On Tue, 26 Apr 2011 11:25:29 +0600
Roman Mamedov wrote:
> Bad news, after a prolonged test there are some errors detected by memtester.
> So it looks like this dual rank module is not completely stable on my system.
> I wonder if it is possible to tweak the RAM controller from a running OS,
> e.g
There drive strength of loongson2F is not good enough to support big memory
in high freequency. That is limited by the chip. So even tune the DDR2
controller setting, perhaps it can not run stable with 2G/4G memories.
2011/4/26 Roman Mamedov
> On Mon, 25 Apr 2011 19:10:07 +0600
> Roman Mamedov
On Mon, 25 Apr 2011 19:10:07 +0600
Roman Mamedov wrote:
> My Fuloong now boots successfully with your patch on 2.6.38.3 and a dual-rank
> 2GB DDR2-800 module from Hynix (8+8=16 chips). It seems to work ok, I am
> running "memtester" now to check stability, no errors so far.
Bad news, after a pro
于 2011年04月26日 00:19, Roman Mamedov 写道:
On Mon, 25 Apr 2011 22:02:13 +0600
Roman Mamedov wrote:
However as I understood the patch expands the possible DMA area from 1GB to
2GB -- so is there be anything harmful in making that 4GB right away, just
in case (maybe someone will also want to test a
On Mon, 25 Apr 2011 22:02:13 +0600
Roman Mamedov wrote:
> However as I understood the patch expands the possible DMA area from 1GB to
> 2GB -- so is there be anything harmful in making that 4GB right away, just
> in case (maybe someone will also want to test a 4GB stick in future)?
> I think I w
On Mon, 25 Apr 2011 23:47:56 +0800
wu zhangjin wrote:
> Does 4GB work on FL6004 with that patch? just want to get more
> information about that patch and then determine if apply it or revisit
> and apply it.
Unfortunately I do not have the 4GB stick anymore, as I returned it.
The problems with
On Mon, Apr 25, 2011 at 9:10 PM, Roman Mamedov wrote:
> On Mon, 25 Apr 2011 16:56:58 +0800
> 项宇 wrote:
>
>> The preliminary conclusions is:
>> 1. Fuloong can't support 2GB RAM with 2 Ranks(up to now, only one
>> special model works fine)
>> 2. Fuloong can support some 2GB RAM with
On Mon, Apr 25, 2011 at 9:14 PM, Roman Mamedov wrote:
> On Mon, 25 Apr 2011 11:23:40 +0800
> wu zhangjin wrote:
>
>> To RM: for the DMA problem, does libata.force=1:pio4 make the IDE work
>> in pio mode?
>
> Sorry I did not get to try that :)
> Because this patch fixed the problem and it now work
On Mon, 25 Apr 2011 11:23:40 +0800
wu zhangjin wrote:
> To RM: for the DMA problem, does libata.force=1:pio4 make the IDE work
> in pio mode?
Sorry I did not get to try that :)
Because this patch fixed the problem and it now works with DMA too.
On Mon, 25 Apr 2011 16:56:58 +0800
项宇 wrote:
> -
On Mon, 25 Apr 2011 16:56:58 +0800
项宇 wrote:
> The preliminary conclusions is:
> 1. Fuloong can't support 2GB RAM with 2 Ranks(up to now, only one
> special model works fine)
> 2. Fuloong can support some 2GB RAM with 1 Rank.(but no warranty
> about the stablility)
>
> The 2.6
Hello, RM:
I do some experiments with 2GB RAM on Fuloong Box。
The preliminary conclusions is:
1. Fuloong can't support 2GB RAM with 2 Ranks(up to now, only one
special model works fine)
2. Fuloong can support some 2GB RAM with 1 Rank.(but no warranty
about the stablility)
The
I test 4 2GB ddr2 memory sticks with diffrent brands(sumsung, Adata, MT,
SCS) right now，they all work fine on my Fuloong6003
with pmon version 1.3.6 and kernel version 2.6.27.1 。
But the 2.6.38.4 kernel not work, I will give the root cause and try to
fix that later。
于 2011年04月25日 11:23, wu zhangj
2011/4/25 晏华 :
> the kernel
> 1. Does not set the memory map properly. or
> 2. The address window not set up properly.
I think so.
>
> Xiangyu:
>Could you help to look into this issue.
These two files may be related: arch/mips/loongson/common/mem.c and
arch/mips/loongson/common/pci.c
the kernel
1. Does not set the memory map properly. or
2. The address window not set up properly.
Xiangyu:
Could you help to look into this issue.
2011/4/24 Roman Mamedov
> Hello,
>
> I am continuing my experiments with pushing the limits of installed RAM
> size on
> Fuloong/Yeeloong
14 matches
Mail list logo
