The following oops occurs in 2.6.9-rc4 when trying to run something which is in cramfs. The cramfs files are on a IDE flash disk and mounted in /etc/init.d/rcS, whenever I try to execute something that is on the mounted cramfs an oops occurs. However, when doing a network boot and the cramfs files are on the initrd instead of the flash, everything runs fine. btw, kernel v2.4.27 runs fine on both situations. INIT: Entering runlevel: 5 Unable to handle kernel NULL pointer dereference at virtual address 0000000c printing eip: c0248b35 *pde = 00000000 Oops: 0000 [#1] Modules linked in: CPU: 0 EIP: 0060:[] Not tainted VLI EFLAGS: 00010002 (2.6.9-rc4) EIP is at __down+0x45/0xd0 eax: 00000000 ebx: 00000000 ecx: 00000000 edx: 00000001 esi: c74faaa0 edi: 00000246 ebp: 00000008 esp: c7535e3c ds: 007b es: 007b ss: 0068 Process bash (pid: 785, threadinfo=c7535000 task=c74faaa0) Stack: 00000001 c74faaa0 c0113290 00000008 00000000 00036810 c7535e68 c0113311 c10d0280 00000019 c6814000 c028bea0 c0248ccc 00000000 00004000 06814000 c0177a1e 00036810 c112c800 00000000 00000000 c10d0280 00000000 c6fc7658 Call Trace: [] default_wake_function+0x0/0x10 [] __wake_up+0x11/0x20 [] __down_failed+0x8/0xc [] .text.lock.inode+0x7d/0x8f [] read_cache_page+0x5e/0x140 [] cramfs_readpage+0x0/0x160 [] page_getlink+0x1c/0x70 [] page_follow_link_light+0xb/0x20 [] link_path_walk+0x6e8/0xbe0 [] path_lookup+0x73/0x140 [] open_namei+0x8b/0x590 [] filp_open+0x2a/0x50 [] common_interrupt+0x18/0x20 [] sys_open+0x31/0x80 [] syscall_call+0x7/0xb Code: 34 fc 31 c0 f3 ab 89 74 24 04 c7 44 24 08 90 32 11 c0 c7 06 02 00 00 00 9c 5f fa 8b 14 24 8d 6b 08 83 ca 01 89 14 24 89 6c 24 0c <8b> 45 04 8d 54 24 0c 89 55 04 89 44 24 10 89 10 8b 43 04 40 89