Skip to contentSkip to conversation

ftrace: to kill a daemon

PATCH Bodo Eggert 3 participants 7 messages Aug 7, 2008 – Aug 9, 2008 13 messages not in the archive

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

#1 lore raw
Quoted 2 lines · Steven Rostedt

The idea I came
up with was to make a tmp.s file for every object just after it is compiled.

make -j would cause these files to overwrite each other, wouldn't it?

#2 lore raw
Quoted 9 lines · Bodo Eggert
Quoted 3 lines · Steven Rostedt

The idea I came
up with was to make a tmp.s file for every object just after it is compiled.

make -j would cause these files to overwrite each other, wouldn't it?

I'd assume the name is actually derived from the real filename...

J

#3 lore raw
In reply to Bodo Eggert
Quoted 6 lines · Bodo Eggert
Quoted 2 lines · Steven Rostedt

The idea I came
up with was to make a tmp.s file for every object just after it is compiled.

make -j would cause these files to overwrite each other, wouldn't it?

I simplified my explanation. I really used something like this:

for init/main.o:

init/.tmp_mc_main.s : for the list file init/.tmp_mc_main.o : for the list compiled object. init/.tmp_gl_main.o : for the object converted static to global. init/.tmp_mx_main.o : for the merged list and global object.

The tmp.s was to keep the focus on the concept. But the actual implementation handles multi-threaded builds.

I use distcc and build with -j40. I had no issues.

Thanks,

-- Steve

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

#16 lore raw
Quoted 5 lines · Steven Rostedt
Quoted 4 lines · Andi Kleen
Quoted 1 line · Steven Rostedt

I'm stubborn, I want to get it right _and_ keep it fast.

Quoted 7 lines
Quoted 4 lines

For me it would seem better to just not use two part 5 byte nops instead of adding such hacks. I doubt there are that many of them anyways. I bet you won't be able to measure any difference between the different nop types in any macro benchmark.

I wish we had a true 5 byte nop. The alternative is a jmp 0, which is measurable.

Did you try short jumps? (0xeb 0x03 0x?? 0x?? 0x??)

#17 lore raw
Quoted 11 lines · Bodo Eggert
Quoted 10 lines · Steven Rostedt
Quoted 8 lines · Andi Kleen
Quoted 1 line · Steven Rostedt

I'm stubborn, I want to get it right _and_ keep it fast.

For me it would seem better to just not use two part 5 byte nops instead of adding such hacks. I doubt there are that many of them anyways. I bet you won't be able to measure any difference between the different nop types in any macro benchmark.

Quoted 4 lines
Quoted 2 lines

I wish we had a true 5 byte nop. The alternative is a jmp 0, which is measurable.

Did you try short jumps? (0xeb 0x03 0x?? 0x?? 0x??)

What would those last three bytes be?

-- Steve

#18 lore raw
Quoted 4 lines · Steven Rostedt
Quoted 2 lines

Did you try short jumps? (0xeb 0x03 0x?? 0x?? 0x??)

What would those last three bytes be?

Ah, I see, you jump 3 bytes forward.

I'm pretty sure that it is still more expensive than a nop, on some archs. But I'm not 100% sure on that.

-- Steve

#19 lore raw
In reply to Steven Rostedt
Quoted 12 lines · Steven Rostedt
Quoted 11 lines · Bodo Eggert
Quoted 10 lines · Steven Rostedt
Quoted 8 lines · Andi Kleen
Quoted 1 line · Steven Rostedt

I'm stubborn, I want to get it right _and_ keep it fast.

For me it would seem better to just not use two part 5 byte nops instead of adding such hacks. I doubt there are that many of them anyways. I bet you won't be able to measure any difference between the different nop types in any macro benchmark.

Quoted 6 lines
Quoted 4 lines
Quoted 2 lines

I wish we had a true 5 byte nop. The alternative is a jmp 0, which is measurable.

Did you try short jumps? (0xeb 0x03 0x?? 0x?? 0x??)

What would those last three bytes be?

Anything, since the CPU will ignore them.

My hope is that different kinds of jump will behaver differently, but I fear the side effect of the jmp (reread memory in case of self-modifying code) will cause the CPU to slow down anyway.

Signature · 1 line
Is reading in the bathroom considered Multitasking?