kernel/linux.git/drivers/soc/bcm, branch v5.10.258

pmdomain: bcm: bcm2835-power: Increase ASB control timeout

2026-04-18T08:31:00+00:00

[ Upstream commit b826d2c0b0ecb844c84431ba6b502e744f5d919a ] The bcm2835_asb_control() function uses a tight polling loop to wait for the ASB bridge to acknowledge a request. During intensive workloads, this handshake intermittently fails for V3D's master ASB on BCM2711, resulting in "Failed to disable ASB master for v3d" errors during runtime PM suspend. As a consequence, the failed power-off leaves V3D in a broken state, leading to bus faults or system hangs on later accesses. As the timeout is insufficient in some scenarios, increase the polling timeout from 1us to 5us, which is still negligible in the context of a power domain transition. Also, replace the open-coded ktime_get_ns()/ cpu_relax() polling loop with readl_poll_timeout_atomic(). Cc: stable@vger.kernel.org Fixes: 670c672608a1 ("soc: bcm: bcm2835-pm: Add support for power domains under a new binding.") Signed-off-by: Maíra Canal Reviewed-by: Stefan Wahren Signed-off-by: Ulf Hansson [ adapted unified bcm2835_asb_control() function changes to separate bcm2835_asb_enable() and bcm2835_asb_disable() functions ] Signed-off-by: Sasha Levin Signed-off-by: Greg Kroah-Hartman

pmdomain: bcm: bcm2835-power: Fix broken reset status read

2026-04-18T08:30:59+00:00

[ Upstream commit 550bae2c0931dbb664a61b08c21cf156f0a5362a ] bcm2835_reset_status() has a misplaced parenthesis on every PM_READ() call. Since PM_READ(reg) expands to readl(power->base + (reg)), the expression: PM_READ(PM_GRAFX & PM_V3DRSTN) computes the bitwise AND of the register offset PM_GRAFX with the bitmask PM_V3DRSTN before using the result as a register offset, reading from the wrong MMIO address instead of the intended PM_GRAFX register. The same issue affects the PM_IMAGE cases. Fix by moving the closing parenthesis so PM_READ() receives only the register offset, and the bitmask is applied to the value returned by the read. Fixes: 670c672608a1 ("soc: bcm: bcm2835-pm: Add support for power domains under a new binding.") Signed-off-by: Maíra Canal Reviewed-by: Florian Fainelli Reviewed-by: Stefan Wahren Cc: stable@vger.kernel.org Signed-off-by: Ulf Hansson Signed-off-by: Sasha Levin Signed-off-by: Greg Kroah-Hartman

soc: brcmstb: pm-arm: Fix refcount leak and __iomem leak bugs

2022-09-15T09:32:04+00:00

[ Upstream commit 1085f5080647f0c9f357c270a537869191f7f2a1 ] In brcmstb_pm_probe(), there are two kinds of leak bugs: (1) we need to add of_node_put() when for_each__matching_node() breaks (2) we need to add iounmap() for each iomap in fail path Fixes: 0b741b8234c8 ("soc: bcm: brcmstb: Add support for S2/S3/S5 suspend states (ARM)") Signed-off-by: Liang He Link: https://lore.kernel.org/r/20220707015620.306468-1-windhl@126.com Signed-off-by: Florian Fainelli Signed-off-by: Sasha Levin

soc: bcm: brcmstb: pm: pm-arm: Fix refcount leak in brcmstb_pm_probe

2022-06-29T06:59:54+00:00

commit 37d838de369b07b596c19ff3662bf0293fdb09ee upstream. of_find_matching_node() returns a node pointer with refcount incremented, we should use of_node_put() on it when not need anymore. Add missing of_node_put() to avoid refcount leak. In brcmstb_init_sram, it pass dn to of_address_to_resource(), of_address_to_resource() will call of_find_device_by_node() to take reference, so we should release the reference returned by of_find_matching_node(). Fixes: 0b741b8234c8 ("soc: bcm: brcmstb: Add support for S2/S3/S5 suspend states (ARM)") Signed-off-by: Miaoqian Lin Reviewed-by: Andy Shevchenko Signed-off-by: Florian Fainelli Signed-off-by: Greg Kroah-Hartman

soc: bcm: brcmstb: biuctrl: Change RAC data line prefetching after 4 consecutive lines

2020-09-06T19:43:02+00:00

Change the RACPREFDATA(x) setting to prefetch the next 256-byte line after 4 consecutive lines have been used, instead of after 2 consecutive lines. This does improve the synthetic memcpy benchmark by an additional +0.5% on top of the previous change for Cortex-A72 CPUs. Signed-off-by: Florian Fainelli

soc: bcm: brcmstb: biuctrl: Change RAC prefetch distance from +/-1 to +/- 2

2020-09-04T20:48:49+00:00

Change the RAC prefetch distance from +/- 1 to +/- 2 for Cortex-A72 CPUs since this provides an average of a 3.8% performance increase for synthetic memcpy benchmarks. Signed-off-by: Florian Fainelli

soc: bcm: brcmstb: biuctrl: Tune MCP settings for 72165

2020-09-04T20:48:42+00:00

72165 uses a Brahma-B53 CPU and its Bus Interface Unit, tune it according to the existing values we have. Signed-off-by: Florian Fainelli

soc: bcm: brcmstb: biuctrl: Tune MCP settings for 72164

2020-09-04T20:44:32+00:00

72164 uses a Brahma-B53 CPU and its Bus Interface Unit, tune it according to the existing values we have. Signed-off-by: Florian Fainelli

soc: bcm: add BCM63xx power domain driver

2020-08-17T16:14:04+00:00

BCM6318, BCM6328, BCM6362 and BCM63268 SoCs have a power domain controller to enable/disable certain components in order to save power. Signed-off-by: Álvaro Fernández Rojas Reviewed-by: Florian Fainelli Signed-off-by: Florian Fainelli

soc: bcm: brcmstb: biuctrl: Enable Read-ahead cache

2020-08-17T16:14:04+00:00

Brahma-B53 and Cortex-A72 CPUs integrated on Broadcom STB SoCs feature a read-ahead cache that performs cache line size adaptation between the bus interface unit and the memory controller. On 32-bit ARM kernels we have to resort to a full featured read-ahead cache driver under arch/arm/mm/cache-b15-rac.c (CONFIG_CACHE_B15_RAC) because there are still cache maintenance operations by set/ways/index that cannot be transparently handled by the ARM Coherency Extension that the read-ahead cache interfaces to. The 64-bit ARM kernel however has long deprecated all of those, so this is simply a one time configuration. Signed-off-by: Florian Fainelli