soc/tegra: cbb: Bug fixes for CBB 2.0 driver
cover soc/tegra: cbb: Bug fixes for CBB 2.0 driver
Bug fixes for Tegra CBB 2.0 driver addressing resume handling, fabric lookup table sizing, and cross-fabric timeout lookup.
Sumit Gupta (3):
soc/tegra: cbb: set ERD on resume for err interrupt soc/tegra: cbb: fix incorrect ARRAY_SIZE in fabric lookup tables soc/tegra: cbb: fix cross-fabric target timeout lookup
drivers/soc/tegra/cbb/tegra234-cbb.c | 35 +++++++++++++++++++++++++--- 1 file changed, 32 insertions(+), 3 deletions(-)
Signature · 1 line
2.34.1
PATCH 1/3 soc/tegra: cbb: set ERD on resume for err interrupt
Set the Error Response Disable (ERD) bit to mask SError responses and use interrupt-based error reporting. When the ERD bit is set, inband error responses to the initiator via SError are suppressed, and fabric errors are reported via an interrupt instead.
The register is set during boot but the info is lost during system suspend and needs to be set again on resume.
--- drivers/soc/tegra/cbb/tegra234-cbb.c | 4 ++++ 1 file changed, 4 insertions(+)
drivers/soc/tegra/cbb/tegra234-cbb.c +4 -0
@@ -1586,6 +1586,10 @@ static int __maybe_unused tegra234_cbb_resume_noirq(struct device *dev) {struct tegra234_cbb *cbb =dev_get_drvdata (dev); +/* set ERD bit to mask SError and generate interrupt to report error */ +if (cbb->fabric->off_mask_erd) +tegra234_cbb_mask_serror (cbb); +tegra234_cbb_error_enable (&cbb->base);dev_dbg (dev,"%s resumed\n" , cbb->fabric->fab_list[cbb->fabric->fab_id].name);
Signature · 1 line
2.34.1
PATCH 2/3 soc/tegra: cbb: fix incorrect ARRAY_SIZE in fabric lookup tables
In reply to Sumit GuptaFix incorrect ARRAY_SIZE usage in fabric lookup tables which could cause out-of-bounds access during target timeout lookup.
--- drivers/soc/tegra/cbb/tegra234-cbb.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/soc/tegra/cbb/tegra234-cbb.c +2 -2
@@ -881,7 +881,7 @@ static const struct tegra234_fabric_lookup tegra234_cbb_fab_list[] = { ARRAY_SIZE(tegra234_common_target_map) }, [T234_AON_FABRIC_ID ] = {"aon-fabric" , true, tegra234_aon_target_map, - ARRAY_SIZE(tegra234_bpmp_target_map) }, + ARRAY_SIZE(tegra234_aon_target_map) }, [T234_PSC_FABRIC_ID] = { "psc-fabric" }, [T234_BPMP_FABRIC_ID ] = {"bpmp-fabric" , true, tegra234_bpmp_target_map,
@@ -1160,7 +1160,7 @@ static const struct tegra234_fabric_lookup tegra241_cbb_fab_list[] = { [T234_CBB_FABRIC_ID ] = {"cbb-fabric" , true, tegra241_cbb_target_map, ARRAY_SIZE(tegra241_cbb_target_map) }, [T234_BPMP_FABRIC_ID ] = {"bpmp-fabric" , true, - tegra241_bpmp_target_map, ARRAY_SIZE(tegra241_cbb_target_map) }, + tegra241_bpmp_target_map, ARRAY_SIZE(tegra241_bpmp_target_map) }, };static const struct tegra234_cbb_fabric tegra241_cbb_fabric = { .fab_id =T234_CBB_FABRIC_ID ,
Signature · 1 line
2.34.1
PATCH 3/3 soc/tegra: cbb: fix cross-fabric target timeout lookup
In reply to Sumit GuptaWhen a fabric receives an error interrupt, the error may have occurred on a different fabric. The target timeout lookup was using the wrong base address (cbb->regs) with offsets from a different fabric's target map, causing a kernel page fault.
Unable to handle kernel paging request at virtual address ffff80000954cc00 pc : tegra234_cbb_get_tmo_slv+0xc/0x28 Call trace: tegra234_cbb_get_tmo_slv+0xc/0x28 print_err_notifier+0x6c0/0x7d0 tegra234_cbb_isr+0xe4/0x1b4
Add tegra234_cbb_get_fabric() to look up the correct fabric device using fab_id, and use its base address for accessing target timeout registers.
--- drivers/soc/tegra/cbb/tegra234-cbb.c | 27 ++++++++++++++++++++++++++- 1 file changed, 26 insertions(+), 1 deletion(-)
drivers/soc/tegra/cbb/tegra234-cbb.c +26 -1
@@ -313,12 +313,37 @@ static void tegra234_cbb_lookup_apbslv(struct seq_file *file, const char *target } } +static struct tegra234_cbb *tegra234_cbb_get_fabric (u8 fab_id) +{ +struct tegra_cbb *entry; + +list_for_each_entry (entry, &cbb_list, node) { +struct tegra234_cbb *priv =to_tegra234_cbb (entry); + +if (priv->fabric->fab_id == fab_id) +return priv; + } + +return NULL ; +} +static void tegra234_sw_lookup_target_timeout (struct seq_file *file,struct tegra234_cbb *cbb,u8 target_id,u8 fab_id) {const struct tegra234_target_lookup *map = cbb->fabric->fab_list[fab_id].target_map; +struct tegra234_cbb *target_cbb =NULL ;void __iomem *addr; +if (fab_id == cbb->fabric->fab_id) + target_cbb = cbb; +else + target_cbb =tegra234_cbb_get_fabric (fab_id); + +if (!target_cbb) { +dev_err (cbb->base.dev,"could not find fabric for fab_id:%d\n" , fab_id); +return ; + } +if (target_id >= cbb->fabric->fab_list[fab_id].max_targets) {tegra_cbb_print_err (file,"\t Invalid target_id:%d\n" , target_id);return ;
@@ -341,7 +366,7 @@ static void tegra234_sw_lookup_target_timeout(struct seq_file *file, struct tegr* e) Goto step-a till all bits are set. */ - addr = cbb->regs + map[target_id].offset; + addr = target_cbb->regs + map[target_id].offset;if (strstr (map[target_id].name,"AXI2APB" )) { addr +=APB_BLOCK_TMO_STATUS_0 ;
Signature · 1 line
2.34.1
Quoted 10 lines · Sumit Gupta
Bug fixes for Tegra CBB 2.0 driver addressing resume handling, fabric lookup table sizing, and cross-fabric timeout lookup.
Sumit Gupta (3):
soc/tegra: cbb: set ERD on resume for err interrupt soc/tegra: cbb: fix incorrectARRAY_SIZEin fabric lookup tables soc/tegra: cbb: fix cross-fabric target timeout lookupdrivers/soc/tegra/cbb/tegra234-cbb.c | 35 +++++++++++++++++++++++++--- 1 file changed, 32 insertions(+), 3 deletions(-)
Applied, thanks.
Thierry