linux-kernel - Re: [PATCH] add PATA host controller support for Cirrus Logic's EP93xx CPUs

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Thu, 26 Nov 2009 17:03:39 +0000
From:	Alan Cox <alan@...rguk.ukuu.org.uk>
To:	Bartlomiej Zolnierkiewicz <bzolnier@...il.com>
Cc:	linux-ide@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
	linux-kernel@...r.kernel.org,
	Sergei Shtylyov <sshtylyov@...montavista.com>,
	Joao Ramos <joao.ramos@...v.pt>,
	H Hartley Sweeten <hsweeten@...ionengravers.com>,
	Ryan Mallon <ryan@...ewatersys.com>
Subject: Re: [PATCH] add PATA host controller support for Cirrus Logic's
 EP93xx CPUs


> +static void pata_ep93xx_set_piomode(struct ata_port *ap,
> +				    struct ata_device *adev)
> +{
> +	void __iomem *base = ap->host->private_data;
> +	struct ata_device *pair = ata_dev_pair(adev);
> +	const struct ata_timing *t = ata_timing_find_mode(adev->pio_mode);
> +	const struct ata_timing *cmd_t = t;
> +	u32 reg = IDECFG_IDEEN | IDECFG_PIO;
> +	u8 pio = adev->pio_mode - XFER_PIO_0;
> +
> +	if (pair && pair->pio_mode < adev->pio_mode)
> +		cmd_t = ata_timing_find_mode(pair->pio_mode);
> +
> +	/*
> +	 * store the adequate PIO mode timings, to be used later
> +	 * by pata_ep93xx_{read,write}
> +	 */
> +	adev->private_data = (void *)t;
> +	ap->private_data = (void *)cmd_t;
> +
> +	/* reconfigure IDE controller according to PIO mode (?) */

This would probably be better in both drivers as a simple table. It would
then not be a switch but collapse the entire pile into a single

	writel(pio_timing[pio], base + IDECFG);


> +static u16 __pata_ep93xx_read(void __iomem *base, void *addr,
> +			      struct ata_timing *t)
> +{
> +	u32 reg;
> +
> +	reg = (unsigned long)addr | IDECTRL_DIORN | IDECTRL_DIOWN;
> +	writel(reg, base + IDECTRL);
> +	ndelay(t->setup);

This implies a bus with no posting, which seems to the case from the
other driver code.

> +	while (!pata_ep93xx_check_iordy(base))
> +		cpu_relax();

And if the drive was hot unplugged this is doom incorporated (plus we
might not be using an IORDY mode.

>
> +	while (!pata_ep93xx_check_iordy(base))
> +		cpu_relax();

Ditto


I don't see anything wrong in the conversion. 

The size thing is true - libata tried to abstract at a higher level which
in turn avoids lots of computed branches to one liners and lets the
compiler inline stuff better (eg all the private read/write methods into
a single call free tf_* etc). It could I guess provide an "awkward
case" skeleton tf_* and other sets of functions that were passed inb/outb
ops. One for Jeff to ponder ?



Alan
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/