Qingqing Zhou <zhouqq(at)cs(dot)toronto(dot)edu> writes:
> There is no LWLock protecting the spinlock in UnpinBuffer(),
I don't believe this is necessary, and if it is necessary the patch
is wrong anyway. Because the code uses _NoHoldoff, there won't be
any check of InterruptPending in that segment of code. Furthermore,
any callers who don't have their own interrupt holdoff in force are
probably broken --- it's unlikely that the state of bufmgr is globally
correct if the caller of UnpinBuffer loses control.
Exactly which path of control are you worried about?
regards, tom lane