[Bug] [SUSFS] Stuck on bootanimation after merged SUSFS patch in day 4/3 to present.

[prpjzz]

Bug Description

Here is a log when stucking on bootanimation. I think its error due to ROM, but i flashed another rom with older kernel (with susfs) still work, newer kernel without susfs still working.

Kernel version: 4.9
CI: here
DebugAssistant.log

Reproduction Steps

No response

Expected Behavior

No response

Actual Behavior

No response

Additional Context

No response

[prpjzz]

This is basically an elimination method and I've eliminated almost all possibilities.

[JackA1ltman]

I think there is indeed a problem on 4.9 kernel devices, I did scroll through a fair amount of error logs when I was watching my 4.9 kernel device boot up with duplicate errors (I flashed the LiveBoot module), but it eventually booted up. This is not unrelated to the latest patch, and I'll take the time to look into the source of the problem.

[prpjzz]

new log, kernel build from 26/3 bootlog.zip

[JackA1ltman]

boot.log

01-18 01:39:23.566   945   945 W ServiceManagerCppClient: Waited one second for activity (is service started? Number of threads started in the threadpool: 16. Are binder threads started and available?)
01-18 01:39:23.879  1317  1353 W Installer: installd not found; trying again

I've noticed that this error occurs several times, and it seems that the package manager fails to start up

01-18 01:39:09.875   778   778 W KernelSU: ksud::android::init_event: KPM: Failed to start KPM watcher: Inappropriate ioctl for device (os error 25)

Here it shows that your KPM function failed to start up.

--------- beginning of crash
01-18 01:39:11.235   917   917 F libc    : Fatal signal 6 (SIGABRT), code -1 (SI_QUEUE) in tid 917 (vendor.lineage.), pid 917 (vendor.lineage.)
01-18 01:39:11.239   887   887 I HidlServiceManagement: Registered android.hardware.audio.effect@6.0::IEffectsFactory/default

From here on, the boot failure seems to be related to the vendor, which should have nothing to do with SuSFS itself. Are you using a ROM with your own packaged LineageOS?

[prpjzz]

yes, I'm a builder and currently building rom based LineageOS with Device tree, vendor source and kernel source from LineageOS and use KernelSU without SUSFS but not any issue.

I building without KPM, and this error you can skip.

01-18 01:39:09.875   778   778 W KernelSU: ksud::android::init_event: KPM: Failed to start KPM watcher: Inappropriate ioctl for device (os error 25)

I pretty sure this issue from SuSFS, but don't know and how to error.

This is a log when unlocked device, and still have this:

--------- beginning of crash
01-18 01:39:11.235   917   917 F libc    : Fatal signal 6 (SIGABRT), code -1 (SI_QUEUE) in tid 917 (vendor.lineage.), pid 917 (vendor.lineage.)
01-18 01:39:11.239   887   887 I HidlServiceManagement: Registered android.hardware.audio.effect@6.0::IEffectsFactory/default

bootlog-20260326-134059.tar.gz

[JackA1ltman]

I checked that you don't seem to have resolved this .rej file fixing error.
I'm not sure if this will help solve the problem though, but task_mmu.c is where SUSFS hides the more critical ones.

--- fs/proc/task_mmu.c
+++ fs/proc/task_mmu.c
@@ -387,12 +397,41 @@ show_map_vma(struct seq_file *m, struct vm_area_struct *vma, int is_pid)
 	dev_t dev = 0;
 	const char *name = NULL;
 	struct dentry *dentry = NULL;
+#ifdef CONFIG_KSU_SUSFS_OPEN_REDIRECT
+	char *spoofed_redirected_name = NULL;
+#endif // #ifdef CONFIG_KSU_SUSFS_OPEN_REDIRECT
 
 	if (file) {
 		struct inode *inode = file_inode(vma->vm_file);
+#ifdef CONFIG_KSU_SUSFS_OPEN_REDIRECT
+		if (SUSFS_IS_INODE_OPEN_REDIRECT(inode)) {
+			if (!susfs_open_redirect_spoof_show_map_vma(inode, &ino, &dev, spoofed_redirected_name)) {
+				pgoff = ((loff_t)vma->vm_pgoff) << PAGE_SHIFT;
+				goto orig_flow;
+			}
+		}
+#endif // #ifdef CONFIG_KSU_SUSFS_OPEN_REDIRECT
+#ifdef CONFIG_KSU_SUSFS_SUS_MAP
+		if (SUSFS_IS_INODE_SUS_MAP(inode)) {
+			seq_setwidth(m, 25 + sizeof(void *) * 6 - 1);
+			seq_printf(m, "%08lx-%08lx %c%c%c%c %08llx %02x:%02x %lu ",
+				start,
+				end,
+				'-',
+				'-',
+				'-',
+				'p',
+				pgoff,
+				MAJOR(dev), MINOR(dev), ino);
+			goto done;
+		}
+#endif // #ifdef CONFIG_KSU_SUSFS_SUS_MAP
 		dev = inode->i_sb->s_dev;
 		ino = inode->i_ino;
 		pgoff = ((loff_t)vma->vm_pgoff) << PAGE_SHIFT;
+#ifdef CONFIG_KSU_SUSFS_SUS_KSTAT
+		susfs_sus_kstat_spoof_show_map_vma(inode, &dev, &ino);
+#endif
         dentry = file->f_path.dentry;
         if (dentry) {
         	const char *path = (const char *)dentry->d_name.name; 
@@ -403,7 +442,7 @@ show_map_vma(struct seq_file *m, struct vm_area_struct *vma, int is_pid)
             	name = "/dev/ashmem (deleted)";
 		goto done;
             	 	}
-            	if (strstr(path, "jit-zygote-cache")) { 
+            	if (strstr(path, "jit-zygote-cache")) {
 	  	start = vma->vm_start;
 		end = vma->vm_end;
 		show_vma_header_prefix_fake(m, start, end, flags, pgoff, dev, ino);

[prpjzz]

See a.patch

I currently patched some function to work. But maybe kernel issue, i will merge function and remove susfs.

[JackA1ltman]

I think it's time for anyone to give up on SuSFS, he's uncompetitive in this day and age and causes tons of compatibility issues.

[prpjzz]

yea, SuSFS it's not effective after KernelSU 3.0.0 and later on NonGKI.

[prpjzz]

I will closed this issue.