hello, every one!
I wrote a filesystem based on fuse, then i changed some args in fuse to
improve the file sytem's performance.
fuse_i.h #define FUSE_MAX_PAGES_PER_REQ 256
fuse_kern_chan.c #define MIN_BUFSIZE 0x101000
and I use "-o direct_io" to mount my filesystem.
I made this changes to enable bigger write like 1m, and i read the fuse's
code, found the write process is just concerned with these two argurements,
but the test result show, it can just up to 128k.
I want to ask? Did the linux kernel limit the write size ?
just like readahead size limitation in include/linux/mm.h to limit the
sequential read size
#define VM_MAX_READAHEAD 128 /* kbytes */
#define VM_MIN_READAHEAD 16 /* kbytes (includes current page) */
after i changed VM_MAX_READAHEAD to 1024
both read and write can up to 1m .
i want to know what limits the read size ?
View this message in context: http://old.nabble.com/the-direct_io-option-in-fuse-tp27390831p27390831.html
Sent from the Fuse - Dev mailing list archive at Nabble.com.