| Message ID | 4AA9C0F1.6060708@np.css.fujitsu.com (mailing list archive) |
|---|---|
| State | Accepted |
| Headers | show |
Hi Jin, On Fri, Sep 11, 2009 at 12:16:01PM +0900, Jin Dongming wrote: > (This patch fixes bug of commit 1c49cd1bbb6e41e97079d389e86063ba5de825ed > titled "Input: i8042 - bypass AUX IRQ delivery test on laptops") > > When dmi_platform_init is called, the table i8042_dmi_laptop_table[] will > be checked. Because the end of this table is not "NULL", dmi_check_system > will lead to pagefault and oops will be happened. > > In fact I got following oops on boot: > Call Trace: > [<ffffffff813f86b8>] panic+0x7a/0x131 > [<ffffffff8105b29b>] ? exit_ptrace+0x94/0x114 > [<ffffffff81054acc>] do_exit+0x7a/0x695 > [<ffffffff813fbaec>] oops_end+0xb9/0xc1 > [<ffffffff81034037>] no_context+0x1f6/0x205 > [<ffffffff81034212>] __bad_area_nosemaphore+0x1cc/0x1f2 > [<ffffffff811f1739>] ? string+0x40/0x9f > [<ffffffff811f20cf>] ? vsnprintf+0x8b/0x426 > [<ffffffff813fac0a>] ?_spin_unlock_irqrestore+0x29/0x41 > [<ffffffff8103424b>] bad_area_nosemaphore+0x13/0x15 > [<ffffffff813fcf9d>] do_page_fault+0x154/0x2a5 > [<ffffffff813faef5>] page_fault+0x25/0x30 > [<ffffffff81335895>] ? dmi_check_system+0x30/0x4b > [<ffffffff816eee5f>] i8042_init+0x2ca/0x3c1 > [<ffffffff816ee5eb>] ? uhci_hcd_init+0x88/0xc0 > [<ffffffff816eeb95>] ? i8042_init+0x0/0x3c1 > [<ffffffff8100a069>] do_one_initcall+0x5e/0x15e > [<ffffffff816be167>] kernel_init+0x170/0x1ca > [<ffffffff81012c5a>] child_rip+0xa/0x20 > [<ffffffff810125bd>] ? restore_args+0x0/0x30 > [<ffffffff81bbdff7>] ? kernel_init+0x0/0x1ca > [<ffffffff81012c50>] ? child_rip+0x0/0x20 > > I confirmed this patch fix this problem. > Thank you for the patch, I will ford it into the original change so that we don't crash the box in case someone will be bisecting...
diff --git a/drivers/input/serio/i8042-x86ia64io.h b/drivers/input/serio/i8042-x86ia64io.h index e1b9ee4..6e244a0 100644 --- a/drivers/input/serio/i8042-x86ia64io.h +++ b/drivers/input/serio/i8042-x86ia64io.h @@ -483,7 +483,7 @@ static struct dmi_system_id __initdata i8042_dmi_laptop_table[] = { DMI_MATCH(DMI_CHASSIS_TYPE, "14"), /* Sub-Notebook */ }, }, - + { } }; #endif
(This patch fixes bug of commit 1c49cd1bbb6e41e97079d389e86063ba5de825ed titled "Input: i8042 - bypass AUX IRQ delivery test on laptops") When dmi_platform_init is called, the table i8042_dmi_laptop_table[] will be checked. Because the end of this table is not "NULL", dmi_check_system will lead to pagefault and oops will be happened. In fact I got following oops on boot: Call Trace: [<ffffffff813f86b8>] panic+0x7a/0x131 [<ffffffff8105b29b>] ? exit_ptrace+0x94/0x114 [<ffffffff81054acc>] do_exit+0x7a/0x695 [<ffffffff813fbaec>] oops_end+0xb9/0xc1 [<ffffffff81034037>] no_context+0x1f6/0x205 [<ffffffff81034212>] __bad_area_nosemaphore+0x1cc/0x1f2 [<ffffffff811f1739>] ? string+0x40/0x9f [<ffffffff811f20cf>] ? vsnprintf+0x8b/0x426 [<ffffffff813fac0a>] ?_spin_unlock_irqrestore+0x29/0x41 [<ffffffff8103424b>] bad_area_nosemaphore+0x13/0x15 [<ffffffff813fcf9d>] do_page_fault+0x154/0x2a5 [<ffffffff813faef5>] page_fault+0x25/0x30 [<ffffffff81335895>] ? dmi_check_system+0x30/0x4b [<ffffffff816eee5f>] i8042_init+0x2ca/0x3c1 [<ffffffff816ee5eb>] ? uhci_hcd_init+0x88/0xc0 [<ffffffff816eeb95>] ? i8042_init+0x0/0x3c1 [<ffffffff8100a069>] do_one_initcall+0x5e/0x15e [<ffffffff816be167>] kernel_init+0x170/0x1ca [<ffffffff81012c5a>] child_rip+0xa/0x20 [<ffffffff810125bd>] ? restore_args+0x0/0x30 [<ffffffff81bbdff7>] ? kernel_init+0x0/0x1ca [<ffffffff81012c50>] ? child_rip+0x0/0x20 I confirmed this patch fix this problem. Signed-off-by: Jin Dongming <jin.dongming@np.css.fujitsu.com> --- drivers/input/serio/i8042-x86ia64io.h | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-)