phram/slram

phram/slram

[Manfred Gruber]

hi !

I have a 2MB nvram on my ARM board.
slram and phram driver is possible to load and nvram works.

but phram driver is very slow (needs 20 seconds to load on insmod), slram is 
fast. does anyone know why ?

another question is i have changed in this drivers for a default erase size to 
get a jffs2 fs running on this. i know that it is not the right way to use a 
flash fs on ram but another fs was not possible to run (ext2).

is it correct to use a fs on ram chips? i will do this because then i have not 
to change my application :-)

thanks for comments regards  

manfred


 

Re: phram/slram

[Michal Schulz]

On Thu, 14 Oct 2004, Manfred Gruber wrote:

> another question is i have changed in this drivers for a default erase size to
> get a jffs2 fs running on this. i know that it is not the right way to use a
> flash fs on ram but another fs was not possible to run (ext2).

why couldn't you use ext2? It works without any problems with slram

> is it correct to use a fs on ram chips? i will do this because then i have not
> to change my application :-)

http://hedera.linuxnews.pl/_news/2002/09/03/_long/1445.html

-- 
Michal Schulz

phram/slram

[Jörn Engel]

Sorry, I wasn't subscribed to the list for a while.

Regarding your insmod problems with phram, I cannot reproduce that on
my notebook.  What's your kernel?  Any patches on it?

Jörn

-- 
He who knows others is wise.
He who knows himself is enlightened.
-- Lao Tsu

Re: phram/slram

[Manfred Gruber]

Am Freitag, 15. Oktober 2004 15:05 schrieben Sie:
> Sorry, I wasn't subscribed to the list for a while.
>
> Regarding your insmod problems with phram, I cannot reproduce that on
> my notebook.  What's your kernel?  Any patches on it?
>
> Jörn

2.6.9 -rc1. When none have the same problem with phram. Than it is a problem 
on my arm board :-(

thanks regards manfred

Re: phram/slram

[Jörn Engel]

On Mon, 18 October 2004 09:06:07 +0200, Manfred Gruber wrote:
> 
> 2.6.9 -rc1. When none have the same problem with phram. Than it is a problem 
> on my arm board :-(

Most people have problems with neither phram or slram working.  Below
is a patch for 2.6.8 (should still apply with -rc1) that fixes the
problem.

Will send it to akpm when 2.6.9 is out.

Jörn

-- 
It's not whether you win or lose, it's how you place the blame.
-- unknown


Finally make phram work again:
o Add simple usage example.
o Fix up unit handling (k, M. G).
o Use correct size limit for setup string.
o Remove pointless printk message.
o Fix mtdblock to not allocate a read-modify-write buffer for CAP_RAM
  devices (and fail to do so, returning an error).

Signed-off-by: Jörn Engel <joern@wohnheim.fh-wedel.de>
---

 devices/phram.c |   10 ++++++----
 mtdblock.c      |    2 +-
 2 files changed, 7 insertions(+), 5 deletions(-)


--- linux-2.6.8cow/drivers/mtd/devices/phram.c~phram_fix	2004-09-05 11:57:14.000000000 +0200
+++ linux-2.6.8cow/drivers/mtd/devices/phram.c	2004-09-05 11:58:00.000000000 +0200
@@ -18,6 +18,9 @@
  * by "k", "M" or "G", the numbers will be interpreted as kilo, mega or
  * gigabytes.
  *
+ * Example:
+ *	phram=swap,896M,110M phram=test,1006M,1M
+ *
  */
 
 #include <asm/io.h>
@@ -184,7 +187,7 @@
 		result *= 1024;
 	case 'k':
 		result *= 1024;
-		endp++;
+		(*endp)++;
 	}
 	return result;
 }
@@ -235,7 +238,7 @@
 	uint32_t len;
 	int i, ret;
 
-	if (strnlen(val, sizeof(str)) >= sizeof(str))
+	if (strnlen(val, sizeof(buf)) >= sizeof(buf))
 		parse_err("parameter too long\n");
 
 	strcpy(str, val);
@@ -283,7 +286,7 @@
 	if (!val || !val[0])
 		parse_err("no arguments to \"slram=\"\n");
 
-	if (strnlen(val, sizeof(str)) >= sizeof(str))
+	if (strnlen(val, sizeof(buf)) >= sizeof(buf))
 		parse_err("parameter too long\n");
 
 	strcpy(str, val);
@@ -342,7 +345,6 @@
 
 int __init init_phram(void)
 {
-	printk(KERN_ERR "phram loaded\n");
 	return 0;
 }
 
--- linux-2.6.8cow/drivers/mtd/mtdblock.c~phram_fix	2004-09-05 11:57:14.000000000 +0200
+++ linux-2.6.8cow/drivers/mtd/mtdblock.c	2004-09-05 11:58:00.000000000 +0200
@@ -248,7 +248,7 @@
 			      unsigned long block, char *buf)
 {
 	struct mtdblk_dev *mtdblk = mtdblks[dev->devnum];
-	if (unlikely(!mtdblk->cache_data)) {
+	if (unlikely(!mtdblk->cache_data && mtdblk->cache_size)) {
 		mtdblk->cache_data = vmalloc(mtdblk->mtd->erasesize);
 		if (!mtdblk->cache_data)
 			return -EINTR;