| Message ID | 1446552832-17592-1-git-send-email-fbarrat@linux.vnet.ibm.com (mailing list archive) |
|---|---|
| State | Changes Requested |
| Headers | show |
Acked-by: Ian Munsie <imunsie@au1.ibm.com>
On 03/11/15 23:13, Frederic Barrat wrote: > When the cxl driver creates a context, it stores the pid of the > calling task, and mistakenly increments the reference count of the > struct pid twice, once through get_task_pid(), once through > get_pid(). The reference count is only decremented once on detach. As > a result, a struct pid is leaked and it reduces the number of > processes that can run simultaneously by one. > The fix is to simply remove the call to get_pid(). > > There is a separate patch for the kernel API, since it goes to > different kernel versions. > > Signed-off-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com> > Cc: <stable@vger.kernel.org> # 3.18+ Reviewed-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
On Tue, 2015-03-11 at 12:13:52 UTC, Frederic Barrat wrote: > When the cxl driver creates a context, it stores the pid of the > calling task, and mistakenly increments the reference count of the > struct pid twice, once through get_task_pid(), once through > get_pid(). The reference count is only decremented once on detach. As > a result, a struct pid is leaked and it reduces the number of > processes that can run simultaneously by one. > The fix is to simply remove the call to get_pid(). > > There is a separate patch for the kernel API, since it goes to > different kernel versions. > > Signed-off-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com> > Cc: <stable@vger.kernel.org> # 3.18+ > Acked-by: Ian Munsie <imunsie@au1.ibm.com> > Reviewed-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com> This no longer applies, since we merged 7b8ad495d592 ("cxl: Fix DSI misses when the context owning task exits"). Please resend if it's still applicable. cheers
Le 26/01/2016 04:11, Michael Ellerman a écrit : > This no longer applies, since we merged 7b8ad495d592 ("cxl: Fix DSI misses when > the context owning task exits"). Yes, on 4.5, it has been superseded by 7b8ad495d592. It may still be worth considering for stable releases though. Fred
diff --git a/drivers/misc/cxl/file.c b/drivers/misc/cxl/file.c index 7ccd299..97003ee 100644 --- a/drivers/misc/cxl/file.c +++ b/drivers/misc/cxl/file.c @@ -199,7 +199,7 @@ static long afu_ioctl_start_work(struct cxl_context *ctx, * behalf of another process, so the AFU's mm gets bound to the process * that performs this ioctl and not the process that opened the file. */ - ctx->pid = get_pid(get_task_pid(current, PIDTYPE_PID)); + ctx->pid = get_task_pid(current, PIDTYPE_PID); trace_cxl_attach(ctx, work.work_element_descriptor, work.num_interrupts, amr);
When the cxl driver creates a context, it stores the pid of the calling task, and mistakenly increments the reference count of the struct pid twice, once through get_task_pid(), once through get_pid(). The reference count is only decremented once on detach. As a result, a struct pid is leaked and it reduces the number of processes that can run simultaneously by one. The fix is to simply remove the call to get_pid(). There is a separate patch for the kernel API, since it goes to different kernel versions. Signed-off-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com> Cc: <stable@vger.kernel.org> # 3.18+ --- drivers/misc/cxl/file.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)