[parisc-linux] TODO list

[parisc-linux] TODO list

[Matthew Wilcox]

Here's some things which need to get fixed, courtesy of Alan Cox:

 * Documentation/parisc/unwritten appears out of date

 * cache_init wants some printk DEBUG levels adding
   [In general grep printk(\" is useful to check]

 * arch/parisc/kernel/irq.c has an unchecked kmalloc around line 550.

 * led_get_net_stats() is interrupt-unsafe.

 * ptrace is chock-full of intrusive debug code.  either remove it or
   macroise it.

 * smp_call_function has problems:
	If the calling thread is real time it wont yield - you need to use
	a timeout

	How do we handle IPI failures or retransmits?

 * Why does the superio driver BUG and thus oops rather than printking an
   error and continuing without that driver ?

 * binfmt_som still uses the old strlen_user and needs revisiting to solve
   DoS issues

 * extern asmlinkage long sys32_sysctl(struct __sysctl_args32 *args) is
   a little strange.  Also shouldnt oldlen32 be unsigned ?

-- 
Revolutions do not require corporate support.

Re: [parisc-linux] TODO list

[Matthew Wilcox]

Removing the ones which are fixed (or found to be not-bugs) leaves:

>  * Documentation/parisc/unwritten appears out of date

Grant has this one in progress.

>  * cache_init wants some printk DEBUG levels adding
>    [In general grep printk(\" is useful to check]

Helge in progress.

>  * led_get_net_stats() is interrupt-unsafe.

Helge in progress.

>  * ptrace is chock-full of intrusive debug code.  either remove it or
>    macroise it.
> 
>  * Why does the superio driver BUG and thus oops rather than printking an
>    error and continuing without that driver ?

-- 
Revolutions do not require corporate support.

Re: [parisc-linux] TODO list

[Grant Grundler]

Matthew Wilcox wrote:
> 
> Here's some things which need to get fixed, courtesy of Alan Cox:
> 
>  * Documentation/parisc/unwritten appears out of date

Here's what I have so far. I'll try to finish revivewing the rest
tonight (or tomorrow) and submit bugs for what doesn't get done by the end
of the week.

grant


sys_idle	./include/asm-parisc/unistd.h:  return sys_idle();
	sys_idle() is obsolete - remove it.

reboot_setup()	arch/parisc/kernel/process.c - delete this function

machine_halt		NOP. needs someone else to review.
machine_power_on	obsolete - delete it

exit_thread		NOP
flush_thread		needs someone else to review

die
die_if_no_fixup		Do we need to export either of these?
			Don't think so. We have "die_if_kernel".

strlen_user	#define strlen_user(str) lstrnlen_user(str, 0x7fffffffL)
			Kind of klugy...is that ok?

do_check_pgt_cache	NOP. needs someone else to review.

set_pte_phys		orphaned? needs someone else to review.

ffz		./include/asm-parisc/bitops.h:extern __inline__ unsigned long ffz(unsigned long word)
		I think "static __inline__" is meant...


check_bugs	NOP. TBD.

I haven't looked at these yet:
csum_partial
csum_partial_copy
csum_partial_copy_generic
csum_partial_nocheck
csum_partial_copy_from_user
csum_tcpudp_nofold
csum_tcpudp_magic
sigaddset
sigdelset
get_cycles