Messages in this thread | | | Date | Thu, 26 May 2011 09:08:05 +0800 | From | Li Zefan <> | Subject | Re: [PATCH 1/2] tracing: add __print_symbolic_u64 to avoid warnings on 32bit machine |
| |
Steven Rostedt wrote: > On Wed, 2011-05-25 at 09:12 -0700, Randy Dunlap wrote: >> >>> But what I sent out previously is not going into 2.6.40 anyway. Ingo >>> feels that it's too late in the merge window to pull those changes and >>> wants to wait till the next merge window. Since there's some fixes in >>> that pull request, I'll need to cherry pick them out and rebase the rest >>> for the next window. >> >> Sorry to hear all of that. This fix is needed in mainline. >> > > Hmm, it is a warning fix. Well, I could add this to the cherry picking > I'm doing (fixes only). > > I'm currently making a fix for a triple fault in the function tracer > caused by PROVE_RCU causing recursive tracing. When I get this fixed, > I'll be pushing the "fixes" next. >
Actually it's more than a compile warning fix.
After a variable is cast from u64 to u32 and passed to __print_symbolic(), we'll see incorrect result in the trace output.
| |