Skip to content

GitLab

Switch branch/tag
  1. 20 Jan, 2020 6 commits
  3. 17 Jan, 2020 11 commits
  5. 15 Jan, 2020 11 commits
  7. 13 Jan, 2020 1 commit
  9. 12 Jan, 2020 2 commits
  11. 10 Jan, 2020 2 commits
    • Deepika Bhavnani's avatar
      Unify type of "cpu_idx" across PSCI module. · 5b33ad17
      Deepika Bhavnani authored 
      

NOTE for platform integrators:
   API `plat_psci_stat_get_residency()` third argument
   `last_cpu_idx` is changed from "signed int" to the
   "unsigned int" type.

Issue / Trouble points
1. cpu_idx is used as mix of `unsigned int` and `signed int` in code
with typecasting at some places leading to coverity issues.

2. Underlying platform API's return cpu_idx as `unsigned int`
and comparison is performed with platform specific defines
`PLAFORM_xxx` which is not consistent

Misra Rule 10.4:
The value of a complex expression of integer type may only be cast to
a type that is narrower and of the same signedness as the underlying
type of the expression.

Based on above points, cpu_idx is kept as `unsigned int` to match
the API's and low-level functions and platform defines are updated
where ever required
Signed-off-by: default avatarDeepika Bhavnani <deepika.bhavnani@arm.com>
Change-Id: Ib26fd16e420c35527204b126b9b91e8babcc3a5c
      5b33ad17
    • Alexei Fedorov's avatar
      FVP: Remove re-definition of topology related build options · 94f1c959
      Alexei Fedorov authored 
      

This patch removes re-definition of the following FVP build
options from plat\arm\board\fvp\fvp_def.h:
 'FVP_CLUSTER_COUNT'
 'FVP_MAX_CPUS_PER_CLUSTER'
 'FVP_MAX_PE_PER_CPU'
which are set in platform.mk.

This fixes a potential problem when a build option set in
platform.mk file can be re-defined in fvp_def.h header file
used by other build component with a different makefile which
does not set this option.
Ref. GENFW-3505.

Change-Id: I4288629920516acf2c239c7b733f92a0c5a812ff
Signed-off-by: default avatarAlexei Fedorov <Alexei.Fedorov@arm.com>
      94f1c959
  13. 09 Jan, 2020 3 commits
  15. 08 Jan, 2020 1 commit
  17. 07 Jan, 2020 3 commits
    • Rajan Vaja's avatar
      zynqmp: pm: Remove CLK_TOPSW_LSBUS from invalid clock list · 20fdf0b0
      Rajan Vaja authored 
      

CLK_TOPSW_LSBUS is parent of WDT clock. Clock from invalid
clock list would not be registered to CCF framework and so
cannot be used as parent of other clocks.

WDT clock has default parent as CLK_TOPSW_LSBUS(APB clock).
If CLK_TOPSW_LSBUS is not registered, CCF would not recognize
that clock and hence rate of WDT clock would be calculated to
be 0 by CCF(as parent rate is considered 0).

So it is necessary to allow registration of CLK_TOPSW_LSBUS
clock.
Signed-off-by: default avatarRajan Vaja <rajan.vaja@xilinx.com>
Signed-off-by: default avatarJolly Shah <jolly.shah@xilinx.com>
Change-Id: Iceaba0f137784fc5fd666e66ffc4c143381c6ccc
      20fdf0b0
    • Mounika Grace Akula's avatar
      zynqmp: pm: Add LPD WDT clock to the pm_clock structure · b3ce966a
      Mounika Grace Akula authored 
      

This patch adds LPD WDT clock node to the pm_clock clocks structure list
so that LPD WDT can be used from Linux.

Also this patch removes the CLK_LPD_LSBUS from invalid clock list to
allow the registration of this clock to CCF framework as it is the
parent of LPD WDT.
Signed-off-by: default avatarMounika Grace Akula <mounika.grace.akula@xilinx.com>
Signed-off-by: default avatarJolly Shah <jolly.shah@xilinx.com>
Change-Id: Iea065aa8150eaba4bb4b42bc6be1fd4b7fe7b403
      b3ce966a
    • Mirela Simonovic's avatar
      zynqmp: pm: Fix clock models and IDs of GEM-related clocks · 06ad9803
      Mirela Simonovic authored 
      

GEM-related clock models were incorrect and are fixed as follows
(documented below for GEM0, but the same holds for any GEM ID):

- CLK_GEM0_REF_UNGATED represents clock that has DIV0/1 divisors and
 the multiplexer controllable in GEM0_REF_CTRL (CRL_APB). The ID of this
 clock is newly introduced in this patch.

- CLK_GEM0_REF models the clock mux that selects the reference clock
 for Tx, i.e. selects CLK_GEM0_REF_UNGATED or external Tx clock. This
 mux is controllable via GEM_CLK_CTRL (IOU_SLCR), bit GEM0_REF_SRC_SEL.
 Note that the routing of external clock to the mux is not modelled
 and is assumed to be configured by the FSBL if required, and not
 changeable at runtime. The ID of this clock is introduced in this patch.

- CLK_GEM0_TX models clock with only a gate that is controlled via
 bit 25 in GEM0_REF_CTRL (CRL_APB). The parent of this clock is
 CLK_GEM0_REF. The clock ID of CLK_GEM0_TX matches the previous ID
 value of CLK_GEM0_REF. This is done in order to fix the clock models
 and incorrect binding without requiring to change device-tree (binding
 of clock IDs to GEM interface).

- CLK_GEM0_RX models clock that has only gate controlled via RX_CLKACT
 bit (26) in GEM0_REF_CTRL (CRL_APB). Parent of this clock is sourced
 from external RGMII PHY (via MIO or EMIO). We do not model the whole
 clock path to the Rx gate, since this is configured by the FSBL and
 never changed at runtime (and there is no mechanism to change the
 path at runtime). The clock ID of CLK_GEM0_RX clock is equal to the
 previous ID value of CLK_GEM0_TX clock. This is done because the TX/RX
 were swapped in device tree, so by fixing the IDs this way there is no
 need for device tree fix.

Rates of the external RX/TX clocks can be specified in device tree if
needed. Right now, that's not necessary because Tx clock is sourced
from an on-chip PLL (via CLK_GEM0_REF_UNGATED/CLK_GEM0_REF), whereas
the Rx clock is sourced from external reference and the driver never
attempts to get/get clock rate (only to enable it). If this changes in
future, ATF clock model doesn't need to be changed. Instead, the clock
rates for gem0_tx_ext and gem0_rx_ext have to be specified in device
tree.
Signed-off-by: default avatarMirela Simonovic <mirela.simonovic@aggios.com>
Acked-by: default avatarWill Wong <will.wong@xilinx.com>
Signed-off-by: default avatarJolly Shah <jolly.shah@xilinx.com>
Change-Id: I6497d4309e92205c527bd81b3aa932f4474f5b79
      06ad9803