If I remember correctly, some of you have connected a SATA drive to the OMAP5432EVM?
Does it work with the mainline 3.17 kernel or does it need additional drivers/configs/udev rules?
BR, Nikolaus
On 09/11/2014 02:17 PM, Dr. H. Nikolaus Schaller wrote:
If I remember correctly, some of you have connected a SATA drive to the OMAP5432EVM?
Does it work with the mainline 3.17 kernel or does it need additional drivers/configs/udev rules?
Im running a kernel based on these scripts/config, which is pretty much mainline.
https://github.com/RobertCNelson/armv7-lpae-multiplatform/tree/v3.17.x
It doesn't need any special changes, the sata driver is compiled into the kernel itself, so root=/dev/sda1 is all that's needed to have a rootfs on sata.
Hi,
Am 11.09.2014 um 15:29 schrieb aTc:
On 09/11/2014 02:17 PM, Dr. H. Nikolaus Schaller wrote:
If I remember correctly, some of you have connected a SATA drive to the OMAP5432EVM?
Does it work with the mainline 3.17 kernel or does it need additional drivers/configs/udev rules?
Im running a kernel based on these scripts/config, which is pretty much mainline.
https://github.com/RobertCNelson/armv7-lpae-multiplatform/tree/v3.17.x
It doesn't need any special changes, the sata driver is compiled into the kernel itself, so root=/dev/sda1 is all that's needed to have a rootfs on sata.
That is good to know.
I have tried to understand, and it looks as if it is just the kernel.org with only some modified config. I have not found any patches that he applies.
In other words, this means that the kernel.org tree already supports the OMAP5 SATA hardware, although I had not yet found a trace of OMAP5 specific code...
Now with this background information I have found it. It is defined mainly in omap5.dtsi as
sata: { compatible = "snps,dwc-ahci" }
So my kernel should also have this. Looks as if I don't have the right .config yet.
Anyways, this means that we will be able to test the USB+SATA hardware as soon as we have the Pyra CPU board and new main board done. Without additional kernel tweaking.
thanks for these infos, Nikolaus
FYI: I have found the final missing CONFIG and now I have eSATA on the OMAP5432EVM on the GTA04 3.17.0 kernel. I have not tested more than that I can read /dev/sdc since my disk has an incompatible partition table, i.e. I need to re-format it firstly.
Anyways, this will be good to test the Pyra hardware.
Here is the latest version/patch:
http://git.openpandora.org/cgi-bin/gitweb.cgi?p=pyra-kernel.git;a=shortlog;h...
BR, Nikolaus
Am 19.09.2014 um 09:34 schrieb Dr. H. Nikolaus Schaller:
Hi,
Am 11.09.2014 um 15:29 schrieb aTc:
On 09/11/2014 02:17 PM, Dr. H. Nikolaus Schaller wrote:
If I remember correctly, some of you have connected a SATA drive to the OMAP5432EVM?
Does it work with the mainline 3.17 kernel or does it need additional drivers/configs/udev rules?
Im running a kernel based on these scripts/config, which is pretty much mainline.
https://github.com/RobertCNelson/armv7-lpae-multiplatform/tree/v3.17.x
It doesn't need any special changes, the sata driver is compiled into the kernel itself, so root=/dev/sda1 is all that's needed to have a rootfs on sata.
That is good to know.
I have tried to understand, and it looks as if it is just the kernel.org with only some modified config. I have not found any patches that he applies.
In other words, this means that the kernel.org tree already supports the OMAP5 SATA hardware, although I had not yet found a trace of OMAP5 specific code...
Now with this background information I have found it. It is defined mainly in omap5.dtsi as
sata: { compatible = "snps,dwc-ahci" }
So my kernel should also have this. Looks as if I don't have the right .config yet.
Anyways, this means that we will be able to test the USB+SATA hardware as soon as we have the Pyra CPU board and new main board done. Without additional kernel tweaking.
thanks for these infos, Nikolaus _______________________________________________ Firmware mailing list Firmware@pyra-handheld.com http://pyra-handheld.com/cgi-bin/mailman/listinfo/firmware
On 10/07/2014 11:56 AM, Dr. H. Nikolaus Schaller wrote:
FYI: I have found the final missing CONFIG and now I have eSATA on the OMAP5432EVM on the GTA04 3.17.0 kernel. I have not tested more than that I can read /dev/sdc since my disk has an incompatible partition table, i.e. I need to re-format it firstly.
Anyways, this will be good to test the Pyra hardware.
Here is the latest version/patch:
http://git.openpandora.org/cgi-bin/gitweb.cgi?p=pyra-kernel.git;a=shortlog;h...
I compiled and ran it on the EVM, but /dev/sda still doesn't show up.
The kernel I normally run on the board finds it at boot and everything works fine. , the config that one is using is here : https://github.com /RobertCNelson/armv7-lpae-multiplatform/blob/v3.17.x/patches/defconfig
the gta04 kernel itself works, although even when booting from /dev/mmcblk1p2 , /dev/sda doesn't show up.
Hi,
Am 08.10.2014 um 20:53 schrieb aTc:
On 10/07/2014 11:56 AM, Dr. H. Nikolaus Schaller wrote:
FYI: I have found the final missing CONFIG and now I have eSATA on the OMAP5432EVM on the GTA04 3.17.0 kernel. I have not tested more than that I can read /dev/sdc since my disk has an incompatible partition table, i.e. I need to re-format it firstly.
Anyways, this will be good to test the Pyra hardware.
Here is the latest version/patch:
http://git.openpandora.org/cgi-bin/gitweb.cgi?p=pyra-kernel.git;a=shortlog;h...
I compiled and ran it on the EVM, but /dev/sda still doesn't show up.
The kernel I normally run on the board finds it at boot and everything works fine. ,
Yes, that is what the gta04 kernel does now for me. It even supports hot-plugging.
the config that one is using is here : https://github.com /RobertCNelson/armv7-lpae-multiplatform/blob/v3.17.x/patches/defconfig
did you try a diff with http://git.openpandora.org/cgi-bin/gitweb.cgi?p=pyra-kernel.git;a=blob;f=arc...
the gta04 kernel itself works, although even when booting from /dev/mmcblk1p2 , /dev/sda doesn't show up.
Did you run a depmod -a? It might not find the new module.
At the moment this is the only idea I have. The patch itself was to include the pipe3 phy driver for USB3 and eSATA.
BR, Nikolaus
On 10/08/2014 09:13 PM, Dr. H. Nikolaus Schaller wrote:
Hi,
Am 08.10.2014 um 20:53 schrieb aTc:
On 10/07/2014 11:56 AM, Dr. H. Nikolaus Schaller wrote:
FYI: I have found the final missing CONFIG and now I have eSATA on the OMAP5432EVM on the GTA04 3.17.0 kernel. I have not tested more than that I can read /dev/sdc since my disk has an incompatible partition table, i.e. I need to re-format it firstly.
Anyways, this will be good to test the Pyra hardware.
Here is the latest version/patch:
http://git.openpandora.org/cgi-bin/gitweb.cgi?p=pyra-kernel.git;a=shortlog;h...
I compiled and ran it on the EVM, but /dev/sda still doesn't show up.
The kernel I normally run on the board finds it at boot and everything works fine. ,
Yes, that is what the gta04 kernel does now for me. It even supports hot-plugging.
the config that one is using is here : https://github.com /RobertCNelson/armv7-lpae-multiplatform/blob/v3.17.x/patches/defconfig
did you try a diff with http://git.openpandora.org/cgi-bin/gitweb.cgi?p=pyra-kernel.git;a=blob;f=arc...
the gta04 kernel itself works, although even when booting from /dev/mmcblk1p2 , /dev/sda doesn't show up.
Did you run a depmod -a? It might not find the new module.
It didn't have any problems loading the other modules, so I assume that all worked fine.
I did run both config through diffuse, but couldn't find anything obviously "wrong" (so there probably is :)
I did find CONFIG_UEVENT_HELPER_PATH="/sbin/hotplug" , which doesn't exist on debian to begin with, and isn't set in the lpea kernel, but that didn't seem to make any difference either when i disabled it.
At the moment this is the only idea I have. The patch itself was to include the pipe3 phy driver for USB3 and eSATA.
I'll try recompiling the kernel and playing with some options around those parts.
Am 08.10.2014 um 21:49 schrieb aTc atc@k-n-p.org:
On 10/08/2014 09:13 PM, Dr. H. Nikolaus Schaller wrote:
Hi,
Am 08.10.2014 um 20:53 schrieb aTc:
On 10/07/2014 11:56 AM, Dr. H. Nikolaus Schaller wrote:
FYI: I have found the final missing CONFIG and now I have eSATA on the OMAP5432EVM on the GTA04 3.17.0 kernel. I have not tested more than that I can read /dev/sdc since my disk has an incompatible partition table, i.e. I need to re-format it firstly.
Anyways, this will be good to test the Pyra hardware.
Here is the latest version/patch:
http://git.openpandora.org/cgi-bin/gitweb.cgi?p=pyra-kernel.git;a=shortlog;h...
I compiled and ran it on the EVM, but /dev/sda still doesn't show up.
The kernel I normally run on the board finds it at boot and everything works fine. ,
Yes, that is what the gta04 kernel does now for me. It even supports hot-plugging.
the config that one is using is here : https://github.com /RobertCNelson/armv7-lpae-multiplatform/blob/v3.17.x/patches/defconfig
did you try a diff with http://git.openpandora.org/cgi-bin/gitweb.cgi?p=pyra-kernel.git;a=blob;f=arc...
the gta04 kernel itself works, although even when booting from /dev/mmcblk1p2 , /dev/sda doesn't show up.
Did you run a depmod -a? It might not find the new module.
It didn't have any problems loading the other modules, so I assume that all worked fine.
I did run both config through diffuse, but couldn't find anything obviously "wrong" (so there probably is :)
I did find CONFIG_UEVENT_HELPER_PATH="/sbin/hotplug" , which doesn't exist on debian to begin with, and isn't set in the lpea kernel, but that didn't seem to make any difference either when i disabled it.
At the moment this is the only idea I have. The patch itself was to include the pipe3 phy driver for USB3 and eSATA.
I'll try recompiling the kernel and playing with some options around those parts.
Can you send me a complete boot log - maybe this reveals something when comparing with mine (failed probing or so).
BR, Nikolaus
On 10/08/2014 10:26 PM, Dr. H. Nikolaus Schaller wrote:
I'll try recompiling the kernel and playing with some options around those parts.
Can you send me a complete boot log - maybe this reveals something when comparing with mine (failed probing or so).
might it have something to do with # CONFIG_USB_XHCI_HCD is not set
On 10/08/2014 11:03 PM, aTc wrote:
On 10/08/2014 10:26 PM, Dr. H. Nikolaus Schaller wrote:
I'll try recompiling the kernel and playing with some options around those parts.
Can you send me a complete boot log - maybe this reveals something when comparing with mine (failed probing or so).
might it have something to do with # CONFIG_USB_XHCI_HCD is not set
Never mind, I must have mixed up some kernel/modules on my boot partition. I did a clean compile of the kernel, and /dev/sda now shows up when booting from mmc.
Having the rootfs on it doesn't work yet, but that's probably just a matter of compiling in some modules.
Am 09.10.2014 um 00:01 schrieb aTc:
On 10/08/2014 11:03 PM, aTc wrote:
On 10/08/2014 10:26 PM, Dr. H. Nikolaus Schaller wrote:
I'll try recompiling the kernel and playing with some options around those parts.
Can you send me a complete boot log - maybe this reveals something when comparing with mine (failed probing or so).
might it have something to do with # CONFIG_USB_XHCI_HCD is not set
Never mind, I must have mixed up some kernel/modules on my boot partition. I did a clean compile of the kernel, and /dev/sda now shows up when booting from mmc.
Ok, fine!
Having the rootfs on it doesn't work yet, but that's probably just a matter of compiling in some modules.
Ah, I see.
Well, then we *must* configure the eSATA related drivers into the kernel and not as modules.
Usually I try to keep the CONFIG=y modules as low as possible so that the kernel binary can fit into a small partition on some NAND flash based system (OMAP3). Therefore, as many CONFIG=m as possible. But this is a special case.
I think a lsmod will tell which modules are dynamically loaded and which dependencies exist.
BR, Nikolaus