Re: [OpenWrt-Devel] [PATCH 0/7] kirkwood 4.4
On Mon, Mar 07, 2016 at 09:02:49PM +0100, Alexander Couzens wrote: > Based on previous patch series ubi for dockstar. > > To install the new ubi style on goflexnet: > - update bootloader (use the same bootloader as of goflexhome) > The goflexnet has a not as old SoC like the dockstar, > which means a recovery of a broken bootloader is easy using > uart recovery via kwboot [...] Thanks, tested on my gloflexHome and works. bye MM -- A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing in e-mail? signature.asc Description: Digital signature ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] current state of kirkwood target in trunk?
Hi, what is the state of the kirkwood targets in openwrt trunk? At the moment I'm trying to get an Segate GoFlex Home running with trunk. I had it previously running with atitude adjustment and wanted to to upgrade it to trunk. Since it seems ubifs is the new way to go, I flashed a new u-boot namely the openwrt-kirkwood-goflexhome-u-boot.kwb. It was missing the bootz command support, so I had to change 200-openwrt-config.patch since it didn't include "openwrt-kirkwood-common.h" for the goflex. diff --git a/package/boot/uboot-kirkwood/patches/200-openwrt-config.patch b/package/boot/uboot-kirkwood/patches/200-openwrt-config.patch index 0e84d3b..92e2b75 100644 --- a/package/boot/uboot-kirkwood/patches/200-openwrt-config.patch +++ b/package/boot/uboot-kirkwood/patches/200-openwrt-config.patch @@ -108,3 +108,12 @@ +#include "openwrt-kirkwood-common.h" + #endif /* _CONFIG_SHEEVAPLUG_H */ +--- a/include/configs/goflexhome.h b/include/configs/goflexhome.h +@@ -133,4 +133,6 @@ + #define CONFIG_RTC_MV + #endif /* CONFIG_CMD_DATE */ + ++#include "openwrt-kirkwood-common.h" ++ + #endif /* _CONFIG_GOFLEXHOME_H */ After flashing the rootfs: GoFlexHome> nand erase.part root NAND erase.part: device 0 offset 0x70, size 0xf90 Skipping bad block at 0x0d92 Erasing at 0xffe -- 100% complete. OK GoFlexHome> ubi part root UBI: attaching mtd1 to ubi0 UBI: scanning is finished UBI: empty MTD device detected UBI: attached mtd1 (name "mtd=2", size 249 MiB) to ubi0 UBI: PEB size: 131072 bytes (128 KiB), LEB size: 129024 bytes UBI: min./max. I/O unit sizes: 2048/2048, sub-page size 512 UBI: VID header offset: 512 (aligned 512), data offset: 2048 UBI: good PEBs: 1991, bad PEBs: 1, corrupted PEBs: 0 UBI: user volume: 0, internal volumes: 1, max. volumes count: 128 UBI: max/mean erase counter: 1/0, WL threshold: 4096, image sequence number: 0 UBI: available PEBs: 1948, total reserved PEBs: 43, PEBs reserved for bad PEB handling: 39 GoFlexHome> ubi create rootfs No size specified -> Using max size (251338752) Creating dynamic volume rootfs of size 251338752 GoFlexHome> tftpboot 0x80 dd-gfh/openwrt-kirkwood-goflexhome-rootfs.ubifs Using egiga0 device TFTP from server 10.10.10.3; our IP address is 10.10.10.6 Filename 'dd-gfh/openwrt-kirkwood-goflexhome-rootfs.ubifs'. Load address: 0x80 Loading: # # # # # # # # # # # # # # # # # # # # # # # 529.3 KiB/s done Bytes transferred = 7612416 (742800 hex) GoFlexHome> ubi write 0x80 rootfs 0x742800 7612416 bytes written to volume rootfs GoFlexHome> ubifsmount ubi:rootfs GoFlexHome> ubifsls 3848 Mon Feb 29 13:56:58 2016 bin 160 Sun Feb 28 13:21:54 2016 dev 3256 Sun Feb 28 13:21:54 2016 etc 1720 Sun Feb 28 13:21:54 2016 lib 160 Sun Feb 28 13:21:54 2016 mnt 224 Mon Feb 29 13:56:58 2016 rom 160 Mon Feb 29 13:56:59 2016 tmp 160 Sun Feb 28 13:21:54 2016 sys 4 Mon Feb 29 13:56:58 2016 var 416 Sun Feb 28 13:21:54 2016 usr 160 Sun Feb 28 13:21:54 2016 www 608 Mon Feb 29 13:59:33 2016 boot 160 Sun Feb 28 13:21:54 2016 proc 3024 Mon Feb 29 13:56:58 2016 sbin 160 Sun Feb 28 13:21:54 2016 root
Re: [OpenWrt-Devel] [PATCH][RFC] uboot-kirkwood: upgrade to 2012.10
Hi Luka, On Thu, Oct 18, 2012 at 10:59:00PM +0200, open...@lukaperkov.net wrote: > All our kirkwood uboot patches have been upstreamed with this release ;) > > Board owners please test this patch and give feedback for the boards: > > * dockstar Sorry it took me a while to test your patch. I applied it to trunk r33391 and can confirm, that uboot works fine on my dockstar when installed as primary bootloader to flash. bye MM -- A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing in e-mail? ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] RFC: [patch 1/1] add bql support to kirkwood mv643xx_eth
Hi,
I tried to add bql support to mv643xx_eth. Is this the correct way to
add it? It compiles fine and the kernel works, but I'm unsure how to
test it really.
Signed-off-by: Martin Mueller
Index: target/linux/kirkwood/patches/301-mv643xx-bql.patch
===
--- target/linux/kirkwood/patches/301-mv643xx-bql.patch (revision 0)
+++ target/linux/kirkwood/patches/301-mv643xx-bql.patch (revision 0)
@@ -0,0 +1,47 @@
+--- a/drivers/net/ethernet/marvell/mv643xx_eth.c
b/drivers/net/ethernet/marvell/mv643xx_eth.c
+@@ -902,6 +902,8 @@ static netdev_tx_t mv643xx_eth_xmit(stru
+ netif_tx_stop_queue(nq);
+ }
+
++ netdev_tx_sent_queue(nq, length);
++
+ return NETDEV_TX_OK;
+ }
+
+@@ -937,6 +939,7 @@ static int txq_reclaim(struct tx_queue *
+ struct mv643xx_eth_private *mp = txq_to_mp(txq);
+ struct netdev_queue *nq = netdev_get_tx_queue(mp->dev, txq->index);
+ int reclaimed;
++ unsigned int pkts_compl = 0, bytes_compl = 0;
+
+ __netif_tx_lock(nq, smp_processor_id());
+
+@@ -982,6 +985,9 @@ static int txq_reclaim(struct tx_queue *
+ }
+
+ if (skb != NULL) {
++ pkts_compl++;
++ bytes_compl += skb->len;
++
+ if (skb_queue_len(&mp->rx_recycle) <
+ mp->rx_ring_size &&
+ skb_recycle_check(skb, mp->skb_size))
+@@ -991,6 +997,7 @@ static int txq_reclaim(struct tx_queue *
+ }
+ }
+
++ netdev_tx_completed_queue(nq, pkts_compl, bytes_compl);
+ __netif_tx_unlock(nq);
+
+ if (reclaimed < budget)
+@@ -2014,7 +2021,9 @@ static int txq_init(struct mv643xx_eth_p
+ static void txq_deinit(struct tx_queue *txq)
+ {
+ struct mv643xx_eth_private *mp = txq_to_mp(txq);
++ struct netdev_queue *nq = netdev_get_tx_queue(mp->dev, txq->index);
+
++ netdev_tx_reset_queue(nq);
+ txq_disable(txq);
+ txq_reclaim(txq, txq->tx_ring_size, 1);
+
bye
MM
--
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing in e-mail?
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] [PATCH] [package] php5: create missing directory
Hi, for some reason building php5 fails for me with the following error: make -C /openwrt/kirkwood/trunk/build_dir/target-arm_v5te_uClibc-0.9.33.2_eabi/php-5.4.5 install INSTALL_ROOT=/openwrt/kirkwood/trunk/build_dir/target-arm_v5te_uClibc-0.9.33.2_eabi/php-5.4.5/staging make[4]: Entering directory `/openwrt/kirkwood/trunk/build_dir/target-arm_v5te_uClibc-0.9.33.2_eabi/php-5.4.5' Installing shared extensions: /openwrt/kirkwood/trunk/build_dir/target-arm_v5te_uClibc-0.9.33.2_eabi/php-5.4.5/staging/usr/lib/php/extensions/no-debug-non-zts-20100525/ Installing PHP CGI binary: /openwrt/kirkwood/trunk/build_dir/target-arm_v5te_uClibc-0.9.33.2_eabi/php-5.4.5/staging/usr/bin/ cp: cannot create regular file `/openwrt/kirkwood/trunk/build_dir/target-arm_v5te_uClibc-0.9.33.2_eabi/php-5.4.5/staging/usr/bin/#INST@8653#': No such file or directory make[4]: *** [install-cgi] Error 1 make[4]: Leaving directory `/openwrt/kirkwood/trunk/build_dir/target-arm_v5te_uClibc-0.9.33.2_eabi/php-5.4.5' make[3]: *** [/openwrt/kirkwood/trunk/staging_dir/target-arm_v5te_uClibc-0.9.33.2_eabi/stamp/.php_installed] Error 2 make[3]: Leaving directory `/openwrt/kirkwood/trunk/feeds/packages/lang/php5' make[2]: *** [package/feeds/packages/php5/compile] Error 2 make[2]: Leaving directory `/openwrt/kirkwood/trunk' make[1]: *** [/openwrt/kirkwood/trunk/staging_dir/target-arm_v5te_uClibc-0.9.33.2_eabi/stamp/.package_compile] Error 2 I needed the following patch: Signed-off-by: Martin Mueller Index: lang/php5/Makefile === --- lang/php5/Makefile (revision 33391) +++ lang/php5/Makefile (working copy) @@ -493,6 +493,7 @@ endef define Build/InstallDev + mkdir -p $(PKG_BUILD_DIR)/staging/usr/bin make -C $(PKG_BUILD_DIR) install INSTALL_ROOT=$(PKG_BUILD_DIR)/staging rm -f $(PKG_BUILD_DIR)/staging/usr/bin/php $(CP) $(PKG_BUILD_DIR)/staging/* $(STAGING_DIR_HOST) bye MM -- A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing in e-mail? ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] corrected: [PATCH 1/1] Add TDMA support to mv_cesa on kirkwood
Hi Luka, On Wed, Jul 11, 2012 at 01:13:16AM +0200, Luka Perkov wrote: > Hi Martin! Hi Daniel! > > On Wed, Jun 06, 2012 at 01:56:03PM +0300, Daniel Golle wrote: > > Hi Martin, > > thanks for bringing that up, sounds like a good improvement. > > Can you give us some numbers? I'm attaching two output files (before and > after applying this patch). I have also rebuilt openssl so it uses hw > crypto support. It's not so easy to use openssl with the mv_cesa engine. I used cryptodev-linux and patches to openssl to test the TDMA stuff, but perfomance with tinc and openssl was worse than with software crypto. Probably because of small packetsize and kernel/userspace transfers. Without cryptodev-linux and patches to openssl, the ioctls on /dev/crypto fail and openssl falls back to soaftware crypto, which you can see in your figures. The main thing the TDMA patch accelerates and for what I use it is encryted block devices with dm-crypt. bye MM -- A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing in e-mail? ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] Seagate FreeAgent GoFlex Home / SATA not working / patch missing?
On Fri, Jun 29, 2012 at 08:27:47PM +0100, Bastian Bittorf wrote: > Is this patch from february still missing? > http://patchwork.openwrt.org/patch/734/ Yes. bye MM -- A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing in e-mail? ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] Seagate FreeAgent GoFlex Home / klog-line missing / Dockstart VS GoFlex
Hi Bastian, On Fri, Jun 29, 2012 at 08:12:59PM +0100, Bastian Bittorf wrote: > > dmesg: > > [0.00] Booting Linux on physical CPU 0 > > [0.00] Linux version 3.3.8 (fnord@tschunk) (gcc version > > 4.6.3 20120201 (prerelease) (Linaro GCC 4.6-2012.02) ) #1 Fri Jun > > 29 03:26:51 UTC 2012 > > [0.00] CPU: Feroceon 88FR131 [56251311] revision 1 > > (ARMv5TE), cr=00053977 > > [0.00] CPU: VIVT data cache, VIVT instruction cache > > by accident one important line was missing: > > [0.00] Machine: Seagate FreeAgent DockStar > > > [0.00] Memory policy: ECC disabled, Data cache writeback > > [0.00] On node 0 totalpages: 32768 > > maybe this means, that the device is wrong detected > as "Dockstar", but it is a "GoFlex". Or is this not related? You need to compile with the following patch to get all features ins the goflex to work: http://patchwork.openwrt.org/patch/2247/ If you planning on using encrypted disks I recommend: http://patchwork.openwrt.org/patch/2245/ There is no need to update the bootloader on the goflex, infact if you use a dockstar bootloader on the goflex, SATA won't work. bye MM -- A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing in e-mail? ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] Patch [1/1]: fix uci-defaults/leds on kirkwood
Hi, there's a typo in the leds configuration. Index: target/linux/kirkwood/base-files/etc/uci-defaults/leds === --- target/linux/kirkwood/base-files/etc/uci-defaults/leds (revision 32125) +++ target/linux/kirkwood/base-files/etc/uci-defaults/leds (working copy) @@ -3,7 +3,7 @@ # Copyright (C) 2012 OpenWrt.org # -. /lib/fuctions/uci-defaults.sh +. /lib/functions/uci-defaults.sh . /lib/kirkwood.sh hardware=$(kirkwood_hardware_name) bye MM -- A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing in e-mail? ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH v2][RFC] uboot-kirkwood: upgrade to 2012.04.01
On Mon, Jun 11, 2012 at 11:44:33PM +0200, Luka Perkov wrote: > On Mon, Jun 11, 2012 at 11:29:53PM +0200, Martin Mueller wrote: > > Hi Luka, > > > > On Mon, Jun 11, 2012 at 10:37:15PM +0200, Luka Perkov wrote: > > > > > > Do you have printenv output from old uboot? > > > > Yes, so everything is fine. After restoring my old environment the > > dockstar boots it's old image again. I guess I was just confused by > > the ubi-stuff. > > > > > You should have uboot environment like above. Do you know how to do it > > > yourself > > > or I need to write the commands? > > > > No, the information, that just the environment needs to get readjusted > > and the old image should be compatible was enough. > > Can you share which uboot environment worked on new uboot? I had to reset the bootargs and bootcmd values. Of course there is a lot missing, as HW-address, tftp ip, netmask etc. This is the minimal env, that works for me to boot the device from nand: DockStar> printenv baudrate=115200 bootargs=console=ttyS0,115200 mtdparts=orion_nand:1M(u-boot),4M@2M(kernel),32M@4M(rootfs),219M@37M(data) rw root=/dev/mtdblock2 rw rootfstype=jffs2 bootargs_root=ubi.mtd=1 root=ubi0:root rootfstype=ubifs ro bootcmd=nand read 0x640 0x20 0x40; bootm 0x640; bootdelay=3 console=console=ttyS0,115200 ethact=egiga0 initrd=/boot/uInitrd kernel=/boot/uImage mtdids=nand0=orion_nand mtdparts=mtdparts=orion_nand:1m(uboot),-(root) stderr=serial stdin=serial stdout=serial bye MM -- A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing in e-mail? ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH v2][RFC] uboot-kirkwood: upgrade to 2012.04.01
Hi Luka, On Mon, Jun 11, 2012 at 10:37:15PM +0200, Luka Perkov wrote: > > Do you have printenv output from old uboot? Yes, so everything is fine. After restoring my old environment the dockstar boots it's old image again. I guess I was just confused by the ubi-stuff. > You should have uboot environment like above. Do you know how to do it > yourself > or I need to write the commands? No, the information, that just the environment needs to get readjusted and the old image should be compatible was enough. Thanks! bye MM -- A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing in e-mail? ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH v2][RFC] uboot-kirkwood: upgrade to 2012.04.01
On Mon, Jun 11, 2012 at 07:25:25PM +0200, Luka Perkov wrote: > On Mon, Jun 11, 2012 at 06:30:31PM +0200, Martin Mueller wrote: > > > And can you please also flash it on the device? As I understand you have > > > loaded it over JTAG only... > > > > Yes, that should be the same as loading it via serial to RAM. I'm > > relucant to flash the bootloader, as I have never flashed on before. > > Which image would i need to flash where? Your bootloader seems to live > > at 0x60 where as the original "Cloud engine uboot 1.1.4" seems to > > live at 0x80. > > Where did you see that it lives on 0x60? Sorry I did a grep for TEXT_BASE, but looked at the iconnect line instead of the dockstar one. > This is how I flash my kirkwood device (0x62118 is the size of loaded > binary): > > tftpboot 0x080 u-boot.kwb > nand erase 0x0 0x7 ; nand write 0x080 0x0 0x62118 > > You should flash it on the beginning of the nand chip. Ok, I flashed the new uboot, here is the output: U-Boot 2012.04.01 (Jun 11 2012 - 19:41:27) Seagate FreeAgent DockStar SoC: Kirkwood 88F6281_A0 DRAM: 128 MiB WARNING: Caches not enabled NAND: 256 MiB *** Warning - bad CRC, using default environment In:serial Out: serial Err: serial Net: egiga0 88E1116 Initialized on egiga0 Hit any key to stop autoboot: 0 DockStar> It still can't boot my old image on the dockstar, so I suppose I need to flash a new one. It want's to bbot some ubi-stuff, which I don't seem to have. The build just produces jffs2 images. Where to go from here? bye MM -- A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing in e-mail? ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH v2][RFC] uboot-kirkwood: upgrade to 2012.04.01
Hi Luka, On Mon, Jun 11, 2012 at 05:49:07PM +0200, Luka Perkov wrote: > > And can you please also flash it on the device? As I understand you have > loaded it over JTAG only... Yes, that should be the same as loading it via serial to RAM. I'm relucant to flash the bootloader, as I have never flashed on before. Which image would i need to flash where? Your bootloader seems to live at 0x60 where as the original "Cloud engine uboot 1.1.4" seems to live at 0x80. bye MM -- A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing in e-mail? ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH v2][RFC] uboot-kirkwood: upgrade to 2012.04.01
Hi Daniel, On Mon, Jun 11, 2012 at 06:41:56PM +0300, Daniel Golle wrote: > Hi Martin! > > On 11/06/12 18:21, Martin Mueller wrote: > > Here is the output: > > > > U-Boot 2012.04.01 (Jun 08 2012 - 13:05:57) > > Seagate FreeAgent DockStar > > > > SoC: Kirkwood 88F6281_A0 > > DRAM: 128 MiB > > WARNING: Caches not enabled > > NAND: 256 MiB > > NAND read from offset 6 failed -74 > > *** Warning - readenv() failed, using default environment > > In:serial > > Out: serial > > Err: serial > > Net: egiga0 > > 88E1116 Initialized on egiga0 > > Hit any key to stop autoboot: 0 > > DockStar> > > seems like there is some problem with the NAND... > or the environment of the original bootloader just lives at a different > location? > Can you nand load stuff (like linux uImage) using that loader? It doesn't looks like that it'd work: DockStar> setenv arcNumber 2998 DockStar> setenv mainlineLinux yes DockStar> setenv bootcmd nand read 0x640 0x10 0x40\; bootm 0x640\; DockStar> setenv bootargs console=ttyS0,115200 mtdparts=orion_nand:1M(u-boot),4M@1M(kernel),32M@4M(rootfs),219M@37M(data) rw root=/dev/mtdblock2 rw rootfstype=jffs2 DockStar> boot NAND read: device 0 offset 0x10, size 0x40 NAND read from offset 10 failed -74 0 bytes read: ERROR Wrong Image Format for bootm command ERROR: can't get kernel image! DockStar> bye MM -- A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing in e-mail? ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH v2][RFC] uboot-kirkwood: upgrade to 2012.04.01
Hi Luka, On Fri, Jun 08, 2012 at 03:32:58AM +0200, Luka Perkov wrote: > Board owners please test this patch and give feedback for the boards: > > * dockstar It seems to work fine on the dockstar. The serial didn't work for me either, so I had to test today with jtag. I loaded the image.bin into RAM at 0x60 and used resume 0x60 via jtag. Here is the output: U-Boot 2012.04.01 (Jun 08 2012 - 13:05:57) Seagate FreeAgent DockStar SoC: Kirkwood 88F6281_A0 DRAM: 128 MiB WARNING: Caches not enabled NAND: 256 MiB NAND read from offset 6 failed -74 *** Warning - readenv() failed, using default environment In:serial Out: serial Err: serial Net: egiga0 88E1116 Initialized on egiga0 Hit any key to stop autoboot: 0 DockStar> bye MM -- A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing in e-mail? ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] [PATCH 1/1] openwrt support for seagate goflex net/home
Hi,
here is my goflex patch again. The setup code is from
https://raw.github.com/archlinuxarm/PKGBUILDs/master/core/linux/archlinuxarm.patch
with some minor fiddeling to make it fit for linux-3.3.8 as is in trunk.
I compiled an image an tested the same image on dockstar and goflex.
bye
MM
--
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing in e-mail?
Index: target/linux/kirkwood/files/arch/arm/mach-kirkwood/goflexnet-setup.c
===
--- target/linux/kirkwood/files/arch/arm/mach-kirkwood/goflexnet-setup.c (revision 0)
+++ target/linux/kirkwood/files/arch/arm/mach-kirkwood/goflexnet-setup.c (revision 0)
@@ -0,0 +1,177 @@
+/*
+ * arch/arm/mach-kirkwood/goflexnet-setup.c
+ *
+ * Seagate GoFlex Net Setup
+ *
+ * This file is licensed under the terms of the GNU General Public
+ * License version 2. This program is licensed "as is" without any
+ * warranty of any kind, whether express or implied.
+ */
+
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+#include "common.h"
+#include "mpp.h"
+
+static struct mtd_partition goflexnet_nand_parts[] = {
+ {
+ .name = "u-boot",
+ .offset = 0,
+ .size = SZ_1M
+ }, {
+ .name = "uImage",
+ .offset = MTDPART_OFS_NXTBLK,
+ .size = SZ_4M
+ }, {
+ .name = "rootfs",
+ .offset = MTDPART_OFS_NXTBLK,
+ .size = SZ_32M
+ }, {
+ .name = "data",
+ .offset = MTDPART_OFS_NXTBLK,
+ .size = MTDPART_SIZ_FULL
+ },
+};
+
+static struct mv643xx_eth_platform_data goflexnet_ge00_data = {
+ .phy_addr = MV643XX_ETH_PHY_ADDR(0),
+};
+
+static struct mv_sata_platform_data goflexnet_sata_data = {
+ .n_ports = 2,
+};
+
+static struct gpio_led goflexnet_led_pins[] = {
+ {
+ .name = "status:green:health",
+ .default_trigger = "default-on",
+ .gpio = 46, // 0x4000
+ .active_low = 1,
+ },
+ {
+ .name = "status:orange:fault",
+ .default_trigger = "none",
+ .gpio = 47, // 0x8000
+ .active_low = 1,
+ },
+ {
+ .name = "status:white:left0",
+ .default_trigger = "none",
+ .gpio = 42, // 0x0400
+ .active_low = 0,
+ },
+ {
+ .name = "status:white:left1",
+ .default_trigger = "none",
+ .gpio = 43, // 0x0800
+ .active_low = 0,
+ },
+ {
+ .name = "status:white:left2",
+ .default_trigger = "none",
+ .gpio = 44, // 0x1000
+ .active_low = 0,
+ },
+ {
+ .name = "status:white:left3",
+ .default_trigger = "none",
+ .gpio = 45, // 0x2000
+ .active_low = 0,
+ },
+ {
+ .name = "status:white:right0",
+ .default_trigger = "none",
+ .gpio = 38, // 0x0040
+ .active_low = 0,
+ },
+ {
+ .name = "status:white:right1",
+ .default_trigger = "none",
+ .gpio = 39, // 0x0080
+ .active_low = 0,
+ },
+ {
+ .name = "status:white:right2",
+ .default_trigger = "none",
+ .gpio = 40, // 0x0100
+ .active_low = 0,
+ },
+ {
+ .name = "status:white:right3",
+ .default_trigger = "none",
+ .gpio = 41, // 0x0200
+ .active_low = 0,
+ }
+};
+
+static struct gpio_led_platform_data goflexnet_led_data = {
+ .leds = goflexnet_led_pins,
+ .num_leds = ARRAY_SIZE(goflexnet_led_pins),
+};
+
+static struct platform_device goflexnet_leds = {
+ .name = "leds-gpio",
+ .id = -1,
+ .dev = {
+ .platform_data = &goflexnet_led_data,
+ }
+};
+
+static unsigned int goflexnet_mpp_config[] __initdata = {
+ MPP29_GPIO, /* USB Power Enable */
+ MPP47_GPIO, /* LED Orange */
+ MPP46_GPIO, /* LED Green */
+ MPP45_GPIO, /* LED Left Capacity 3 */
+ MPP44_GPIO, /* LED Left Capacity 2 */
+ MPP43_GPIO, /* LED Left Capacity 1 */
+ MPP42_GPIO, /* LED Left Capacity 0 */
+ MPP41_GPIO, /* LED Right Capacity 3 */
+ MPP40_GPIO, /* LED Right Capacity 2 */
+ MPP39_GPIO, /* LED Right Capacity 1 */
+ MPP38_GPIO, /* LED Right Capacity 0 */
+ 0
+};
+
+static void __init goflexnet_init(void)
+{
+ /*
+ * Basic setup. Needs to be called early.
+ */
+ kirkwood_init();
+
+ /* setup gpio pin select */
+ kirkwood_mpp_conf(goflexnet_mpp_config);
+
+ kirkwood_uart0_init();
+ kirkwood_nand_init(ARRAY_AND_SIZE(goflexnet_nand_parts), 40);
+
+ if (gpio_request(29, "USB Power Enable") != 0 ||
+ gpio_direction_output(29, 1) != 0)
+ printk(KERN_ERR "can't set up GPIO 29 (USB Power Enable)\n");
+ kirkwood_ehci_init();
+ kirkwood_ge00_init(&goflexnet_ge00_data);
+ kirkwood_sata_init(&goflexnet_sata_data);
+
+ platform_device_register(&goflexnet_leds);
+}
+
+MACHINE_START(GOFLEXNET, "Seagate GoFlex Net")
+ /* Maintainer: Peter Carmichael */
+ .atag_offset = 0x100,
+ .init_machine = goflexnet_init,
+ .map_io = kirkwood_map_io,
+ .init_early = kirkwood_init_early,
+ .init_irq = kirkwood_init_irq,
+ .timer = &kirkwood_timer,
+ .restart = kirkwood_restart,
+MACHINE_END
+
Index: target/linux/kirkwood/files/arch/arm/mach-kirkwood/goflexhome-setup.c
===
--- target/linux/kirkwood/files/arch/arm/mach-ki
[OpenWrt-Devel] corrected: [PATCH 1/1] Add TDMA support to mv_cesa on kirkwood
On Wed, Jun 06, 2012 at 01:01:54AM +0200, Martin Mueller wrote:
>
> here is mv_cesa TDMA support for the kirkwood target. I only adapted
> the patches from Phil Sutter on the linux-crypto list and adjusted the
> kernel modules Makefile.
I had accidently deleted a line from package/kernel/modules/crypto.mk,
so here is the new version.
Index: target/linux/kirkwood/patches-3.3/300-mv_cesa-tdma.patch
===
--- target/linux/kirkwood/patches-3.3/300-mv_cesa-tdma.patch(revision 0)
+++ target/linux/kirkwood/patches-3.3/300-mv_cesa-tdma.patch(revision 0)
@@ -0,0 +1,1495 @@
+--- a/arch/arm/mach-kirkwood/common.c
b/arch/arm/mach-kirkwood/common.c
+@@ -268,9 +268,42 @@ void __init kirkwood_uart1_init(void)
+ /*
+ * Cryptographic Engines and Security Accelerator (CESA)
+ /
++static struct resource kirkwood_tdma_res[] = {
++ {
++ .name = "regs deco",
++ .start = CRYPTO_PHYS_BASE + 0xA00,
++ .end= CRYPTO_PHYS_BASE + 0xA24,
++ .flags = IORESOURCE_MEM,
++ }, {
++ .name = "regs control and error",
++ .start = CRYPTO_PHYS_BASE + 0x800,
++ .end= CRYPTO_PHYS_BASE + 0x8CF,
++ .flags = IORESOURCE_MEM,
++ }, {
++ .name = "crypto error",
++ .start = IRQ_KIRKWOOD_TDMA_ERR,
++ .end= IRQ_KIRKWOOD_TDMA_ERR,
++ .flags = IORESOURCE_IRQ,
++ },
++};
++
++static u64 mv_tdma_dma_mask = 0xUL;
++
++static struct platform_device kirkwood_tdma_device = {
++ .name = "mv_tdma",
++ .id = -1,
++ .dev= {
++ .dma_mask = &mv_tdma_dma_mask,
++ .coherent_dma_mask = 0x,
++ },
++ .num_resources = ARRAY_SIZE(kirkwood_tdma_res),
++ .resource = kirkwood_tdma_res,
++};
++
+ void __init kirkwood_crypto_init(void)
+ {
+ kirkwood_clk_ctrl |= CGC_CRYPTO;
++ platform_device_register(&kirkwood_tdma_device);
+ orion_crypto_init(CRYPTO_PHYS_BASE, KIRKWOOD_SRAM_PHYS_BASE,
+ KIRKWOOD_SRAM_SIZE, IRQ_KIRKWOOD_CRYPTO);
+ }
+--- a/arch/arm/mach-kirkwood/include/mach/irqs.h
b/arch/arm/mach-kirkwood/include/mach/irqs.h
+@@ -51,6 +51,7 @@
+ #define IRQ_KIRKWOOD_GPIO_HIGH_16_23 41
+ #define IRQ_KIRKWOOD_GE00_ERR 46
+ #define IRQ_KIRKWOOD_GE01_ERR 47
++#define IRQ_KIRKWOOD_TDMA_ERR 49
+ #define IRQ_KIRKWOOD_RTC53
+
+ /*
+--- a/arch/arm/plat-orion/common.c
b/arch/arm/plat-orion/common.c
+@@ -911,9 +911,15 @@ static struct resource orion_crypto_reso
+ },
+ };
+
++static u64 mv_crypto_dmamask = DMA_BIT_MASK(32);
++
+ static struct platform_device orion_crypto = {
+ .name = "mv_crypto",
+ .id = -1,
++ .dev= {
++ .dma_mask = &mv_crypto_dmamask,
++ .coherent_dma_mask = DMA_BIT_MASK(32),
++ },
+ };
+
+ void __init orion_crypto_init(unsigned long mapbase,
+--- /dev/null
b/drivers/crypto/dma_desclist.h
+@@ -0,0 +1,79 @@
++#ifndef __DMA_DESCLIST__
++#define __DMA_DESCLIST__
++
++struct dma_desc {
++ void *virt;
++ dma_addr_t phys;
++};
++
++struct dma_desclist {
++ struct dma_pool *itempool;
++ struct dma_desc *desclist;
++ unsigned long length;
++ unsigned long usage;
++};
++
++#define DESCLIST_ITEM(dl, x) ((dl).desclist[(x)].virt)
++#define DESCLIST_ITEM_DMA(dl, x) ((dl).desclist[(x)].phys)
++#define DESCLIST_FULL(dl) ((dl).length == (dl).usage)
++
++static inline int
++init_dma_desclist(struct dma_desclist *dl, struct device *dev,
++ size_t size, size_t align, size_t boundary)
++{
++#define STRX(x) #x
++#define STR(x) STRX(x)
++ dl->itempool = dma_pool_create(
++ "DMA Desclist Pool at "__FILE__"("STR(__LINE__)")",
++ dev, size, align, boundary);
++#undef STR
++#undef STRX
++ if (!dl->itempool)
++ return 1;
++ dl->desclist = NULL;
++ dl->length = dl->usage = 0;
++ return 0;
++}
++
++static inline int
++set_dma_desclist_size(struct dma_desclist *dl, unsigned long nelem)
++{
++ /* need to increase size first if requested */
++ if (nelem > dl->length) {
++ struct dma_desc *newmem;
++ int newsize = nelem * sizeof(struct dma_desc);
++
++ newmem = krealloc(dl->desclist, newsize, GFP_KERNEL);
++ if (!newmem)
++ return -ENOMEM;
++ dl->desclist = newmem;
++ }
++
++ /* allocate/free dma descriptors,
[OpenWrt-Devel] [PATCH 1/1] Add TDMA support to mv_cesa on kirkwood
Hi,
here is mv_cesa TDMA support for the kirkwood target. I only adapted
the patches from Phil Sutter on the linux-crypto list and adjusted the
kernel modules Makefile.
The speed improvement on a dockstar is only around 25% from 13MB/s to
16MB/s on an encrypted file system.
The original patches can be found here:
http://www.mail-archive.com/linux-crypto@vger.kernel.org/msg07101.html
http://www.mail-archive.com/linux-crypto@vger.kernel.org/msg07115.html
Index: target/linux/kirkwood/patches-3.3/300-mv_cesa-tdma.patch
===
--- target/linux/kirkwood/patches-3.3/300-mv_cesa-tdma.patch(revision 0)
+++ target/linux/kirkwood/patches-3.3/300-mv_cesa-tdma.patch(revision 0)
@@ -0,0 +1,1495 @@
+--- a/arch/arm/mach-kirkwood/common.c
b/arch/arm/mach-kirkwood/common.c
+@@ -268,9 +268,42 @@ void __init kirkwood_uart1_init(void)
+ /*
+ * Cryptographic Engines and Security Accelerator (CESA)
+ /
++static struct resource kirkwood_tdma_res[] = {
++ {
++ .name = "regs deco",
++ .start = CRYPTO_PHYS_BASE + 0xA00,
++ .end= CRYPTO_PHYS_BASE + 0xA24,
++ .flags = IORESOURCE_MEM,
++ }, {
++ .name = "regs control and error",
++ .start = CRYPTO_PHYS_BASE + 0x800,
++ .end= CRYPTO_PHYS_BASE + 0x8CF,
++ .flags = IORESOURCE_MEM,
++ }, {
++ .name = "crypto error",
++ .start = IRQ_KIRKWOOD_TDMA_ERR,
++ .end= IRQ_KIRKWOOD_TDMA_ERR,
++ .flags = IORESOURCE_IRQ,
++ },
++};
++
++static u64 mv_tdma_dma_mask = 0xUL;
++
++static struct platform_device kirkwood_tdma_device = {
++ .name = "mv_tdma",
++ .id = -1,
++ .dev= {
++ .dma_mask = &mv_tdma_dma_mask,
++ .coherent_dma_mask = 0x,
++ },
++ .num_resources = ARRAY_SIZE(kirkwood_tdma_res),
++ .resource = kirkwood_tdma_res,
++};
++
+ void __init kirkwood_crypto_init(void)
+ {
+ kirkwood_clk_ctrl |= CGC_CRYPTO;
++ platform_device_register(&kirkwood_tdma_device);
+ orion_crypto_init(CRYPTO_PHYS_BASE, KIRKWOOD_SRAM_PHYS_BASE,
+ KIRKWOOD_SRAM_SIZE, IRQ_KIRKWOOD_CRYPTO);
+ }
+--- a/arch/arm/mach-kirkwood/include/mach/irqs.h
b/arch/arm/mach-kirkwood/include/mach/irqs.h
+@@ -51,6 +51,7 @@
+ #define IRQ_KIRKWOOD_GPIO_HIGH_16_23 41
+ #define IRQ_KIRKWOOD_GE00_ERR 46
+ #define IRQ_KIRKWOOD_GE01_ERR 47
++#define IRQ_KIRKWOOD_TDMA_ERR 49
+ #define IRQ_KIRKWOOD_RTC53
+
+ /*
+--- a/arch/arm/plat-orion/common.c
b/arch/arm/plat-orion/common.c
+@@ -911,9 +911,15 @@ static struct resource orion_crypto_reso
+ },
+ };
+
++static u64 mv_crypto_dmamask = DMA_BIT_MASK(32);
++
+ static struct platform_device orion_crypto = {
+ .name = "mv_crypto",
+ .id = -1,
++ .dev= {
++ .dma_mask = &mv_crypto_dmamask,
++ .coherent_dma_mask = DMA_BIT_MASK(32),
++ },
+ };
+
+ void __init orion_crypto_init(unsigned long mapbase,
+--- /dev/null
b/drivers/crypto/dma_desclist.h
+@@ -0,0 +1,79 @@
++#ifndef __DMA_DESCLIST__
++#define __DMA_DESCLIST__
++
++struct dma_desc {
++ void *virt;
++ dma_addr_t phys;
++};
++
++struct dma_desclist {
++ struct dma_pool *itempool;
++ struct dma_desc *desclist;
++ unsigned long length;
++ unsigned long usage;
++};
++
++#define DESCLIST_ITEM(dl, x) ((dl).desclist[(x)].virt)
++#define DESCLIST_ITEM_DMA(dl, x) ((dl).desclist[(x)].phys)
++#define DESCLIST_FULL(dl) ((dl).length == (dl).usage)
++
++static inline int
++init_dma_desclist(struct dma_desclist *dl, struct device *dev,
++ size_t size, size_t align, size_t boundary)
++{
++#define STRX(x) #x
++#define STR(x) STRX(x)
++ dl->itempool = dma_pool_create(
++ "DMA Desclist Pool at "__FILE__"("STR(__LINE__)")",
++ dev, size, align, boundary);
++#undef STR
++#undef STRX
++ if (!dl->itempool)
++ return 1;
++ dl->desclist = NULL;
++ dl->length = dl->usage = 0;
++ return 0;
++}
++
++static inline int
++set_dma_desclist_size(struct dma_desclist *dl, unsigned long nelem)
++{
++ /* need to increase size first if requested */
++ if (nelem > dl->length) {
++ struct dma_desc *newmem;
++ int newsize = nelem * sizeof(struct dma_desc);
++
++ newmem = krealloc(dl->desclist, newsize, GFP_KERNEL);
++ if (!newmem)
++ return -ENOMEM;
++ dl->desclist = newmem;
++ }
++
++ /* allocate/free dma descriptors, adjusting dl->len
Re: [OpenWrt-Devel] [PATCH 2/3] uboot-kirkwood: backport dockstar board support to 2012.04.01
On Tue, Jun 05, 2012 at 11:33:08PM +0200, Luka Perkov wrote: > On Tue, Jun 05, 2012 at 11:13:04PM +0200, Martin Mueller wrote: > > On Tue, Jun 05, 2012 at 10:44:16PM +0200, Luka Perkov wrote: > > > On Tue, Jun 05, 2012 at 10:30:51PM +0200, Martin Mueller wrote: > > > > On Tue, Jun 05, 2012 at 09:42:03PM +0200, Luka Perkov wrote: > > > > > > > > > > Questions for all: > > > > > > > > > > 1) Anyone here who has dockstar board and would like to do some > > > > > tests? > > > > > > > > I have a dockstar and be willing to test. I also sent patches for > > > > seagate goflex-net support some monhts ago, but nobody seemed > > > > interested. > > > > > > Did you send it upstream? > > > > No. > > > > > > > > Can you post links of the patch here please? > > > > http://patchwork.openwrt.org/patch/734/ > > That is target not uboot patch... Ah, sorry. On the goflex there is no need to change the factory bootloader. bye MM -- A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing in e-mail? ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH 2/3] uboot-kirkwood: backport dockstar board support to 2012.04.01
On Tue, Jun 05, 2012 at 10:44:16PM +0200, Luka Perkov wrote: > On Tue, Jun 05, 2012 at 10:30:51PM +0200, Martin Mueller wrote: > > On Tue, Jun 05, 2012 at 09:42:03PM +0200, Luka Perkov wrote: > > > > > > Questions for all: > > > > > > 1) Anyone here who has dockstar board and would like to do some tests? > > > > I have a dockstar and be willing to test. I also sent patches for > > seagate goflex-net support some monhts ago, but nobody seemed > > interested. > > Did you send it upstream? No. > > Can you post links of the patch here please? http://patchwork.openwrt.org/patch/734/ bye MM -- A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing in e-mail? ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH 2/3] uboot-kirkwood: backport dockstar board support to 2012.04.01
On Tue, Jun 05, 2012 at 09:42:03PM +0200, Luka Perkov wrote: > > Questions for all: > > 1) Anyone here who has dockstar board and would like to do some tests? I have a dockstar and be willing to test. I also sent patches for seagate goflex-net support some monhts ago, but nobody seemed interested. bye MM -- A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing in e-mail? ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCHv2 0/3] ubifs on kirkwood (and others)
Hi Daniel, On Fri, Aug 05, 2011 at 02:28:54PM +0200, Daniel Golle wrote: > This updates and syncs tools/mtd-utils and packages/utils/mtd-utils to one > common > codebase. In order to prevent kernel panics I had to use a recent version of > mtd-utils to generate an ubi image on the host and flash it using ubiformat on > the target (kirkwood in my case). Thanks for the patches, before I try them, I'd like to ask: Do I loose the ability to flash the images from the original uboot bootloader, that came with the device? At the moment I flash via serial cable and tftp. Thanks in advance! bye MM -- A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing in e-mail? ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH] seagate goflex support for dockstar target
On Tue, Jul 05, 2011 at 05:56:03PM +0200, Jan Willies wrote: > 2011/7/5 Martin Mueller > > > > > Yes, but he seems to have some Problems of his own with it. I > > remember reading a post form the original developer of mv_cesa that he > > tried to implement DMA, but was disapointed with the speed gain it > > brought an so abandoned it. > > > > Hm, too bad. Do you remember his name? The source states " > Sebastian Andrzej Siewior". Yes, correct. This is the post: http://lists.debian.org/debian-arm/2009/11/msg00113.html bye MM -- A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing in e-mail? ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH] seagate goflex support for dockstar target
Hi Jan, On Tue, Jul 05, 2011 at 09:52:34AM +0200, Jan Willies wrote: > > Apparently someone is working now on TDMA support for mv_cesa: > http://code.google.com/p/openrd/issues/detail?id=18#c0 > > Would be nice to get this going. Yes, but he seems to have some Problems of his own with it. I remember reading a post form the original developer of mv_cesa that he tried to implement DMA, but was disapointed with the speed gain it brought an so abandoned it. Thanks for the info, let's see if soemthing shows up. bye MM -- A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing in e-mail? ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH] seagate goflex support for dockstar target
Hi Jan, On Fri, Feb 25, 2011 at 01:21:40PM +0100, Jan Willies wrote: > > 2011/2/23 Martin Mueller : > > I added support for the seagate GoFlex Net/Home to the dockstar > > target. The kernel will get slightly larger, but I think this is > > acceptable on this platform, since it has enough resources. > > That's great! I guess that'd be a huge performance boost to the > Dockstar since no USB is involved anymore, if I read the product page > right. Yes I'm quite satisfied with the performance: root@OpenWrt:/# for i in sda sdb md0 ; do hdparm -tT /dev/$i ; done /dev/sda: Timing cached reads: 432 MB in 2.00 seconds = 215.58 MB/sec Timing buffered disk reads: 252 MB in 3.02 seconds = 83.38 MB/sec /dev/sdb: Timing cached reads: 428 MB in 2.00 seconds = 213.96 MB/sec Timing buffered disk reads: 254 MB in 3.01 seconds = 84.49 MB/sec /dev/md0: Timing cached reads: 418 MB in 2.00 seconds = 208.89 MB/sec Timing buffered disk reads: 246 MB in 3.00 seconds = 81.97 MB/sec > Do you happen to know whether the GoFlex Net has the same crypto > hardware acceleration as the Dockstar > (http://wiki.openwrt.org/toh/seagate/dockstar#crypto.hardware.acceleration), > ie is it common about the Kirkwood SoCs? Yes it is and it works just like on the dockstar. Performance is not the best as the mv_cesa still doesn't use DMA. Unfortunatelly this seems out of my knowledge to implement it, but it would be nice. root@OpenWrt:/# hdparm -t /dev/mapper/crypt /dev/mapper/crypt: Timing buffered disk reads: 40 MB in 3.03 seconds = 13.19 MB/sec > > --- target/linux/kirkwood/config-default (revision 25666) > > +++ target/linux/kirkwood/config-default (working copy) > > @@ -71,7 +71,10 @@ > > # CONFIG_I2C_MV64XXX is not set > > CONFIG_INET_LRO=y > > CONFIG_INITRAMFS_SOURCE="" > > -# CONFIG_IP_ADVANCED_ROUTER is not set > > +CONFIG_IP_ADVANCED_ROUTER=y > > +CONFIG_IP_MROUTE=y > > +CONFIG_IP_MULTIPLE_TABLES=y > > +CONFIG_IP_ROUTE_FWMARK=y > > Why is this needed? I'm sorry this is not needed. It's just what I added since I did some fancy DSL-routing on the device. It can be left out (but it doesn't hurt either). bye MM -- A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing in e-mail? ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] [PATCH] seagate goflex support for dockstar target
Hi,
I added support for the seagate GoFlex Net/Home to the dockstar
target. The kernel will get slightly larger, but I think this is
acceptable on this platform, since it has enough resources.
No change of the bootloader is required, but you must set the u-boot
environment variables correctly. These mainly are "mainlineLinux" and
"arcNumber". "mainlineLinux" mut be set to "yes" for the u-boot to
export the "arcNumber" variable to the kernel, and the arcNumber must
be set for the kernel to correctly distinguish between a dockstar and a
goflex.
To flash openwrt to factory default goflex net, the following steps
are neccessary:
- Attach a serial console to the goflex and at the bootloader prompt
enter the following commands:
setenv arcNumber 3089
setenv mainlineLinux yes
setenv bootcmd nand read 0x640 0x10 0x40\; bootm 0x640\;
setenv bootargs console=ttyS0,115200 root=/dev/mtdblock2 rw rootfstype=jffs2
saveenv
flash kernel:
mw 0x640 0x 0x40
tftpboot 0x640 openwrt-kirkwood-uImage
nand erase 0x10 0x40
nand write.e 0x640 0x10 0x40
flash rootfs:
mw 0x640 0x 0x40
tftpboot 0x640 openwrt-kirkwood-Dockstar-jffs2-128k.img
nand erase 0x50 0xfb0
nand write.e 0x640 0x50 0x40
boot
That's it. I hope you apply my patch to openwrt.
Signed-off-by: Martin Mueller
Index: target/linux/kirkwood/patches/110-goflex.patch
===
--- target/linux/kirkwood/patches/110-goflex.patch (revision 0)
+++ target/linux/kirkwood/patches/110-goflex.patch (revision 0)
@@ -0,0 +1,151 @@
+--- a/arch/arm/mach-kirkwood/Makefile
b/arch/arm/mach-kirkwood/Makefile
+@@ -6,6 +6,7 @@
+ obj-$(CONFIG_MACH_MV88F6281GTW_GE)+= mv88f6281gtw_ge-setup.o
+ obj-$(CONFIG_MACH_SHEEVAPLUG) += sheevaplug-setup.o
+ obj-$(CONFIG_MACH_DOCKSTAR) += dockstar-setup.o
++obj-$(CONFIG_MACH_GOFLEX) += goflex-setup.o
+ obj-$(CONFIG_MACH_ESATA_SHEEVAPLUG) += sheevaplug-setup.o
+ obj-$(CONFIG_MACH_GURUPLUG) += guruplug-setup.o
+ obj-$(CONFIG_MACH_TS219) += ts219-setup.o tsx1x-common.o
+--- a/arch/arm/mach-kirkwood/Kconfig
b/arch/arm/mach-kirkwood/Kconfig
+@@ -105,6 +105,12 @@
+ Say 'Y' here if you want your kernel to support the
+ Seagate FreeAgent DockStar.
+
++config MACH_GOFLEX
++ bool "Seagate GoFlex Net/Home"
++ help
++Say 'Y' here if you want your kernel to support the
++Seagate GoFlex Net/Home.
++
+ endmenu
+
+ endif
+--- /dev/null
b/arch/arm/mach-kirkwood/goflex-setup.c
+@@ -0,0 +1,123 @@
++/*
++ * arch/arm/mach-kirkwood/goflex-setup.c
++ *
++ * Seagate GoFlex Net/Home
++ *
++ * This file is licensed under the terms of the GNU General Public
++ * License version 2. This program is licensed "as is" without any
++ * warranty of any kind, whether express or implied.
++ */
++
++#include
++#include
++#include
++#include
++#include
++#include
++#include
++#include
++#include
++#include
++#include
++#include
++#include "common.h"
++#include "mpp.h"
++
++static struct mtd_partition goflex_nand_parts[] = {
++ {
++ .name = "u-boot",
++ .offset = 0,
++ .size = SZ_1M
++ }, {
++ .name = "uImage",
++ .offset = MTDPART_OFS_NXTBLK,
++ .size = SZ_2M + SZ_2M,
++ }, {
++ .name = "rootfs",
++ .offset = MTDPART_OFS_NXTBLK,
++ .size = SZ_32M,
++ }, {
++ .name = "data",
++ .offset = MTDPART_OFS_NXTBLK,
++ .size = MTDPART_SIZ_FULL
++ },
++};
++
++static struct mv643xx_eth_platform_data goflex_ge00_data = {
++ .phy_addr = MV643XX_ETH_PHY_ADDR(0),
++};
++
++static struct gpio_led goflex_led_pins[] = {
++ {
++ .name = "goflex:green:health",
++ .default_trigger= "default-on",
++ .gpio = 46,
++ .active_low = 1,
++ },
++ {
++ .name = "goflex:orange:misc",
++ .default_trigger= "none",
++ .gpio = 47,
++ .active_low = 1,
++ },
++};
++
++static struct gpio_led_platform_data goflex_led_data = {
++ .leds = goflex_led_pins,
++ .num_leds = ARRAY_SIZE(goflex_led_pins),
++};
++
++static struct platform_device goflex_leds = {
++ .name = "leds-gpio",
++ .id = -1,
++ .dev= {
++ .platform_data = &goflex_led_data,
++ }
++};
++
++static struct mv_sata_platform_data goflex_esata_sata_data = {
++ .n_ports
