This is a generic review and I haven't looked into the PAPR spec for
htmdump hcall and it's interface.
Madhavan Srinivasan writes:
> This patch adds debugfs interface to export Hardware Trace Macro (HTM)
> function data in a LPAR. New hypervisor call "H_HTM" has been
> defined to setup, configure, control and dump the HTM data.
> This patch supports only dumping of HTM data in a LPAR.
> New debugfs folder called "htmdump" has been added under
> /sys/kernel/debug/arch path which contains files need to
> pass required parameters for the H_HTM dump function. New Kconfig
> option called "CONFIG_HTMDUMP" has been in platform/pseries for the same.
>
> With patch series applied and booted, list of files in debugfs path
>
> # pwd
> /sys/kernel/debug/powerpc/htmdump
> # ls
> coreindexonchip htmtype nodalchipindex nodeindex trace
>
> Signed-off-by: Madhavan Srinivasan
> ---
> arch/powerpc/platforms/pseries/Kconfig | 8 ++
> arch/powerpc/platforms/pseries/Makefile | 1 +
> arch/powerpc/platforms/pseries/htmdump.c | 130 +++
> 3 files changed, 139 insertions(+)
> create mode 100644 arch/powerpc/platforms/pseries/htmdump.c
>
> diff --git a/arch/powerpc/platforms/pseries/Kconfig
> b/arch/powerpc/platforms/pseries/Kconfig
> index afc0f6a61337..46c0ea605e33 100644
> --- a/arch/powerpc/platforms/pseries/Kconfig
> +++ b/arch/powerpc/platforms/pseries/Kconfig
> @@ -128,6 +128,14 @@ config CMM
> will be reused for other LPARs. The interface allows firmware to
> balance memory across many LPARs.
>
> +config HTMDUMP
> + tristate "PHYP HTM data dumper"
Not sure if we can make machine_device_initcall() as a tristate?
Did we try compiling it as a module?
It we would like to keep this as a module - then why not use module_init
call and then make it depend upon...
depends on PPC_PSERIES && DEBUG_FS (??)
> + default y
and then since this is mostly a debug trace facility, then we need not enable
it by default right?
> + help
> + Select this option, if you want to enable the kernel debugfs
> + interface to dump the Hardware Trace Macro (HTM) function data
> + in the LPAR.
> +
> config HV_PERF_CTRS
> bool "Hypervisor supplied PMU events (24x7 & GPCI)"
> default y
> diff --git a/arch/powerpc/platforms/pseries/Makefile
> b/arch/powerpc/platforms/pseries/Makefile
> index 7bf506f6b8c8..3f3e3492e436 100644
> --- a/arch/powerpc/platforms/pseries/Makefile
> +++ b/arch/powerpc/platforms/pseries/Makefile
> @@ -19,6 +19,7 @@ obj-$(CONFIG_HVC_CONSOLE) += hvconsole.o
> obj-$(CONFIG_HVCS) += hvcserver.o
> obj-$(CONFIG_HCALL_STATS)+= hvCall_inst.o
> obj-$(CONFIG_CMM)+= cmm.o
> +obj-$(CONFIG_HTMDUMP)+= htmdump.o
> obj-$(CONFIG_IO_EVENT_IRQ) += io_event_irq.o
> obj-$(CONFIG_LPARCFG)+= lparcfg.o
> obj-$(CONFIG_IBMVIO) += vio.o
> diff --git a/arch/powerpc/platforms/pseries/htmdump.c
> b/arch/powerpc/platforms/pseries/htmdump.c
> new file mode 100644
> index ..540cdb7e069c
> --- /dev/null
> +++ b/arch/powerpc/platforms/pseries/htmdump.c
> @@ -0,0 +1,130 @@
> +// SPDX-License-Identifier: GPL-2.0-or-later
> +/*
> + * Copyright (C) IBM Corporation, 2024
> + */
> +
> +#define pr_fmt(fmt) "htmdump: " fmt
> +
> +#include
> +#include
> +#include
> +#include
> +#include
> +#include
> +#include
> +#include
> +#include
> +#include
> +#include
> +#include
> +#include
Do we need all of the above?
e.g. slab, memory_hotplug etc are not needed IMO.
Maybe only?
#include
#include
#include
#include
#include
#include
(module.h depending upon if we make it module_init())
> +
> +/* This enables us to keep track of the memory removed from each node. */
> +struct htmdump_entry {
> + void *buf;
> + struct dentry *dir;
> + char name[16];
> +};
> +
> +static u32 nodeindex = 0;
> +static u32 nodalchipindex = 0;
> +static u32 coreindexonchip = 0;
> +static u32 htmtype = 0;
> +
> +#define BUFFER_SIZE PAGE_SIZE
> +
> +static ssize_t htmdump_read(struct file *filp, char __user *ubuf,
> + size_t count, loff_t *ppos)
> +{
> + struct htmdump_entry *ent = filp->private_data;
> + unsigned long page, read_size, available;
> + loff_t offset;
> + long rc;
> +
> + page = ALIGN_DOWN(*ppos, BUFFER_SIZE);
> + offset = (*ppos) % BUFFER_SIZE;
> +
> + rc = htm_get_dump_hardware(nodeindex, nodalchipindex, coreindexonchip,
> +htmtype, virt_to_phys(ent->buf),
> BUFFER_SIZE, page);
> +
> + switch(rc) {
> + case H_SUCCESS:
> + case H_PARTIAL:
> + break;
> + case H_NOT_AVAILABLE:
> + return 0;
> + case H_BUSY:
> + case H_LONG_BUSY_ORDER_1_MSEC:
> + case H_LONG_BUSY_ORDER_10_MSEC:
> + case H_LONG_BUSY_ORDER_100_MSEC:
> + case H_LONG_BUSY_ORDER_1_SEC:
> + case H_LONG_BUSY_ORDER_10_SEC:
> + case H_LONG_BUSY_ORDER_100_SEC:
> +
