| To: | "Ralf Baechle" <ralf@linux-mips.org>, <linux-mips@linux-mips.org> |
|---|---|
| Subject: | [help] exec_usermodehelper() then content of CURRENT is destroyed |
| From: | "??" <kevin@gv.com.tw> |
| Date: | Fri, 20 Dec 2002 11:37:26 +0800 |
| Cc: | <cms@gv.com.tw> |
| Original-recipient: | rfc822;linux-mips@linux-mips.org |
| References: | <011c01c29c54$7aa64c60$e20310ac@gv.com.tw> <20021205132358.A5634@linux-mips.org> |
| Sender: | linux-mips-bounce@linux-mips.org |
dear all, 1)in our 2.4.18 mips kernel , when we call exec_usermodehelper() (in kernel/kmod.c) the content of CURRENT is destroyed during [page_fault handler loading pages from storage-device to ram] which kind of error may cause this problem? any hint/idea? 2)btw, in some mips linux port, [syscall] is done by [scall_o32.S] when [handle_sys] is done, it will return by [o32_ret_from_sys_call], [o32_ret_from_sys_call] will do [reschedule] and [signal], but the question is here it doesn't check whether [syscall] is from User-space or Kernel-space if we use [syscall] in kernel-space, and do [signal] after [syscall] return, won't it cause problem? is this bug of linux code? thanks alot in advanced! |
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | Re: [PATCH]: for poor sools with old I2 & 64 bits kernel, Ralf Baechle |
|---|---|
| Next by Date: | hpt366, Pete Popov |
| Previous by Thread: | Re: enlarge KERNEL_STACK_SIZE, Ralf Baechle |
| Next by Thread: | EJTAG and NMI handler broken, Carsten Langgaard |
| Indexes: | [Date] [Thread] [Top] [All Lists] |