kernel/linux.git, branch v4.4.85

Linux 4.4.85

2017-08-30T08:19:42+00:00

ACPI / APEI: Add missing synchronize_rcu() on NOTIFY_SCI removal

2017-08-30T08:19:29+00:00

commit 7d64f82cceb21e6d95db312d284f5f195e120154 upstream. When removing a GHES device notified by SCI, list_del_rcu() is used, ghes_remove() should call synchronize_rcu() before it goes on to call kfree(ghes), otherwise concurrent RCU readers may still hold this list entry after it has been freed. Signed-off-by: James Morse Reviewed-by: "Huang, Ying" Fixes: 81e88fdc432a (ACPI, APEI, Generic Hardware Error Source POLL/IRQ/NMI notification type support) Signed-off-by: Rafael J. Wysocki Signed-off-by: Greg Kroah-Hartman

ACPI: ioapic: Clear on-stack resource before using it

2017-08-30T08:19:29+00:00

commit e3d5092b6756b9e0b08f94bbeafcc7afe19f0996 upstream. The on-stack resource-window 'win' in setup_res() is not properly initialized. This causes the pointers in the embedded 'struct resource' to contain stale addresses. These pointers (in my case the ->child pointer) later get propagated to the global iomem_resources list, causing a #GP exception when the list is traversed in iomem_map_sanity_check(). Fixes: c183619b63ec (x86/irq, ACPI: Implement ACPI driver to support IOAPIC hotplug) Signed-off-by: Joerg Roedel Signed-off-by: Rafael J. Wysocki Signed-off-by: Greg Kroah-Hartman

ntb_transport: fix bug calculating num_qps_mw

2017-08-30T08:19:29+00:00

commit 8e8496e0e9564b66165f5219a4e8ed20b0d3fc6b upstream. A divide by zero error occurs if qp_count is less than mw_count because num_qps_mw is calculated to be zero. The calculation appears to be incorrect. The requirement is for num_qps_mw to be set to qp_count / mw_count with any remainder divided among the earlier mws. For example, if mw_count is 5 and qp_count is 12 then mws 0 and 1 will have 3 qps per window and mws 2 through 4 will have 2 qps per window. Thus, when mw_num < qp_count % mw_count, num_qps_mw is 1 higher than when mw_num >= qp_count. Signed-off-by: Logan Gunthorpe Fixes: e26a5843f7f5 ("NTB: Split ntb_hw_intel and ntb_transport drivers") Acked-by: Allen Hubbe Signed-off-by: Jon Mason Signed-off-by: Greg Kroah-Hartman

ntb_transport: fix qp count bug

2017-08-30T08:19:28+00:00

commit cb827ee6ccc3e480f0d9c0e8e53eef55be5b0414 upstream. In cases where there are more mw's than spads/2-2, the mw count gets reduced to match the limitation. ntb_transport also tries to ensure that there are fewer qps than mws but uses the full mw count instead of the reduced one. When this happens, the math in 'ntb_transport_setup_qp_mw' will get confused and result in a kernel paging request bug. This patch fixes the bug by reducing qp_count to the reduced mw count instead of the full mw count. Signed-off-by: Logan Gunthorpe Fixes: e26a5843f7f5 ("NTB: Split ntb_hw_intel and ntb_transport drivers") Acked-by: Allen Hubbe Signed-off-by: Jon Mason Signed-off-by: Greg Kroah-Hartman

ASoC: rsnd: don't call update callback if it was NULL

2017-08-30T08:19:28+00:00

commit d7289565483c65094d0473555625a4acd89567d3 upstream. Signed-off-by: Kuninori Morimoto Signed-off-by: Mark Brown Signed-off-by: Thong Ho Signed-off-by: Nhan Nguyen Signed-off-by: Greg Kroah-Hartman

ASoC: rsnd: ssi: 24bit data needs right-aligned settings

2017-08-30T08:19:28+00:00

commit f46a93b820eb3707faf238cd769a004e2504515f upstream. Data left/right aligned is controlled by PDTA bit on SSICR. But default is left-aligned. Thus 24bit sound will be very small sound without this patch. Signed-off-by: Kuninori Morimoto Signed-off-by: Mark Brown Signed-off-by: Thong Ho Signed-off-by: Nhan Nguyen Signed-off-by: Greg Kroah-Hartman

ASoC: rsnd: Add missing initialization of ADG req_rate

2017-08-30T08:19:28+00:00

commit 8b27418f300cafbdbbb8cfa9c29d398ed34d6723 upstream. If the "clock-frequency" DT property is not found, req_rate is used uninitialized, and the "audio_clkout" clock will be created with an arbitrary clock rate. This uninitialized kernel stack data may leak to userspace through /sys/kernel/debug/clk/clk_summary, cfr. the value in the "rate" column: clock enable_cnt prepare_cnt rate accuracy phase -------------------------------------------------------------------- audio_clkout 0 0 4001836240 0 0 Signed-off-by: Geert Uytterhoeven Acked-by: Kuninori Morimoto Signed-off-by: Mark Brown Signed-off-by: Thong Ho Signed-off-by: Nhan Nguyen Signed-off-by: Greg Kroah-Hartman

ASoC: rsnd: avoid pointless loop in rsnd_mod_interrupt()

2017-08-30T08:19:28+00:00

commit 2daf71ad8da6cb57f919c9c876ee7e42530371df upstream. Current Renesas sound driver doesn't have 1:1 relationship between stream <-> mod because it is supporting MIX. Because of this reason rsnd_mod_interrupt() is searching correspond mod by for loop. But this loop is not needed, because each mod has own type. This patch avoid pointless loop by using mod->type. This patch is good for SSI-parent support, because stream might have 2 SSI as SSI-parent/child. SSI interrupt handler will be called twice if stream has SSI-parent without this patch. Signed-off-by: Kuninori Morimoto Signed-off-by: Mark Brown Signed-off-by: Thong Ho Signed-off-by: Nhan Nguyen Signed-off-by: Greg Kroah-Hartman

ASoC: rsnd: disable SRC.out only when stop timing

2017-08-30T08:19:27+00:00

commit b761bf272bce6dff4d8a7ccf4385c9f3d4018094 upstream. Because SRC is connected to DMA and DMA want to keep dreq when stop timing. This patch makes SRC stop SRC.out only when stop timing. And it stops both SRC.out/SRC.in when quit timing Signed-off-by: Kuninori Morimoto Signed-off-by: Mark Brown Signed-off-by: Thong Ho Signed-off-by: Nhan Nguyen Signed-off-by: Greg Kroah-Hartman