64, redzone request must be 128, 256, 512, 1024, etc.
145 -06:16- : alexh@: yah, k
146 -06:16- : dillon@: original request 32, current redzone code would be 32+128 which is WRONG.
147 -06:16- : alexh@: how big is PAGE_SIZE ?
148 -06:16- : dillon@: 4096 on i386 and amd64
149 -06:17- : alexh@: and one single malloc can't be bigger than that?
150 -06:17- : dillon@: I'm fairly sure our kmalloc does not guarantee alignment past PAGE_SIZE (that is,
151 the alignment will be only PAGE_SIZE eve if you allocate PAGE_SIZE*2)
152 -06:17- : dillon@: a single kmalloc can be larger then PAGE_SIZe
153 -06:18- : dillon@: it will use the zone up to around 1/2 the zone size (~64KB I think), after which it
154 allocates pages directly with the kernel kvm allocator
155 -06:18- : dillon@: if you look at the kmalloc code you will see the check for oversized allocations
156 -06:18- : alexh@: yah, saw that
157 -06:18- : alexh@: "handle large allocations directly"
158 -06:19- : alexh@: not sure how to do this, really, as the size is obviously also changed in
160 -06:20- : alexh@: but kmem_slab_alloc isn't called always, is it?
161 -06:20- : alexh@: only if the req doesn't fit into an existant zone
162 -06:20- : dillon@: right
163 -06:20- : dillon@: you don't want to redzone the zone allocation itself