| 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
| ||
|
Message-ID: <5121F161.2090503@ti.com>
Date: Mon, 18 Feb 2013 10:16:17 +0100
From: Peter Ujfalusi <peter.ujfalusi@...com>
To: Chen Gang <gang.chen@...anux.com>
CC: <sameo@...ux.intel.com>, Tony Lindgren <tony@...mide.com>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [Suggestion] drivers/mfd/twl*: about power_count in twl6040_remove
On 02/16/2013 10:07 AM, Chen Gang wrote:
>
> if necessary to treat 'power_count' as a counter
> when power_count is more than 1, we do not really power off it.
> this may cause issue:
> after finish calling twl6040_remove, another modules still use it.
> suggest to give additional check, or synchronisation for other modules.
The module can be only removed when all depending modules has been already
removed.
If we remove the core MFD driver we _should_ power off the chip no matter what.
>
> else (not necessary to treat it as a counter)
> better to use it as a boolean value.
> it seems 'power_count' may still have chance to be more than 1:
> if it is true, we still may can not power off it in twl6040_remove.
No, we have child driver for the chip (audio, vibra, GPO, clock) we need to
count the active users in order to keep it powered whenever at least one of
the child driver is using the chip.
>
> another question:
> may 'power_count' be a large number (which may cause value overflow) ?
Given that we have four child drivers and they do their power request balanced
it is never going to happen.
>
>
> thanks.
>
> gchen.
>
>
> ------------------------------------------------------------------------------
>
> relative calls:
>
> drivers/mfd/twl6040.c:212: twl6040_power(twl6040, 0);
> drivers/mfd/twl6040.c:215: twl6040_power(twl6040, 1);
> drivers/mfd/twl6040.c:707: twl6040_power(twl6040, 0);
> drivers/input/misc/twl6040-vibra.c:100: twl6040_power(info->twl6040, 1);
> drivers/input/misc/twl6040-vibra.c:128: twl6040_power(info->twl6040, 0);
> sound/soc/codecs/twl6040.c:911: ret = twl6040_power(twl6040, 1);
> sound/soc/codecs/twl6040.c:926: twl6040_power(twl6040, 0);
> drivers/clk/clk-twl6040.c:51: ret = twl6040_power(twl6040_clk->twl6040, 1);
> drivers/clk/clk-twl6040.c:64: ret = twl6040_power(twl6040_clk->twl6040, 0);
>
>
>
> drivers/mfd/twl6040.c:
>
> 244 int twl6040_power(struct twl6040 *twl6040, int on)
> 245 {
> 246 int ret = 0;
> 247
> 248 mutex_lock(&twl6040->mutex);
> 249
> 250 if (on) {
> 251 /* already powered-up */
> 252 if (twl6040->power_count++)
> 253 goto out;
> 254
> 255 if (gpio_is_valid(twl6040->audpwron)) {
> 256 /* use automatic power-up sequence */
> 257 ret = twl6040_power_up_automatic(twl6040);
> 258 if (ret) {
> 259 twl6040->power_count = 0;
> 260 goto out;
> 261 }
> 262 } else {
> 263 /* use manual power-up sequence */
> 264 ret = twl6040_power_up_manual(twl6040);
> 265 if (ret) {
> 266 twl6040->power_count = 0;
> 267 goto out;
> 268 }
> 269 }
> 270 /* Default PLL configuration after power up */
> 271 twl6040->pll = TWL6040_SYSCLK_SEL_LPPLL;
> 272 twl6040->sysclk = 19200000;
> 273 twl6040->mclk = 32768;
> 274 } else {
> 275 /* already powered-down */
> 276 if (!twl6040->power_count) {
> 277 dev_err(twl6040->dev,
> 278 "device is already powered-off\n");
> 279 ret = -EPERM;
> 280 goto out;
> 281 }
> 282
> 283 if (--twl6040->power_count)
> 284 goto out;
> 285
> 286 if (gpio_is_valid(twl6040->audpwron)) {
> 287 /* use AUDPWRON line */
> 288 gpio_set_value(twl6040->audpwron, 0);
> 289
> 290 /* power-down sequence latency */
> 291 usleep_range(500, 700);
> 292 } else {
> 293 /* use manual power-down sequence */
> 294 twl6040_power_down_manual(twl6040);
> 295 }
> 296 twl6040->sysclk = 0;
> 297 twl6040->mclk = 0;
> 298 }
> 299
> 300 out:
> 301 mutex_unlock(&twl6040->mutex);
> 302 return ret;
> 303 }
> 304 EXPORT_SYMBOL(twl6040_power);
> ...
>
> 702 static int twl6040_remove(struct i2c_client *client)
> 703 {
> 704 struct twl6040 *twl6040 = i2c_get_clientdata(client);
> 705
> 706 if (twl6040->power_count)
> 707 twl6040_power(twl6040, 0);
> 708
> 709 if (gpio_is_valid(twl6040->audpwron))
> 710 gpio_free(twl6040->audpwron);
> 711
> 712 free_irq(twl6040->irq_ready, twl6040);
> 713 free_irq(twl6040->irq_th, twl6040);
> 714 regmap_del_irq_chip(twl6040->irq, twl6040->irq_data);
> 715
> 716 mfd_remove_devices(&client->dev);
> 717 i2c_set_clientdata(client, NULL);
> 718
> 719 regulator_bulk_disable(TWL6040_NUM_SUPPLIES, twl6040->supplies);
> 720 regulator_bulk_free(TWL6040_NUM_SUPPLIES, twl6040->supplies);
> 721
> 722 return 0;
> 723 }
>
> -----------------------------------------------------------------------------------
>
>
>
--
Péter
--
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/
Powered by blists - more mailing lists