> > > Hmm, instead of tracing nothing, as this is already a "conditional > > > trace event", why not add to that condition: > > > > > > TP_CONDITION(res >= 0 && read_write == I2C_SMBUS_READ), > > > > > > Unless you want to still trace some data on failure. > > > > John, any comment to this? > > > > The issue we were dealing with was access to uninitialized memory on the stack. > The change '30f939feaeee i2c: fix kernel memory disclosure in dev interface' > does the initialization, so the tracing code is no longer affected. > We just didn't > have that change in the particular kernel we were testing. > > Still, Steven's suggestion seems fine to me. Would you like me to create a new > patch based on that? I am not a tracing expert, but all that was said here tell me that it would be better to have the change. So, yes, please, that would be appreciated.
Attachment:
signature.asc
Description: PGP signature
