Re: [PATCH 14/14] scripts/sorttable: ftrace: Do not add weak functions to available_filter_functions

From
Peter Zijlstra <peterz@infradead.org>
Date
2025-01-02 20:48:14
Also in
bpf, linux-kbuild, lkml
Reply to
Steven Rostedt, 2025-01-02
On Thu, Jan 02, 2025 at 03:41:46PM -0500, Steven Rostedt wrote:
On Thu, 2 Jan 2025 21:32:00 +0100
Peter Zijlstra [off-list ref] wrote:
quoted
This is quite the insane interface -- but whatever. I still feel
strongly you should fix kallsyms so that we can all deal more sanely
with the weak crap.
Question about fixing kallsyms, which I would like done too. I guess an
invisible place holder for weak functions may be best. Saving the size of
all functions could be memory wasteful. As there are a lot of functions:

 # wc -l /proc/kallsyms 
 207126 /proc/kallsyms
IIRC the vast majority of space is taken up by the actual symbol names
-- and rust is only making that *way* worse.
What would be best? To add a placeholder where weak functions are, but they
would not be printed in /proc/kallsyms?  If a lookup occurs, and it lands
on one of theses functions, to return "not found"?
Placeholder yes -- ideally the toolchain itself would not erase the
symbol, but instead mangle it in a well defined way (eg.
<symname>.weak.# or somesuch)

Not printing in kallsyms, I'm not sure, by not printing them it becomes
impossible for userspace consumers of kallsyms to do the same, eg. they
will trip over these same 'holes'.

Default lookup might indeed be best served by returning as if not found.

There's patches out there doing much of the above IIRC.
Thread (36 messages) 36 messages, 4 authors, last reply 2025-01-03
Keyboard shortcuts
hback out one level
jnext message in thread / next series-diff position
kprevious message in thread / previous series-diff position
ldrill in
Escclose help / fold thread tree
?toggle this help