I have encountered a BUG_ON in VFS kernel version: 2.6.35.14 Architecture: ARM Test Case: Tried to remove a file from a corrupted filesystem. >From the trigger BUG_ON point it points that the inode which is to be removed - has not matching parent directory from which it is to be removed. Even though -it is reasonable - to verify this parent/child behaviour, but I have a doubt about the placement of BUG_ON at this place. Because this leads to 'halt' of complete system - even though it is related to particular filesystem(file). So, instead of BUG_ON -> WARN_ON could have been used to highlight the issue. Please share your opinion - also is this really related with the FS corruption thing or can be some issue with the dentry caching part? kernel BUG at fs/namei.c:1320! Unable to handle kernel NULL pointer dereference at virtual address 00000000 pgd = deb5c000 [00000000] *pgd=a577d031, *pte=00000000, *ppte=00000000 ... ... PC is at __bug+0x2c/0x38 LR is at sub_preempt_count+0x4c/0x68 pc : [<c003f4b8>] lr : [<c038f8ac>] psr: 60000013 sp : deb39ed0 ip : deb39df0 fp : deb39edc r10: 00000000 r9 : deb38000 r8 : c003b708 r7 : 00000028 r6 : 00000001 r5 : e401c9d0 r4 : e41a5220 r3 : 00000000 r2 : 40000000 r1 : c04b3ce4 r0 : 00000034 ... ... Backtrace(CPU 1): [<c003f48c>] (__bug+0x0/0x38) from [<c0142cb0>] (may_delete+0x44/0x134) [<c0142c6c>] (may_delete+0x0/0x134) from [<c014446c>] (vfs_rmdir+0x24/0xdc) r7:00000028 r6:0000000c r5:e401cbb8 r4:e41a5220 [<c0144448>] (vfs_rmdir+0x0/0xdc) from [<c0146068>] (do_rmdir+0xac/0xf8) r6:0000000c r5:00000000 r4:e41a5220 r3:e41a5220 [<c0145fbc>] (do_rmdir+0x0/0xf8) from [<c014610c>] (sys_rmdir+0x20/0x24) r5:ffffffff r4:00994388 [<c01460ec>] (sys_rmdir+0x0/0x24) from [<c003b4c0>] (ret_fast_syscall+0x0/0x48) Code: e59f0010 e1a01003 eb0d28f0 e3a03000 (e5833000) Thanks & Regards, Amit Sahrawat -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
