drivers/mic/scif: convert get_user_pages() --> pin_user_pages()
authorJohn Hubbard <jhubbard@nvidia.com>
Mon, 18 May 2020 04:13:07 +0000 (21:13 -0700)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Tue, 19 May 2020 14:46:12 +0000 (16:46 +0200)
commit5459ceedb3940df460a6048762cc49d549049d6f
treeb693af059ba486f1134b7b78d0fd67227ed4dd2a
parent6a0953ce7de918b06f7fed561754ad114fd964c1
drivers/mic/scif: convert get_user_pages() --> pin_user_pages()

This code was using get_user_pages*(), in a "Case 2" scenario
(DMA/RDMA), using the categorization from [1]. That means that it's
time to convert the get_user_pages*() + put_page() calls to
pin_user_pages*() + unpin_user_pages() calls.

There is some helpful background in [2]: basically, this is a small
part of fixing a long-standing disconnect between pinning pages, and
file systems' use of those pages.

Note that this effectively changes the code's behavior as well: it now
ultimately calls set_page_dirty_lock(), instead of SetPageDirty(). This
is probably more accurate.

As Christoph Hellwig put it, "set_page_dirty() is only safe if we are
dealing with a file backed page where we have reference on the inode it
hangs off." [3]

[1] Documentation/core-api/pin_user_pages.rst

[2] "Explicit pinning of user-space pages":
    https://lwn.net/Articles/807108/

[3] https://lore.kernel.org/r/20190723153640.GB720@lst.de

Signed-off-by: John Hubbard <jhubbard@nvidia.com>
Link: https://lore.kernel.org/r/20200518041307.1987328-1-jhubbard@nvidia.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/misc/mic/scif/scif_rma.c