Shrink_page_list expects all pages come from a same zone
but it's too limited to use.
This patch removes the dependency so next patch can use
shrink_page_list with pages from multiple zones.
Change-Id: I34469b7f0a79f2b79e30e40033ba8b3e1dd5f2d0
Signed-off-by: Minchan Kim <minchan@kernel.org>
Patch-mainline: linux-mm @ 9 May 2013 16:21:25
[vinmenon@codeaurora.org: changes for node based lrus]
Signed-off-by: Vinayak Menon <vinmenon@codeaurora.org>
These day, there are many platforms available in the embedded market
and they are smarter than kernel which has very limited information
about working set so they want to involve memory management more heavily
like android's lowmemory killer and ashmem or recent many lowmemory
notifier.
One of the simple imagine scenario about userspace's intelligence is that
platform can manage tasks as forground and background so it would be
better to reclaim background's task pages for end-user's *responsibility*
although it has frequent referenced pages.
This patch adds new knob "reclaim under proc/<pid>/" so task manager
can reclaim any target process anytime, anywhere. It could give another
method to platform for using memory efficiently.
It can avoid process killing for getting free memory, which was really
terrible experience because I lost my best score of game I had ever
after I switch the phone call while I enjoyed the game.
Reclaim file-backed pages only.
echo file > /proc/PID/reclaim
Reclaim anonymous pages only.
echo anon > /proc/PID/reclaim
Reclaim all pages
echo all > /proc/PID/reclaim
Change-Id: Iabdb7bc2ef3dc4d94e3ea005fbe18f4cd06739ab
Signed-off-by: Minchan Kim <minchan@kernel.org>
Patch-mainline: linux-mm @ 9 May 2013 16:21:24
[vinmenon@codeaurora.org: merge conflict fixes]
Signed-off-by: Vinayak Menon <vinmenon@codeaurora.org>
Currently, ucsi_qti_notify() is checking the received data from
charger firmware (PPM) to see if there is any connector status
response with partner accessory information (e.g. analog audio).
It does this based on setting a flag (cmd_requested_flags) when
UCSI_GET_CONNECTOR_STATUS command is sent from UCSI framework to
PPM so that the response for that command can be read and clients
registered for getting notification on partner accessory can get
notified. However, in a certain scenario, multiple response for
different commands can be received from PPM. See a possible
example below.
1. UCSI framework sends UCSI_GET_CONNECTOR_STATUS command
2. ucsi_glink driver sets cmd_requested_flags, waits for response
3. UCSI framework sends UCSI_GET_CURRENT_CAM command
4. ucsi_glink driver gets a response for UCSI_GET_CONNECTOR_STATUS
and schedules notify_work since ucsi_qti_notify() is called
with cmd_requested_flags set.
5. ucsi_glink driver gets a response for UCSI_GET_CURRENT_CAM and
this can end up reading "flags" because cmd_requested_flags is
not cleared yet. However, this causes a buffer overflow because
ucsi_qti_read() passes "val" pointer passed from UCSI framework
which is of 1 byte length and ucsi_qti_notify() ends up reading
"flags" which is at an offset 2 of length 9 bytes.
6. ucsi_qti_notify_work() clears cmd_requested_flags.
Fix this by checking the message length in ucsi_qti_notify()
to ensure that status->flags is read only when a response is
received for UCSI_GET_CONNECTOR_STATUS. Also, relocate the
clearing of "cmd_requested_flags" flag.
CRs-Fixed: 2678391
Change-Id: Iac1d5c58e1ed2fd0f3bc153da23cddf0700cc097
Signed-off-by: Subbaraman Narayanamurthy <subbaram@codeaurora.org>
Register driver work handler will only power up device, resume PCIe
link and start MHI to download firmware. It may only block there
for a timeout, so there is no need to post it as killable event.
Change-Id: Ib82050358dc8df4a8163ecb43702aa86ed34f23d
Signed-off-by: Yue Ma <yuem@codeaurora.org>
For uniprocessor systems, add a NULL definition of tick_broadcast to
add support for tick broadcast.
Change-Id: I7dc47984cb6b60c9814171888f6adbc480a1e990
Signed-off-by: Mahesh Sivasubramanian <msivasub@codeaurora.org>
