I'm not sure I understand the problem. Load_elf_binary also
uses sys_close for recovery in case of error. Can you please
give me more details on the problems you see with using sys_close ?
Thanks,
Yoav.
Yoav Zach
IA-32 Execution Layer
Performance Tools Lab
Intel Corp.
>-----Original Message-----
>From: Arjan van de Ven [mailto:[email protected]]
>Sent: Saturday, June 19, 2004 17:39
>To: Linux Kernel Mailing List
>Cc: Zach, Yoav; [email protected]
>Subject: Re: [PATCH] Handle non-readable binfmt_misc executables
>
>
>> +_error_close_file:
>> + if (fd_binary > 0) {
>> + sys_close (fd_binary);
>> + fd_binary = -1;
>> + bprm->file = NULL;
>> +
>
>
>ewww sys_close.... there HAS to be a better way to do that...
>
On Sun, Jun 20, 2004 at 01:41:30PM +0300, Zach, Yoav wrote:
> I'm not sure I understand the problem. Load_elf_binary also
> uses sys_close for recovery in case of error. Can you please
> give me more details on the problems you see with using sys_close ?
for one, sys_close, while currently exported, shouldn't be really.
(it is exported right now for a few drivers that have invalid firmware
loaders that haven't been converted to the firmware loading framework).
In addition it's way overkill, you created the fd so half the safety
precautions shouldn/t be needed