umh && creds: kill sub_info->cred
cover umh && creds: kill sub_info->cred
(on top of kmod-replace-call_usermodehelper_pipe-with-use-of-umh-init-function-and-resolve-limit.patch)
David, all.
I was going to do some minor fixes in kmod.c, but looking at this code I am really puzzled by subprocess_info->cred complications. I know absoulutely nothing about creds/keys, but at first glance this all looks completely unnecessary?
IOW, please review these 2 simple patches. And sorry, I have no idea how to really test these changes.
In case you are not aware of recent call_usermodehelper() changes in -mm: struct subprocess_info has the new members,
int (*init)(struct subprocess_info *info); void (*cleanup)(struct subprocess_info *info); void *data;
info->init() is called by ____call_usermodehelper() right before kernel_execve(), info->cleanup() is called by call_usermodehelper_freeinfo(). info->data is obvious.
And we have the new helper, call_usermodehelper_setfns(init, cleanup, data) which merely initializes these members.
Oleg.
Quoted 1 line · Oleg Nesterov
(on top of kmod-replace-call_usermodehelper_pipe-with-use-of-umh-init-function-and-resolve-limit.patch)
Can you send me this patch? Your patches as they stand don't apply to the latest Linus tree.
David
Quoted 7 lines · David Howells
Quoted 1 line · Oleg Nesterov
(on top of kmod-replace-call_usermodehelper_pipe-with-use-of-umh-init-function-and-resolve-limit.patch)
Can you send me this patch? Your patches as they stand don't apply to the latest Linus tree.
I guess you need 5 patches from -mm, I'll send you them privately in a minute...
Oleg.