Unanswered topics | Active topics


Board index » »

All times are UTC-06:00



Post new topic  Reply to topic  Page 1 of 2
  [ 17 posts ] 
  Previous topic | Next topic 
Author Message
Fossett
PostPosted: Sat May 15, 2010 11:03 am 
Offline

Joined: Thu May 13, 2010 4:34 pm
Posts: 5
Hello everybody,

I tried to use the search function but it ignores the "AND" keyword. Only the string after "AND" keyword will be searched for. So I got no usuable results. Searched for "sata AND efika" and got no useful outputs, so I have to put my question here. An individual answer same as a link to another thread would be fine, of course.

I got an Efika that runs MorphOS 2.4 at the moment. I would like to stick with it as it is quite fast in general operation speed but it lacks too many important things that make me think of a switch to Linux: Virtual graphics driver is not working, there is no SATA support (which of course implies working Virtual driver) and IDE speed is extremely slow due to no DMA.
So I think of installing Linux, ripping the Radeon out and putting a PCI SATA card in it to connect 2 or more SATA drives. I know that PCI is slow and most likely will be the bottleneck in this setup. Nevertheless I want to know if somebody runs such setup, which SATA card he uses, and what performance he gets out of it netwise. It shall run as a small storage server, no real workload would be needed.
What speed can I expect to get out of a pair or bunch of RAID disks connected to a SiI controller or the like? Plan is to pack 2 or 4 pieces of 1TB drives in it. From their current 80MB/s and upwards I am sure I won't see much at the other side of the cable. But how much throughput would I achieve from my desktop using SMB, NFS and such? Will PCI be the bummer or even the 400MHz PPC?
Top
   
Reply with quote  
Fossett
 Post subject:
PostPosted: Thu May 20, 2010 6:35 am 
Offline

Joined: Thu May 13, 2010 4:34 pm
Posts: 5
So using SATA devices on EFIKA seems to be very uncommon.
Top
   
Reply with quote  
JVid
 Post subject:
PostPosted: Thu May 20, 2010 6:51 am 
Offline

Joined: Thu Aug 27, 2009 1:07 pm
Posts: 14
Location: Bohemia
Yeah. Unfortunately, EFIKAs only have a single PCI slot, which is used by GFX on desktops or by a second networking card on servers, so most people are left with the painfully slow ATA. What's more, SATA expansion card costs more than a whole new EFIKA, so most people buy a new Atom-powered board rather than extending the unsupported, underpowered and overpriced product from Genesi.

Yes, I'm very disappointed by the state of drivers for this board, which degrades the otherwise well-performing 603e to a snail. No digital audio availability is also a problem to many users, fortunately I don't need it.
Top
   
Reply with quote  
bbrv
 Post subject:
PostPosted: Thu May 20, 2010 8:02 am 
Offline
Genesi

Joined: Fri Sep 24, 2004 1:39 am
Posts: 1422
Hi Folks, just a few words about the 5200B based Efika...

The board was first introduced at FTF 2005 (see last paragraph). This board we revised with the advent of the 5200B, which solved some of the design problems found in its predecessor, but not all of them. The Efika 5200B began to ship in quantity in November 2006. Freescale themselves became a reseller through DevTool Direct. By the summer of 2008, the Efika had shipped to 95 countries and was supported by many Linux distributions, BSDs, QNX and MorphOS. The Efika sold for $99. Many noteworthy projects were inspired by the board. Here is a list of the best of them: Power Projects. To this day there are many users and developers that devote significant effort to the platform. Of course, we do our best here also.

Please feel free to send us an email and let us know how you think we could do better.

Sincerely,
R&B :)

_________________
http://bbrv.blogspot.com
Top
   
Reply with quote  
markos
 Post subject:
PostPosted: Thu May 20, 2010 8:25 am 
Offline

Joined: Wed Oct 13, 2004 7:26 am
Posts: 348
Quote:
What's more, SATA expansion card costs more than a whole new EFIKA, so most people buy a new Atom-powered board rather than extending the unsupported, underpowered and overpriced product from Genesi.
I won't comment on the rest, but I have this one:

DELOCK 70154 PCI 4 PORT RAID SATA CONTROLLER (link in Greek sorry)

on the Pegasos. I got it for 27EUR, which is much less than the Efika ever cost... It works fine on Linux with the sil_sata driver (80-100MB/s is typical and within the limits of SATA1 and PCI bus). It's small so it fits the efika case, but it doesn't offer any e-SATA ports so you'd have to find a bigger case, or another model with e-SATA. This is just to prove that you were mistaken wrt SATA cards and prices.
Top
   
Reply with quote  
JVid
 Post subject:
PostPosted: Thu May 20, 2010 10:19 am 
Offline

Joined: Thu Aug 27, 2009 1:07 pm
Posts: 14
Location: Bohemia
Quote:
I have this one:

DELOCK 70154 PCI 4 PORT RAID SATA CONTROLLER (link in Greek sorry)

on the Pegasos. I got it for 27EUR, which is much less than the Efika ever cost...
Wow, this one is cool! I was even lucky enough to find it in one of the "local" webshops, so they will hopefully ship it tomorrow :)
I was looking for a good SATA controller since the day I bought that damned machine, because I didn't know about the BestComm fiasco before buying and I originally planned to make a cheap server and PPC-assembly devel board out of it, but I never used it for actually serving anything because of the PIO-only HDD accesses.
The last time I checked was about a year ago, but I was unable to find anything with 3.3V PCI under ~120 $.

The EFIKA had the potential to become a very good solution to all my needs and the amount I paid for it seemed more than good, but it is mostly unusable without proper driver support. I actually bought something overpriced and slow, even if it is called "Efficient".


Sorry, I forgot to comment on the second post:
Quote:
let us know how you think we could do better.
For example, you shouldn't tell people it has "powerful DMA controller" (I don't remember where I read that), "DMA to the ATA controller is handled by the integrated BestComm DMA controller." (here) and similar stuff, which is obviously not true. I know there isn't much you can do about the drivers themselves, without proper docs and with flawed DMA engine, but you should tell your customers the full truth.


Last edited by JVid on Thu May 20, 2010 11:30 am, edited 1 time in total.

Top
   
Reply with quote  
acrux
 Post subject:
PostPosted: Thu May 20, 2010 3:20 pm 
Offline

Joined: Thu Oct 19, 2006 7:19 am
Posts: 271
Location: Italy/Greece
...well, i quickly mounted a SIL3114 based pci card (the same cheap one as above).
Sorry but i don't have another spare HDD to build a RAID and not much interested to understand if this card is able to build an hw raid or simply a sw one.

With linux-2.6.34 BestComm & FEC DMA 'd work quite fine, i cannot test it on a 2.5" HDD 'cause i'm running my os from a 4GB micro SDHC class 6. Although a class 6 i have only ~4MB I/O.
Sound now works too but not dested digital output.
Code:
root@efika:~# uname -a
Linux efika 2.6.34-efika #1 Thu May 20 22:57:01 CEST 2010 ppc G2_LE Efika 5200B PowerPC System GNU/Linux
root@efika:~# lspci
00:18.0 RAID bus controller: Silicon Image, Inc. SiI 3114 [SATALink/SATARaid] Serial ATA Controller (rev 02)
00:1c.0 Bridge: Motorola MPC5200B
root@efika:~# hdparm -T -t /dev/sdb

/dev/sdb:
Timing cached reads: 176 MB in 2.01 seconds = 87.52 MB/sec
Timing buffered disk reads: 116 MB in 3.02 seconds = 38.42 MB/sec
Code:
root@efika:~# dmesg
Using Efika machine description
Linux version 2.6.34-efika (root@pegasos2) (gcc version 4.4.4 (CRUX PPC) (GCC) ) #1 Thu May 20 22:57:01 CEST 2010
bootconsole [udbg0] enabled
Efika: PCI bus 0 controlled by /pci@80000000

PCI host bridge /pci@80000000 ranges:
IO 0x00000000f8000000..0x00000000f800ffff -> 0x0000000000000000
MEM 0x0000000080000000..0x00000000bfffffff -> 0x0000000080000000
Top of RAM: 0x8000000, Total RAM: 0x8000000
Memory hole size: 0MB
Zone PFN ranges:
DMA 0x00000000 -> 0x00008000
Normal empty
Movable zone start PFN for each node
early_node_map[1] active PFN ranges
0: 0x00000000 -> 0x00008000
On node 0 totalpages: 32768
free_area_init_node: node 0, pgdat c0583624, node_mem_map c05b6000
DMA zone: 256 pages used for memmap
DMA zone: 0 pages reserved
DMA zone: 32512 pages, LIFO batch:7
Built 1 zonelists in Zone order, mobility grouping on. Total pages: 32512
Kernel command line: root=/dev/sda3 video=nvidiafb:800x600-8@60
PID hash table entries: 512 (order: -1, 2048 bytes)
Dentry cache hash table entries: 16384 (order: 4, 65536 bytes)
Inode-cache hash table entries: 8192 (order: 3, 32768 bytes)
Memory: 124044k/131072k available (5412k kernel code, 7028k reserved, 276k data, 147k bss, 196k init)
Kernel virtual memory layout:
* 0xfffdf000..0xfffff000 : fixmap
* 0xfdfee000..0xfe000000 : early ioremap
* 0xc9000000..0xfdfee000 : vmalloc & ioremap
SLUB: Genslabs=13, HWalign=32, Order=0-3, MinObjects=0, CPUs=1, Nodes=1
Hierarchical RCU implementation.
NR_IRQS:512
MPC52xx PIC is up and running!
time_init: decrementer frequency = 33.000000 MHz
time_init: processor frequency = 396.000000 MHz
clocksource: timebase mult[79364d9] shift[22] registered
clockevent: decrementer mult[872b020] shift[32] cpu[0]
Console: colour dummy device 80x25
console [tty0] enabled, bootconsole disabled
Mount-cache hash table entries: 512
NET: Registered protocol family 16
PCI: Probing PCI hardware
pci 0000:00:18.0: reg 10: [io 0x1000-0x1007]
pci 0000:00:18.0: reg 14: [io 0x1008-0x100b]
pci 0000:00:18.0: reg 18: [io 0x1010-0x1017]
pci 0000:00:18.0: reg 1c: [io 0x100c-0x100f]
pci 0000:00:18.0: reg 20: [io 0x1020-0x102f]
pci 0000:00:18.0: reg 24: [mem 0x80000000-0x800003ff]
pci 0000:00:18.0: reg 30: [mem 0x80080000-0x800fffff pref]
pci 0000:00:18.0: supports D1 D2
irq: irq 0 on host /builtin@F0000000/pic@F0000500 mapped to virtual irq 16
pci_bus 0000:00: resource 0 [io 0xfdfee000-0xfdffdfff]
pci_bus 0000:00: resource 1 [mem 0x80000000-0xbfffffff]
DMA: MPC52xx BestComm driver
DMA: MPC52xx BestComm engine @f0001200 ok !
irq: irq 73 on host /builtin@F0000000/pic@F0000500 mapped to virtual irq 73
mpc52xx-gpt f0000600.gpt: can function as watchdog
irq: irq 74 on host /builtin@F0000000/pic@F0000500 mapped to virtual irq 74
irq: irq 75 on host /builtin@F0000000/pic@F0000500 mapped to virtual irq 75
irq: irq 76 on host /builtin@F0000000/pic@F0000500 mapped to virtual irq 76
irq: irq 77 on host /builtin@F0000000/pic@F0000500 mapped to virtual irq 77
irq: irq 78 on host /builtin@F0000000/pic@F0000500 mapped to virtual irq 78
irq: irq 79 on host /builtin@F0000000/pic@F0000500 mapped to virtual irq 79
irq: irq 80 on host /builtin@F0000000/pic@F0000500 mapped to virtual irq 80
bio: create slab <bio-0> at 0
vgaarb: loaded
SCSI subsystem initialized
libata version 3.00 loaded.
usbcore: registered new interface driver usbfs
usbcore: registered new interface driver hub
usbcore: registered new device driver usb
Advanced Linux Sound Architecture Driver Version 1.0.22.1.
Bluetooth: Core ver 2.15
NET: Registered protocol family 31
Bluetooth: HCI device and connection manager initialized
Bluetooth: HCI socket layer initialized
cfg80211: Calling CRDA to update world regulatory domain
Switching to clocksource timebase
NET: Registered protocol family 2
IP route cache hash table entries: 1024 (order: 0, 4096 bytes)
TCP established hash table entries: 4096 (order: 3, 32768 bytes)
TCP bind hash table entries: 4096 (order: 2, 16384 bytes)
TCP: Hash tables configured (established 4096 bind 4096)
TCP reno registered
UDP hash table entries: 256 (order: 0, 4096 bytes)
UDP-Lite hash table entries: 256 (order: 0, 4096 bytes)
NET: Registered protocol family 1
RPC: Registered udp transport module.
RPC: Registered tcp transport module.
RPC: Registered tcp NFSv4.1 backchannel transport module.
PCI: CLS 32 bytes, default 32
Thermal assist unit not available
mpc52xx watchdog on GPT0: watchdog device registered
msgmni has been set to 242
alg: No test for stdrng (krng)
Block layer SCSI generic (bsg) driver version 0.4 loaded (major 253)
io scheduler noop registered
io scheduler deadline registered (default)
Generic non-volatile memory driver v1.1
Linux agpgart interface v0.103
Serial: MPC52xx PSC UART driver
irq: irq 129 on host /builtin@F0000000/pic@F0000500 mapped to virtual irq 129
f0002000.serial: ttyPSC0 at MMIO 0xf0002000 (irq = 129) is a MPC52xx PSC
brd: module loaded
loop: module loaded
ata: MPC52xx IDE/ATA libata driver
irq: irq 135 on host /builtin@F0000000/pic@F0000500 mapped to virtual irq 135
irq: irq 192 on host /builtin@F0000000/pic@F0000500 mapped to virtual irq 192
scsi0 : mpc52xx_ata
ata1: PATA max PIO4 ata_regs 0xf0003a00 irq 135
Fixed MDIO Bus: probed
usbmon: debugfs is not available
ohci_hcd: USB 1.1 'Open' Host Controller (OHCI) Driver
irq: irq 134 on host /builtin@F0000000/pic@F0000500 mapped to virtual irq 134
ppc-of-ohci f0001000.usb: OF OHCI
ppc-of-ohci f0001000.usb: new USB bus registered, assigned bus number 1
ppc-of-ohci f0001000.usb: irq 134, io mem 0xf0001000
usb usb1: New USB device found, idVendor=1d6b, idProduct=0001
usb usb1: New USB device strings: Mfr=3, Product=2, SerialNumber=1
usb usb1: Product: OF OHCI
usb usb1: Manufacturer: Linux 2.6.34-efika ohci_hcd
usb usb1: SerialNumber: PPC-OF USB
hub 1-0:1.0: USB hub found
hub 1-0:1.0: 2 ports detected
Initializing USB Mass Storage driver...
usbcore: registered new interface driver usb-storage
USB Mass Storage support registered.
mice: PS/2 mouse device common for all mice
rtc-generic rtc-generic: rtc core: registered rtc-generic as rtc0
i2c /dev entries driver
Linux video capture interface: v2.00
usbcore: registered new interface driver uvcvideo
USB Video Class driver (v0.1.0)
device-mapper: ioctl: 4.17.0-ioctl (2010-03-05) initialised: dm-devel@redhat.com
ALSA device list:
No soundcards found.
TCP cubic registered
NET: Registered protocol family 10
NET: Registered protocol family 17
ata1.00: CFA: Memory Card Adapter, 67281306, max UDMA/66
ata1.00: 8028160 sectors, multi 0: LBA
ata1.00: configured for PIO4
scsi 0:0:0:0: Direct-Access ATA Memory Card Adap 6728 PQ: 0 ANSI: 5
sd 0:0:0:0: Attached scsi generic sg0 type 0
sd 0:0:0:0: [sda] 8028160 512-byte logical blocks: (4.11 GB/3.82 GiB)
sd 0:0:0:0: [sda] Write Protect is off
sd 0:0:0:0: [sda] Mode Sense: 00 3a 00 00
sd 0:0:0:0: [sda] Write cache: disabled, read cache: enabled, doesn't support DPO or FUA
sda: sda1 sda2 sda3
sd 0:0:0:0: [sda] Attached SCSI removable disk
REISERFS (device sda3): found reiserfs format "3.6" with standard journal
REISERFS (device sda3): using ordered data mode
REISERFS (device sda3): journal params: device sda3, size 8192, journal first block 18, max trans len 1024, max batch 900, max commit age 30, max trans age 30
REISERFS (device sda3): checking transaction log (sda3)
REISERFS (device sda3): Using r5 hash to sort names
VFS: Mounted root (reiserfs filesystem) readonly on device 8:3.
Freeing unused kernel memory: 196k init
sata_sil 0000:00:18.0: version 2.4
sata_sil 0000:00:18.0: enabling device (0000 -> 0003)
sata_sil 0000:00:18.0: Applying R_ERR on DMA activate FIS errata fix
mpc52xx MII bus: probed
scsi1 : sata_sil
scsi2 : sata_sil
scsi3 : sata_sil
scsi4 : sata_sil
ata2: SATA max UDMA/100 mmio m1024@0x80000000 tf 0x80000080 irq 16
ata3: SATA max UDMA/100 mmio m1024@0x80000000 tf 0x800000c0 irq 16
ata4: SATA max UDMA/100 mmio m1024@0x80000000 tf 0x80000280 irq 16
ata5: SATA max UDMA/100 mmio m1024@0x80000000 tf 0x800002c0 irq 16
irq: irq 193 on host /builtin@F0000000/pic@F0000500 mapped to virtual irq 193
irq: irq 194 on host /builtin@F0000000/pic@F0000500 mapped to virtual irq 194
irq: irq 133 on host /builtin@F0000000/pic@F0000500 mapped to virtual irq 133
ata2: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
ata2.00: ATA-7: ST380215AS, 3.AAD, max UDMA/133
ata2.00: 156301488 sectors, multi 0: LBA48 NCQ (depth 0/32)
ata2.00: configured for UDMA/100
scsi 1:0:0:0: Direct-Access ATA ST380215AS 3.AA PQ: 0 ANSI: 5
sd 1:0:0:0: Attached scsi generic sg1 type 0
sd 1:0:0:0: [sdb] 156301488 512-byte logical blocks: (80.0 GB/74.5 GiB)
sd 1:0:0:0: [sdb] Write Protect is off
sd 1:0:0:0: [sdb] Mode Sense: 00 3a 00 00
sd 1:0:0:0: [sdb] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA
sdb:
sd 1:0:0:0: [sdb] Attached SCSI disk
ata3: SATA link down (SStatus 0 SControl 310)
ata4: SATA link down (SStatus 0 SControl 310)
ata5: SATA link down (SStatus 0 SControl 310)
irq: irq 130 on host /builtin@F0000000/pic@F0000500 mapped to virtual irq 130
irq: irq 195 on host /builtin@F0000000/pic@F0000500 mapped to virtual irq 195
irq: irq 196 on host /builtin@F0000000/pic@F0000500 mapped to virtual irq 196
Adding 262324k swap on /dev/sda1. Priority:-1 extents:1 across:262324k
STAC9766 SoC Audio Codec 0.10
asoc: stac9766 analog <-> AC97 mapping ok
asoc: stac9766 IEC958 <-> SPDIF mapping ok
PHY: f0003000:10 - Link is Up - 100/Full
eth0: no IPv6 routers present

_________________
acrux _at_ linuxmail _dot_ org
Top
   
Reply with quote  
Neko
 Post subject:
PostPosted: Thu May 20, 2010 6:18 pm 
Offline
Site Admin

Joined: Fri Sep 24, 2004 1:39 am
Posts: 1589
Location: Austin, TX
Quote:
For example, you shouldn't tell people it has "powerful DMA controller" (I don't remember where I read that), "DMA to the ATA controller is handled by the integrated BestComm DMA controller." (here) and similar stuff, which is obviously not true. I know there isn't much you can do about the drivers themselves, without proper docs and with flawed DMA engine, but you should tell your customers the full truth.
Actually the first drivers out there were written by bplan, and were as fast as they get. That is to say, actually pretty slow (PIO4)

The DMA controller - BestComm - is actually pretty powerful. It is also the ONLY way you can get DMA involved in disk access reliably. The MPC5200B ATA state engine has a horrible bug in it that means that as a buffer underrun or overrun (and the FIFO is only 512 bytes so, one disk block) it does not inform the ATA DMA engine, which will simply keep pulling garbage data out of the FIFO. What it should do is insert a wait state until the underrun condition is over. Added to that the FIFO does not trigger alarms on those conditions to userspace - basically, the mechanisms to support making sure the data in the FIFO is valid and not being trampled with garbage are broken on both sides.

What BestComm allows you to do is submit a PIO job (not much CPU time needed to do this, a few IO register writes) and then use BestComm to fill and empty the FIFO in 512-byte blocks.

Yes there is ATA DMA support in the current Linux driver, but good luck getting it to work.

_________________
Matt Sealey
Top
   
Reply with quote  
JVid
 Post subject:
PostPosted: Thu May 20, 2010 10:28 pm 
Offline

Joined: Thu Aug 27, 2009 1:07 pm
Posts: 14
Location: Bohemia
Quote:
Yes there is ATA DMA support in the current Linux driver, but good luck getting it to work.
Yes, I know about that. I tried it in 2.6.32 (?probably, can't remember now) without success, the kernel didn't even boot.
Top
   
Reply with quote  
acrux
 Post subject:
PostPosted: Fri May 21, 2010 2:13 am 
Offline

Joined: Thu Oct 19, 2006 7:19 am
Posts: 271
Location: Italy/Greece
Quote:
Yes there is ATA DMA support in the current Linux driver, but good luck getting it to work.

linux-2.6.34/drivers/ata/pata_mpc52xx.c
Code:
/*
* By default, all DMA modes are disabled for the MPC5200. Some
* boards don't have the required signals routed to make DMA work.
* Also, the MPC5200B has a silicon bug that causes data corruption
* with UDMA if it is used at the same time as the LocalPlus bus.
*
* Instead of trying to guess what modes are usable, check the
* ATA device tree node to find out what DMA modes work on the board.
* UDMA/MWDMA modes can also be forced by adding "libata.force=<mode>"
* to the kernel boot parameters.
*
* The MPC5200 ATA controller supports MWDMA modes 0, 1 and 2 and
* UDMA modes 0, 1 and 2.
*/

_________________
acrux _at_ linuxmail _dot_ org
Top
   
Reply with quote  
Neko
 Post subject:
PostPosted: Fri May 21, 2010 10:26 pm 
Offline
Site Admin

Joined: Fri Sep 24, 2004 1:39 am
Posts: 1589
Location: Austin, TX
Quote:
Quote:
Yes there is ATA DMA support in the current Linux driver, but good luck getting it to work.

linux-2.6.34/drivers/ata/pata_mpc52xx.c
Code:
/*
* By default, all DMA modes are disabled for the MPC5200. Some
* boards don't have the required signals routed to make DMA work.
* Also, the MPC5200B has a silicon bug that causes data corruption
* with UDMA if it is used at the same time as the LocalPlus bus.
*
* Instead of trying to guess what modes are usable, check the
* ATA device tree node to find out what DMA modes work on the board.
* UDMA/MWDMA modes can also be forced by adding "libata.force=<mode>"
* to the kernel boot parameters.
*
* The MPC5200 ATA controller supports MWDMA modes 0, 1 and 2 and
* UDMA modes 0, 1 and 2.
*/
Not entirely true.

1) The Efika has the signals routed. DMA "works" if you turn it on

2) LocalPlus and ATA aren't used together on the Efika (LocalPlus and PCI are mutually exclusive, and PCI doesn't cause the problem) so this is not causing broken DMA.

3) the ATA state engine bug is NOT a documented errata. Freescale's semi-official stance on it is, yes, it is broken, but who cares, this is an automotive chip and you probably won't be using ATA drives on it anyway. If you get it to work, good luck!

Remember ATA DMA is not there to improve the speed of the transfer but to work around the IDE cable being badly designed (basically UDMA is DDR MWDMA), and reduce CPU usage which helps to keep a reliable speed under load. MWDMA is not any faster than PIO, and UDMA2 is 33MB/s which is actually faster than most 2.5" magnetic media operates. If you factor in access latency on the drive and random access, you will not see much improved performance in these modes whether you use DMA-assisted-PIO (bplan style driver) or UDMA.

_________________
Matt Sealey
Top
   
Reply with quote  
Fossett
 Post subject:
PostPosted: Sat May 22, 2010 1:50 pm 
Offline

Joined: Thu May 13, 2010 4:34 pm
Posts: 5
Thanks for your comments although they do not answer my questions. Of course a SiI card or from any other manufacturer will work in EFIKA. I expected that as there are linux drivers for them. That is not the point. The question is what performance I can expect. Pegasos II is a lot faster than EFIKA, so I doubt that I can achieve 80MB/s on EFIKA. Maybe it is possible from SATA drive to SATA drive as long as all traffic happens on the SATA card. But what about SATA to onboard IDE or vice versa? Also, I am really unsure on how to estimate the throughput to other stations. Can I at least get full Fast Ethernet performance or will it be slowed down by any component? I still cannot judge or estimate on the e300 here and all the setup around. Network on MorphOS is really disappointing, so I hope it to be better on Linux.

If onboard IDE is too slow: is it possible to boot off SATA card?

And not to ask the last question: as IDE connector has DMA problems, what speed gain can I expect for the system itself running from SATA drives? Not that much is needed but can someone estimate and say 10% or 2% or something like that?

@JVid
The driver problem is always the same with Genesi products. First they start and then they do not bring it to an end software/driver-wise. They always promise more than they deliver, always premature und undecided products/strategies. Even in times when MorphOS was directly shipped with Pegasos, has there ever been accurate drivers for Firewire or Gigabit? Never. It happened again with EFIKA and I doubt it will ever improve. But the problem of Genesi not knowing what they want and presenting a clear strategy was not my point here, so I leave it that way.

@ bbrv
If you really want to know what to do and want to do something then contact BestComm to get ALL the necessary documents and information to release them to all developers such that those bugs in state-machine can be worked around. This is the least that you owe your customers from my POV if you ever want to be seen as reliable business partner. And as a business partner with BestComm it is your absolute right to access all those information in cases like this where obvious bugs prevent YOUR customers from using the device YOU delievered as it should be used. THis is what you owe to your customers both for Linux and MoerphOS. I even think BestComm is the one whom you need to urge to work on a solution for that problem but as stated above you often miss perspective and strategy. Sell an item and off you go. The job is done for you. Not the best behaviour to be taken seriously in future.
Top
   
Reply with quote  
JVid
 Post subject:
PostPosted: Sat May 22, 2010 2:40 pm 
Offline

Joined: Thu Aug 27, 2009 1:07 pm
Posts: 14
Location: Bohemia
Quote:
Pegasos II is a lot faster than EFIKA, so I doubt that I can achieve 80MB/s on EFIKA.
Sure you can, PCI has DMA in a working state (and acrux measured it in this very thread).
Quote:
But what about SATA to onboard IDE or vice versa?
4.7 MBps on my machine.
Quote:
Also, I am really unsure on how to estimate the throughput to other stations. Can I at least get full Fast Ethernet performance or will it be slowed down by any component?
Yeah, I think you'll get 100 Mbps minus protocol overhead. If it works for memory to net, why shouldn't it work for PCI to net? (The answer to this question is BestComm, if you really want to know :)
Quote:
If onboard IDE is too slow: is it possible to boot off SATA card?
Dunno. Not a problem for Linux, but what about Aura? Your best bet would be to boot your kernel from IDE, then mount the root FS from SATA.
Quote:
And not to ask the last question: as IDE connector has DMA problems, what speed gain can I expect for the system itself running from SATA drives? Not that much is needed but can someone estimate and say 10% or 2% or something like that?
Depends. It won't be much faster for standard OS workload (many small reads, random access), but you can get over 20× more for large data transfers.
Quote:
@JVid
The driver problem is always the same with Genesi products. First they start and then they do not bring it to an end software/driver-wise. They always promise more than they deliver, always premature und undecided products/strategies. Even in times when MorphOS was directly shipped with Pegasos, has there ever been accurate drivers for Firewire or Gigabit? Never. It happened again with EFIKA and I doubt it will ever improve. But the problem of Genesi not knowing what they want and presenting a clear strategy was not my point here, so I leave it that way.
I didn't own Pegasos and I don't use MorphOS, so I can't really tell, but this attitude seems to be improving with EfikaMX. We can only hope.
Quote:
Sell an item and off you go. The job is done for you. Not the best behaviour to be taken seriously in future.
Definitely.
Top
   
Reply with quote  
Neko
 Post subject:
PostPosted: Sat May 22, 2010 4:58 pm 
Offline
Site Admin

Joined: Fri Sep 24, 2004 1:39 am
Posts: 1589
Location: Austin, TX
Quote:
@ bbrv
If you really want to know what to do and want to do something then contact BestComm to get ALL the necessary documents and information to release them to all developers such that those bugs in state-machine can be worked around. This is the least that you owe your customers from my POV if you ever want to be seen as reliable business partner. And as a business partner with BestComm
BestComm is the name of the DMA engine..

It is not broken. The bug is in the ATA controller state machine which is Freescale's responsibility and it's burned into the silicon. It can't be fixed or worked around except to poll the FIFO incessently OR make sure you never initiate any transfer larger than the FIFO size.
Quote:
information in cases like this where obvious bugs prevent YOUR customers from using the device YOU delievered as it should be used. This is what you owe to your customers both for Linux and MorphOS.
The drivers work fine. We fulfilled our commitment to customers by implementing the best workaround possible - DMA assisted PIO. That is the best option you can expect from the MPC5200B.

_________________
Matt Sealey
Top
   
Reply with quote  
Fossett
 Post subject:
PostPosted: Sun May 23, 2010 6:24 am 
Offline

Joined: Thu May 13, 2010 4:34 pm
Posts: 5
Quote:
PCI has DMA in a working state (and acrux measured it in this very thread).
So you say PCI-wise I can expect the same speed in EFIKA as in Pegasos II? I think I've read somewhere that both use the same PCI devices but I am not sure anymore.
Quote:
4.7 MBps on my machine.
4.7MB, so it's an upper case 'B', right? I hjust want to be sure.
Quote:
Not a problem for Linux, but what about Aura? Your best bet would be to boot your kernel from IDE, then mount the root FS from SATA.
Aura is the new firmware design, right? I think it is only planned for EfikaMX but not EFIKA (PPC). So I think it won't matter. But right question though: what about OpenFirmware in EFIKA? Can it boot off SATA drives? It has a lot of problems with booting off USB devices so maybe only onboard IDE is working correctly.
Quote:
I didn't own Pegasos and I don't use MorphOS, so I can't really tell, but this attitude seems to be improving with EfikaMX. We can only hope.
EfikaMX is not an option for me but I think that's the way they go as PPC gets more and more uninteresting.
Top
   
Reply with quote  
Display posts from previous:  Sort by  
Post new topic  Reply to topic  Page 1 of 2
  [ 17 posts ] 

Board index » »

All times are UTC-06:00

Who is online

Users browsing this forum: No registered users and 16 guests

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum

Search for:
Jump to:  
PowerDeveloper.org: Copyright © 2004-2012, Genesi USA, Inc. The Power Architecture and Power.org wordmarks and the Power and Power.org logos and related marks are trademarks and service marks licensed by Power.org.
All other names and trademarks used are property of their respective owners. Privacy Policy
Powered by phpBB® Forum Software © phpBB Group