From: "Randy.Dunlap" <rdunlap@xenotime.net>
To: Heiko J Schick <schihei@de.ibm.com>
Cc: openib-general@openib.org, RAISCH@de.ibm.com, HNGUYEN@de.ibm.com,
MEDER@de.ibm.com, schihei@de.ibm.com,
linux-kernel@vger.kernel.org, linuxppc-dev@ozlabs.org
Subject: Re: [PATCH 14/16] ehca: queue page table handling
Date: Mon, 15 May 2006 13:54:19 -0700 [thread overview]
Message-ID: <20060515135419.f77a1d8b.rdunlap@xenotime.net> (raw)
In-Reply-To: <4468BDAD.3020701@de.ibm.com>
On Mon, 15 May 2006 19:43:09 +0200 Heiko J Schick wrote:
> Signed-off-by: Heiko J Schick <schickhj@de.ibm.com>
>
>
> drivers/infiniband/hw/ehca/ipz_pt_fn.c | 177 ++++++++++++++++++++++
> drivers/infiniband/hw/ehca/ipz_pt_fn.h | 254 +++++++++++++++++++++++++++++++++
> 2 files changed, 431 insertions(+)
>
>
>
> --- linux-2.6.17-rc2-orig/drivers/infiniband/hw/ehca/ipz_pt_fn.h 1970-01-01 01:00:00.000000000 +0100
> +++ linux-2.6.17-rc2/drivers/infiniband/hw/ehca/ipz_pt_fn.h 2006-05-12 12:25:43.000000000 +0200
> @@ -0,0 +1,254 @@
> +/* return current Queue Page , increment Queue Page iterator from
> + * page to page in struct ipz_queue, last increment will return 0! and
> + * NOT wrap
> + * returns address (kv) of Queue Page
> + * warning don't use in parallel with ipz_QE_get_inc()
> + */
When not using kernel-doc format... the preferred multi-line
comment format is:
/*
* foo foo foo ........
* bar bar bar .......
* blahz ..........
*/
repeat below.
> +void *ipz_qpageit_get_inc(struct ipz_queue *queue);
> +
> +/* return current Queue Entry, increment Queue Entry iterator by one
> + * step in struct ipz_queue, will wrap in ringbuffer
> + * @returns address (kv) of Queue Entry BEFORE increment
> + * warning don't use in parallel with ipz_qpageit_get_inc()
> + * warning unpredictable results may occur if steps>act_nr_of_queue_entries
> + */
> +static inline void *ipz_qeit_get_inc(struct ipz_queue *queue)
> +{
> + void *ret = NULL;
> +
> + ret = ipz_qeit_get(queue);
> + queue->current_q_offset += queue->qe_size;
> + if (queue->current_q_offset >= queue->queue_length) {
> + queue->current_q_offset = 0;
> + /* toggle the valid flag */
> + queue->toggle_state = (~queue->toggle_state) & 1;
> + }
> +
> + EDEB(7, "queue=%p ret=%p new current_q_addr=%lx qe_size=%x",
> + queue, ret, queue->current_q_offset, queue->qe_size);
> +
> + return ret;
> +}
> +
> +/* return current Queue Entry, increment Queue Entry iterator by one
> + * step in struct ipz_queue, will wrap in ringbuffer
> + * returns address (kv) of Queue Entry BEFORE increment
> + * returns 0 and does not increment, if wrong valid state
> + * warning don't use in parallel with ipz_qpageit_get_inc()
> + * warning unpredictable results may occur if steps>act_nr_of_queue_entries
> + */
> +static inline void *ipz_qeit_get_inc_valid(struct ipz_queue *queue)
> +{
> + struct ehca_cqe *cqe = ipz_qeit_get(queue);
> + u32 cqe_flags = cqe->cqe_flags;
> +
> + if ((cqe_flags >> 7) != (queue->toggle_state & 1))
> + return NULL;
> +
> + ipz_qeit_get_inc(queue);
> + return cqe;
> +}
> +
> +/* destructor for a ipz_queue_t
> + * -# free queue
> + * see ipz_queue_ctor()
> + * returns true if ok, false if queue was NULL-ptr of free failed
> + */
> +int ipz_queue_dtor(struct ipz_queue *queue);
> +
> +/* constructor for a ipz_qpt_t,
> + * placement new for struct ipz_queue, new for all dependent datastructors
> + *
> + * all QP Tables are the same,
> + * flow:
> + * -# allocate+pin queue
> + * -# initialise ptcb
> + * -# allocate+pin PTs
> + * -# link PTs to a ring, according to HCA Arch, set bit62 id needed
> + * -# the ring must have room for exactly nr_of_PTEs
> + * see ipz_qpt_ctor()
> + */
> +void ipz_qpt_ctor(struct ipz_qpt *qpt,
> + const u32 nr_of_QEs,
> + const u32 pagesize,
> + const u32 qe_size,
> + const u8 lowbyte, const u8 toggle,
> + u32 * act_nr_of_QEs, u32 * act_nr_of_pages);
> +
> +/* return current Queue Entry, increment Queue Entry iterator by one
> + * step in struct ipz_queue, will wrap in ringbuffer
> + * returns address (kv) of Queue Entry BEFORE increment
> + * warning don't use in parallel with ipz_qpageit_get_inc()
> + * warning unpredictable results may occur if steps>act_nr_of_queue_entries
> + *
> + * fix EQ page problems
> + */
> +void *ipz_qeit_eq_get_inc(struct ipz_queue *queue);
> +
> +/* return current Event Queue Entry, increment Queue Entry iterator
> + * by one step in struct ipz_queue if valid, will wrap in ringbuffer
> + * returns address (kv) of Queue Entry BEFORE increment
> + * returns 0 and does not increment, if wrong valid state
> + * warning don't use in parallel with ipz_queue_QPageit_get_inc()
> + * warning unpredictable results may occur if steps>act_nr_of_queue_entries
> + */
> +static inline void *ipz_eqit_eq_get_inc_valid(struct ipz_queue *queue)
> +{
> + void *ret = ipz_qeit_get(queue);
> + u32 qe = *(u8 *) ret;
> + EDEB(7, "ipz_QEit_EQ_get_inc_valid qe=%x", qe);
> + if ((qe >> 7) == (queue->toggle_state & 1))
> + ipz_qeit_eq_get_inc(queue); /* this is a good one */
> + else
> + ret = NULL;
> + return ret;
> +}
---
~Randy
prev parent reply other threads:[~2006-05-15 20:51 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-05-15 17:43 Heiko J Schick
2006-05-15 20:54 ` Randy.Dunlap [this message]
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=20060515135419.f77a1d8b.rdunlap@xenotime.net \
--to=rdunlap@xenotime.net \
--cc=HNGUYEN@de.ibm.com \
--cc=MEDER@de.ibm.com \
--cc=RAISCH@de.ibm.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linuxppc-dev@ozlabs.org \
--cc=openib-general@openib.org \
--cc=schihei@de.ibm.com \
/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
all inboxes | Powered by JetHome®