New: wrong #endif
A message that is not in the archive (sent off-list or not yet received): [email protected]
Jesper Nilsson
Acked-by Mikael StarvikQuoted 103 lines · Andrew Morton
Begin forwarded message:
Date: Sun, 7 Jun 2009 14:02:50 GMT
From: [email protected]
To: [email protected]
Subject: [Bugme-new] [Bug 13476] New: wrong #endifhttp://bugzilla.kernel.org/show_bug.cgi?id=13476
Summary: wrong #endif Product: Other Version: 2.5 Kernel Version: 2.6.29.4 Platform: All OS/Version: Linux Tree: Mainline Status: NEW Severity: normal Priority: P1 Component: Other AssignedTo: [email protected] ReportedBy: [email protected] Regression: NoHello,
i've detected a wrong #ifdef ...#endif sequence. This was detected by cppcheck, a static code analysis tool.
At file :
linux-2.6.29.4/arch/cris/arch-v32/kernel/irq.cTake a look at the code, to function
voidcrisv32_do_multiple(struct pt_regs* regs){ int cpu; int mask; int masked[NBR_REGS]; int bit; int i; cpu = smp_processor_id(); /* An extra irq_enter here to prevent softIRQs to run after * each do_IRQ. This will decrease the interrupt latency. */ irq_enter(); for (i = 0; i < NBR_REGS; i++) { /* Get which IRQs that happend. */ masked[i] = REG_RD_INT_VECT(intr_vect, irq_regs[cpu], r_masked_vect, i); /* Calculate new IRQ mask with these IRQs disabled. */ mask = REG_RD_INT_VECT(intr_vect, irq_regs[cpu], rw_mask, i); mask &= ~masked[i]; /* Timer IRQ is never masked */ #ifdef TIMER_VECT1 if ((i == 1) && (masked[0] & TIMER_MASK)) mask |= TIMER_MASK; #else if ((i == 0) && (masked[0] & TIMER_MASK)) mask |= TIMER_MASK; #endif /* Block all the IRQs */ REG_WR_INT_VECT(intr_vect, irq_regs[cpu], rw_mask, i, mask); /* Check for timer IRQ and handle it special. */ #ifdef TIMER_VECT1 if ((i == 1) && (masked[i] & TIMER_MASK)) { masked[i] &= ~TIMER_MASK; do_IRQ(TIMER0_INTR_VECT, regs); } #else if ((i == 0) && (masked[i] & TIMER_MASK)) { masked[i] &= ~TIMER_MASK; do_IRQ(TIMER0_INTR_VECT, regs); } }#endif
.....
Here, the last #endif is at the wrong place. It has to be before the bracket is closed. This is the corrected version:
#ifdef TIMER_VECT1 if ((i == 1) && (masked[i] & TIMER_MASK)) { masked[i] &= ~TIMER_MASK; do_IRQ(TIMER0_INTR_VECT, regs); } #else if ((i == 0) && (masked[i] & TIMER_MASK)) { masked[i] &= ~TIMER_MASK; do_IRQ(TIMER0_INTR_VECT, regs); } #endif }
Quoted 8 lines
Best regards
Martin Ettl
Signature · 3 lines
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
I'm unsure how to handle sign-off-by in this case, but I've queued the following for the CRIS tree:
Subject: [PATCH] CRISv32: irq.c - Move end brace outside #endif
The end brace for a larger for statement was placed inside the #else part of #ifdef TIMER_VECT1. However, for all current chips, the define TIMER_VECT1 is always unset, and the error was never triggered.
Move the brace down below the #endif.
Fixes:
http://bugzilla.kernel.org/show_bug.cgi?id=13476
--- arch/cris/arch-v32/kernel/irq.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-)
arch/cris/arch-v32/kernel/irq.c +1 -1
@@ -428,8 +428,8 @@ crisv32_do_multiple(struct pt_regs* regs) masked[i] &= ~TIMER_MASK ; do_IRQ(TIMER0_INTR_VECT, regs); } - }#endif + }#ifdef IGNORE_MASK /* Remove IRQs that can't be handled as multiple. */
Signature · 9 lines
1.6.1
Unless anyone have objections I'll push this to the cris for-next branch
in a day or two.
/^JN - Jesper Nilsson
--
Jesper Nilsson -- [email protected]