Re: [PATCH x86/urgent] bpf: emulate push insns for uprobe on x86
From: Yonghong Song Date: Wed, 8 Nov 2017 13:37:12 -0800 > Uprobe is a tracing mechanism for userspace programs. > Typical uprobe will incur overhead of two traps. > First trap is caused by replaced trap insn, and > the second trap is to execute the original displaced > insn in user space. ... I don't understand how this is bpf related, and if it is you don't explain it well in the commit message. Thank you.
Re: [PATCH x86/urgent] bpf: emulate push insns for uprobe on x86
On 11/8/17 4:06 PM, David Miller wrote: From: Yonghong Song Date: Wed, 8 Nov 2017 13:37:12 -0800 Uprobe is a tracing mechanism for userspace programs. Typical uprobe will incur overhead of two traps. First trap is caused by replaced trap insn, and the second trap is to execute the original displaced insn in user space. ... I don't understand how this is bpf related, and if it is you don't explain it well in the commit message. Right. This is not related to bpf. Will remove the "bpf" from the subject line in the next revision. Thank you.
Re: [PATCH x86/urgent] bpf: emulate push insns for uprobe on x86
On Wed, 8 Nov 2017, Yonghong Song wrote: > On 11/8/17 4:06 PM, David Miller wrote: > > From: Yonghong Song > > Date: Wed, 8 Nov 2017 13:37:12 -0800 > > > > > Uprobe is a tracing mechanism for userspace programs. > > > Typical uprobe will incur overhead of two traps. > > > First trap is caused by replaced trap insn, and > > > the second trap is to execute the original displaced > > > insn in user space. > > ... > > > > I don't understand how this is bpf related, and if it is you don't > > explain it well in the commit message. > > Right. This is not related to bpf. Will remove the "bpf" from the subject line > in the next revision. The proper subject is something like: [PATCH] uprobes/x86: ... which you can figure out by looking at the subsystem prefixes via git log arch/x86/kernel/uprobes.c Note, that it says [PATCH} and nothing else. That patch is a nice performance improvement, but certainly not x86/urgent material. x86/urgent is for bug and regression fixes. Thanks, tglx
Re: [PATCH x86/urgent] bpf: emulate push insns for uprobe on x86
On 11/8/17 10:53 PM, Thomas Gleixner wrote: On Wed, 8 Nov 2017, Yonghong Song wrote: On 11/8/17 4:06 PM, David Miller wrote: From: Yonghong Song Date: Wed, 8 Nov 2017 13:37:12 -0800 Uprobe is a tracing mechanism for userspace programs. Typical uprobe will incur overhead of two traps. First trap is caused by replaced trap insn, and the second trap is to execute the original displaced insn in user space. ... I don't understand how this is bpf related, and if it is you don't explain it well in the commit message. Right. This is not related to bpf. Will remove the "bpf" from the subject line in the next revision. The proper subject is something like: [PATCH] uprobes/x86: ... Thanks, Thomas, I will fix the subject etc. Previously, I added x86/urgent as the branch I did my test on top of it (similar to net-next). I will add that information in the comments and re-submit. which you can figure out by looking at the subsystem prefixes via git log arch/x86/kernel/uprobes.c Note, that it says [PATCH} and nothing else. That patch is a nice performance improvement, but certainly not x86/urgent material. x86/urgent is for bug and regression fixes. Thanks, tglx