| 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: <20220628191124.qvto5tyfe63htxxr@bsd-mbp.dhcp.thefacebook.com>
Date: Tue, 28 Jun 2022 12:11:24 -0700
From: Jonathan Lemon <jonathan.lemon@...il.com>
To: Vadim Fedorenko <vfedorenko@...ek.ru>
Cc: Jakub Kicinski <kuba@...nel.org>,
Arkadiusz Kubalewski <arkadiusz.kubalewski@...el.com>,
Vadim Fedorenko <vadfed@...com>, Aya Levin <ayal@...dia.com>,
netdev@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
linux-clk@...r.kernel.org
Subject: Re: [RFC PATCH v2 3/3] ptp_ocp: implement DPLL ops
On Mon, Jun 27, 2022 at 11:13:22PM +0100, Vadim Fedorenko wrote:
> On 27.06.2022 20:34, Jonathan Lemon wrote:
> > On Sun, Jun 26, 2022 at 10:24:44PM +0300, Vadim Fedorenko wrote:
> > > From: Vadim Fedorenko <vadfed@...com>
> > >
> > > Implement DPLL operations in ptp_ocp driver.
> > >
> > > Signed-off-by: Vadim Fedorenko <vadfed@...com>
> > > ---
> > > drivers/ptp/Kconfig | 1 +
> > > drivers/ptp/ptp_ocp.c | 169 ++++++++++++++++++++++++++++++--------
> > > include/uapi/linux/dpll.h | 2 +
> > > 3 files changed, 136 insertions(+), 36 deletions(-)
> > >
> > > diff --git a/drivers/ptp/Kconfig b/drivers/ptp/Kconfig
> > > index 458218f88c5e..f74846ebc177 100644
> > > --- a/drivers/ptp/Kconfig
> > > +++ b/drivers/ptp/Kconfig
> > > @@ -176,6 +176,7 @@ config PTP_1588_CLOCK_OCP
> > > depends on !S390
> > > depends on COMMON_CLK
> > > select NET_DEVLINK
> > > + select DPLL
> > > help
> > > This driver adds support for an OpenCompute time card.
> > > diff --git a/drivers/ptp/ptp_ocp.c b/drivers/ptp/ptp_ocp.c
> > > index e59ea2173aac..f830625a63a3 100644
> > > --- a/drivers/ptp/ptp_ocp.c
> > > +++ b/drivers/ptp/ptp_ocp.c
> > > @@ -21,6 +21,7 @@
> > > #include <linux/mtd/mtd.h>
> > > #include <linux/nvmem-consumer.h>
> > > #include <linux/crc16.h>
> > > +#include <uapi/linux/dpll.h>
> > > #define PCI_VENDOR_ID_FACEBOOK 0x1d9b
> > > #define PCI_DEVICE_ID_FACEBOOK_TIMECARD 0x0400
> > > @@ -336,6 +337,7 @@ struct ptp_ocp {
> > > struct ptp_ocp_signal signal[4];
> > > struct ptp_ocp_sma_connector sma[4];
> > > const struct ocp_sma_op *sma_op;
> > > + struct dpll_device *dpll;
> > > };
> > > #define OCP_REQ_TIMESTAMP BIT(0)
> > > @@ -660,18 +662,19 @@ static DEFINE_IDR(ptp_ocp_idr);
> > > struct ocp_selector {
> > > const char *name;
> > > int value;
> > > + int dpll_type;
> > > };
> > > static const struct ocp_selector ptp_ocp_clock[] = {
> > > - { .name = "NONE", .value = 0 },
> > > - { .name = "TOD", .value = 1 },
> > > - { .name = "IRIG", .value = 2 },
> > > - { .name = "PPS", .value = 3 },
> > > - { .name = "PTP", .value = 4 },
> > > - { .name = "RTC", .value = 5 },
> > > - { .name = "DCF", .value = 6 },
> > > - { .name = "REGS", .value = 0xfe },
> > > - { .name = "EXT", .value = 0xff },
> > > + { .name = "NONE", .value = 0, .dpll_type = 0 },
> > > + { .name = "TOD", .value = 1, .dpll_type = 0 },
> > > + { .name = "IRIG", .value = 2, .dpll_type = 0 },
> > > + { .name = "PPS", .value = 3, .dpll_type = 0 },
> > > + { .name = "PTP", .value = 4, .dpll_type = 0 },
> > > + { .name = "RTC", .value = 5, .dpll_type = 0 },
> > > + { .name = "DCF", .value = 6, .dpll_type = 0 },
> > > + { .name = "REGS", .value = 0xfe, .dpll_type = 0 },
> > > + { .name = "EXT", .value = 0xff, .dpll_type = 0 },
> >
> > No need for zeros, or are they just temp stubs?
>
> These are just temp stubs for now.
>
> >
> > > @@ -680,37 +683,37 @@ static const struct ocp_selector ptp_ocp_clock[] = {
> > > #define SMA_SELECT_MASK GENMASK(14, 0)
> > > static const struct ocp_selector ptp_ocp_sma_in[] = {
> > > - { .name = "10Mhz", .value = 0x0000 },
> > > - { .name = "PPS1", .value = 0x0001 },
> > > - { .name = "PPS2", .value = 0x0002 },
> > > - { .name = "TS1", .value = 0x0004 },
> > > - { .name = "TS2", .value = 0x0008 },
> > > - { .name = "IRIG", .value = 0x0010 },
> > > - { .name = "DCF", .value = 0x0020 },
> > > - { .name = "TS3", .value = 0x0040 },
> > > - { .name = "TS4", .value = 0x0080 },
> > > - { .name = "FREQ1", .value = 0x0100 },
> > > - { .name = "FREQ2", .value = 0x0200 },
> > > - { .name = "FREQ3", .value = 0x0400 },
> > > - { .name = "FREQ4", .value = 0x0800 },
> > > - { .name = "None", .value = SMA_DISABLE },
> > > + { .name = "10Mhz", .value = 0x0000, .dpll_type = DPLL_TYPE_EXT_10MHZ },
> > > + { .name = "PPS1", .value = 0x0001, .dpll_type = DPLL_TYPE_EXT_1PPS },
> > > + { .name = "PPS2", .value = 0x0002, .dpll_type = DPLL_TYPE_EXT_1PPS },
> > > + { .name = "TS1", .value = 0x0004, .dpll_type = DPLL_TYPE_CUSTOM },
> > > + { .name = "TS2", .value = 0x0008, .dpll_type = DPLL_TYPE_CUSTOM },
> > > + { .name = "IRIG", .value = 0x0010, .dpll_type = DPLL_TYPE_CUSTOM },
> > > + { .name = "DCF", .value = 0x0020, .dpll_type = DPLL_TYPE_CUSTOM },
> > > + { .name = "TS3", .value = 0x0040, .dpll_type = DPLL_TYPE_CUSTOM },
> > > + { .name = "TS4", .value = 0x0080, .dpll_type = DPLL_TYPE_CUSTOM },
> > > + { .name = "FREQ1", .value = 0x0100, .dpll_type = DPLL_TYPE_CUSTOM },
> > > + { .name = "FREQ2", .value = 0x0200, .dpll_type = DPLL_TYPE_CUSTOM },
> > > + { .name = "FREQ3", .value = 0x0400, .dpll_type = DPLL_TYPE_CUSTOM },
> > > + { .name = "FREQ4", .value = 0x0800, .dpll_type = DPLL_TYPE_CUSTOM },
> > > + { .name = "None", .value = SMA_DISABLE, .dpll_type = DPLL_TYPE_NONE },
> >
> > 80-column limit (here and throughout the file)
>
> I thought this rule was relaxed up to 100-columns?
Only in exceptional cases, IIRC. checkpatch complains too.
> > > static int
> > > ptp_ocp_probe(struct pci_dev *pdev, const struct pci_device_id *id)
> > > {
> > > @@ -3768,6 +3855,14 @@ ptp_ocp_probe(struct pci_dev *pdev, const struct pci_device_id *id)
> > > ptp_ocp_info(bp);
> > > devlink_register(devlink);
> > > +
> > > + bp->dpll = dpll_device_alloc(&dpll_ops, ARRAY_SIZE(bp->sma), ARRAY_SIZE(bp->sma), bp);
> > > + if (!bp->dpll) {
> > > + dev_err(&pdev->dev, "dpll_device_alloc failed\n");
> > > + return 0;
> > > + }
> > > + dpll_device_register(bp->dpll);
> > > +
> > > return 0;
> >
> > 80 cols, and this should be done before ptp_ocp_complete()
> > Also, should 'goto out', not return 0 and leak resources.
>
> I don't think we have to go with error path. Driver itself can work without
> DPLL device registered, there is no hard dependency. The DPLL device will
> not be registered and HW could not be configured/monitored via netlink, but
> could still be usable.
Not sure I agree with that - the DPLL device is selected in Kconfig, so
users would expect to have it present. I think it makes more sense to
fail if it cannot be allocated.
--
Jonathan
Powered by blists - more mailing lists