From 33a38f7453a5c2b4afca8729c081bda79afa2a44 Mon Sep 17 00:00:00 2001 From: "Kenneth D. Merry" Date: Fri, 7 Dec 2012 23:48:54 +0000 Subject: [PATCH] Fix a panic during CAM EDT traversal. The problem was a race condition between the EDT traversal used by things like 'camcontrol devlist', and CAM peripheral driver removal. The EDT traversal code holds the CAM topology lock, and wants to show devices that have been invalidated. It acquires a reference to the peripheral to make sure the peripheral it is examining doesn't go away. However, because the peripheral removal code in camperiphfree() drops the CAM topology lock to call the peripheral's destructor routine, we can run into a situation where the EDT traversal increments the peripheral reference count after free process is already in progress. At that point, the reference count is ignored, because it was 0 when we started the process. Fix this race by setting a flag, CAM_PERIPH_FREE, that I previously added and checked in xptperiphtraverse() and xptpdperiphtravsere(), but failed to use. If the EDT traversal code sees that flag, it will know that the peripheral free process has already started, and that it should not access that peripheral. Also, fix an inconsistency in the locking between xptpdperiphtraverse() and xptperiphtraverse(). They now both hold the CAM topology lock while calling the peripheral traversal function. cam_xpt.c: Change xptperiphtraverse() to hold the CAM topology lock across calls to the traversal function. Take out the comment in xptpdperiphtraverse() that referenced the locking inconsistency. cam_periph.c: Set the CAM_PERIPH_FREE flag when we are in the process of freeing a peripheral driver. Sponsored by: Spectra Logic Corporation MFC after: 1 week --- sys/cam/cam_periph.c | 8 ++++++++ sys/cam/cam_xpt.c | 16 ++-------------- 2 files changed, 10 insertions(+), 14 deletions(-) diff --git a/sys/cam/cam_periph.c b/sys/cam/cam_periph.c index b6e2b1088b88..523e549a52e7 100644 --- a/sys/cam/cam_periph.c +++ b/sys/cam/cam_periph.c @@ -614,6 +614,14 @@ camperiphfree(struct cam_periph *periph) return; } + /* + * We need to set this flag before dropping the topology lock, to + * let anyone who is traversing the list that this peripheral is + * about to be freed, and there will be no more reference count + * checks. + */ + periph->flags |= CAM_PERIPH_FREE; + /* * The peripheral destructor semantics dictate calling with only the * SIM mutex held. Since it might sleep, it should not be called diff --git a/sys/cam/cam_xpt.c b/sys/cam/cam_xpt.c index b184dd104ee8..304d8c790b3c 100644 --- a/sys/cam/cam_xpt.c +++ b/sys/cam/cam_xpt.c @@ -2178,8 +2178,8 @@ xptperiphtraverse(struct cam_ed *device, struct cam_periph *start_periph, * invalidated, but not peripherals that are scheduled to * be freed. So instead of calling cam_periph_acquire(), * which will fail if the periph has been invalidated, we - * just check for the free flag here. If it is free, we - * skip to the next periph. + * just check for the free flag here. If it is in the + * process of being freed, we skip to the next periph. */ if (periph->flags & CAM_PERIPH_FREE) { next_periph = SLIST_NEXT(periph, periph_links); @@ -2192,15 +2192,8 @@ xptperiphtraverse(struct cam_ed *device, struct cam_periph *start_periph, */ periph->refcount++; - xpt_unlock_buses(); - retval = tr_func(periph, arg); - /* - * We need the lock for list traversal. - */ - xpt_lock_buses(); - /* * Grab the next peripheral before we release this one, so * our next pointer is still valid. @@ -2283,11 +2276,6 @@ xptpdperiphtraverse(struct periph_driver **pdrv, */ periph->refcount++; - /* - * XXX KDM we have the toplogy lock here, but in - * xptperiphtraverse(), we drop it before calling the - * traversal function. Which is correct? - */ retval = tr_func(periph, arg); /*