xfs: allow disabling xfs tracepoints via Kconfig
PATCH 1/1 xfs: allow disabling xfs tracepoints via Kconfig
linux/tracepoints.h allows individual subsystems to disable their tracepoints. Add such a knob for xfs. Disabling XFS_TRACEPOINTS reduces the resident size of xfs.ko by about a third, or ~350 KiB.
$ size /tmp/xfs/{a,b}/xfs.ko
text data bss dec hex filename
671726 4129 632 676487 a5287 /tmp/xfs/a/xfs.ko
893192 166737 632 1060561 102ed1 /tmp/xfs/b/xfs.ko--- fs/xfs/Kconfig | 9 +++++++++ fs/xfs/xfs_trace.h | 4 ++++ 2 files changed, 13 insertions(+)
fs/xfs/Kconfig +9 -0
@@ -141,3 +141,12 @@ config XFS_ASSERT_FATAL result in warnings. This behavior can be modified at runtime via sysfs. + +config XFS_TRACEPOINTS + bool "XFS tracepoints" + default y + depends on XFS_FS && TRACEPOINTS + help + Say Y here to build XFS with tracepoints. + + You can say N here to reduce the size of the kernel image.
fs/xfs/xfs_trace.h +4 -0
@@ -6,6 +6,10 @@#undef TRACE_SYSTEM#define TRACE_SYSTEM xfs +#ifndef CONFIG_XFS_TRACEPOINTS +#define NOTRACE +#endif +#if !defined(_TRACE_XFS_H) || defined(TRACE_HEADER_MULTI_READ )#define _TRACE_XFS_H
Signature · 1 line
2.20.1
Quoted 3 lines · Rasmus Villemoes
linux/tracepoints.hallows individual subsystems to disable their tracepoints. Add such a knob for xfs. DisablingXFS_TRACEPOINTSreduces the resident size of xfs.ko by about a third, or ~350 KiB.
LOL. :)
Quoted 9 lines
$ size /tmp/xfs/{a,b}/xfs.ko text data bss dec hex filename 671726 4129 632 676487 a5287 /tmp/xfs/a/xfs.ko 893192 166737 632 1060561 102ed1 /tmp/xfs/b/xfs.koSigned-off-by Rasmus Villemoes <[email protected]>--- fs/xfs/Kconfig | 9 +++++++++ fs/xfs/xfs_trace.h | 4 ++++
I think a similar thing ought to apply to fs/xfs/scrub/trace.h, right?
More general questions:
(a) Do we want to be doing this on a per-subsystem basis?
(b) Since xfs_scrub is a separate trace domain, do we want a separate Kconfig item for toggling scrub trace, or is one big one enough?
(I can't think of why I'd want one on and the other off.)
--D
Quoted 37 lines
2 files changed, 13 insertions(+)fs/xfs/Kconfig +9
-0@@ -141,3 +141,12 @@ config XFS_ASSERT_FATAL result in warnings. This behavior can be modified at runtime via sysfs. + +config XFS_TRACEPOINTS + bool "XFS tracepoints" + default y + depends on XFS_FS && TRACEPOINTS + help + Say Y here to build XFS with tracepoints. + + You can say N here to reduce the size of the kernel image.fs/xfs/xfs_trace.h +4
-0@@ -6,6 +6,10 @@#undef TRACE_SYSTEM#define TRACE_SYSTEM xfs +#ifndef CONFIG_XFS_TRACEPOINTS +#define NOTRACE +#endif +#if !defined(_TRACE_XFS_H) || defined(TRACE_HEADER_MULTI_READ )#define _TRACE_XFS_HSignature · 1 line
2.20.1
Quoted 3 lines · Rasmus Villemoes
linux/tracepoints.hallows individual subsystems to disable their tracepoints. Add such a knob for xfs. DisablingXFS_TRACEPOINTSreduces the resident size of xfs.ko by about a third, or ~350 KiB.
Ok, now we can't debug typical problems on live production systems if tracepoints are turned off on the user/distro kernels. So under what circumstances would we ever want to turn off tracepoints on XFS?
Cheers,
Dave.
Signature · 2 lines
Dave Chinner [email protected]
Quoted 9 lines · Dave Chinner
Quoted 3 lines · Rasmus Villemoes
linux/tracepoints.hallows individual subsystems to disable their tracepoints. Add such a knob for xfs. DisablingXFS_TRACEPOINTSreduces the resident size of xfs.ko by about a third, or ~350 KiB.Ok, now we can't debug typical problems on live production systems if tracepoints are turned off on the user/distro kernels. So under what circumstances would we ever want to turn off tracepoints on XFS?
I don't expect any mainstream distros to turn it off. But for embedded systems that use a hand-tuned .config, being able to shave off 100s of K of the kernel image is quite valuable. Tracing _is_ useful, also/especially when doing embedded development, which is why "just turn off CONFIG_TRACING" isn't really an option.
Would the knob be more acceptable if it was under CONFIG_EXPERT?
Rasmus
Quoted 6 lines · Rasmus Villemoes
Quoted 5 lines · Dave Chinner
Quoted 3 lines · Rasmus Villemoes
linux/tracepoints.hallows individual subsystems to disable their tracepoints. Add such a knob for xfs. DisablingXFS_TRACEPOINTSreduces the resident size of xfs.ko by about a third, or ~350 KiB.
Quoted 8 lines
Quoted 4 lines
Ok, now we can't debug typical problems on live production systems if tracepoints are turned off on the user/distro kernels. So under what circumstances would we ever want to turn off tracepoints on XFS?
I don't expect any mainstream distros to turn it off. But for embedded systems that use a hand-tuned .config, being able to shave off 100s of K of the kernel image is quite valuable.
However, small embedded systems don't use XFS because of it's size, dependence on 64 bit capability (ie. 64BIT || LBDAF), its CPU and memory overhead in comparison to other filesystems, etc.
Increasing kconfig options increases the test matrix (even just the "does this changeset compile" matrix) so these things don't come for free. XFS is firmly focussed on the other end of the scale (big, lots, fast), so I'm not sure that increasing the kconfig combination matrix to cater for really low end embedded systems is really that worthwhile.
Quoted 3 lines
Tracing _is_ useful,
also/especially when doing embedded development, which is why "just turn offCONFIG_TRACING" isn't really an option.
Do you have a specific need for this, or is it just "I noticed this, here's a patch"? If you need it and will use it, great, let's add it. But config options that don't ever get used tend to rot....
Cheers,
Dave.
Signature · 2 lines
Dave Chinner [email protected]
Quoted 9 lines · Dave Chinner
Quoted 5 lines · Rasmus Villemoes
Quoted 4 lines · Dave Chinner
Quoted 3 lines · Rasmus Villemoes
linux/tracepoints.hallows individual subsystems to disable their tracepoints. Add such a knob for xfs. DisablingXFS_TRACEPOINTSreduces the resident size of xfs.ko by about a third, or ~350 KiB.Do you have a specific need for this, or is it just "I noticed this, here's a patch"?
A little of both, actually. I noticed it, did the patch for the kernel I use on my laptop, and thought others might find it useful. But no, I don't currently deal with any customer boards where it'd be relevant.
Quoted 2 lines
If you need it and will use it, great, let's add
it. But config options that don't ever get used tend to rot....
Fair enough, I'll just keep it privately and resend if I encounter a use case in the wild.
Rasmus