From: "Michael Büsch" <mb@bu3sch.de>
To: "Rafał Miłecki" <zajec5@gmail.com>
Cc: linux-wireless@vger.kernel.org, b43-dev@lists.infradead.org
Subject: [RFC][PATCH] b43: N-PHY: prepare for 2056 radio tables writing
Date: Sat, 18 Dec 2010 07:10:47 +0100 [thread overview]
Message-ID: <1292652647.7646.1.camel@maggie> (raw)
In-Reply-To: <1292628161-6201-1-git-send-email-zajec5@gmail.com> (sfid-20101217_182318_257450_0DD1BC69)
On Sat, 2010-12-18 at 00:22 +0100, Rafa? Mi?ecki wrote:
> ---
> Here is the way I'd like to handle many different tables, 3 for every PHY rev.
>
> This adds some helper structure, pointers and magic macro. I understand this
> but is this readable for others?
>
> Advantage of this crazy layout is that we avoid mess like:
> b2056_find_syn_table()
> {
> if (phy == 3)
> return table_phy3;
> else if (phy == 4)
> return table_phy4;
> ...
> else if (phy == 9)
> return table_phy9;
> }
> ... same for tx and rx.
>
> Any objections?
> ---
> drivers/net/wireless/b43/radio_2056.c | 82 +++++++++++++++++++++++++++++++++
> drivers/net/wireless/b43/radio_2056.h | 3 +
> 2 files changed, 85 insertions(+), 0 deletions(-)
>
> diff --git a/drivers/net/wireless/b43/radio_2056.c b/drivers/net/wireless/b43/radio_2056.c
> index 0cdf6a4..d30c10a 100644
> --- a/drivers/net/wireless/b43/radio_2056.c
> +++ b/drivers/net/wireless/b43/radio_2056.c
> @@ -24,6 +24,47 @@
> #include "radio_2056.h"
> #include "phy_common.h"
>
> +struct b2056_inittab_entry {
> + /* Value to write if we use the 5GHz band. */
> + u16 ghz5;
> + /* Value to write if we use the 2.4GHz band. */
> + u16 ghz2;
> + /* Flags */
> + u8 flags;
> +#define B2056_INITTAB_ENTRY_OK 0x01
> +#define B2056_INITTAB_UPLOAD 0x02
> +};
> +#define UPLOAD .flags = B2056_INITTAB_ENTRY_OK | B2056_INITTAB_UPLOAD
> +#define NOUPLOAD .flags = B2056_INITTAB_ENTRY_OK
> +
> +struct b2056_inittabs_pts {
> + const struct b2056_inittab_entry *syn;
> + unsigned int syn_length;
> + const struct b2056_inittab_entry *tx;
> + unsigned int tx_length;
> + const struct b2056_inittab_entry *rx;
> + unsigned int rx_length;
> +};
> +
> +static const struct b2056_inittab_entry b2056_inittab_rev3_syn[] = {
> +};
> +static const struct b2056_inittab_entry b2056_inittab_rev3_tx[] = {
> +};
> +static const struct b2056_inittab_entry b2056_inittab_rev3_rx[] = {
> +};
> +
> +#define INITTABSPTS(prefix) \
> + .syn = prefix##_syn, \
> + .syn_length = ARRAY_SIZE(prefix##_syn), \
> + .tx = prefix##_tx, \
> + .tx_length = ARRAY_SIZE(prefix##_tx), \
> + .rx = prefix##_rx, \
> + .rx_length = ARRAY_SIZE(prefix##_rx)
> +
> +struct b2056_inittabs_pts b2056_inittabs[] = {
> + [3] = { INITTABSPTS(b2056_inittab_rev3) },
> +};
> +
> #define RADIOREGS3(r00, r01, r02, r03, r04, r05, r06, r07, r08, r09, \
> r10, r11, r12, r13, r14, r15, r16, r17, r18, r19, \
> r20, r21, r22, r23, r24, r25, r26, r27, r28, r29, \
> @@ -6045,6 +6086,47 @@ static const struct b43_nphy_channeltab_entry_rev3 b43_nphy_channeltab_rev8[] =
> },
> };
>
> +static void b2056_upload_inittab(struct b43_wldev *dev, bool ghz5,
> + bool ignore_uploadflag, u16 routing,
> + const struct b2056_inittab_entry *e,
> + unsigned int length)
> +{
> + unsigned int i;
> + u16 value;
> +
> + for (i = 0; i < length; i++) {
> + if (!(e->flags & B2056_INITTAB_ENTRY_OK))
> + continue;
> + if ((e->flags & B2056_INITTAB_UPLOAD) || ignore_uploadflag) {
> + if (ghz5)
> + value = e->ghz5;
> + else
> + value = e->ghz2;
> + b43_radio_write(dev, routing | i, value);
> + }
> + e++;
> + }
> +}
> +
> +void b2056_upload_inittabs(struct b43_wldev *dev,
> + bool ghz5, bool ignore_uploadflag)
> +{
> + struct b2056_inittabs_pts *pts;
> +
> + pts = &b2056_inittabs[dev->phy.rev];
I'm OK with this, except that this really needs more sanity checking
here. You need to check the array boundaries and check for
uninitialized (all-zero) entries and print a big fat warn-on.
> +
> + b2056_upload_inittab(dev, ghz5, ignore_uploadflag,
> + B2056_SYN, pts->syn, pts->syn_length);
> + b2056_upload_inittab(dev, ghz5, ignore_uploadflag,
> + B2056_TX0, pts->tx, pts->tx_length);
> + b2056_upload_inittab(dev, ghz5, ignore_uploadflag,
> + B2056_TX1, pts->tx, pts->tx_length);
> + b2056_upload_inittab(dev, ghz5, ignore_uploadflag,
> + B2056_RX0, pts->rx, pts->rx_length);
> + b2056_upload_inittab(dev, ghz5, ignore_uploadflag,
> + B2056_RX1, pts->rx, pts->rx_length);
> +}
> +
> /* TODO: add support for rev4+ devices by searching in rev4+ tables */
> const struct b43_nphy_channeltab_entry_rev3 *
> b43_nphy_get_chantabent_rev3(struct b43_wldev *dev, u16 freq)
> diff --git a/drivers/net/wireless/b43/radio_2056.h b/drivers/net/wireless/b43/radio_2056.h
> index 302600c..d601f6e 100644
> --- a/drivers/net/wireless/b43/radio_2056.h
> +++ b/drivers/net/wireless/b43/radio_2056.h
> @@ -1114,4 +1114,7 @@ struct b43_nphy_channeltab_entry_rev3 {
> struct b43_phy_n_sfo_cfg phy_regs;
> };
>
> +void b2056_upload_inittabs(struct b43_wldev *dev,
> + bool ghz5, bool ignore_uploadflag);
> +
> #endif /* B43_RADIO_2056_H_ */
--
Greetings Michael.
prev parent reply other threads:[~2010-12-18 6:10 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-12-17 23:22 [RFC][PATCH] b43: N-PHY: prepare for 2056 radio tables writing Rafał Miłecki
2010-12-18 6:10 ` Michael Büsch [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=1292652647.7646.1.camel@maggie \
--to=mb@bu3sch.de \
--cc=b43-dev@lists.infradead.org \
--cc=linux-wireless@vger.kernel.org \
--cc=zajec5@gmail.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;
as well as URLs for NNTP newsgroup(s).