ia64/linux-2.6.18-xen.hg

view Documentation/networking/net-modules.txt @ 897:329ea0ccb344

balloon: try harder to balloon up under memory pressure.

Currently if the balloon driver is unable to increase the guest's
reservation it assumes the failure was due to reaching its full
allocation, gives up on the ballooning operation and records the limit
it reached as the "hard limit". The driver will not try again until
the target is set again (even to the same value).

However it is possible that ballooning has in fact failed due to
memory pressure in the host and therefore it is desirable to keep
attempting to reach the target in case memory becomes available. The
most likely scenario is that some guests are ballooning down while
others are ballooning up and therefore there is temporary memory
pressure while things stabilise. You would not expect a well behaved
toolstack to ask a domain to balloon to more than its allocation nor
would you expect it to deliberately over-commit memory by setting
balloon targets which exceed the total host memory.

This patch drops the concept of a hard limit and causes the balloon
driver to retry increasing the reservation on a timer in the same
manner as when decreasing the reservation.

Also if we partially succeed in increasing the reservation
(i.e. receive less pages than we asked for) then we may as well keep
those pages rather than returning them to Xen.

Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
author Keir Fraser <keir.fraser@citrix.com>
date Fri Jun 05 14:01:20 2009 +0100 (2009-06-05)
parents 831230e53067
children
line source
1 Wed 2-Aug-95 <matti.aarnio@utu.fi>
3 Linux network driver modules
5 Do not mistake this for "README.modules" at the top-level
6 directory! That document tells about modules in general, while
7 this one tells only about network device driver modules.
9 This is a potpourri of INSMOD-time(*) configuration options
10 (if such exists) and their default values of various modules
11 in the Linux network drivers collection.
13 Some modules have also hidden (= non-documented) tunable values.
14 The choice of not documenting them is based on general belief, that
15 the less the user needs to know, the better. (There are things that
16 driver developers can use, others should not confuse themselves.)
18 In many cases it is highly preferred that insmod:ing is done
19 ONLY with defining an explicit address for the card, AND BY
20 NOT USING AUTO-PROBING!
22 Now most cards have some explicitly defined base address that they
23 are compiled with (to avoid auto-probing, among other things).
24 If that compiled value does not match your actual configuration,
25 do use the "io=0xXXX" -parameter for the insmod, and give there
26 a value matching your environment.
28 If you are adventurous, you can ask the driver to autoprobe
29 by using the "io=0" parameter, however it is a potentially dangerous
30 thing to do in a live system. (If you don't know where the
31 card is located, you can try autoprobing, and after possible
32 crash recovery, insmod with proper IO-address..)
34 --------------------------
35 (*) "INSMOD-time" means when you load module with
36 /sbin/insmod you can feed it optional parameters.
37 See "man insmod".
38 --------------------------
41 8390 based Network Modules (Paul Gortmaker, Nov 12, 1995)
42 --------------------------
44 (Includes: smc-ultra, ne, wd, 3c503, hp, hp-plus, e2100 and ac3200)
46 The 8390 series of network drivers now support multiple card systems without
47 reloading the same module multiple times (memory efficient!) This is done by
48 specifying multiple comma separated values, such as:
50 insmod 3c503.o io=0x280,0x300,0x330,0x350 xcvr=0,1,0,1
52 The above would have the one module controlling four 3c503 cards, with card 2
53 and 4 using external transceivers. The "insmod" manual describes the usage
54 of comma separated value lists.
56 It is *STRONGLY RECOMMENDED* that you supply "io=" instead of autoprobing.
57 If an "io=" argument is not supplied, then the ISA drivers will complain
58 about autoprobing being not recommended, and begrudgingly autoprobe for
59 a *SINGLE CARD ONLY* -- if you want to use multiple cards you *have* to
60 supply an "io=0xNNN,0xQQQ,..." argument.
62 The ne module is an exception to the above. A NE2000 is essentially an
63 8390 chip, some bus glue and some RAM. Because of this, the ne probe is
64 more invasive than the rest, and so at boot we make sure the ne probe is
65 done last of all the 8390 cards (so that it won't trip over other 8390 based
66 cards) With modules we can't ensure that all other non-ne 8390 cards have
67 already been found. Because of this, the ne module REQUIRES an "io=0xNNN"
68 argument passed in via insmod. It will refuse to autoprobe.
70 It is also worth noting that auto-IRQ probably isn't as reliable during
71 the flurry of interrupt activity on a running machine. Cards such as the
72 ne2000 that can't get the IRQ setting from an EEPROM or configuration
73 register are probably best supplied with an "irq=M" argument as well.
76 ----------------------------------------------------------------------
77 Card/Module List - Configurable Parameters and Default Values
78 ----------------------------------------------------------------------
80 3c501.c:
81 io = 0x280 IO base address
82 irq = 5 IRQ
83 (Probes ports: 0x280, 0x300)
85 3c503.c:
86 io = 0 (It will complain if you don't supply an "io=0xNNN")
87 irq = 0 (IRQ software selected by driver using autoIRQ)
88 xcvr = 0 (Use xcvr=1 to select external transceiver.)
89 (Probes ports: 0x300, 0x310, 0x330, 0x350, 0x250, 0x280, 0x2A0, 0x2E0)
91 3c505.c:
92 io = 0
93 irq = 0
94 dma = 6 (not autoprobed)
95 (Probes ports: 0x300, 0x280, 0x310)
97 3c507.c:
98 io = 0x300
99 irq = 0
100 (Probes ports: 0x300, 0x320, 0x340, 0x280)
102 3c509.c:
103 io = 0
104 irq = 0
105 ( Module load-time probing Works reliably only on EISA, ISA ID-PROBE
106 IS NOT RELIABLE! Compile this driver statically into kernel for
107 now, if you need it auto-probing on an ISA-bus machine. )
109 8390.c:
110 (No public options, several other modules need this one)
112 a2065.c:
113 Since this is a Zorro board, it supports full autoprobing, even for
114 multiple boards. (m68k/Amiga)
116 ac3200.c:
117 io = 0 (Checks 0x1000 to 0x8fff in 0x1000 intervals)
118 irq = 0 (Read from config register)
119 (EISA probing..)
121 apricot.c:
122 io = 0x300 (Can't be altered!)
123 irq = 10
125 arcnet.c:
126 io = 0
127 irqnum = 0
128 shmem = 0
129 num = 0
130 DO SET THESE MANUALLY AT INSMOD!
131 (When probing, looks at the following possible addresses:
132 Suggested ones:
133 0x300, 0x2E0, 0x2F0, 0x2D0
134 Other ones:
135 0x200, 0x210, 0x220, 0x230, 0x240, 0x250, 0x260, 0x270,
136 0x280, 0x290, 0x2A0, 0x2B0, 0x2C0,
137 0x310, 0x320, 0x330, 0x340, 0x350, 0x360, 0x370,
138 0x380, 0x390, 0x3A0, 0x3E0, 0x3F0 )
140 ariadne.c:
141 Since this is a Zorro board, it supports full autoprobing, even for
142 multiple boards. (m68k/Amiga)
144 at1700.c:
145 io = 0x260
146 irq = 0
147 (Probes ports: 0x260, 0x280, 0x2A0, 0x240, 0x340, 0x320, 0x380, 0x300)
149 atari_bionet.c:
150 Supports full autoprobing. (m68k/Atari)
152 atari_pamsnet.c:
153 Supports full autoprobing. (m68k/Atari)
155 atarilance.c:
156 Supports full autoprobing. (m68k/Atari)
158 atp.c: *Not modularized*
159 (Probes ports: 0x378, 0x278, 0x3BC;
160 fixed IRQs: 5 and 7 )
162 cops.c:
163 io = 0x240
164 irq = 5
165 nodeid = 0 (AutoSelect = 0, NodeID 1-254 is hand selected.)
166 (Probes ports: 0x240, 0x340, 0x200, 0x210, 0x220, 0x230, 0x260,
167 0x2A0, 0x300, 0x310, 0x320, 0x330, 0x350, 0x360)
169 de4x5.c:
170 io = 0x000b
171 irq = 10
172 is_not_dec = 0 -- For non-DEC card using DEC 21040/21041/21140 chip, set this to 1
173 (EISA, and PCI probing)
175 de600.c:
176 de600_debug = 0
177 (On port 0x378, irq 7 -- lpt1; compile time configurable)
179 de620.c:
180 bnc = 0, utp = 0 <-- Force media by setting either.
181 io = 0x378 (also compile-time configurable)
182 irq = 7
184 depca.c:
185 io = 0x200
186 irq = 7
187 (Probes ports: ISA: 0x300, 0x200;
188 EISA: 0x0c00 )
190 dummy.c:
191 No options
193 e2100.c:
194 io = 0 (It will complain if you don't supply an "io=0xNNN")
195 irq = 0 (IRQ software selected by driver)
196 mem = 0 (Override default shared memory start of 0xd0000)
197 xcvr = 0 (Use xcvr=1 to select external transceiver.)
198 (Probes ports: 0x300, 0x280, 0x380, 0x220)
200 eepro.c:
201 io = 0x200
202 irq = 0
203 (Probes ports: 0x200, 0x240, 0x280, 0x2C0, 0x300, 0x320, 0x340, 0x360)
205 eexpress.c:
206 io = 0x300
207 irq = 0 (IRQ value read from EEPROM)
208 (Probes ports: 0x300, 0x270, 0x320, 0x340)
210 eql.c:
211 (No parameters)
213 ewrk3.c:
214 io = 0x300
215 irq = 5
216 (With module no autoprobing!
217 On EISA-bus does EISA probing.
218 Static linkage probes ports on ISA bus:
219 0x100, 0x120, 0x140, 0x160, 0x180, 0x1A0, 0x1C0,
220 0x200, 0x220, 0x240, 0x260, 0x280, 0x2A0, 0x2C0, 0x2E0,
221 0x300, 0x340, 0x360, 0x380, 0x3A0, 0x3C0)
223 hp-plus.c:
224 io = 0 (It will complain if you don't supply an "io=0xNNN")
225 irq = 0 (IRQ read from configuration register)
226 (Probes ports: 0x200, 0x240, 0x280, 0x2C0, 0x300, 0x320, 0x340)
228 hp.c:
229 io = 0 (It will complain if you don't supply an "io=0xNNN")
230 irq = 0 (IRQ software selected by driver using autoIRQ)
231 (Probes ports: 0x300, 0x320, 0x340, 0x280, 0x2C0, 0x200, 0x240)
233 hp100.c:
234 hp100_port = 0 (IO-base address)
235 (Does EISA-probing, if on EISA-slot;
236 On ISA-bus probes all ports from 0x100 thru to 0x3E0
237 in increments of 0x020)
239 hydra.c:
240 Since this is a Zorro board, it supports full autoprobing, even for
241 multiple boards. (m68k/Amiga)
243 ibmtr.c:
244 io = 0xa20, 0xa24 (autoprobed by default)
245 irq = 0 (driver cannot select irq - read from hardware)
246 mem = 0 (shared memory base set at 0xd0000 and not yet
247 able to override thru mem= parameter.)
249 lance.c: *Not modularized*
250 (PCI, and ISA probing; "CONFIG_PCI" needed for PCI support)
251 (Probes ISA ports: 0x300, 0x320, 0x340, 0x360)
253 loopback.c: *Static kernel component*
255 ne.c:
256 io = 0 (Explicitly *requires* an "io=0xNNN" value)
257 irq = 0 (Tries to determine configured IRQ via autoIRQ)
258 (Probes ports: 0x300, 0x280, 0x320, 0x340, 0x360)
260 net_init.c: *Static kernel component*
262 ni52.c: *Not modularized*
263 (Probes ports: 0x300, 0x280, 0x360, 0x320, 0x340
264 mems: 0xD0000, 0xD2000, 0xC8000, 0xCA000,
265 0xD4000, 0xD6000, 0xD8000 )
267 ni65.c: *Not modularized* **16MB MEMORY BARRIER BUG**
268 (Probes ports: 0x300, 0x320, 0x340, 0x360)
270 pi2.c: *Not modularized* (well, NON-STANDARD modularization!)
271 Only one card supported at this time.
272 (Probes ports: 0x380, 0x300, 0x320, 0x340, 0x360, 0x3A0)
274 plip.c:
275 io = 0
276 irq = 0 (by default, uses IRQ 5 for port at 0x3bc, IRQ 7
277 for port at 0x378, and IRQ 2 for port at 0x278)
278 (Probes ports: 0x278, 0x378, 0x3bc)
280 ppp.c:
281 No options (ppp-2.2+ has some, this is based on non-dynamic
282 version from ppp-2.1.2d)
284 seeq8005.c: *Not modularized*
285 (Probes ports: 0x300, 0x320, 0x340, 0x360)
287 skeleton.c: *Skeleton*
289 slhc.c:
290 No configuration parameters
292 slip.c:
293 slip_maxdev = 256 (default value from SL_NRUNIT on slip.h)
296 smc-ultra.c:
297 io = 0 (It will complain if you don't supply an "io=0xNNN")
298 irq = 0 (IRQ val. read from EEPROM)
299 (Probes ports: 0x200, 0x220, 0x240, 0x280, 0x300, 0x340, 0x380)
301 tulip.c: *Partial modularization*
302 (init-time memory allocation makes problems..)
304 tunnel.c:
305 No insmod parameters
307 wavelan.c:
308 io = 0x390 (Settable, but change not recommended)
309 irq = 0 (Not honoured, if changed..)
311 wd.c:
312 io = 0 (It will complain if you don't supply an "io=0xNNN")
313 irq = 0 (IRQ val. read from EEPROM, ancient cards use autoIRQ)
314 mem = 0 (Force shared-memory on address 0xC8000, or whatever..)
315 mem_end = 0 (Force non-std. mem. size via supplying mem_end val.)
316 (eg. for 32k WD8003EBT, use mem=0xd0000 mem_end=0xd8000)
317 (Probes ports: 0x300, 0x280, 0x380, 0x240)
319 znet.c: *Not modularized*
320 (Only one device on Zenith Z-Note (notebook?) systems,
321 configuration information from (EE)PROM)