https://bugzilla.kernel.org/show_bug.cgi?id=201199 Bug ID: 201199 Summary: hpsa blocking boot on HP smart Array P410 Product: IO/Storage Version: 2.5 Kernel Version: 4.18.3 Hardware: Intel OS: Linux Tree: Mainline Status: NEW Severity: normal Priority: P1 Component: SCSI Assignee: linux-scsi@xxxxxxxxxxxxxxx Reporter: seblu@xxxxxxxxx Regression: No Hello, I lease an HP Proliant server, with an Intel Xeon E31220 @3.10Ghz, 16GHz RAM and a Hewlett-Packard Smart Array P410. The operating system is an (up to date) Arch Linux. The Smart Array P410 firmware version is 6.64 (the last one, from 2015), the BIOS was updated to the last version (from 2018). I have an issue I cannot understand an my investigations are currently stalled. I cannot boot any Arch or vanilla kernels newer than 4.9.x. The last and currently booted is a 4.9.128 (config will be attached). Any higher version I tested, up to 4.19-rc4, are stalled because disks are not detected. As a side note, the issue looks like bug #199703, but the solution was to upgrade the controller firmware to the last version, which is my case, didn't fix the issue. I will attach the error screenshots from the console, but basically, looks like the hpsa didn't detect the hard drives and systemd-udevd is stalled. Not sure it's important, but it's not possible to rmmod hpsa module with the initramfs fallback shell. I tried to play with hpsa_allow_any=1, scsi_mod.scan=sync, iommu=off but nothing helped. Then, I tried to boot with a live cd of ubuntu 18.04 (kernel 4.15.0-22), and it worked. So, I used the config file from this kernel and built a vanilla kernel (ubuntu 4.15.0-22 is based on a vanilla 4.15.17) to boot on. Unfortunately it didn't worked, still stalled at boot. So, I took the ubuntu bionic 4.15.0-22 kernel sources and I built a kernel to boot my Arch Linux with. It didn't work neither. So, I took the already compiled Ubuntu kernel and initramfs from the livecd to boot with. And it booted! There were expected issues with missing modules but the hpsa worked and the drives was detected. That's weird, because there is a difference between my built from bionic sources and the one ubunbu did I cannot explain. I don't think the toolchain version, like GCC recent version require a patch to built objtool, should be a reason of booting or not. So, I'm stuck with no idea to where I should dig. I need some help, do you have suggestions? -- You are receiving this mail because: You are the assignee for the bug.
