Kernel 4.5

[zzd10h]

@ddni

Ah ok, 3.19... As you posted in 4.5 thread, I thought that you finally booted the 4.5 kernel.

[ddni]

Sorry, my bad I should have started a different thread for my woes.

[mechanic]

I assume you have the modules installed.

Probably the first thing you will want to do is update the system. Mint is pretty good at keeping the installed distro up to date but it's a good idea to do the update.

In the System menu item there should be an Update Manager. Click it. If there are updates just install them. After it finishes do not shutdown the program right away. Give it a few (20) seconds as it may go looking for more updates that depended on any updates it just installed.

Also, you may want to try the 4.1.13 or 4.1.15 kernel and modules. I would stay away from anything newer until you get a good handle on Linux.

Do not use the Software Manager. it is crap.

Time to veg out for a while, then off to bed.

[ddni]

Cheers, away with work all day today so no tinkering. I started a new thread.
http://forum.hyperion-entertainment.biz ... =35&t=3324

[xeno74]

zzd10h wrote:
@xeno74

Tested RC5 & RC6, both the same problem.

Thanks a lot for testing!

I have read your serial log file. Unfortunately I don't see the kernel messages. Could you please boot the Alpha5 and Alpha6 kernel with ‘console=ttyS0,115200’ appended to your boot arguments?

[xeno74]

Hi All,

I have released some Alpha7 kernels for testing today.

New: Linux Git log

FYI:

vmlinux-4.5-a7-without-radeon -> Alpha7 with AMDGPU support but without Radeon support
vmlinux-4.5-a7-without-radeon_with_fb -> Alpha7 with AMDGPU support, without Radeon support but with FB support
vmlinux-4.5 -> Alpha7 with AMDGPU, Radeon, and FB support

Download: vmlinux-4.5-alpha7-AmigaONE_X1000.tar.gz

I tested all these kernels with my XFX Radeon HD6870 today. The kernels vmlinux-4.5-a7-without-radeon and vmlinux-4.5-a7-without-radeon_with_fb boot but without screen output. I was able to login via SSH. Maybe the R7 and the R9 users have a screen output.
The kernel vmlinux-4.5 boots with screen output.



Cheers,

Christian

[xeno74]

zzd10h wrote:
Luigi & Christian

I just tested vmlinux-4.5-a7-without-radeon & vmlinux-4.5 (alpha7) with command

boot -elf -noints -fatfs usbdisk0:vmlinux-4.5-a7-without-radeon console=ttys0,115200
and
boot -elf -noints -fatfs usbdisk0:vmlinux-4.5 console=ttys0,115200

=> no more serial output and the same black screen.

Is Console option are to be included in the boot command ?

thx

Thank you for testing. You need console=ttyS0,115200 in your boot arguments.

For example:

Code: Select all

setenv bootargs "root=/dev/sdb1 console=ttyS0,115200"

[xeno74]

zzd10h wrote:
@Xeno74

Right, it works better with ttyS0.

With your standard kernel, crashs

http://zzd10h.amiga-ng.org/X1000/Linux/ ... Serial.txt

With "without-radeon" kernel
it seems to boot fine but I don't have display on my monitor (I retested without serial bootargs, it doesn't work, no linux startup messages)

http://zzd10h.amiga-ng.org/X1000/Linux/ ... Serial.txt

Many thanks for testing. I have compiled the Alpha8 yesterday. I know that the kernel 4.4 final works with your R7-250X but the kernel 4.5 Alpha1 doesn't work with your R7-250X. Between the final of kernel 4.4 and the first alpha of kernel 4.5 there was a big DRM update for the kernel 4.5. I removed this update (reverted) and after that I compiled the kernel vmlinux-4.5-a8-without-drm-next-4.5. Could you please test this kernel?

Thank you very much for your help!

Download: vmlinux-4.5-alpha8-AmigaONE_X1000.tar.gz

New:

- Linux Git log

FYI:

vmlinux-4.5-a8-without-drm-next-4.5 -> Alpha8 without branch 'drm-next-4.5' (git revert b483666bb4939ee14501da9978e5f87e4aa22228 -m 1)
vmlinux-4.5 -> Alpha8 with all Linux Git updates

Spectre660 wrote:
@Xeno74 & zzd10h

Seeing the same thing on my Sam460ex and Kernel 4.4 .

Having an isssue when adding Radeon TAHITI_vce.bin firmware to kernel.
When not compiled into 4.4.0rc kernels system boots but with error messages :

Direct firmware load for radeon/TAHITI_vce.bin failed with error -2
Can't load firmware "radeon/TAHITI_vce.bin

When I add the TAHITI_vce.bin firmware into the kernel it results in an oops:

Unable to handle kernel paging request for data at address 0xf000001c

Thanks a lot for the hint! Since kernel 4.4-rc4 I integrate the firmwares from the Debian package firmware-amd-graphics (20151018-2). Since this firmware update, Srtest wasn't able to boot his X1000 with the R9 270. But zzd10h could boot his X1000 with R7-250X till/including kernel 4.4 final. That's strange.
I will try to install an old firmware package and integrate it to the Alpha9.

alexdeucher wrote:
I think the firmware is little endian, so you'll probably need to byte
swap it before interpreting it in the driver. Something like the
attached patch should avoid the crash, but I'm not sure if that is
enough for the firmware to load properly.

Alex

Code: Select all

0001-drm-radeon-properly-byte-swap-vce-firmware-setup.patch

From 88ee458e1258c73dae7d2956a134caacf0fe160f Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Fri, 22 Jan 2016 00:13:15 -0500
Subject: [PATCH] drm/radeon: properly byte swap vce firmware setup

Firmware is LE.  Need to properly byteswap some of the fields
so they are interpreted correctly by the driver on BE systems.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/radeon/vce_v1_0.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/drivers/gpu/drm/radeon/vce_v1_0.c b/drivers/gpu/drm/radeon/vce_v1_0.c
index 07a0d37..a01efe3 100644
--- a/drivers/gpu/drm/radeon/vce_v1_0.c
+++ b/drivers/gpu/drm/radeon/vce_v1_0.c
@@ -178,12 +178,12 @@ int vce_v1_0_load_fw(struct radeon_device *rdev, uint32_t *data)
 		return -EINVAL;
 	}
 
-	for (i = 0; i < sign->num; ++i) {
-		if (sign->val[i].chip_id == chip_id)
+	for (i = 0; i < le32_to_cpu(sign->num); ++i) {
+		if (le32_to_cpu(sign->val[i].chip_id) == chip_id)
 			break;
 	}
 
-	if (i == sign->num)
+	if (i == le32_to_cpu(sign->num))
 		return -EINVAL;
 
 	data += (256 - 64) / 4;
@@ -191,18 +191,18 @@ int vce_v1_0_load_fw(struct radeon_device *rdev, uint32_t *data)
 	data[1] = sign->val[i].nonce[1];
 	data[2] = sign->val[i].nonce[2];
 	data[3] = sign->val[i].nonce[3];
-	data[4] = sign->len + 64;
+	data[4] = cpu_to_le32(le32_to_cpu(sign->len) + 64);
 
 	memset(&data[5], 0, 44);
 	memcpy(&data[16], &sign[1], rdev->vce_fw->size - sizeof(*sign));
 
-	data += data[4] / 4;
+	data += le32_to_cpu(data[4]) / 4;
 	data[0] = sign->val[i].sigval[0];
 	data[1] = sign->val[i].sigval[1];
 	data[2] = sign->val[i].sigval[2];
 	data[3] = sign->val[i].sigval[3];
 
-	rdev->vce.keyselect = sign->val[i].keyselect;
+	rdev->vce.keyselect = le32_to_cpu(sign->val[i].keyselect);
 
 	return 0;
 }
-- 

[xeno74]

zzd10h wrote:
@xeno74

Sorry, still crash !

http://zzd10h.amiga-ng.org/X1000/Linux/ ... Serial.txt

Thanks for testing. I am already compiling the Alpha9 with the old firmware package "firmware-nonfree_0.44".

[xeno74]

zzd10h wrote:
@xeno74

Wait a little bit.

The latest 4.4 kernel that work for me is 4.4.rc2 and not 4.4 final RC.

http://zzd10h.amiga-ng.org/X1000/Linux/ ... l44rc2.png

I will have to download all your 4.4 and find the very latest that work...

Fantastic!!!!! With the kernel 4.4-rc4 I updated the firmware package from "firmware-nonfree_0.44" to "firmware-amd-graphics (20151018-2)" It seems it is a firmware problem. Many thanks for the hint!!!!!!!!!!!!