On Wednesday 25 March 2009 06:30:16 Max Filippov wrote:
> Instead of firmware itself p54_upload_firmware was sending to the device content of struct firmware
> and the following random garbage. Notice '&' before priv->firmware->data at p54spi_spi_write.
> But simple removing of '&' sign triggered BUG_ON at dma_cache_maint. Thus kmemdup - kfree workaround.
>
> Signed-off-by: Max Filippov <jcmvbkbc@xxxxxxxxx>
> ---
> drivers/net/wireless/p54/p54spi.c | 11 +++++++++--
> 1 files changed, 9 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/net/wireless/p54/p54spi.c b/drivers/net/wireless/p54/p54spi.c
> index d13268f..4ef57c5 100644
> --- a/drivers/net/wireless/p54/p54spi.c
> +++ b/drivers/net/wireless/p54/p54spi.c
> @@ -265,8 +265,15 @@ static int p54spi_upload_firmware(struct ieee80211_hw *dev)
> p54spi_write32(priv, SPI_ADRS_DMA_WRITE_BASE,
> cpu_to_le32(fw_addr));
>
> - p54spi_spi_write(priv, SPI_ADRS_DMA_DATA,
> - &priv->firmware->data, _fw_len);
> + {
> + void *fw = kmemdup(priv->firmware->data, _fw_len, GFP_KERNEL);
> + if (fw)
> + {
> + p54spi_spi_write(priv, SPI_ADRS_DMA_DATA,
> + fw, _fw_len);
> + kfree(fw);
> + }
> + }
>
> fw_len -= _fw_len;
> fw_addr += _fw_len;
---
checkpatch.pl complains
WARNING: line over 80 characters
#94: FILE: drivers/net/wireless/p54/p54spi.c:269:
+ void *fw = kmemdup(priv->firmware->data, _fw_len, GFP_KERNEL);
ERROR: that open brace { should be on the previous line
#95: FILE: drivers/net/wireless/p54/p54spi.c:270:
+ if (fw)
+ {
what do you think about the attached proposal?
Regards,
Chr
---
diff --git a/drivers/net/wireless/p54/p54spi.c b/drivers/net/wireless/p54/p54spi.c
index d13268f..7d32a2f 100644
--- a/drivers/net/wireless/p54/p54spi.c
+++ b/drivers/net/wireless/p54/p54spi.c
@@ -228,8 +228,15 @@ static int p54spi_request_eeprom(struct ieee80211_hw *dev)
static int p54spi_upload_firmware(struct ieee80211_hw *dev)
{
struct p54s_priv *priv = dev->priv;
- unsigned long fw_len, fw_addr;
- long _fw_len;
+ unsigned long fw_len, _fw_len;
+ unsigned int offset = 0;
+ int err = 0;
+ void *fw;
+
+ fw_len = priv->firmware->size;
+ fw = kmemdup(priv->firmware->data, fw_len, GFP_KERNEL);
+ if (!fw)
+ return -ENOMEM;
/* stop the device */
p54spi_write16(priv, SPI_ADRS_DEV_CTRL_STAT, cpu_to_le16(
@@ -244,9 +251,6 @@ static int p54spi_upload_firmware(struct ieee80211_hw *dev)
msleep(TARGET_BOOT_SLEEP);
- fw_addr = ISL38XX_DEV_FIRMWARE_ADDR;
- fw_len = priv->firmware->size;
-
while (fw_len > 0) {
_fw_len = min_t(long, fw_len, SPI_MAX_PACKET_SIZE);
@@ -257,23 +261,20 @@ static int p54spi_upload_firmware(struct ieee80211_hw *dev)
cpu_to_le32(HOST_ALLOWED)) == 0) {
dev_err(&priv->spi->dev, "fw_upload not allowed "
"to DMA write.");
- return -EAGAIN;
+ err = -EAGAIN;
+ goto out;
}
p54spi_write16(priv, SPI_ADRS_DMA_WRITE_LEN,
cpu_to_le16(_fw_len));
- p54spi_write32(priv, SPI_ADRS_DMA_WRITE_BASE,
- cpu_to_le32(fw_addr));
+ p54spi_write32(priv, SPI_ADRS_DMA_WRITE_BASE, cpu_to_le32(
+ ISL38XX_DEV_FIRMWARE_ADDR + offset));
p54spi_spi_write(priv, SPI_ADRS_DMA_DATA,
- &priv->firmware->data, _fw_len);
+ (fw + offset), _fw_len);
fw_len -= _fw_len;
- fw_addr += _fw_len;
-
- /* FIXME: I think this doesn't work if firmware is large,
- * this loop goes to second round. fw->data is not
- * increased at all! */
+ offset += _fw_len;
}
BUG_ON(fw_len != 0);
@@ -292,7 +293,10 @@ static int p54spi_upload_firmware(struct ieee80211_hw *dev)
p54spi_write16(priv, SPI_ADRS_DEV_CTRL_STAT, cpu_to_le16(
SPI_CTRL_STAT_HOST_OVERRIDE | SPI_CTRL_STAT_RAM_BOOT));
msleep(TARGET_BOOT_SLEEP);
- return 0;
+
+out:
+ kfree(fw);
+ return err;
}
static void p54spi_power_off(struct p54s_priv *priv)
--
To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
