Huxy

This is pretty close to it. I spent some time in the summer working on a kernel build unique to my hardware. Part of which was running the guest using PVHVM drivers. If I remember correctly the first part of the problem is that Xen PV disk drivers don't use the SCSI drivers that synology relies on. Instead it uses a special block device driver and remaps the drives to xvd[] style mappings. The second is that Xen block drivers aren't seen as systemdisks. Once I made these change I had issues with remapping of the xvd[] devices conflicting with the sd[] device numbering. I had to create a new boot param which would add an offset to the sd[] drives. i.e. 2 xen drives would mean the first drive using SCSI drivers would now utilise sdc and not sda. Preventing the system hangs on boot. Finally I tweaked the kernel configs and rebuilt it from scratch. After that I was able to use 5.0 with PVHVM drivers (with a heavily customised boot image). It is possible to run PVHVM and 5.0, it just takes a lot of effort. Additional Notes: I was also never able to get the drives to shutdown reliably when rebooting. I had to repair the storage on boot, however, none of the data was ever compromised. Additionally the dom0 logs were flooded with this message, regardless of my memory settings in the config file. I think this is a bug in dom0's kernel though as I've seen it in other DomU's. [8200018.808510] xen:balloon: reserve_additional_memory: add_memory() failed: -17

I managed to track down the issue I was having. Now I can get the install to complete, but I'm getting an issue where I get a segfault and then the system shuts down before I can check the log files to see what's causing the issue. I've attached a screenshot. Again, any ideas? I'm not sure if scemd is raid related? Which could make sense as I've had issues with the xen pv frontend driver. Interestingly, If I boot gnoboot it updates all the modules and then I can boot using my custom kernel thereafter (albeit with errors due to toolchain differences). Cheers

I've been able to compile and boot my own kernel, but I'm having an issue when installing DSM. Every reboot it just states that the configuration has been lost. Searching around the forum I think this might be related to installing DSM directly on SATA/SAS hardware? I can see some patch notes from trant0r that says he changed his synoboot to allow this, but I can't find any instructions on how this was done. Any help would be appreciated! Cheers

Thanks! This worked perfectly.

With regards to this driver, v1 and v2 boot, but only show one drive. v3 and v4 hang in an infinite loop on boot with the following error messages: ata7.15: SATA link up 3Gbps (SStatus 123 SControl 0) ata7: Apply Synology fast PMP boot ata7.x failed to resume link ata7.x COMRESET failed (errno=-5) ata7.x COMRESET fail, set COMERSET fail flag ata7.x failed to read SCR2 (Emask=0x40) ata7.x failed to read SCR2 (Emask=0x40) ata7.x hard resetting link ata7.x reset failed (errno=-85), retying in 10secs) And it just loops, moving from ata7.00 to ata7.15 Unfortunately, I can't get a dmesg printout as boot never progresses. I'm using the Sil 3132 eSata controller and the ICYBOX IB-3640 eSata bay. Ubuntu and Debian both detect and load the port multiplier fine, which is what's consistently been said around the forums. Not sure if any of that info helps.