From: David Carrillo-Cisneros <davidcc@google.com>
To: Vikas Shivappa <vikas.shivappa@linux.intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>,
Alexander Shishkin <alexander.shishkin@linux.intel.com>,
Arnaldo Carvalho de Melo <acme@kernel.org>,
Ingo Molnar <mingo@redhat.com>, Tony Luck <tony.luck@intel.com>,
Stephane Eranian <eranian@google.com>,
Paul Turner <pjt@google.com>,
x86@kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH 26/32] perf/x86/intel/cqm: integrate CQM cgroups with scheduler
Date: Fri, 29 Apr 2016 14:26:13 -0700 [thread overview]
Message-ID: <CALcN6mjYfKHFuR0pj1wtvHH6uoXJ9a5C0HR5727UerCCCPn6eg@mail.gmail.com> (raw)
In-Reply-To: <alpine.DEB.2.10.1604291358530.18257@vshiva-Udesk>
if __intel_cqm_no_event_sched_in does nothing, the PQR_ASSOC msr is
still updated if state->rmid != state->next_rmid in __pqr_update,
even if next_rmid_mode == PQR_RMID_MODE_NOEVENT .
On Fri, Apr 29, 2016 at 2:01 PM, Vikas Shivappa
<vikas.shivappa@linux.intel.com> wrote:
>
>
> On Fri, 29 Apr 2016, David Carrillo-Cisneros wrote:
>
>> Not sure I see the problem you point here. In step 3, PQR_ASSOC is
>> updated with RMID1, __pqr_update is the one called using the scheduler
>> hook, right after perf sched_in .
>>
>> On Fri, Apr 29, 2016 at 1:25 PM, Vikas Shivappa
>> <vikas.shivappa@linux.intel.com> wrote:
>>>
>>>
>>>
>>> On Thu, 28 Apr 2016, David Carrillo-Cisneros wrote:
>>>
>>>> Allow monitored cgroups to update the PQR MSR during task switch even
>>>> without an associated perf_event.
>>>>
>>>> The package RMID for the current monr associated with a monitored
>>>> cgroup is written to hw during task switch (after perf_events is run)
>>>> if perf_event did not write a RMID for an event.
>>>>
>>>> perf_event and any other caller of pqr_cache_update_rmid can update the
>>>> CPU's RMID using one of two modes:
>>>> - PQR_RMID_MODE_NOEVENT: A RMID that do not correspond to an event.
>>>> e.g. the RMID of the root pmonr when no event is scheduled.
>>>> - PQR_RMID_MODE_EVENT: A RMID used by an event. Set during pmu::add
>>>> unset on pmu::del. This mode prevents from using a non-event
>>>> cgroup RMID.
>>>>
>>>> This patch also introduces caching of writes to PQR MSR within the
>>>> per-pcu
>>>> pqr state variable. This interface to update RMIDs and CLOSIDs will be
>>>> also utilized in upcoming versions of Intel's MBM and CAT drivers.
>>>>
>>>> Reviewed-by: Stephane Eranian <eranian@google.com>
>>>> Signed-off-by: David Carrillo-Cisneros <davidcc@google.com>
>>>> ---
>>>> arch/x86/events/intel/cqm.c | 65
>>>> +++++++++++++++++++++++++++++----------
>>>> arch/x86/events/intel/cqm.h | 2 --
>>>> arch/x86/include/asm/pqr_common.h | 53 +++++++++++++++++++++++++++----
>>>> arch/x86/kernel/cpu/pqr_common.c | 46 +++++++++++++++++++++++----
>>>> 4 files changed, 135 insertions(+), 31 deletions(-)
>>>>
>>>> diff --git a/arch/x86/events/intel/cqm.c b/arch/x86/events/intel/cqm.c
>>>> index daf9fdf..4ece0a4 100644
>>>> --- a/arch/x86/events/intel/cqm.c
>>>> +++ b/arch/x86/events/intel/cqm.c
>>>> @@ -198,19 +198,6 @@ static inline int cqm_prmid_update(struct prmid
>>>> *prmid)
>>>> return __cqm_prmid_update(prmid, __rmid_min_update_time);
>>>> }
>>>>
>>>> -/*
>>>> - * Updates caller cpu's cache.
>>>> - */
>>>> -static inline void __update_pqr_prmid(struct prmid *prmid)
>>>> -{
>>>> - struct intel_pqr_state *state = this_cpu_ptr(&pqr_state);
>>>> -
>>>> - if (state->rmid == prmid->rmid)
>>>> - return;
>>>> - state->rmid = prmid->rmid;
>>>> - wrmsr(MSR_IA32_PQR_ASSOC, prmid->rmid, state->closid);
>>>> -}
>>>> -
>>>> static inline bool __valid_pkg_id(u16 pkg_id)
>>>> {
>>>> return pkg_id < PQR_MAX_NR_PKGS;
>>>> @@ -2531,12 +2518,11 @@ static inline bool cqm_group_leader(struct
>>>> perf_event *event)
>>>> static inline void __intel_cqm_event_start(
>>>> struct perf_event *event, union prmid_summary summary)
>>>> {
>>>> - u16 pkg_id = topology_physical_package_id(smp_processor_id());
>>>> if (!(event->hw.state & PERF_HES_STOPPED))
>>>> return;
>>>> -
>>>> event->hw.state &= ~PERF_HES_STOPPED;
>>>> - __update_pqr_prmid(__prmid_from_rmid(pkg_id,
>>>> summary.sched_rmid));
>>>> +
>>>> + pqr_cache_update_rmid(summary.sched_rmid, PQR_RMID_MODE_EVENT);
>>>> }
>>>>
>>>> static void intel_cqm_event_start(struct perf_event *event, int mode)
>>>> @@ -2566,7 +2552,7 @@ static void intel_cqm_event_stop(struct perf_event
>>>> *event, int mode)
>>>> /* Occupancy of CQM events is obtained at read. No need to read
>>>> * when event is stopped since read on inactive cpus succeed.
>>>> */
>>>> - __update_pqr_prmid(__prmid_from_rmid(pkg_id,
>>>> summary.sched_rmid));
>>>> + pqr_cache_update_rmid(summary.sched_rmid,
>>>> PQR_RMID_MODE_NOEVENT);
>>>> }
>>>>
>>>> static int intel_cqm_event_add(struct perf_event *event, int mode)
>>>> @@ -2977,6 +2963,8 @@ static void intel_cqm_cpu_starting(unsigned int
>>>> cpu)
>>>>
>>>> state->rmid = 0;
>>>> state->closid = 0;
>>>> + state->next_rmid = 0;
>>>> + state->next_closid = 0;
>>>>
>>>> /* XXX: lock */
>>>> /* XXX: Make sure this case is handled when hotplug happens. */
>>>> @@ -3152,6 +3140,12 @@ static int __init intel_cqm_init(void)
>>>> pr_info("Intel CQM monitoring enabled with at least %u rmids per
>>>> package.\n",
>>>> min_max_rmid + 1);
>>>>
>>>> + /* Make sure pqr_common_enable_key is enabled after
>>>> + * cqm_initialized_key.
>>>> + */
>>>> + barrier();
>>>> +
>>>> + static_branch_enable(&pqr_common_enable_key);
>>>> return ret;
>>>>
>>>> error_init_mutex:
>>>> @@ -3163,4 +3157,41 @@ error:
>>>> return ret;
>>>> }
>>>>
>>>> +/* Schedule task without a CQM perf_event. */
>>>> +inline void __intel_cqm_no_event_sched_in(void)
>>>> +{
>>>> +#ifdef CONFIG_CGROUP_PERF
>>>> + struct monr *monr;
>>>> + struct pmonr *pmonr;
>>>> + union prmid_summary summary;
>>>> + u16 pkg_id = topology_physical_package_id(smp_processor_id());
>>>> + struct pmonr *root_pmonr = monr_hrchy_root->pmonrs[pkg_id];
>>>> +
>>>> + /* Assume CQM enabled is likely given that PQR is enabled. */
>>>> + if (!static_branch_likely(&cqm_initialized_key))
>>>> + return;
>>>> +
>>>> + /* Safe to call from_task since we are in scheduler lock. */
>>>> + monr = monr_from_perf_cgroup(perf_cgroup_from_task(current,
>>>> NULL));
>>>> + pmonr = monr->pmonrs[pkg_id];
>>>> +
>>>> + /* Utilize most up to date pmonr summary. */
>>>> + monr_hrchy_get_next_prmid_summary(pmonr);
>>>> + summary.value = atomic64_read(&pmonr->prmid_summary_atomic);
>>>> +
>>>> + if (!prmid_summary__is_mon_active(summary))
>>>> + goto no_rmid;
>>>> +
>>>> + if (WARN_ON_ONCE(!__valid_rmid(pkg_id, summary.sched_rmid)))
>>>> + goto no_rmid;
>>>> +
>>>> + pqr_cache_update_rmid(summary.sched_rmid,
>>>> PQR_RMID_MODE_NOEVENT);
>>>> + return;
>>>> +
>>>> +no_rmid:
>>>> + summary.value =
>>>> atomic64_read(&root_pmonr->prmid_summary_atomic);
>>>> + pqr_cache_update_rmid(summary.sched_rmid,
>>>> PQR_RMID_MODE_NOEVENT);
>>>> +#endif
>>>> +}
>>>> +
>>>> device_initcall(intel_cqm_init);
>>>> diff --git a/arch/x86/events/intel/cqm.h b/arch/x86/events/intel/cqm.h
>>>> index 0f3da94..e1f8bd0 100644
>>>> --- a/arch/x86/events/intel/cqm.h
>>>> +++ b/arch/x86/events/intel/cqm.h
>>>> @@ -82,8 +82,6 @@ union prmid_summary {
>>>> };
>>>> };
>>>>
>>>> -# define INVALID_RMID (-1)
>>>> -
>>>> /* A pmonr in (U)state has no sched_rmid, read_rmid can be 0 or
>>>> INVALID_RMID
>>>> * depending on whether monitoring is active or not.
>>>> */
>>>> diff --git a/arch/x86/include/asm/pqr_common.h
>>>> b/arch/x86/include/asm/pqr_common.h
>>>> index f770637..abbb235 100644
>>>> --- a/arch/x86/include/asm/pqr_common.h
>>>> +++ b/arch/x86/include/asm/pqr_common.h
>>>> @@ -3,31 +3,72 @@
>>>>
>>>> #if defined(CONFIG_INTEL_RDT)
>>>>
>>>> +#include <linux/jump_label.h>
>>>> #include <linux/types.h>
>>>> #include <asm/percpu.h>
>>>> +#include <asm/msr.h>
>>>>
>>>> #define MSR_IA32_PQR_ASSOC 0x0c8f
>>>> +#define INVALID_RMID (-1)
>>>> +#define INVALID_CLOSID (-1)
>>>> +
>>>> +
>>>> +extern struct static_key_false pqr_common_enable_key;
>>>> +
>>>> +enum intel_pqr_rmid_mode {
>>>> + /* RMID has no perf_event associated. */
>>>> + PQR_RMID_MODE_NOEVENT = 0,
>>>> + /* RMID has a perf_event associated. */
>>>> + PQR_RMID_MODE_EVENT
>>>> +};
>>>>
>>>> /**
>>>> * struct intel_pqr_state - State cache for the PQR MSR
>>>> - * @rmid: The cached Resource Monitoring ID
>>>> - * @closid: The cached Class Of Service ID
>>>> + * @rmid: Last rmid written to hw.
>>>> + * @next_rmid: Next rmid to write to hw.
>>>> + * @next_rmid_mode: Next rmid's mode.
>>>> + * @closid: The current Class Of Service ID
>>>> + * @next_closid: The Class Of Service ID to use.
>>>> *
>>>> * The upper 32 bits of MSR_IA32_PQR_ASSOC contain closid and the
>>>> * lower 10 bits rmid. The update to MSR_IA32_PQR_ASSOC always
>>>> * contains both parts, so we need to cache them.
>>>> *
>>>> - * The cache also helps to avoid pointless updates if the value does
>>>> - * not change.
>>>> + * The cache also helps to avoid pointless updates if the value does
>>>> not
>>>> + * change. It also keeps track of the type of RMID set (event vs no
>>>> event)
>>>> + * used to determine when a cgroup RMID is required.
>>>> */
>>>> struct intel_pqr_state {
>>>> - u32 rmid;
>>>> - u32 closid;
>>>> + u32 rmid;
>>>> + u32 next_rmid;
>>>> + enum intel_pqr_rmid_mode next_rmid_mode;
>>>> + u32 closid;
>>>> + u32 next_closid;
>>>> };
>>>>
>>>> DECLARE_PER_CPU(struct intel_pqr_state, pqr_state);
>>>>
>>>> #define PQR_MAX_NR_PKGS 8
>>>>
>>>> +void __pqr_update(void);
>>>> +
>>>> +inline void __intel_cqm_no_event_sched_in(void);
>>>> +
>>>> +inline void pqr_cache_update_rmid(u32 rmid, enum intel_pqr_rmid_mode
>>>> mode);
>>>> +
>>>> +inline void pqr_cache_update_closid(u32 closid);
>>>> +
>>>> +static inline void pqr_update(void)
>>>> +{
>>>> + if (static_branch_unlikely(&pqr_common_enable_key))
>>>> + __pqr_update();
>>>> +}
>>>> +
>>>> +#else
>>>> +
>>>> +static inline void pqr_update(void)
>>>> +{
>>>> +}
>>>> +
>>>> #endif
>>>> #endif
>>>> diff --git a/arch/x86/kernel/cpu/pqr_common.c
>>>> b/arch/x86/kernel/cpu/pqr_common.c
>>>> index 9eff5d9..d91c127 100644
>>>> --- a/arch/x86/kernel/cpu/pqr_common.c
>>>> +++ b/arch/x86/kernel/cpu/pqr_common.c
>>>> @@ -1,9 +1,43 @@
>>>> #include <asm/pqr_common.h>
>>>>
>>>> -/*
>>>> - * The cached intel_pqr_state is strictly per CPU and can never be
>>>> - * updated from a remote CPU. Both functions which modify the state
>>>> - * (intel_cqm_event_start and intel_cqm_event_stop) are called with
>>>> - * interrupts disabled, which is sufficient for the protection.
>>>> - */
>>>> DEFINE_PER_CPU(struct intel_pqr_state, pqr_state);
>>>> +
>>>> +DEFINE_STATIC_KEY_FALSE(pqr_common_enable_key);
>>>> +
>>>> +inline void pqr_cache_update_rmid(u32 rmid, enum intel_pqr_rmid_mode
>>>> mode)
>>>> +{
>>>> + struct intel_pqr_state *state = this_cpu_ptr(&pqr_state);
>>>> +
>>>> + state->next_rmid_mode = mode;
>>>> + state->next_rmid = rmid;
>>>> +}
>>>> +
>>>> +inline void pqr_cache_update_closid(u32 closid)
>>>> +{
>>>> + struct intel_pqr_state *state = this_cpu_ptr(&pqr_state);
>>>> +
>>>> + state->next_closid = closid;
>>>> +}
>>>> +
>>>> +/* Update hw's RMID using cgroup's if perf_event did not.
>>>> + * Sync pqr cache with MSR.
>>>> + */
>>>> +inline void __pqr_update(void)
>>>> +{
>>>> + struct intel_pqr_state *state = this_cpu_ptr(&pqr_state);
>>>> +
>>>> + /* If perf_event has set a next_rmid that is used, do not try
>>>> + * to obtain another one from current task.
>>>> + */
>>>> + if (state->next_rmid_mode == PQR_RMID_MODE_NOEVENT)
>>>> + __intel_cqm_no_event_sched_in();
>>>
>>>
>>>
>>> if perf_cgroup is not defined then the state is not updated here so the
>>> state might have state->rmid might have IDs ?
>>>
>>> 1.event1 for PID1 has RMID1,
>>> 2.perf sched_in - state->next_rmid = rmid1
>>> 3.pqr_update - state->rmid = rmid1
>>> 4.sched_out - write PQR_NOEVENT -
>>> 5.next switch_to - state->rmid not reset nothing changes (when no
>>> perf_cgroup) ?
>
>
> Between 4 and 5 say event1 is dead. Basically on the next context switch(#5)
> if perf sched_in wasnt called the PQR still has RMID1.
>
> When you have perf_cgroup you call the __intel_cqm_no_event.. then it sees
> if there is continuous monitoring or you set the next_rmid to 0- but all the
> code there is inside #ifdef PERF_CGROUP, so without cgroup its never reset
> to zero ?
>
>
>
>>>
>>>
>>>> +
>>>> + /* __intel_cqm_no_event_sched_in might have changed next_rmid.
>>>> */
>>>> + if (state->rmid == state->next_rmid &&
>>>> + state->closid == state->next_closid)
>>>> + return;
>>>> +
>>>> + state->rmid = state->next_rmid;
>>>> + state->closid = state->next_closid;
>>>> + wrmsr(MSR_IA32_PQR_ASSOC, state->rmid, state->closid);
>>>> +}
>>>> --
>>>> 2.8.0.rc3.226.g39d4020
>>>>
>>>>
>>>
>>
>
next prev parent reply other threads:[~2016-04-29 21:26 UTC|newest]
Thread overview: 51+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-04-29 4:43 [PATCH 00/32] 2nd Iteration of Cache QoS Monitoring support David Carrillo-Cisneros
2016-04-29 4:43 ` [PATCH 01/32] perf/x86/intel/cqm: temporarily remove MBM from CQM and cleanup David Carrillo-Cisneros
2016-04-29 20:19 ` Vikas Shivappa
2016-04-29 4:43 ` [PATCH 02/32] perf/x86/intel/cqm: remove check for conflicting events David Carrillo-Cisneros
2016-04-29 4:43 ` [PATCH 03/32] perf/x86/intel/cqm: remove all code for rotation of RMIDs David Carrillo-Cisneros
2016-04-29 4:43 ` [PATCH 04/32] perf/x86/intel/cqm: make read of RMIDs per package (Temporal) David Carrillo-Cisneros
2016-04-29 4:43 ` [PATCH 05/32] perf/core: remove unused pmu->count David Carrillo-Cisneros
2016-04-29 4:43 ` [PATCH 06/32] x86/intel,cqm: add CONFIG_INTEL_RDT configuration flag and refactor PQR David Carrillo-Cisneros
2016-04-29 4:43 ` [PATCH 07/32] perf/x86/intel/cqm: separate CQM PMU's attributes from x86 PMU David Carrillo-Cisneros
2016-04-29 4:43 ` [PATCH 08/32] perf/x86/intel/cqm: prepare for next patches David Carrillo-Cisneros
2016-04-29 9:18 ` Peter Zijlstra
2016-04-29 4:43 ` [PATCH 09/32] perf/x86/intel/cqm: add per-package RMIDs, data and locks David Carrillo-Cisneros
2016-04-29 20:56 ` Vikas Shivappa
2016-04-29 4:43 ` [PATCH 10/32] perf/x86/intel/cqm: basic RMID hierarchy with per package rmids David Carrillo-Cisneros
2016-04-29 4:43 ` [PATCH 11/32] perf/x86/intel/cqm: (I)state and limbo prmids David Carrillo-Cisneros
2016-04-29 4:43 ` [PATCH 12/32] perf/x86/intel/cqm: add per-package RMID rotation David Carrillo-Cisneros
2016-04-29 4:43 ` [PATCH 13/32] perf/x86/intel/cqm: add polled update of RMID's llc_occupancy David Carrillo-Cisneros
2016-04-29 4:43 ` [PATCH 14/32] perf/x86/intel/cqm: add preallocation of anodes David Carrillo-Cisneros
2016-04-29 4:43 ` [PATCH 15/32] perf/core: add hooks to expose architecture specific features in perf_cgroup David Carrillo-Cisneros
2016-04-29 4:43 ` [PATCH 16/32] perf/x86/intel/cqm: add cgroup support David Carrillo-Cisneros
2016-04-29 4:43 ` [PATCH 17/32] perf/core: adding pmu::event_terminate David Carrillo-Cisneros
2016-04-29 4:43 ` [PATCH 18/32] perf/x86/intel/cqm: use pmu::event_terminate David Carrillo-Cisneros
2016-04-29 4:43 ` [PATCH 19/32] perf/core: introduce PMU event flag PERF_CGROUP_NO_RECURSION David Carrillo-Cisneros
2016-04-29 4:43 ` [PATCH 20/32] x86/intel/cqm: use PERF_CGROUP_NO_RECURSION in CQM David Carrillo-Cisneros
2016-04-29 4:43 ` [PATCH 21/32] perf/x86/intel/cqm: handle inherit event and inherit_stat flag David Carrillo-Cisneros
2016-04-29 4:43 ` [PATCH 22/32] perf/x86/intel/cqm: introduce read_subtree David Carrillo-Cisneros
2016-04-29 4:43 ` [PATCH 23/32] perf/core: introduce PERF_INACTIVE_*_READ_* flags David Carrillo-Cisneros
2016-04-29 4:43 ` [PATCH 24/32] perf/x86/intel/cqm: use PERF_INACTIVE_*_READ_* flags in CQM David Carrillo-Cisneros
2016-04-29 4:43 ` [PATCH 25/32] sched: introduce the finish_arch_pre_lock_switch() scheduler hook David Carrillo-Cisneros
2016-04-29 8:52 ` Peter Zijlstra
[not found] ` <CALcN6miyq9_4GQfO9=bjFb-X_2LSQdwfWnm+KvT=UrYRCAb6Og@mail.gmail.com>
2016-04-29 18:40 ` David Carrillo-Cisneros
2016-04-29 20:21 ` Vikas Shivappa
2016-04-29 20:50 ` David Carrillo-Cisneros
2016-04-29 4:43 ` [PATCH 26/32] perf/x86/intel/cqm: integrate CQM cgroups with scheduler David Carrillo-Cisneros
2016-04-29 20:25 ` Vikas Shivappa
2016-04-29 20:48 ` David Carrillo-Cisneros
2016-04-29 21:01 ` Vikas Shivappa
2016-04-29 21:26 ` David Carrillo-Cisneros [this message]
2016-04-29 21:32 ` Vikas Shivappa
2016-04-29 21:49 ` David Carrillo-Cisneros
2016-04-29 23:49 ` Vikas Shivappa
2016-04-30 17:50 ` David Carrillo-Cisneros
2016-05-02 13:22 ` Thomas Gleixner
2016-04-29 4:43 ` [PATCH 27/32] perf/core: add perf_event cgroup hooks for subsystem attributes David Carrillo-Cisneros
2016-04-29 4:43 ` [PATCH 28/32] perf/x86/intel/cqm: add CQM attributes to perf_event cgroup David Carrillo-Cisneros
2016-04-29 4:43 ` [PATCH 29/32] perf,perf/x86,perf/powerpc,perf/arm,perf/*: add int error return to pmu::read David Carrillo-Cisneros
2016-04-29 4:43 ` [PATCH 30/32] perf,perf/x86: add hook perf_event_arch_exec David Carrillo-Cisneros
2016-04-29 4:43 ` [PATCH 31/32] perf/stat: fix bug in handling events in error state David Carrillo-Cisneros
2016-04-29 4:43 ` [PATCH 32/32] perf/stat: revamp error handling for snapshot and per_pkg events David Carrillo-Cisneros
2016-04-29 21:06 ` [PATCH 00/32] 2nd Iteration of Cache QoS Monitoring support Vikas Shivappa
2016-04-29 21:10 ` David Carrillo-Cisneros
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=CALcN6mjYfKHFuR0pj1wtvHH6uoXJ9a5C0HR5727UerCCCPn6eg@mail.gmail.com \
--to=davidcc@google.com \
--cc=acme@kernel.org \
--cc=alexander.shishkin@linux.intel.com \
--cc=eranian@google.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@redhat.com \
--cc=peterz@infradead.org \
--cc=pjt@google.com \
--cc=tony.luck@intel.com \
--cc=vikas.shivappa@linux.intel.com \
--cc=x86@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).