summaryrefslogtreecommitdiff
path: root/drivers/ata/ahci_dm816.c
diff options
context:
space:
mode:
authorFlorian Fainelli <f.fainelli@gmail.com>2018-01-12 04:31:08 +0300
committerTejun Heo <tj@kernel.org>2018-01-17 18:52:29 +0300
commiteb73390ae2413bbd5d56a396dfd79fe57de6b5ae (patch)
treebee3bca6ea2e4705347044af06b86d926680d122 /drivers/ata/ahci_dm816.c
parent3e507769d15e4b14b5a3688d7f5cba516cb4e72d (diff)
downloadlinux-eb73390ae2413bbd5d56a396dfd79fe57de6b5ae.tar.xz
ata: ahci_brcm: Recover from failures to identify devices
When powering up, the SATA controller may fail to mount the HDD. The SATA controller will lock up, preventing it from negotiating to a lower speed or transmitting data. Root cause is power supply noise creating resonance at 6 Ghz and 3 GHz frequencies, which causes instability in the Clock-Data Recovery (CDR) frontend module, resulting in false acquisition of the clock at SATA 6G/3G speeds. The SATA controller may fail to mount the HDD and lock up, requiring a power cycle. Broadcom chips suspected of being susceptible to this issue include BCM7445, BCM7439, and BCM7366. The Kernel implements an error recovery mechanism that resets the SATA PHY and digital controller when the controller locks up. During this error recovery process, typically there is less activity on the board and Broadcom STB chip, so that the power supply is less noisy, thus allowing the SATA controller to lock correctly. Signed-off-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Tejun Heo <tj@kernel.org>
Diffstat (limited to 'drivers/ata/ahci_dm816.c')
0 files changed, 0 insertions, 0 deletions