Предыдущая тема :: Следующая тема |
Автор |
Сообщение |
PAMAGITE Гость
|
Добавлено: Сб Июл 17 2004 03:26 Заголовок сообщения: Simple Question-With no Answer???!? |
|
|
Hi my name is Michael and i have specific question about HDD read/write sequence.
Is there any way to uderstand in program way when the data requsted from HDD is taken
from the physical surface or from somewhere else(cahe,buffer,or...)?
Thanks!
Michael |
|
Вернуться к началу |
|
![](templates/subSilver/images/spacer.gif) |
grayrat
Зарегистрирован: 30.06.2003 Сообщения: 189
|
Добавлено: Сб Июл 17 2004 16:55 Заголовок сообщения: |
|
|
I guess the only way to undersand where the data were read from is to estimate time of access |
|
Вернуться к началу |
|
![](templates/subSilver/images/spacer.gif) |
совсем незнакомый
Зарегистрирован: 24.12.2003 Сообщения: 183 Откуда: Israel
|
Добавлено: Сб Июл 17 2004 22:05 Заголовок сообщения: Re: Simple Question-With no Answer???!? |
|
|
PAMAGITE писал(а): | Hi my name is Michael and i have specific question about HDD read/write sequence.
Is there any way to uderstand in program way when the data requsted from HDD is taken
from the physical surface or from somewhere else(cahe,buffer,or...)?
Thanks!
Michael |
if you mean CPU cache & let us assume you're familiar with:
1. OS version & payload for swap in/out of the VM module.
2. IDE controller's model and timings,
3. HDD timings etc.
it is possible.
if you use linux/ free *BSD, and the kernel is compiled with debugging symbols you can debug it and know specifically, so it is considerably easy to accomplish.
in other case [if you don't consider HDD's controller cache as HDD's surface] i don't know if there is a non JTAG way of knowing exactly what was going on, because the delay considerations are probability based,
and you may get really fast response from the surface, while thinking it was from the HDD's cache.
Cheers. |
|
Вернуться к началу |
|
![](templates/subSilver/images/spacer.gif) |
Michael. Гость
|
Добавлено: Сб Июл 17 2004 22:55 Заголовок сообщения: Read/Write Seqence |
|
|
Hi again,
Well no i'm talking only for the hdd-cache,if you are familiar with dos/win98,i need the answer because i need to know how and if Dos/win98 io.sys change this buffer/cache from where hdd takes the requested data.I have some experience in monitoring the boot process and especially the loading of io.sys in RAM and the executing.I'm almost sure that io.sys make some changes to this buffer/cache,'cause i dont see why when the data on the hdd is say(01 01 ff a0...) and you gave this data to io.sys but in modifyed state(via int 13h) the next time when the request is for the same sector int 13h returns the modifyed state or completely diffrent data that is not actually on the hdd surface(pattern).
So how to understand when the data is taken from the physical surface and when not (estimate time of access is not working in this case).
I have to be sure when the data is taken from hdd surface so i can give the io.sys the correct data.
(I guess under hdd surface i mean hdd's controller cache.)Is there any possibility that io.sys manipulate this cache???
Thanks
Michael |
|
Вернуться к началу |
|
![](templates/subSilver/images/spacer.gif) |
|