I applied the 373 patchset and compiled a version with CIFS_FSCACHE enabled.
Same problem. Got ~2GiB into a cp before it started generating 'bad page
state' errors.
It's not mission-critical for this application, but it would be nice to
have.
On 6/17/2013 2:20 AM, David Howells wrote:
Rob Bos <rbos@xxxxxx> wrote:
BUG: Bad page state in process aria2c pfn:78078
page:ffffea0001a41a40 flags:0020000000001000 count:0 mapcount:0 mapping:(null)
index:9 (Tainted: G B --------------- T)
Pid: 2975, comm: aria2c Tainted: G B --------------- T
2.6.32-358.6.2.el6.rbos.1.x86_64 #1
I have a series of patches poised to go into RHEL-6 to clean up a number of
issues in fscache. There's at least one patch in there that might fix this as
a side effect (ie. this particular bug isn't the one the patch is actually
intended to fix). But since it has to do with retrieval operation lifetime
accounting, it might just.
If you are able to try a debugging kernel of slightly more recent vintage, I
have one already prepared for another customer:
http://people.redhat.com/~dhowells/kernel-2.6.32-373.el6.bz928746.2.x86_64.rpm
http://people.redhat.com/~dhowells/kernel-firmware-2.6.32-373.el6.bz928746.2.noarch.rpm
might work for you. The kernel rpm is just the base RHEL kernel plus my
outstanding fscache patches.
David
--
Linux-cachefs mailing list
Linux-cachefs@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/linux-cachefs
