Hi,
Maybe you should verify you don't have an alignment problem when using O_DIRECT.
When you try , from user space , to use O_DIRECT for reading
from a file, you must use page granularity for achieving page alignemnent.
according to the man (2) read , regarding O_DIRECT flag:
"...
Transfer
Hello
Tehuti Networks company is looking for talented programmer
to join Linux Driver group
Job description:
* writing/extending network driver
* performance tuning, optimization, benchmarking
* implementing new features, like RDMA, TCP offloading etc
* creating (semi) automat
On Tue, 2005-08-30 at 14:40 +0300, guy keren wrote:
> On Tue, 30 Aug 2005, Gilboa Davara wrote:
>
> > > Anyway, if you don't know how to use O_DIRECT from inside the kernel,
> > > why don't you take a look at the part in teh kernel that implments
> > > O_DIRECT from user space and do the same thin
On Tue, 2005-08-30 at 14:40 +0300, guy keren wrote:
> On Tue, 30 Aug 2005, Gilboa Davara wrote:
>
> > > Anyway, if you don't know how to use O_DIRECT from inside the kernel,
> > > why don't you take a look at the part in teh kernel that implments
> > > O_DIRECT from user space and do the same thin
On Tue, 30 Aug 2005, Gilboa Davara wrote:
> > Anyway, if you don't know how to use O_DIRECT from inside the kernel,
> > why don't you take a look at the part in teh kernel that implments
> > O_DIRECT from user space and do the same thing?
>
> Already on it; However, this code is fairly complex. I
Gilboa Davara wrote:
Hello all,
Couple of weeks ago I started a thread called "File I/O within kernel
threads?" which ended with a call "for numbers". (Read: Is I/O within
kernel performance increase justifies the development overhead?)
Last week I took my user-land simulator (that takes a file,
On Tue, 2005-08-30 at 12:09 +0300, Gilad Ben-Yossef wrote:
> Gilboa Davara wrote:
> > Hello all,
> >
> > Couple of weeks ago I started a thread called "File I/O within kernel
> > threads?" which ended with a call "for numbers". (Read: Is I/O within
> > kernel performance increase justifies the de
Hi,
I should further point out that simply adding O_DIRECT to flip_open does
not solve the problem.
When I try to flip->read or flip->write with kernel allocated pages I
get a EFAULT from somewhere inside the vmm layer.
I assume that the DIO layer expects the pages to be allocated and mapped
on b
Rami,
While I can pass the O_DIRECT flag to the flip_open, it seems that flip-
>read and flip->write requires me to map the target pages as user pages.
(get_user_pages ?)
I'm trying to find ways to cut through the overhead as much as I can.
In general, I want to be able to control my own caching;