i2c-pnx: Improvements and fixes
Subsystems (2)
This patch set fixes a stop condition issue with transfers, caps the maximum computed divider generated from the parent clock, and disables the I2C peripheral clock when the I2C bus isn't in use (to save power).
Patch 3 was previously posted a few months ago, but I didn't follow up on getting it into the driver.
All code is tested on the lpc3250.
A message that is not in the archive (sent off-list or not yet received): 1268780138-10019-1-git-send-email-wellsk40-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org
PATCH 1/3 i2c-pnx: Limit maximum divider to 1023
From: Kevin Wells <[email protected]>
Limit maximum divider to 0x3ff on divider computations. On high I2C parent clock rates, the divider can exceed 0x3ff. This will help prevent some very odd clock rates.
--- drivers/i2c/busses/i2c-pnx.c | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-)
drivers/i2c/busses/i2c-pnx.c +2 -0
@@ -632,6 +632,8 @@ static int __devinit i2c_pnx_probe(struct platform_device *pdev)*/ tmp = ((freq /1000 ) /I2C_PNX_SPEED_KHZ ) /2 -2 ; +if (tmp >0x3FF ) + tmp =0x3FF ;iowrite32 (tmp,I2C_REG_CKH (alg_data));iowrite32 (tmp,I2C_REG_CKL (alg_data));
Signature · 1 line
1.6.6
PATCH 1/3 i2c-pnx: Limit maximum divider to 1023
In reply to a message not in the archiveFrom: Kevin Wells <[email protected]>
Limit maximum divider to 0x3ff to divider computations. On high I2C parent clock rates, the divider can exceed 0x3ff. This will help prevent some very odd clock rates.
--- drivers/i2c/busses/i2c-pnx.c | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-)
drivers/i2c/busses/i2c-pnx.c +2 -0
@@ -632,6 +632,8 @@ static int __devinit i2c_pnx_probe(struct platform_device *pdev)*/ tmp = ((freq /1000 ) /I2C_PNX_SPEED_KHZ ) /2 -2 ; +if (tmp >0x3FF ) + tmp =0x3FF ;iowrite32 (tmp,I2C_REG_CKH (alg_data));iowrite32 (tmp,I2C_REG_CKL (alg_data));
Signature · 1 line
1.6.6
PATCH 2/3 i2c-pnx: Add stop conditions for end of transfer
In reply to a message not in the archiveFrom: Kevin Wells <[email protected]>
Add a stop condition bit flag to the last byte in the transfer. This will generate an extra clock to handle the stop condition and prevent devices from staying in an ACK'd state.
--- drivers/i2c/busses/i2c-pnx.c | 6 ++++++ 1 files changed, 6 insertions(+), 0 deletions(-)
drivers/i2c/busses/i2c-pnx.c +6 -0
@@ -172,6 +172,9 @@ static int i2c_pnx_master_xmit(struct i2c_pnx_algo_data *alg_data)/* We still have something to talk about... */ val = *alg_data->mif.buf++; +if (alg_data->mif.len ==1 ) + val |= stop_bit; + alg_data->mif.len--;iowrite32 (val,I2C_REG_TX (alg_data));
@@ -245,6 +248,9 @@ static int i2c_pnx_master_rcv(struct i2c_pnx_algo_data *alg_data) __func__);if (alg_data->mif.len ==1 ) { +/* Last byte, do not acknowledge next rcv. */ + val |= stop_bit; +/* * Enable interrupt RFDAIE (data in Rx fifo), * and disable DRMIE (need data for Tx)
Signature · 1 line
1.6.6
PATCH 3/3 i2c-pnx: Disable/enable clocks only when needed to reduce power
In reply to a message not in the archiveFrom: Kevin Wells <[email protected]>
Enable I2C peripheral clocking at the start of a transfer and disable on transfer completion to reduce peripheral power use when bus is not in use.
--- drivers/i2c/busses/i2c-pnx.c | 33 +++++++-------------------------- 1 files changed, 7 insertions(+), 26 deletions(-)
drivers/i2c/busses/i2c-pnx.c +7 -26
@@ -448,6 +448,8 @@ i2c_pnx_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, int num)"%s(): entering: %d messages, stat = %04x.\n" , __func__, num, ioread32(I2C_REG_STS(alg_data))); +clk_enable (alg_data->clk); +bus_reset_if_active (alg_data);/* Process transactions in a loop. */
@@ -516,6 +518,8 @@ i2c_pnx_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, int num)bus_reset_if_active (alg_data); +clk_disable (alg_data->clk); +/* Cleanup to be sure... */ alg_data->mif.buf =NULL ; alg_data->mif.len =0 ;
@@ -539,29 +543,6 @@ static struct i2c_algorithm pnx_algorithm = { .functionality = i2c_pnx_func, }; -#ifdef CONFIG_PM -static int i2c_pnx_controller_suspend (struct platform_device *pdev, -pm_message_t state) -{ -struct i2c_pnx_algo_data *alg_data =platform_get_drvdata (pdev); - -/* FIXME: shouldn't this be clk_disable? */ -clk_enable (alg_data->clk); - -return 0 ; -} - -static int i2c_pnx_controller_resume (struct platform_device *pdev) -{ -struct i2c_pnx_algo_data *alg_data =platform_get_drvdata (pdev); - -return clk_enable (alg_data->clk); -} -#else -#define i2c_pnx_controller_suspend NULL -#define i2c_pnx_controller_resume NULL -#endif -static int __devinit i2c_pnx_probe (struct platform_device *pdev) {unsigned long tmp;
@@ -665,6 +646,9 @@ static int __devinit i2c_pnx_probe(struct platform_device *pdev)dev_dbg (&pdev->dev,"%s: Master at %#8x, irq %d.\n" , alg_data->adapter.name, i2c_pnx->base, i2c_pnx->irq); +/* Disable clock until needed */ +clk_disable (alg_data->clk); +return 0 ; out_irq:
@@ -692,7 +676,6 @@ static int __devexit i2c_pnx_remove(struct platform_device *pdev)free_irq (i2c_pnx->irq, alg_data);i2c_del_adapter (&alg_data->adapter); -clk_disable (alg_data->clk);iounmap (alg_data->ioaddr);release_mem_region (i2c_pnx->base,I2C_PNX_REGION_SIZE );clk_put (alg_data->clk);
@@ -709,8 +692,6 @@ static struct platform_driver i2c_pnx_driver = { }, .probe = i2c_pnx_probe, .remove = __devexit_p(i2c_pnx_remove), - .suspend = i2c_pnx_controller_suspend, - .resume = i2c_pnx_controller_resume, };static int __init i2c_adap_pnx_init (void )
Signature · 1 line
1.6.6
A message that is not in the archive (sent off-list or not yet received): 1268780138-10019-5-git-send-email-wellsk40-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org
Re: [PATCH 3/3] i2c-pnx: Disable/enable clocks only when needed to reduce power
Quoted 5 lines · On Tue, Mar 16, 2010 at 03:55:38PM -0700…
From: Kevin Wells <[email protected]>
Enable I2C peripheral clocking at the start of a transfer and disable on transfer completion to reduce peripheral power use when bus is not in use.
I'm going to leave this one for -next as it isn't really a bugfix.
Quoted 91 lines
Signed-off-by Kevin Wells <[email protected]>--- drivers/i2c/busses/i2c-pnx.c | 33 +++++++-------------------------- 1 files changed, 7 insertions(+), 26 deletions(-)drivers/i2c/busses/i2c-pnx.c +7
-26@@ -448,6 +448,8 @@ i2c_pnx_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, int num)"%s(): entering: %d messages, stat = %04x.\n" , __func__, num, ioread32(I2C_REG_STS(alg_data))); +clk_enable (alg_data->clk); +bus_reset_if_active (alg_data);/* Process transactions in a loop. */ @@ -516,6 +518,8 @@ i2c_pnx_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, int num)bus_reset_if_active (alg_data); +clk_disable (alg_data->clk); +/* Cleanup to be sure... */ alg_data->mif.buf =NULL ; alg_data->mif.len =0 ;@@ -539,29 +543,6 @@ static struct i2c_algorithm pnx_algorithm = { .functionality = i2c_pnx_func, }; -#ifdef CONFIG_PM -static int i2c_pnx_controller_suspend (struct platform_device *pdev, -pm_message_t state) -{ -struct i2c_pnx_algo_data *alg_data =platform_get_drvdata (pdev); - -/* FIXME: shouldn't this be clk_disable? */ -clk_enable (alg_data->clk); - -return 0 ; -} - -static int i2c_pnx_controller_resume (struct platform_device *pdev) -{ -struct i2c_pnx_algo_data *alg_data =platform_get_drvdata (pdev); - -return clk_enable (alg_data->clk); -} -#else -#define i2c_pnx_controller_suspend NULL -#define i2c_pnx_controller_resume NULL -#endif -static int __devinit i2c_pnx_probe (struct platform_device *pdev) {unsigned long tmp;@@ -665,6 +646,9 @@ static int __devinit i2c_pnx_probe(struct platform_device *pdev)dev_dbg (&pdev->dev,"%s: Master at %#8x, irq %d.\n" , alg_data->adapter.name, i2c_pnx->base, i2c_pnx->irq); +/* Disable clock until needed */ +clk_disable (alg_data->clk); +return 0 ; out_irq:@@ -692,7 +676,6 @@ static int __devexit i2c_pnx_remove(struct platform_device *pdev)free_irq (i2c_pnx->irq, alg_data);i2c_del_adapter (&alg_data->adapter); -clk_disable (alg_data->clk);iounmap (alg_data->ioaddr);release_mem_region (i2c_pnx->base,I2C_PNX_REGION_SIZE );clk_put (alg_data->clk);@@ -709,8 +692,6 @@ static struct platform_driver i2c_pnx_driver = { }, .probe = i2c_pnx_probe, .remove = __devexit_p(i2c_pnx_remove), - .suspend = i2c_pnx_controller_suspend, - .resume = i2c_pnx_controller_resume, };static int __init i2c_adap_pnx_init (void )Signature · 6 lines
1.6.6 -- To unsubscribe from this list: send the line "unsubscribe linux-i2c" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
Signature · 3 lines
Ben ([email protected], http://www.fluff.org/) 'a smiley only costs 4 bytes'
Hi Vitaly,
Can you please ack these patches for the i2c-pnx driver? These are all important changes for the i2c-pnx.
thanks,
Kevin
Quoted 11 lines · On Tue, Mar 16, 2010 at 3:55 PM, <[email protected]…
This patch set fixes a stop condition issue with transfers, caps the maximum computed divider generated from the parent clock, and disables the I2C peripheral clock when the I2C bus isn't in use (to save power).
Patch 3 was previously posted a few months ago, but I didn't follow up on getting it into the driver.
All code is tested on the lpc3250.
A message that is not in the archive (sent off-list or not yet received): a08234131003221009j1ec74009jc2eb289386455b65-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org
Quoted 4 lines · Kevin Wells
Hi Vitaly,
Can you please ack these patches for the i2c-pnx driver? These are all important changes for the i2c-pnx.
I'm going to take no news as good news, and apply these unless anyone objects before the pull request.
Quoted 19 lines
thanks,
KevinQuoted 11 lines · On Tue, Mar 16, 2010 at 3:55 PM, <[email protected]…
This patch set fixes a stop condition issue with transfers, caps the maximum computed divider generated from the parent clock, and disables the I2C peripheral clock when the I2C bus isn't in use (to save power).
Patch 3 was previously posted a few months ago, but I didn't follow up on getting it into the driver.
All code is tested on the lpc3250.
--
To unsubscribe from this list: send the line "unsubscribe linux-i2c" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
Signature · 3 lines
Ben ([email protected], http://www.fluff.org/) 'a smiley only costs 4 bytes'