diff --git a/README.md b/README.md
index 48513c4..382c363 100644
--- a/README.md
+++ b/README.md
@@ -93,6 +93,16 @@
 #  https://lists.archlinux.org/pipermail/aur-general/2022-January/036767.html
 #  https://aur.archlinux.org/packages/makedumpfile/#comment-843853
 #
+#  (e) NVMe multipathing introduced a "randomness" level to device naming on
+#  Linux, so "nvme0n1" could be "nvme1n1" in some boots, if we have more than
+#  one NVMe device. There's a kernel parameter to avoid that
+#  ("nvme_core.multipath=0"). So, since we rely in getting the NVMe device name
+#  to be used in kdump during the regular boot process, we could in theory have
+#  different names between regular kernel boot and the kdump one, hence causing
+#  a failure in kdump collection. But this is pretty much safe now since we
+#  don't have multiple NVMe devices, also we could disable multipath in kernel
+#  config (CONFIG_NVME_MULTIPATH) or use the above cmdline.
+#
 #
 #  TODOs
 #  ###########################################################################
@@ -103,10 +113,8 @@
 #  images - it happens due to our package installing files on directory
 #  "/usr/lib/dracut/modules.d" which triggers the unfortunate initramfs rebuild.
 #
-#  * We have a "fragile" way of determining a mount point required for Kdump;
-#  this is something to improve maybe, in order to make the Kdump more reliable.
-#  Also in the list of fragile things, VDF parsing is...complicated. Something
-#  that would be nice to improve as well.
+#  * VDF parsing would benefit from some improvement, it's at least "fragile"
+#  for now, to be generous...but that seems a bit complicated.
 #
 #  * Pstore ramoops back-end has some limitations that we're discussing with
 #  the kernel community - right now we can only collect ONE dmesg and its
diff --git a/kdump-collect.sh b/kdump-collect.sh
index 9f03016..dd14dea 100644
--- a/kdump-collect.sh
+++ b/kdump-collect.sh
@@ -24,8 +24,10 @@ if [ ! -f $VMCORE ]; then
 	reboot -f
 fi
 
+DEVN="$(cat /usr/lib/kdump/kdump.devnode)"
 mkdir -p "/kdump_path"
-if ! mount "/dev/${MOUNT_DEVNODE}" /kdump_path; then
+
+if ! mount "${DEVN}" /kdump_path; then
 	reboot -f
 fi
 
@@ -39,7 +41,7 @@ if [ "${FULL_COREDUMP}" -ne 0 ]; then
 	sync "${KDUMP_FOLDER}/vmcore.compressed"
 fi
 
-umount "/dev/${MOUNT_DEVNODE}"
+umount "${DEVN}"
 sync
 
 reboot -f
diff --git a/kdump-load.sh b/kdump-load.sh
index b6270ac..5cc85f9 100644
--- a/kdump-load.sh
+++ b/kdump-load.sh
@@ -53,8 +53,8 @@ fi
 
 . /etc/default/kdump
 
-#  Fragile way for finding the proper mount point for DEVNODE:
-DEVN_MOUNTED=$(mount |grep "${MOUNT_DEVNODE}" | head -n1 | cut -f3 -d\ )
+#  Find the proper mount point for /home:
+DEVN_MOUNTED="$(findmnt "${MOUNT_DEVNODE}" -fno TARGET)"
 KDUMP_FOLDER="${DEVN_MOUNTED}/${KDUMP_FOLDER}"
 
 echo "${KDUMP_FOLDER}" > "${KDUMP_MNT}"
diff --git a/kdump.etc b/kdump.etc
index 5bfa060..6c8e0a3 100644
--- a/kdump.etc
+++ b/kdump.etc
@@ -9,12 +9,13 @@
 #  /usr/lib/kdump/kdump-load.sh initrd
 #
 
-#  Mount-related options - the DEVNODE must exist and be available during the
-#  kdump script execution. The KDUMP_FOLDER will be create if doesn't exist.
-#  The KDUMP_MNT is just a temporary file that will carry the mounted folder
-#  path across boot-time scripts.
+#  Mount-related options - the DEVNODE points to the /home directory link;
+#  this is used to derive the numerical devnode for kdump, since the link
+#  is not present so early in the system boot. The KDUMP_FOLDER will be
+#  created if doesn't exist. The KDUMP_MNT is just a temporary file that
+#  carries the mounted folder path across boot-time scripts.
 
-MOUNT_DEVNODE="nvme0n1p8"
+MOUNT_DEVNODE="/dev/disk/by-partsets/shared/home"
 KDUMP_FOLDER="/.steamos/offload/var/kdump"
 KDUMP_MNT="/tmp/kdump.mnt"
 
diff --git a/module-setup.sh b/module-setup.sh
index fb9ab7d..d7b714c 100644
--- a/module-setup.sh
+++ b/module-setup.sh
@@ -18,6 +18,13 @@ installkernel() {
 }
 
 install() {
+    #  Having a valid /etc/default/kdump is essential for kdump.
+    if [ ! -f "/etc/default/kdump" ]; then
+        return 1
+    fi
+
+    . /etc/default/kdump
+
     #  First clear all unnecessary firmwares/drivers added by drm in order to
     #  reduce the size of this minimal initramfs being created. This should
     #  be already done via command-line arguments, but let's play safe and delete
@@ -31,6 +38,12 @@ install() {
     inst makedumpfile
 
     mkdir -p $initdir/usr/lib/kdump
+
+    #  Determine the numerical devnode for kdump, and save it on initrd;
+    #  notice that partset link is not available that early in boot time.
+    DEVN="$(readlink -f "${MOUNT_DEVNODE}")"
+    echo "${DEVN}" > $initdir/usr/lib/kdump/kdump.devnode
+
     cp -LR --preserve=all /usr/lib/kdump/* $initdir/usr/lib/kdump/
     cp -LR --preserve=all /etc/default/kdump $initdir/usr/lib/kdump/kdump.etc