Messages in this thread Patch in this message | | | Date | Sat, 10 May 2003 03:59:42 -0400 (EDT) | From | Zwane Mwaikambo <> | Subject | [PATCH][2.5] export cpufreq_driver to fix oops in proc interface |
| |
The proc interface has no way of telling wether there is an active cpufreq driver or not. This means that if you don't have a cpufreq supported processor, this will oops in various possible places.
This hasn't been committed in case there is a cleaner way.
Unable to handle kernel NULL pointer dereference at virtual address 00000020 printing eip: c013589f *pde = 00000000 Oops: 0000 [#1] CPU: 0 EIP: 0060:[<c013589f>] Not tainted EFLAGS: 00010246 EIP is at cpufreq_cpu_get+0x1f/0xe0 eax: 00000000 ebx: 00000000 ecx: cbbf3ba0 edx: cbc21eb0 esi: 00000001 edi: cbc21eb0 ebp: cbc21e68 esp: cbc21e5c ds: 007b es: 007b ss: 0068 Process mc (pid: 300, threadinfo=cbc20000 task=c13ac780) Stack: cbc21e74 00000000 00000000 cbc21e88 c01367e1 00000000 cbc21e88 c01b5a4f 00000000 c59ce045 00000000 cbc21f1c c023434f cbc21eb0 00000000 c02dfaac cbc20000 c02dfaac cbc21ed0 c013d15c c10e0830 00000000 c02e0f30 00000000 Call Trace: [<c01367e1>] cpufreq_get_policy+0x21/0x90 [<c01b5a4f>] sprintf+0x1f/0x30 [<c023434f>] cpufreq_proc_read+0x8f/0x150 [<c013d15c>] buffered_rmqueue+0xbc/0x160 [<c013d28d>] __alloc_pages+0x8d/0x300 [<c02342c0>] cpufreq_proc_read+0x0/0x150 [<c017f2ee>] proc_file_read+0xbe/0x250 [<c0153ae3>] vfs_read+0xd3/0x140 [<c0164784>] do_fcntl+0xd4/0x1c0 [<c0153d8c>] sys_read+0x3c/0x60 [<c010a33b>] syscall_call+0x7/0xb
Code: 8b 50 20 85 d2 74 2b b8 00 e0 ff ff 21 e0 ff 40 14 83 3a 02
Index: linux-2.5.69/include/linux/cpufreq.h =================================================================== RCS file: /build/cvsroot/linux-2.5.69/include/linux/cpufreq.h,v retrieving revision 1.1.1.1 diff -u -p -B -r1.1.1.1 cpufreq.h --- linux-2.5.69/include/linux/cpufreq.h 6 May 2003 12:20:41 -0000 1.1.1.1 +++ linux-2.5.69/include/linux/cpufreq.h 10 May 2003 07:33:01 -0000 @@ -310,5 +309,8 @@ void cpufreq_frequency_table_put_attr(un #endif /* CONFIG_CPU_FREQ_TABLE */ + +/* Currently exported only for the proc interface, remove when that goes */ +extern struct cpufreq_driver *cpufreq_driver; #endif /* _LINUX_CPUFREQ_H */ Index: linux-2.5.69/kernel/cpufreq.c =================================================================== RCS file: /build/cvsroot/linux-2.5.69/kernel/cpufreq.c,v retrieving revision 1.1.1.1 diff -u -p -B -r1.1.1.1 cpufreq.c --- linux-2.5.69/kernel/cpufreq.c 6 May 2003 12:21:33 -0000 1.1.1.1 +++ linux-2.5.69/kernel/cpufreq.c 10 May 2003 07:28:10 -0000 @@ -28,9 +28,11 @@ * level driver of CPUFreq support, and its locking mutex. * cpu_max_freq is in kHz. */ -static struct cpufreq_driver *cpufreq_driver; +struct cpufreq_driver *cpufreq_driver; static DECLARE_MUTEX (cpufreq_driver_sem); +/* required for the proc interface, remove when that goes away */ +EXPORT_SYMBOL_GPL(cpufreq_driver); /** * Two notifier lists: the "policy" list is involved in the Index: linux-2.5.69/drivers/cpufreq/proc_intf.c =================================================================== RCS file: /build/cvsroot/linux-2.5.69/drivers/cpufreq/proc_intf.c,v retrieving revision 1.1.1.1 diff -u -p -B -r1.1.1.1 proc_intf.c --- linux-2.5.69/drivers/cpufreq/proc_intf.c 6 May 2003 12:20:36 -0000 1.1.1.1 +++ linux-2.5.69/drivers/cpufreq/proc_intf.c 10 May 2003 07:15:33 -0000 @@ -209,6 +209,9 @@ static int __init cpufreq_proc_init (voi { struct proc_dir_entry *entry = NULL; + if (!cpufreq_driver) + return -ENODEV; + /* are these acceptable values? */ entry = create_proc_entry("cpufreq", S_IFREG|S_IRUGO|S_IWUSR, &proc_root); -- function.linuxpower.ca - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |